Helmut Mayer (ed.) Celebrating the 50 Years of the Meteorological
Transcription
Helmut Mayer (ed.) Celebrating the 50 Years of the Meteorological
Berichte des Meteorologischen Instituts der Albert-Ludwigs Universität Freiburg Nr. 17 Helmut Mayer (ed.) Celebrating the 50 Years of the Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Freiburg, May 2008 2 ISSN 1435-618X Copyright reserved, particularly rights of reproduction, distribution and translation Self-publishing company of the Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Print: Printing office of the Albert-Ludwigs-University of Freiburg Editor: Prof. Dr. Helmut Mayer Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany Tel.: +49/761/203-3590; Fax: +49/761/203-3586 e-mail: [email protected] Documentation: Ber. Meteor. Inst. Univ. Freiburg Nr. 17, 2008, 227 pp. 3 Editorial The Meteorological Institute, Albert-Ludwigs-University of Freiburg, celebrated its 50th anniversary on 1st April 2008. In view of this occasion, a festive presentation of lectures on - future developments in forest meteorology and applied urban climatology, - future requirements of (i) forest science as well as forestry to forest meteorology and (ii) city planning to applied urban climatology took place in the “Haus zur Lieben Hand” of the Albert-Ludwigs-University of Freiburg. The subjects forest meteorology and urban climatology have been selected for the presentations as they are among the main research fields of the Meteorological Institute since 1992 (Fig. 1). Fig. 1: Research fields of the Meteorological Institute, Albert-Ludwigs-University of Freiburg, since 1992 In order to overview the research at the Meteorological Institute of the Albert-LudwigsUniversity of Freiburg, this report contains articles of current and former institute members dealing with different subjects. In addition, the research data bank of the AlbertLudwigs-University of Freiburg, informs on the research and published results of the Meteorological Institute classified according to years under http://forschdb.verwaltung.uni-freiburg.de/servuni/forschdbuni.fdbfbr1 Research and publications are also documented on the institute's website under http://www.meteo.uni-freiburg.de Helmut Mayer, Editor 4 5 Contents Editorial 3 50 Jahre Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg 7 (A. Kessler and H. Mayer) The forest meteorological experimental site Hartheim of the Meteorological Institute, Albert-Ludwigs-University of Freiburg 17 (H. Mayer, D. Schindler, J. Holst, D. Redepenning and G. Fernbach) Biomechanical properties of forest trees 39 (D. Schindler, J.-P. Egner, J. Schönborn and H. Mayer) Using GIS for mapping storm-endangered forests in Baden-Württemberg 51 (K. Grebhan, D. Schindler and H. Mayer) Analyses of water balance components of beech stands in south-western Germany using BROOK90 61 (J. Holst, Th. Holst and H. Mayer) Ecosystem flux measurements of BVOC at a sub-arctic wetland site 69 (Th. Holst, A. Arneth, S. Hayward and A. Ekberg) Eigenschaften des Klimas in der Überflutungsaue des Oberrheins 81 (D. Ahrens and W. Möhle) KLIMES - a joint research project on human thermal comfort in cities 101 (H. Mayer) Importance of urban meteorological stations - the example of Freiburg, Germany 119 (A. Matzarakis and H. Mayer) Dependence of the thermal urban climate on morphological variables 129 (A. Matzarakis and H. Mayer) Towards urban sustainability: trends and challenges of building environmental assessment methods 141 (F. Ali-Toudert) The assessment of the thermal environment 153 (G. Jendritzky) Temporal patterns of nitrogen oxides measured at selected air quality monitoring stations in south-west Germany 165 (H. Mayer, J. Holst and D. Ahrens) Climatic tourism potential in the North Sea and Black Forest region – a comparison between REMO and DWD data 179 (Ch. Endler and A. Matzarakis) Regional determination of historical heavy precipitation for the reconstruction of extreme flood events 191 (P. Dostal, F. Imbery, K. Bürger and J. Seidel) An analysis of cloud observations from Vernadsky, Antarctica (A. Kirchgäßner) 203 6 Usefulness of seasonal temperature forecasts for Germany 211 (Ch. Koppe) Einladung zur festlichen Vortragsveranstaltung anlässlich des Jubiläums 50 Jahre Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg 221 Photos from the festive presentation on 1st April 2008 in Freiburg 224 7 50 Jahre Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg Albrecht Kessler und Helmut Mayer Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg, Deutschland Zusammenfassung Das Meteorologische Institut der Albert-Ludwigs-Universität Freiburg wurde am 1. April 1958 gegründet. Anlässlich des 50. Jahrestags seiner Gründung wird ein Überblick über die Entwicklung des Meteorologischen Instituts in Lehre und Forschung in den letzten 50 Jahren zu geben. 50 years Meteorological Institute, Albert-Ludwigs-University of Freiburg Abstract The Meteorological Institute, Albert-Ludwigs-University of Freiburg, was established on April 1st, 1958. The 50th anniversary of its establishment prompts the outline on its evolution in teaching and research in the last 50 years. 1. Einleitung Das Jubiläum des Meteorologischen Instituts der Universität Freiburg soll zum Anlass genommen werden, über Forschungen und die Lehre zu berichten, die in den letzten 50 Jahren unter der Leitung der jeweiligen Direktoren durchgeführt worden sind. Seit Gründung des Instituts im Jahre 1958 durch Prof. Dr. Heinz Lossnitzer (19041964) stand die interdisziplinäre Zusammenarbeit in Forschung und Lehre mit anderen Fächern der Universität, insbesondere der Naturwissenschaftlich-Mathematischen Fakultät, später der Geowissenschaftlichen und der Forstwissenschaftlichen Fakultäten und heute der Fakultät für Forst- und Umweltwissenschaften, im Vordergrund der Institutsarbeit. Die Berufungspolitik der verschiedenen Fakultäten, denen das Meteorologische Institut in den letzten 50 Jahren angehörte, war daher bestrebt, diesem Umstand Rechnung zu tragen, auf Einrichtung eines Diplomstudienganges für Meteorologie in Freiburg zu verzichten und mehr auf synergetische Effekte bei interdisziplinärer Kooperation zu setzen. Dies hatte zur Folge, dass bei Berufungen darauf zu achten war, eine möglichst umfassende Vertretung des Faches „Meteorologie und Klimatologie“ zu gewährleisten. Die Forschungsschwerpunkte des Meteorologischen Instituts änderten sich freilich beim Wechsel der Fachvertreter. Über die Entwicklung des Stellenplans des Lehr- und Institutspersonals sowie über Ausstattung und Unterbringung des Instituts wurde bereits an anderer Stelle ausführlich berichtet1), so dass wir uns auf die Schilderung der wichtigsten Punkte beschränken können. 1 ) vergl. Kessler, A.: Das Meteorologische Institut der Universität Freiburg und seine ehemaligen Direktoren. Manuskript 1996, Archiv des Meteorologischen Instituts. – außerdem: Mayer, H. und A. Kessler: Das Meteorologische Institut der Universität Freiburg, in: Promet, 2005, Jg. 31, Nr.1, 60-64. 8 2. Von der Gründung des Instituts bis zum Jahre 1972 Das Institut und der Lehrstuhl für Meteorologie und Klimatologie sind aus bescheidenen Anfängen eines Lehrauftrages für einen Physiker an der ehemaligen Naturwissenschaftlich-Mathematischen Fakultät am Anfang des vorigen Jahrhunderts in einer langwierigen Entwicklung hervorgegangen. Die Verlegung der Forstlichen Lehre von den Universitäten Karlsruhe und Tübingen an die Universität Freiburg im Jahre 1920 und die Einrichtung des Prüfungsfaches „Meteorologie und Klimatologie“ bei der Diplomausbildung für Forstwirte führten schließlich zur Sicherung eines dauerhaften Lehrangebots in diesem Fach. Dadurch entwickelte sich auch eine bis heute anhaltende, enge Verbindung zwischen dem Meteorologischen Institut und Forstwissenschaftlichen Instituten. Ab 1937 führte Dr. Lossnitzer die Lehrveranstaltungen durch. Er habilitierte sich 1947 und wurde 1950 zum Honorarprofessor ernannt. Schließlich gelang es Prof. Lossnitzer, das von ihm geleitete Bioklima-Institut des Deutschen Wetterdienstes am 1. April 1958 als Meteorologisches Institut in die Albert-Ludwigs-Universität Freiburg zu überführen. Prof. Lossnitzer war der erste Direktor des Meteorologischen Instituts. Er war ein vielseitig interessierter Wissenschaftler2). Schwerpunkte seiner Forschungen lagen auf den Gebieten der meteorologischen Strahlungskunde und der Biometeorologie. Er beschäftigte sich vor allem mit medizinmeteorologischen Fragen. Lossnitzer kann daher als Mitbegründer dieser Disziplin in Deutschland angesehen werden. Durch Honorarprofessor Dr. Gerd Jendritzky wird dieses Forschungsgebiet auch heute am Meteorologischen Institut weiter betreut. Lossnitzer erkannte zeitig in diesem Zusammenhang das Problem der Luftverschmutzung und wurde einer der Initiatoren der laufenden Messungen der Luftbeimengungen in unserer Region. Durch den frühen Tod von Prof. Lossnitzer im Jahr 1964 geriet der Ausbau des Meteorologischen Instituts als Universitätseinrichtung ins Stocken. In der kurzen Zeit als Institutsleiter hat er Arbeiten veröffentlicht u.a. über Interferenzfiltermessungen der Sonnenstrahlung3), über spektrale Durchlässigkeit von verschiedenen Filtermaterialien4) und über die Spektralverteilungen optischer Umgebungsstrahlung5). Sein letzter Einsatz galt dem von der Deutschen Forschungsgemeinschaft geförderten Bodensee-Projekt, an dem das Institut beteiligt war. Die Arbeiten blieben unvollendet. Die Berufung eines neuen Institutsleiters gelang erst nach vier Jahren. In der Zwischenzeit wurde die kommissarische Institutsleitung dem Professor der Forstwissenschaft Dr. Gerhard Mitscherlich (1911-2007) übertragen. Diese Entscheidung wirkte sich positiv auf spätere forstmeteorologische Projekte des Meteorologischen Instituts aus. Mitscherlich setzte das Institutspersonal für seine ertragskundlichen Arbeiten im Freiburger Raum6) ein, so dass die Techniker des Meteorologischen Instituts mit den Besonderheiten forstmeteorologischer Messungen vertraut wurden. Dieses Zusammenwirken mündete in die Einrichtung einer gemeinsamen Werkstatt des Meteorologischen Instituts und des Instituts für Forstli2 ) vergl. Neuwirth, R.: Nachruf auf Prof. Dr. Heinz Lossnitzer, in: Meteorol. Rundschau, 1964, Bd.17, 130. 3 ) Archiv für Meteorol., Geophysik und Bioklimat., Serie B, Bd.9, 1958,73-79. 4 ) Archiv für Meteorol., Geophysik und Bioklimat., Serie B, Bd.9, 1959,434-440. 5 ) Archiv für Hygiene und Bakteriologie, Bd. 147, 1963, 1-10. 6 ) vergl. Mitscherlich, G.: Wald, Wachstum und Umwelt. 2 Bände. 2. Aufl., 1978 bzw. 1981. 9 che Ertragskunde und wurde auch über mehrere Umzüge hinweg beibehalten. Vorlesungen für die Forststudenten übernahm Dr. Neuwirth, ein Schüler von Prof. Lossnitzer. 1968 konnte man den Meteorologen und damaligen Dozenten am Institut für Meereskunde der Universität Kiel Dr. Hans Hinzpeter (1921-1999)7) als neuen Institutsdirektor gewinnen, der als Strahlungsfachmann ausgewiesen war. Während seiner kurzen zweijährigen Tätigkeit am Institut – Hinzpeter folgte 1970 einem Ruf auf einen Lehrstuhl an der Universität Mainz – veröffentlichte er Beiträge über die Grenzfläche Ozean-Atmosphäre8), die auf Vorarbeiten an dem Kieler Institut basierten. Im Jahre 1970 kam es zu einer Neuordnung der Fakultäten an der Universität Freiburg. Hinzpeter votierte dafür, das Fach „Meteorologie und Klimatologie“ und das Meteorologische Institut der Geowissenschaftlichen Fakultät zuzuordnen. Diese Grundsatzentscheidung – eine Zuordnung zur Fakultät für Physik war außerdem erwogen worden - führte dem Institut neue Lehraufgaben zu und beeinflusste weitere Berufungen. Im Jahre 1969 begannen Vorbereitungen zusammen mit den Instituten für Forstliche Ertragskunde und Bodenkunde zu dem interdisziplinären Projekt „Ökologie eines Kiefernwaldes im Trockengebiet der südlichen Oberrheinebene bei Hartheim“. Nach dem Weggang von Prof. Hinzpeter wurde das Meteorologische Institut kommissarisch von dem Geographen Prof. Dr. Wolfgang Weischet (1921-1998) geleitet. Die Vertretung des Faches Meteorologie und Klimatologie wurde bis zur Neubesetzung der Institutsleitung durch Lehraufträge an Prof. Hinzpeter, Mainz, und Prof. Kessler, Bonn, wahrgenommen, die auch die Vordiplomprüfung der Studierenden der Forstwissenschaftlichen Fakultät durchführten. 3. Die Entwicklung in den Jahren 1972 bis 1992 Im Jahre 1972 nahm Prof. Dr. Albrecht Kessler (*1930), der eine Abteilung für Klimatologie und Hydrologie am Geographischen Institut der Universität Bonn leitete, den Ruf auf die Stelle eines Abteilungsvorstehers und Wissenschaftlichen Rates an der Geowissenschaftlichen Fakultät der Universität Freiburg an. Ihm wurde auch die Institutsleitung übertragen. Seine Stelle wurde 1980 nach einer Rufabwendung – Kessler hatte 1977 den Ruf auf einen Lehrstuhl an der Universität Köln abgelehnt – in ein Ordinariat umgewandelt und mit ihm besetzt, das er bis zu seinem Ruhestand 1992 innehatte. Nachdem bis Anfang der 70er Jahre das Fach „Meteorologie und Klimatologie“ nur im Studiengang der Forstwissenschaften als Pflichtfach verankert war, gingen die Bemühungen Kesslers dahin, das Fach auch in anderen Studiengängen als Wahlpflichtfach zu etablieren. Diese Anstrengungen waren nach anfänglichen Schwierigkeiten durch Fächerkonkurrenzen und -egoismen schließlich doch erfolgreich, vor allem bei den Disziplinen, die sich mit Umweltproblemen befassten. Besonders mit der Hydrologie, die sich aus einer Professur für Physische Geographie zu einem eigenständigen Fach entwickelte hatte, kam es zu einer engeren Kooperation. Dies wiederum führte dazu, dass der Lehrstoff der Vorlesungen, Übungen und Seminare angepasst und erweitert werden musste. Außerdem eröff- 7 ) vergl. Storch, H. von (Hrsg.): Interview mit Prof. Dr. Hans Hinzpeter, Eigenverlag des Max-PlanckInstituts für Meteorologie, 1995. 8 ) Veröffentlichungen in: Kieler Meeresforschung Bd. 24, 1968; Annalen d. Meteorologie NF, Nr.4, 1969; Meteor Forschungsergebn. B, Nr. 5, 1970. 10 nete sich die Möglichkeit, dass in Diplomstudiengängen mit engerem Bezug zum Fach „Meteorologie und Klimatologie“ Diplomarbeiten am Meteorologischen Institut angefertigt werden konnten. Während Kesslers Institutsleitung standen drei Themenkreise im Vordergrund der Untersuchungen: a) Ausbau der Messstelle im Hartheimer Kiefernwald zu einer Dauermessstation des Strahlungs-, Energie- und Wasserhaushalts, b) Arbeiten zum globalen Klima, c) Arbeiten zum Klima, Wasserhaushalt und Paläoklima des peruanisch-bolivianischen Altiplano. a) Die globale Diskussion über die rezenten Klimaänderungen im Zusammenhang mit dem sogenannten Treibhauseffekt basiert auf langjährigen Messungen vor allem der Lufttemperatur und des Niederschlags. Im krassen Gegensatz dazu fehlen auch im internationalen Rahmen weitgehend Langzeitmessungen der physikalischen Prozesse, die das Klima der erdbodennahen Atmosphäre bedingen9), so dass entsprechende Modellsimulationen mit Messdaten kaum überprüft werden können. Mit dem Projekt „Hartheimer Kiefernwald“ sollte diesem Manko mit langjährigen Strahlungs- und Energiestrommessungen an einem repräsentativen Sonderstandort begegnet werden. Dabei war vor allem die Abhängigkeit der Energie- und Stoffflüsse von der Witterungsund Klimavariabilität zu klären. Weiterhin sollte untersucht werden, wie die genannten Komponenten durch Umweltveränderungen - in dem besonderen Fall durch das natürliche Waldwachstum und forstliche Pflegemaßnahmen - beeinflusst werden. Das bisher gesammelte, mehr als 30 Jahre umfassende Datenmaterial bildet eine fundamentale Basis für die „Global Change“ Debatte. Die ununterbrochenen Messungen laufen seit 1974 und werden seit der Emeritierung von A. Kessler im Jahre 1992 unter Leitung seines Nachfolgers Prof. Dr. H. Mayer (*1947) weitergeführt. Die wissenschaftliche/technische Betreuung der Station lag bis Ende 2002 in den Händen von apl. Prof. Dr. Lutz Jaeger10) und wird seitdem von Dr. Dirk Schindler wahrgenommen. Der Kiefernwald zeigt seit 1974 ein Höhenwachstum von ca. 12 Metern. Das besondere technische Problem des Gesamtprojektes Hartheim lag daher in der laufenden Höhenanpassung der Messfühleranordnung an den Messtürmen, an der Gewährleistung homogener, lückenloser Messdatengewinnung und an einer zeitgemäßen Datenspeicherung, Datensicherung und Weiterverarbeitung. Die Registrierung begann 1972 mit Fallbügelschreibern, später kamen 2 Prozessrechner11) zum Einsatz, während heute mit Datenloggern gearbeitet wird. Die Anpassung an den mess- und datentechnischen Fortschritt in relativ kurzer Zeit stellte eine besondere Herausforderung für das wissenschaftliche und technische Personal dar. Die Hartheimer Messstelle entwickelte sich schließlich mit ihrer technischen Infrastruktur zu einer Ankerstation für interdisziplinäre und internationale Projekte, bei denen das Umfeld zusätzlich von zahlreichen Disziplinen untersucht wurde; zu nennen wären Bodenkunde, Biologie, Hydrologie, Fernerkundung, Forstwissenschaft, Umweltchemie etc. Die Arbeiten des Meteoro9 ) Daten sind zusammengestellt bei Kessler, A.: Heat balance climatology, World Survey of Climatology VOL 1A, Amsterdam, Oxford, New York, Tokyo, 1985, und bei Gilgen, H., M. Wild and A. Ohmura: Global Energy Balance Archive GEBA. Zürcher Geogr. Schriften, H. 74, 1997. 10 ) vergl. Jaeger, L.: Die klimatologische Messstation Hartheim des Meteorologischen Instituts der Universität Freiburg im Brg. Ber. Naturf. Ges. Freiburg, Bd.68, 1978; Mayer, H., L. Jaeger, A. Matzarakis, G. Fernbach und D. Redepenning: Forstmeteorologische Messstelle Hartheim des Meteorologischen Instituts der Universität Freiburg. Ber. Meteorol. Inst. Univ. Freiburg, Nr.5, 2000. 11 ) vergl. Jaeger, L.: Prozessrechner in der Mikroklimatologie. Siemens Energietechnik, 2, 1980. 11 logischen Instituts wurden in mehreren Großprojekten gefördert (u.a. Flugzeugmessprogramm 1976 und Mesoskaliges Klimaprojekt MESOKLIP durch die Deutsche Forschungsgemeinschaft; Klimaforschungsprogramm der Bundesregierung ab 1983; Trinationales Regio-Klima-Projekt REKLIP - Deutschland, Frankreich, Schweiz - durch die Landesregierung). Während REKLIP wurden drei weitere Energiebilanzstationen im Höhenprofil zwischen Rheinebene und dem Feldberg über Rasenoberflächen aufgebaut und für einige Jahre in Dauerbetrieb genommen, um die Höhenabhängigkeit der Strahlungs- und Energieströme zu studieren. Außerdem bot sich die Möglichkeit, den Einfluss unterschiedlicher Erdoberflächentypen (Wald und Rasen) auf die genannten Komponenten klimatologisch zu erforschen. b) Die während Kesslers Freiburger Zeit entstandenen Arbeiten zum globalen Klima beschäftigen sich mit dem Strahlungs-, Energie- und Wasserhaushalt des Planeten Erde. Anfang der 70er Jahre fehlten Niederschlagskarten von den Land- und Wasserflächen der Erde für die einzelnen Monate. Basierend auf einer Vorstudie12) entwickelte Jaeger in seiner Dissertation13) die erste globale Niederschlagsjahresbilanz auf Beobachtungsbasis für Kontinente und die Ozeane, die nach Verbesserungen durch andere Forscher heute vielfach als Bezugsstandard zur Evaluierung von Simulationen des globalen Wasserhaushalts herangezogen wird14). Eine ähnliche Studie zur globalen Verdunstungsjahresbilanz hat Kessler vorgelegt15). Weitere Arbeiten16) zur globalen Klimatologie setzen sich mit den Veränderungen der irdischen Atmosphärenmasse auseinander, die durch den schwankenden atmosphärischen Wasserdampfgehalt hervorgerufen werden. c) Nach einer von der Deutschen Forschungsgemeinschaft finanzierten Reise in die peruanischbolivianischen Hochanden im Jahre 1962 hat sich Kessler immer wieder mit den Klimaproblemen dieses Raumes auseinandergesetzt. Die Auswertung der seit dem Internationalen Geophysikalischen Jahr 1957/58 stark anwachsenden Radiosondendaten führten zur Klärung des Niederschlagsregimes des Altiplano17). Es wurde zum ersten Mal die große Bedeutung der tropischen Ostströmung der freien Atmosphäre postuliert, die auch für die rezenten und quartären Feuchteschwankungen verantwortlich ist. Die außertropische Westströmung spielt in diesem Zusammenhang keine besondere Rolle, wie man zunächst vermutet hatte. Diese These konnte auch von anderen Forschern seitdem immer wieder bestätigt werden. Es folgten Arbeiten über den Einfluss des El-Nino Phänomens und der 26monate-Schwingung (quasi biennual oscillation) auf die Pegelschwankungen des Titicacasees. Kessler beteiligte sich auch an der Diskussion über das spätpleistozäne Pluvial, das zur Bildung eines riesigen Paläosees18) auf dem südlichen Altiplano geführt hat; das Phänomen konnte bisher noch nicht völlig geklärt werden. 12 ) Kessler, A.: Globalbilanzen von Klimaelementen. Ein Beitrag zur allgemeinen Klimatologie der Erde. Ber. Meteorol. Institut der Techn. Universität Hannover, H.3, 1968. 13 ) Jaeger, L.: Globalbilanzen von Niederschlagen. Diss. 1975; Jaeger, L.: Monatskarten des Niederschlags für die ganze Erde. Ber. Deutscher Wetterdienst, Nr. 139, 1976. 14 ) Vergleiche mit den Ergebnissen von 29 globalen Klima-Modellierungen in: Lau, K., J. Kim and Y. Sud: Intercomparison of Hydrologic Processes in AMIP GCMs. Bull. American Met. Soc .Vol.77, 1996. Vergleiche außerdem: Schlosser, C.A. and P.R. Houser: Assessing a Satellite-Era Perspective of the Global Water Cycle. Journal of Climate, Vol. 20, 2007 15 ) Kessler, A.: Heat balance climatology, World Survey of Climatology VOL: I A, Amsterdam, Oxford, New York, Tokyo, 1985 16 ) Kessler, A.: Global sea level pressure and cosmic ray flux. Meteorol. Zeitschrift, Vol.11, 2002. Dazu vergleiche Hantel, M.: A note on the article „Global sea level pressure and cosmic ray flux“ by A. Kessler . Meteorol. Zeitschrift, Vol.11, 2002. 17 ) Kessler, A.: Atmosphärische Zirkulationsanomalien und Spiegelschwankungen des Titicacasees. Bonner Meteorol. Abhandlungen H. 17, 1974, 18 ) Kessler, A.: The palaeohydrology of the Late Pleistocene Lake Tauca on the Bolivian Altiplano and recent climatic fluctuations. In Vogel, J.C.(Ed.) , Late Cainozoic Palaeoclimates of the Southern Hemisphere, Rotterdam, Boston, 1984. 12 Die vollständige Liste von Kesslers Veröffentlichungen bis zum Jahre 2000 findet man in den Berichten des Meteorologischen Instituts der Universität Freiburg Nr.5, 2000. Weitere Titel sind auf der Homepage des Meteorologischen Instituts und auf den Seiten des Meteorologischen Instituts in der Forschungsdatenbank der Albert-Ludwigs-Universität Freiburg zusammengestellt. 4. Die Entwicklung in den Jahren seit 1992 Nach der Emeritierung von Prof. Dr. Kessler im September 1992 folgte Prof. Dr. Helmut Mayer vom Lehrstuhl für Bioklimatologie und Angewandte Meteorologie der Universität München im Oktober 1992 der Berufung auf den Lehrstuhl für Meteorologie und Klimatologie am Meteorologischen Institut der Albert-Ludwigs-Universität Freiburg. Zugleich übernahm er die Funktion des Institutsdirektors. Die Forschungsausrichtung des Meteorologischen Instituts verlagerte sich in die Bereiche (i) Forstliche Meteorologie/Hydrometeorologie, (ii) Umweltmeteorologie mit den Schwerpunkten Stadtklimatologie, Luftreinhaltung und Angewandte Human-Biometeorologie und (iii) physikalische Klimatologie auf regionaler Skala. Das Meteorologische Institut war 1992 im Erdgeschoß des Gebäudes am Werderring 10 untergebracht. Dieses Gebäude hat einen herrlichen Garten, in dem in den 90er Jahren verschiedene Obstbäume für phänologische Beobachtungen gepflanzt wurden Im Rückgebäude am Werderring 6 betrieb das Meteorologische Institut gemeinsam mit dem Institut für Waldwachstum der Albert-Ludwigs-Universität Freiburg eine Feinmechanik-Werkstatt. Die beengte Raumsituation erschwerte eine Drittmittelforschung mit Doktorandenstellen und experimentellen Ansätzen erheblich. Sie wurde im Jahr 1997 mit der Zuweisung von zusätzlichen Büro-, Labor- und Lagerräume im Rückgebäude in der Hebelstraße 27 entspannt. Seit dem Jahr 2003 befindet sich die gemeinsame Feinmechanik-Werkstatt in einem ehemaligen Gebäude des Herder-Verlags (Tennenbacher Straße 4). Die Zusammenführung der drei Institutsteile an einem einzigen Standort stellt eine wichtige Zukunftsaufgabe dar. Sie ist im Ende 2007 erschienenen Bericht zur vergleichenden Evaluation der Geowissenschaftlichen Institute an Universitäten in BadenWürttemberg explizit angegeben. Bei einer erneuten Strukturreform der Fakultäten der Albert-Ludwigs-Universität Freiburg im Jahre 2002 wurde die Geowissenschaftliche Fakultät aufgelöst und das Meteorologische Institut zusammen mit anderen geo- und forstwissenschaftlichen Instituten zu einer Fakultät für Forst- und Umweltwissenschaften vereinigt. Im Rahmen des 550-jährigen Jubiläums der Albert-Ludwigs-Universität Freiburg im Jahr 2007 wurde im September 2007 „Werderring“ in „Werthmannstraße“ umbenannt. 13 4.1 Lehre Bis zur Umstrukturierung der Studiengänge infolge des im Jahr 1999 beschlossenen Bologna-Prozesses gab es an der Albert-Ludwigs-Universität Freiburg keinen Diplomstudiengang Meteorologie, weil er durch das relativ kleine Meteorologische Institut nicht „gestemmt“ werden konnte, und ein Studiengang, der weitgehend auf Lehrimport beruht, als nicht sinnvoll erachtet wurde. Da am deutlich größeren Institut für Meteorologie und Klimaforschung an der Universität Karlsruhe (TH) ein Diplomstudiengang Meteorologie existierte, bestand auch überhaupt kein Bedarf, an einer weiteren Universität in Baden-Württemberg einen vergleichbaren Studiengang einzurichten. Daher hatte das Fach „Meteorologie und Klimatologie“, das vom Meteorologischen Institut vertreten wurde, in der Lehre die Rolle eines Querschnittsfaches mit Lehrveranstaltungen in Grundlagen von Meteorologie und Klimatologie sowie Alleinstellungsmerkmalen in vertiefenden Lehrangeboten aus seiner Forschungsausrichtung. Das Fach „Meteorologie und Klimatologie“ war Pflichtfach im Grundstudium der Diplomstudiengänge für Forstwissenschaft und für Hydrologie sowie im Magister Scientiarum Studiengang. Darüber hinaus war es Wahlpflichtfach im Hauptstudium der Diplomstudiengänge Forstwissenschaft und Hydrologie sowie in den Diplomstudiengängen Informatik und Physik. An der Fakultät für Forst- und Umweltwissenschaften der Albert-Ludwigs-Universität Freiburg wurde im Wintersemester 2005/2006 mit der Umstrukturierung der Studiengänge begonnen. Innerhalb der neu geschaffenen BSc und MSc Studiengänge ist das Querschnittsfach „Meteorologie und Klimatologie“ in unterschiedlicher fachlicher Ausrichtung in Modulen oder Lehrbausteinen innerhalb von Modulen vertreten. Das neue BSc Nebenfach „Meteorologie und Klimatologie“, das gemeinsam mit dem Institut für Physische Geographie angeboten wird, ist vor dem Hintergrund eines berufsqualifizierenden Abschlusses so gestaltet, dass es Kompetenzen im Bereich von umweltmeteorologischen Fragestellungen im globalen und regionalen Klimawandel vermittelt. Lehrangebote des Meteorologischen Instituts werden durch neue Studiengänge aus anderen Fakultäten der Albert-Ludwigs-Universität Freiburg im Rahmen einer Mehrfachnutzung berücksichtigt. 4.2 Forschung Die Arbeitsgruppen des Meteorologischen Instituts der Universität Freiburg bearbeiten aktuelle Probleme innerhalb seiner Forschungsausrichtung mit finanzieller Unterstützung der DFG, des BMBF, der EU oder anderer Drittmittelgeber. Die meisten Untersuchungen werden derzeit im Rahmen von Verbundvorhaben vor dem Hintergrund des regionalen Klimawandels durchgeführt. Die aktuellen Forschungsarbeiten sind (i) auf der Homepage des Meteorologischen Instituts, (ii) auf den Seiten des Meteorologischen Instituts in der Forschungsdatenbank der Albert-Ludwigs-Universität Freiburg und (iii) über Artikel in diesem Band dargestellt, so dass hier ein kurzer Überblick ausreicht. Forstliche Meteorologie/Hydrometeorologie Die Forstmeteorologische Messstelle Hartheim die sich innerhalb eines Kiefernwaldes (Pinus sylvestris L.) ca. 25 km südwestlich von Freiburg in der südlichen Oberrheinebene befindet (siehe Abschnitte zuvor), stellt das zentrale Freiluftlabor des Meteorolo- 14 gischen Instituts der Universität Freiburg dar. Dort werden seit dem Jahre 1974 kontinuierlich alle meteorologischen, hydrologischen und forstlichen Variablen gemessen, die für die langfristige Untersuchung der Wechselwirkungen zwischen der Wuchsdynamik des Waldes (Höhen- und Dickenwachstum, Durchforstungseingriffe) und den Prozessen sowie daraus resultierenden Zuständen seines Strahlungs-, Wärme-, Wasser- und Impulshaushaltes erforderlich sind. Aufgrund der vorhandenen Infrastruktur sowie der horizontalen Gelände- und Bestandeshomogenität fanden an diesem Standort häufig temporäre Untersuchungen zu aktuellen waldbezogenen Fragestellungen (z.B. Hartheim Experiment HartX im Rahmen von REKLIP, BMBF Projekt Bestimmung von Quellstärken von flüchtigen organischen Verbindungen für Waldökosysteme, MWK Landesforschungsprojekt Wasser- und CO2-Haushalt eines Waldes auf einem Trockenstandort, INTEREG IIIA Projekt 3c.10 Auswirkungen von Klimaänderungen auf Pflanzenbestände am Oberrhein) statt, in die auswärtige Forschergruppen einbezogen sind. Der zunehmenden Bedeutung von Buchen (Fagus sylvatica L.) in der Forstwirtschaft wird in interdisziplinären Forschungsarbeiten zum Wasser- und Nährstoffhaushalt von Buchenwäldern Rechnung getragen. Seit dem Jahre 1999 werden die Einflüsse von Exposition und Überschirmung auf die klimatischen Verhältnisse von Buchenbeständen experimentell und über Modellansätze untersucht. Sie befinden sich auf dem NO- und SW-Hang eines engen Tals bei Möhringen auf der südwestlichen Schwäbischen Alb. In Ergänzung zu den experimentellen Arbeiten wird der vor dem Hintergrund des Klimawandels immer bedeutender werdende Wasserhaushalt der Wälder über die Anwendung verschiedener forstliche Wasserhaushaltsmodelle simuliert. Sensitivitätsstudien dienen u.a. der Analyse, inwieweit Wälder atmosphärische Stressbedingungen, wie die extreme Dürre im Sommer 2003, durch Beeinflussung ihrer Bestandesstruktur graduell abpuffern könnten. Zunehmende intensive Sturmereignisse sind ein weiterer forstlich relevanter Risikofaktor, der auf den Klimawandel zurückgeführt wird. Unter diesem Kontext sind GIS gestützte Ansätze zur Ausweisung von potenziell sturmgefährdeten Waldstandorten in Baden-Württemberg und Modellierungen zu den Wechselwirkungen von dynamischen Windlasten und dynamischen Reaktionen von Bäumen unter Berücksichtigung der Kronenkollisionsproblematik einzuordnen, die über experimentelle Fallstudien ergänzt werden. Umweltmeteorologie Die derzeitigen umweltmeteorologischen Forschungsarbeiten beziehen sich auf aktuelle Fragestellungen in der Stadtklimatologie, Luftreinhaltung und Angewandten HumanBiometeorologie. Gemeinsam mit Forschungspartnern werden integrative Methoden zur auf Menschen bezogenen Bewertung der thermischen und lufthygienischen Komponente des Klimas entwickelt und in verschiedenen Skalenebenen von Stadt- und Regionalplanung angewandt. Durch die Kombination des mikroskaligen Modells ENVI-met für die Simulation des dreidimensionalen Feldes von meteorologischen Parametern in urbanen Strukturen mit der modellierten physiologisch äquivalenten Temperatur PET als thermophysiologisch relevanter Bewertungsindex steht eine Methode zur auf Menschen bezogenen Beurteilung von urbanen Freiraumbedingungen zur Verfügung, die für Städte in Mitteleuropa unter derzeitigen und zukünftigen Klimabedingungen, besonders bezogen auf großräumige Hitze, angewandt und validiert wird. Über immissionsstatische 15 Verfahren erfolgt die Analyse regionaler Charakteristika von Luftkomponenten, wobei unterschiedliche Emissionsbedingungen und atmosphärisch geprägte Transmissionszustände Randbedingungen bilden. Dazu zählt auch die Frage, inwieweit sich Schwebstaubkonzentrationen durch lokale Maßnahmen reduzieren lassen Physikalische Klimatologie Retrospektive Analysen von gemessenen Klimavariablen dienen der Beantwortung der Fragestellung, welches Muster und Ausmaß der Klimawandel in Südwestdeutschland angenommen hat. Sie bilden die Grundlage für die Abschätzung der Auswirkungen des regionalen Klimawandels, wobei sich das Meteorologische Institut weitgehend auf die Zielobjekte Menschen und Wälder beschränkt. Außerdem stehen Auswertungen der langfristigen Veränderungen der Strahlungs- und Wärmeströme über dem Hartheimer Kiefernwald im Vordergrund. Das Klima in der südlichen Oberrheinebene in der Zeit vor der Instrumentenbeobachtung wurde über Proxydaten untersucht. Die dabei erprobten Methoden werden bei der Bearbeitung von Witterungsextremen weiter entwickelt, um u.a. historische Hochwasserereignisse in Südwestdeutschland für ein zukünftiges Hochwasser-Risikomanagement zu analysieren. Methoden der Human-Biometeorologie zur graduellen Bewertung der thermischen Freiraumbedingungen durch Menschen wurden für die Konstruktion von hoch aufgelösten Bioklimakarten von Österreich angewendet. Sie dienen der Fortschreibung der atmosphärischen Grundlagen für Fremdenverkehrszwecke. Dafür wurden das am Meteorologischen Institut entwickelte Human-Bioklimamodell RayMan sowie PET als thermischer Bewertungsindex eingesetzt. Diese Verfahren eignen sich als Tools für Untersuchungen zu den Auswirkungen des Klimawandels auf den Tourismus im regionalen Maßstab. 5. Schlussfolgerungen Die Entwicklung des Meteorologischen Instituts der Albert-Ludwigs-Universität Freiburg in seinen verschiedenen Bereichen stellt eine Art von Wissenschaftsgeschichte in den letzten 50 Jahren in einem spezifischen Wissenschaftsbereich dar. Auf übliche Indikatoren, an denen Arbeiten von Hochschulinstituten beurteilt werden, wie z.B. jährliche Anzahl von peer-reviewten Publikationen, Drittmitteleinwerbungen, abgeschlossene Promotions- und Habilitationsverfahren, wurde hier bewusst nicht eingegangen, weil sie in anderen Quellen zur Verfügung stehen. Anschrift der Autoren: Prof. (em.) Dr. Albrecht Kessler ([email protected]) Prof. Dr. Helmut Mayer ([email protected]) Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg Werthmannstraße 10, D-79085 Freiburg 16 17 The forest meteorological experimental site Hartheim of the Meteorological Institute, Albert-Ludwigs-University of Freiburg Helmut Mayer, Dirk Schindler, Jutta Holst, Dirk Redepenning and Gerhard Fernbach Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Abstract The forest meteorological experimental site Hartheim continuously run since 1974 by the Meteorological Institute, Albert-Ludwigs-University of Freiburg, is located within a Scots pine forest (Pinus sylvestris L.) in the southern upper Rhine plain about 25 km southwest of the city of Freiburg. The main objective of the experimental investigation is the long-term analysis of the interactions between the growth dynamics of the forest and its atmospheric environment. They are characterised by processes and resulting phenomena, which are represented by different balances like radiation, heat, water, impulse and matter balance. In addition, temporary investigations are conducted at the Hartheim site aiming to up-to-date problems in forest meteorology like the dependence of the net ecosystem carbon exchange on weather conditions or the physics of damped sways of a collective of adjacent trees due to dynamic wind loads. Taken into account a previous description of the forest meteorological experimental site Hartheim (MAYER et al., 2000), the site is briefly presented including measuring systems and exemplary results. Forstmeteorologische Messstelle Hartheim des Meteorologischen Instituts der Albert-Ludwigs-Universität Freiburg Zusammenfassung Das Meteorologische Institut der Albert-Ludwigs-Universität Freiburg betreibt seit dem Jahr 1974 kontinuierlich die Forstmeteorologische Messstelle Hartheim. Sie befindet sich in der südlichen Oberrheinebene ca. 25 km südwestlich von Freiburg in einem Waldkiefernbestand (Pinus sylvestris L.) innerhalb des Waldes der Gemeinde Hartheim. Das primäre Ziel der experimentellen Untersuchungen an diesem Standort ist die langzeitliche Analyse der Wechselwirkungen zwischen der Wuchsdynamik des Waldes und seiner atmosphärischen Umgebung. Sie sind durch Prozesse und daraus resultierende Zustände gekennzeichnet, die sich in verschiedenen Bilanzen, wie Strahlungs-, Wärme-, Wasser-, Impuls- und Stoffbilanz widerspiegeln. Zusätzlich finden an diesem Waldstandort zeitlich limitierte Untersuchungen zu aktuellen forstmeteorologischen Fragestellungen statt. Dazu zählen die Abhängigkeit des Netto-Ökosystem-austauschs für Kohlenstoff von den Wetter- und Witterungsbedingungen oder die Physik gedämpfter Biegeschwingungen eines Kollektivs von benachbarten Bäumen infolge von dynamischen Windlasten. Unter Berücksichtigung einer früheren Beschreibung der Forstmeteorologischen Messstelle Hartheim (MAYER et al., 2000) wird hier der Standort einschließlich eingesetzter Messsysteme und exemplarischer Ergebnisse erklärt. 1. Introduction In order to investigate the long-term interactions between the growth dynamics of a forest (radial and height growth, thinnings) and its atmospheric environment, the permanent forest meteorological experimental site Hartheim was established in 1974 by the Meteorological Institute, Albert-Ludwigs-University of Freiburg (MAYER et al., 2000; WELLPOTT et al., 2005; SCHINDLER et al., 2006). The objectives of this article are (i) to describe the site and development of the instrumentation, (ii) to point to short-term in- 18 vestigations conducted at the Hartheim site by different research groups besides the long-term experimental basic program, and (iii) to present selected results. 2. Site The forest meteorological experimental site Hartheim (47° 56′ 04′ N, 7° 36′ 02′ E; 201 m a.s.l.) is located within a slowly growing, even-aged Scots pine forest (Pinus sylvestris L.) of the Hartheim municipality in the middle of the flat southern upper Rhine plain, approximately 3 km WSW of Hartheim (Figs. 1 and 2) and approximately 25 km south-west of the city of Freiburg, SW Germany (MAYER et al., 2000; WELLPOTT et al., 2005; SCHINDLER et al., 2006). forest meteorological experimental site Hartheim Fig. 1: Location of the forest meteorological experimental site Hartheim within the Hartheim Scots pine forest The Scots pine forest was planted in NNE-SSW oriented rows in 1963. Its row structure is visible up to now (Fig. 3). Due to its planting and the flat southern upper Rhine plain, the canopy surface of the Scots pine forest is horizontally homogeneous (Fig. 4). Due to the NNE-SSW oriented upper Rhine valley, the Hartheim site has comparatively warm and dry climate conditions. From 1978 to 2007, i.e. over a period of 30 years, the mean annual air temperature was 10.3 °C and the mean annual precipitation reached 629 mm. Analogous values over the same period for the city of Freiburg are: - mean annual air temperature: 11.4 °C (urban heat island effect), - mean annual precipitation: 949 mm (on the windward side of the Black Forest). 19 N meteorological tower, z/h = 1.2 meteorological tower, z/h = 2.0 forest meteorological experimental site Hartheim Fig. 2: Aerial view of the forest meteorological experimental site Hartheim (picture taken on 14 May 2008) Fig. 3: Row structure of the Scots pine forest at the forest meteorological experimental site Hartheim (picture taken on 7 February 2008) 20 Fig. 4: Horizontally homogeneous canopy structure of the Scots pine forest at the forest meteorological experimental site Hartheim (picture taken on 7 February 2008) At the forest meteorological experimental site Hartheim, the upper soil layer, which consists of sandy loam without any stone fraction, has an average depth of 0.4 m. It has a sufficient reserve of nutrients (see MAYER et al., 2000). Therefore, the roots of the trees are well developed in this layer. The field capacity and the permanent wilting point amount to 33 Vol.% and 12 Vol.%, respectively (STURM, 1998). The lower soil is comprised of sandy gravel, which is not always penetrated by the roots. Therefore, mainly the upper soil layer contributes to the water supply of the forest. The maximum available water storage was calculated as 84 mm. Due to various methods of hydraulic engineering applied to the river Rhine, the depth of the ground water table has dropped to 7 m below the surface. Thus, the ground water is not available for the water supply of the Hartheim Scots pine forest. Due to soil properties, elevation and geographic location between the Vosges Mountains in the west and the Black Forest Mountains in the east, the southern upper Rhine plain represents the warmest and driest region in Germany. The forest meteorological experimental site Hartheim is well known for the regular occurrence of dry summer periods characterised by varying intensity and length (STURM, 1998). 3. Experimental design for continuous measurements at the forest meteorological experimental site Hartheim The forest meteorological experimental site Hartheim covers a fenced area of approximately 7000 m2 within the Hartheim forest. The temporal evolution of the mean stand height h and stand density is contained in Fig. 5. The mean stand height has been determined as an average over a collective of approximately 80 trees around the higher 21 meteorological tower since 2003. Based on an inventory on 7 February 2008, the mean stand height was 15.5 m including a standard deviation of 1.1 m. Analysing fish-eye photos from 7 February 2008 taken upwards in 1.2 m a.g.l., the mean plant area index was 1.6 and the sky view factor 0.2. The hernaceous understorey vegetation is dominated by Brachypodium pinnatum, Carex alba, and Carex flacca. Additionally, approximately 60% of the enclosed area is covered by various shrub and tree species. Forest meteorological experimental site Hartheim growth of the Scots pine stand (Pinus sylvestris L.) 21 14000 18 12000 h N 10000 12 8000 9 6000 6 4000 3 2000 0 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 N (tress/ha) h (m) 15 0 2009 year Fig. 5: Evolution of mean stand height h and number N of trees at the forest meteorological experimental site Hartheim from 1969 to 2008 Different forest meteorological facilities are permanently available at the forest meteorological experimental site Hartheim: - two meteorological walk-up towers with meteorological measurement systems (Fig. 6), - measuring devices above the forest floor and within the near-surface soil layer, - wooden hut containing data logger, PCs as well as mechanic, electric and electronic tools, - measurement container for research groups from outside. In combination with ideal site characteristics, the availability of electric power and other technical facilities is a major reason that research groups from other institutions conduct temporarily limited investigations at the forest meteorological experimental site Hartheim. The BREB method represents the scientific basis for the long-term analysis of the energy balance of the Hartheim Scots pine forest. As it was introduced in 1974, it has been applied for the whole investigation period to avoid methodical impacts. In addition, the 22 aerodynamic method and the water balance method have been applied. Therefore, meteorological probes for: Fig. 6: Measuring systems at the forest meteorological experimental site Hartheim - vertical profile of air temperature, i.e. dry-bulb temperature (2.0 m, 6.1 m, 9.9 m, 12.1 m, 15.5 m, 19.1 m, 23.4 m, 29.3 m a.g.l.), - vertical profile of wet-bulb temperature (2.0 m, 6.1 m, 9.9 m, 12.1 m, 15.5 m, 19.1 m, 23.4 m, 29.3 m a.g.l.), - vertical profile of horizontal wind speed (2.1 m, 6.0 m, 9.7 m, 11.7 m, 15.3 m, 18.8 m, 23.5 m, 29.4 m a.g.l.), - photosynthetically active radiation below and above the canopy (3 m, 11 m, 18 m a.g.l.) as well as short- and long-wave radiation flux densities from the upper and the lower hemisphere above the canopy (18 m a.g.l.), - wind direction (30 m a.g.l.), - gross precipitation (30 m a.g.l.) have been installed on the taller of both meteorological towers with a relative height of z/h = 2.0 (Fig. 7). 23 Comparative measurements of gross precipitation are conducted at the top (18 m a.g.l.) of the lower meteorological tower with a relative height of z/h = 1.2 (Fig. 8). Fig. 7: Meteorological tower at the forest meteorological experimental site Hartheim with double stand height (picture taken on 12 June 2005) Fig. 8: Meteorological tower at the forest meteorological experimental site Hartheim with 1.2-fold stand height (picture taken on 16 January 2006) 24 Fig. 9: Vertical profile of psychrometers mounted on the taller meteorological tower at the forest meteorological experimental site Hartheim (picture taken on 7 August 2006) LI-190SA CG1 CM21 CG4 Fig. 10: Different sensors for photosynthetically active radiation (Licor LI-190SA), short-wave radiation flux densities (Kipp & Zonen CM21) and long-wave radiation flux densities (Kipp & Zonen CG1 and CG4) from the upper hemisphere mounted approximately 2 m above mean stand height on the taller meteorological tower at the forest meteorological experimental site Hartheim (picture taken on 16 March 2005) On principle, the lower meteorological tower is provided for short-term investigations on different objectives. As examples for the meteorological measurement systems, Fig. 9 shows the vertical profile of electrically ventilated psychrometers measuring dry- and wet-bulb temperature. Fig. 10 gives an overview on different sensors for radiation flux densities currently used. As basis for the calculation of the soil heat flux, soil temperature is measured in one vertical profile with Pt 100 sensors in the depths 1 cm, 3 cm, 5 cm, 10 cm, 20 cm and 40 25 cm. To determine the soil heat flux using directly, three heat flux plates (Hukseflux, HFP01) are installed in 2 cm depth and two self-calibrating heat flux plates (Hukseflux, HFP01SC) are installed in 4 cm depth. Volumetric soil moisture content of the upper 0.3 m layer of the soil is monitored by three water content reflectometers (TDR probes, Campbell CS615), which are distributed over the area of the forest meteorological experimental site Hartheim. Fig. 11: Datalogger Campbell CR23X (right) including multiplexers (left) used for the long-term registration of forest meteorological variables at the forest meteorological experimental site Hartheim (picture taken on 20 November 2003) Fig. 12: Wooden hut at the forest meteorological experimental site Hartheim (picture taken on 17 August 2006) 26 All sensors are sampled every 30 s. The data acquisition systems (Campbell CR23X including multiplexers, Fig. 11), which calculates 10 min mean values and 10 min totals for the gross precipitation, respectively, is placed in a wooden hut (Fig. 12). For climatic purposes, a traditional weather hut (Stevenson type) is in use (Fig. 13) containing a hygro-thermograph, which mechanically records air temperature and relative humidity. Fig. 13: Traditional weather hut (Stevenson type) at the forest meteorological experimental site Hartheim (picture taken on 27 August 2007) Besides gross precipitation above the forest, throughfall and stem flow are continuously monitored below the canopy. Throughfall is measured by three rain gutters, which are arranged perpendicular and drain the precipitation in tipping-bucket rain gauges (Campbell ARG100, Fig. 14). Spiral sleeves around ten exemplary Scots pine trees representative for the stand are used to capture the stem flow and collect it in PVC bottles (Fig. 14), which are depleted weekly. The stand stem flow is calculated according to the procedure described in DVWK (1986). Throughfall values are recorded every 30 s und aggregated to 10 min totals and averaged over the three rain gutters. Values for stand stem flow are available as monthly totals. In order to analyse the long-term radial growth reaction of the Scots pine stand on the meteorological conditions, the Institute for Forest Growth, Albert-Ludwigs-University of Freiburg, has installed dendrometers (Fig. 15) at six exemplary Scots pine trees. The principle of this self-built dendrometers is described in detail in HAUSER (2003). 27 Fig. 14: Rain gutters including tipping-bucket rain gauges and stem flow measurement devices at the forest meteorological experimental site Hartheim (pictures taken on 27 August 2007, above, and 7 February 2008, below) Fig. 15: Dendrometer (3.0 m a.g.l.) at a Scots pine tree at the forest meteorological experimental site Hartheim (pictures taken on 16 March 1995, left, and 7 February 2008, right) 28 4. Experimental design for specific investigations at the forest meteorological experimental site Hartheim The technical infrastructure at the forest meteorological experimental site Hartheim offers opportunities for further investigations. Examples for recent investigations conducted by the Meteorological Institute, Albert-Ludwigs-University of Freiburg, are: - test of a modified psychrometer type developed by Gerhard Fernbach, Meteorological Institute, Albert-Ludwigs-University of Freiburg, in order to improve and ensure the water supply of the wet-bulb thermometer (Fig. 16), - investigation on the net CO2 flux FCO2 and the net ecosystem C exchange FNEE using a closed path CO2/H2O analyzer (Fig. 17) in order to estimate the C sink or source strength of Scots pine forest under particular climate and weather conditions, - investigation on FCO2 and FNEE using an open path CO2/H2O analyzer (Fig. 18) in order to assess the weather dependent carbon budget of Scots pine forest in contrast to other land use types in the southern upper Rhine plain, - analysis of the turbulent wind field above and within the Scots pine stand based on a vertical profile of sonic anemometers as background conditions for wind-induced tree sways (Fig. 19), - analysis of wind-induced sways of single and a collective of adjacent Scots pine trees using inclinometers (Fig. 20) with the objective to describe and explain physical processes and features of crown collisions. Fig. 16: Comparative investigations between a traditional psychrometer according to Frankenberger (left) and a modified psychrometer according to Fernbach (right) mounted on the taller meteorological tower at the forest meteorological experimental site Hartheim (picture taken on 22 June 2004) 29 Fig. 17: Gill Solent R2 sonic anemometer and inlet for a Licor LI-6262 closed path CO2/H2O analyzer mounted at the relative height z/h=1.3 on the lower meteorological tower at the forest meteorological experimental site Hartheim (picture taken on 18 December 2003) Fig. 18: Campbell CSAT sonic anemometer and Licor LI-7500 open path CO2/H2O analyzer mounted in 21.4 m (z/h=1.4) above the mean stand height on the taller meteorological tower at the forest meteorological experimental site Hartheim (pictures taken on 16 May 2005, left, and 27 August 2007, right) 30 Fig. 19: Sonic anemometers mounted on a meteorological tower adjacent to the forest meteorological experimental site Hartheim from 2005 to 2007 (pictures taken on 13 October 2005, left, and 16 January 2007, right) Fig. 20: Inclinometers mounted on Scots pine trees adjacent to the forest meteorological experimental site Hartheim to monitor wind-induced tree sways (pictures taken on 13 October 2005, left, and 4 May 2006, right) 31 Fig. 21: Sap flow measurements implemented by the Chair for Tree Physiology, AlbertLudwigs-University of Freiburg, at the forest meteorological experimental site Hartheim (picture taken on 12 September 2003) Fig. 22: Instrumentation within the Hartheim Scots pine forest for the advective experiment (HX06) of groups from the University of Basel and Berlin in spring 2006 (picture taken on 15 April 2006) Examples for recent experimental investigations carried out by other institutions at the forest meteorological experimental site Hartheim are: 32 - analysis of the sap flow based transpiration of Scots pine trees as indicator for their water state, conducted by the Chair of Tree Physiology, Albert-Ludwigs-University of Freiburg (Fig. 21), - investigation on the turbulent advection and advective CO2 fluxes below the canopy of the Scots pine stand, performed in spring 2006 by the Institute for Meteorology, Climatology and Remote Sensing, University of Basel, and the Chair Climatology, Technical University of Berlin (Fig. 22). 5. Complication of experimental investigations due to weather impacts Continuous measurements of forest meteorological variables are exposed weather impacts, which can impair the data quality unless daily maintenances are conducted during specific weather conditions (e.g. heat waves in summer, snowfall, formation of hoarfrost on probes after cloudless nights in the winter half year or formation of dew on probes after cloudless nights during the summer half year). Fig. 23: Influences of winter weather on meteorological measurement systems at the forest meteorological experimental site Hartheim (pictures taken on 16 January 2006) 33 Fig. 24: Inlets of two tubes of an electrically ventilated psychrometer blocked by hoarfrost (left) and snow in a rain gutter (right) at the forest meteorological experimental site Hartheim (pictures taken on 16 January 2006, left, and 6 March 2006, right) Examples for the formation of hoarfrost on meteorological probes are contained in Fig. 23, while Fig. 24 shows how the inlets of the tubes of an electrically ventilated psychrometer can be blocked by hoarfrost and snow completely fills a rain gutter, respectively. 6. Exemplary results Scientific research at the forest meteorological experimental site Hartheim is documented by various publications. They are compiled in a detailed description of this site, which is available at the website of the Meteorological Institute under: http://www.meteo.uni-freiburg.de/forschung/publikationen/hartheimfuehrer Therefore, this article only points out to exemplary results. As climatic background information, the evolution of annual mean air temperature Ta and annual totals of gross precipitation P at the forest meteorological experimental site Hartheim in the 30-yr period 1978 to 2007 (Fig. 25) is characterized by a statistically significant increase of Ta, which is relatively pronounced, and a slight, statistically insignificant increase of P. The dependence of the roughness length z0 and the zero-point displacement d on the mean stand height (Fig. 26) is positively correlated. For the specific site and stand conditions, it can be deduced that z0 and d are approximately 10% and 78% of the mean stand height, respectively. This shows, that it is not advised to use constant z0 and d values for forests. 34 Forest meteorological experimental site Hartheim Scots pine forest (Pinus sylvestris L.) 1000 13 Ta trend: 0.51 K/dec. P trend: 20.8 mm/dec. 900 P 11 Ta 700 10 600 9 500 8 400 7 300 6 1978 1982 1986 1990 1994 1998 2002 Ta (°C) P (mm) 800 12 2006 year Fig. 25: Annual mean values of air temperature Ta and annual totals of gross precipitation P above the Hartheim Scots pine including trends from 1978 to 2007 Forest meteorological experimental site Hartheim Scots pine forest (Pinus sylvestris L.) 12 10 d = 0.78*h - 1.2 2 R = 0.902 z0, d (m) 8 6 4 z0 = 0.097*h + 0.38 2 R = 0.364 2 0 5 7 9 11 13 15 h (m) Fig. 26: Roughness length z0 and zero-point displacement d of the Hartheim Scots pine forest as a function of the mean stand height h (according to IMBERY, 2005) 35 Forest meteorological experimental site Hartheim Scots pine forest (Pinus sylvestris L.) 1.2 Tact/Tpot, Tact/P 1.0 Tact/Tpot 0.8 0.6 Tact/P 0.4 0.2 0.0 1978 1981 1984 1987 1990 1993 1996 1999 year Fig. 27: Simulated annual transpiration index Tact/Tpot and ratio Tact/P (Tact: actual transpiration, Tpot: potential transpiration, P: gross precipitation) of the Hartheim Scots pine forest (according to WELLPOTT et al., 2005) Fig. 27 presents results for the transpiration index Tact/Tpot (Tact: and Tpot: actual and potential transpiration) and the ratio Tact/P obtained by retrospective simulations using the forest water balance model BROOK90. The transpiration index can be used to assess the water availability for the forest. For example, it was reduced in 1991 indicated by a relatively low value for Tact/Tpot. The forest meteorological experimental site Hartheim is the only forest site in SW Germany, where the carbon sink or source strength of a forest is quantified by use of the eddy-covariance approach. To determine FCO2, measuring systems were used, which differ in the types of sonic anemometers and gas analyzers (Figs. 17 and 18). Fig. 26 shows that the Hartheim Scots pine forest was a clear CO2 source in August 2003. The first half of this month was characterized by an extreme heat wave over Central Europe. The mean annual pattern of FNEE, which is typical of a forest at a warm and dry site, is exhibited in Fig. 27. In addition to Fig. 26, FNEE was also positive in July 2006 caused by low water availability and a high air temperature from June to end of July 2006. The climatic growth conditions for the Hartheim Scots pine forest are reflected by its energy flux densities (Fig. 28). Monthly mean values of net radiation indicate that this site does not suffer from available energy. The limitation is caused by an insufficient water supply, which leads to only slight differences between the turbulent fluxes of sensible and latent heat. Averaged over the investigation period from March 1992 to September 1996 mean net radiation was 65 W m-2, mean turbulent sensible heat flux was 30 W m-2, mean turbulent latent heat flux was 37 W m-2and mean sum of soil heat flux and storage heat flux was -2 W m-2. The Bowen ratio amounted to 0.8. 36 Forest meteorological experimental site Hartheim Scots pine forest (Pinus sylvestris L.) 40 -2 -1 FNEE (g C m month ) 20 0 -20 -40 -60 -80 -100 -120 -140 Aug03 Oct03 Dec03 Feb04 Apr04 Jun04 Aug04 month Fig. 26: Monthly totals of the net ecosystem carbon exchange FNEE of the Hartheim Scots pine forest from August 2003 to September 2004 determined by use of a Licor LI-6262 closed path CO2/H2O analyzer (according to SCHINDLER et al., 2006) Forest meteorological experimental site Hartheim Scots pine forest (Pinus sylvestris L.) 40 -2 -1 FNEE (g C m month ) 20 0 -20 -40 -60 -80 -100 -120 -140 Jan05 May05 Sep05 Jan06 May06 Sep06 month Fig. 27: Monthly totals of the net ecosystem carbon exchange FNEE of the Hartheim Scots pine forest from January 2005 to December 2006 determined by use of a Licor LI-7500 open path CO2/H2O analyzer (according to HOLST et al., 2008) 37 Forest meteorological experimental site Hartheim Scots pine forest (Pinus sylvestris L.), March 1992 - September 1996 180 net radiation energy flux densities (W m-2) 160 140 120 100 turbulent flux of sensible heat 80 turbulent flux of latent heat 60 40 20 0 soil heat flux + storage heat flux -20 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec month Fig. 28: Monthly mean values of energy flux densities of the Hartheim Scots pine forest from March 1992 to September 1996 (according to ROST and MAYER, 2006) Acknowledgement The authors are indebted to the people who worked at the forest meteorological experimental site Hartheim in different manner, particularly to Prof. Dr. Albrecht Kessler being responsible for this site from 1972 to 1992. References DVWK, 1986: Ermittlung des Interzeptionsverlustes in Waldbeständen bei Regen. DVWKMerkblätter zur Wasserwirtschaft 211/1986. – Hamburg, Verlag Paul Parey. HAUSER, S., 2003: Dynamik hochaufgelöster radialer Schaftveränderungen und des Dickenwachstums bei Buchen (Fagus sylvatica L.) der Schwäbischen Alb unter dem Einfluß von Witterung und Bewirtschaftung. – PhD thesis, Institute for Forest Growth, Albert-LudwigsUniversity of Freiburg. HOLST, J., R. BARNARD, N. BUCHMANN, L. JAEGER, 2008: Impact of water shortage during summer on carbon fluxes above a Scots pine forest in the southern upper Rhine plain. - Agric. Forest Meteorol. 151, in press. IMBERY, F., 2005: Langjährige Variabilität der aerodynamischen Oberflächenrauhigkeit und Energieflüsse eines Kiefernwaldes in der südlichen Oberrheinebene (Hartheim). - Ber. Meteor. Inst. Univ. Freiburg Nr. 14. MAYER, H., L. JAEGER, A. MATZARAKIS, G. FERNBACH, D. REDEPENNING, 2000: Forstmeteorologische Messstelle Hartheim des Meteorologischen Instituts der Universität Freiburg. - Ber. Meteor. Inst. Univ. Freiburg Nr. 5, 55-83. ROST, J., H. MAYER, 2006: Comparative analysis of albedo and surface energy balance of a grassland site and an adjacent Scots pine forest. - Clim. Res. 30, 227-237. 38 SCHINDLER, D., M. TÜRK, H. MAYER, 2006: CO2 fluxes of a Scots pine forest growing in the warm and dry southern upper Rhine plain, SW Germany. - Eur. J. Forest Res. 125, 201-212. STURM, N., 1998: Steuerung, Skalierung und Umsatz der Wasserflüsse im Hartheimer Kiefernforst (Pinus sylvestris L.). - Bayreuther Forum Ökologie, Band 63. WELLPOTT A., F. IMBERY, D. SCHINDLER, H. MAYER, 2005: Simulation of drought for a Scots pine forest (Pinus sylvestris L.) in the southern upper Rhine plain. - Meteorol. Zeitschrift 14, 143-150. Authors' address: Prof. Dr. Helmut Mayer ([email protected]) Dr. Dirk Schindler ([email protected]) Dr. Jutta Holst ([email protected]) Dirk Redepenning ([email protected]) Gerhard Fernbach ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 39 Biomechanical properties of forest trees Dirk Schindler, Jan-Philipp Egner, Jochen Schönborn, Helmut Mayer Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Abstract Besides collective tree responses to near-surface airflow conditions, the mechanical stability of forests against wind loading is determined by mechanical properties of individual trees. Therefore, tree-pulling experiments were carried out at the forest meteorological experimental site Hartheim over the period June to September 2006. The static and dynamic mechanical properties of 25 pulled Scots pine trees were analysed. This paper presents biomechanical properties of the Scots pine trees and demonstrates their importance for the mechanical stability against wind loading. Biegemechanische Eigenschaften von Waldbäumen Zusammenfassung Die mechanische Stabilität von Wäldern gegenüber Windeinwirkung wird neben kollektiven Baumreaktionen auf bodennahe Strömungsverhältnisse durch biegemechanische Eigenschaften von Einzelbäumen bestimmt. Vor diesem Hintergrund wurden zwischen Juni und September 2006 Zugversuche an 25 Waldkiefern auf der Forstmeteorologischen Messstelle Hartheim durchgeführt. Basierend auf den Ergebnissen der Zugversuche wurden statische und dynamische biegemechanische Kenngrößen der Waldkiefern ermittelt. Die vorliegende Arbeit zeigt die Bedeutung von Baumkenngrößen und daraus für die mechanische Stabilität von Einzelbäumen gegenüber Windeinwirkung abgeleitete biegemechanische Zusammenhänge auf. 1. Introduction During the last decades, storms accounted for the largest part of total damage among the abiotic disturbances in European forests (SCHELHAAS et al., 2003). In the state of Baden-Wuerttemberg, the percentage of random yield loss of forest harvesting varied between 10 and 90% during the period 1986-2005 (FVA, 2006). After the storms Vivian and Wiebke in February 1990 (SCHÜEPP et al., 1994; GOYETTE et al., 2001) and after the severe storm Lothar in December 1999 (MAYER and SCHINDLER, 2002), the percentage of random yield loss of forest harvesting increased significantly in the years after the storms. Damages in forests due to abiotic and biotic causes lead to problems in implementation of regular forestry practices. Beside the scheduled and sustained utilization of growing stock, forest recreation and protective functions of forests take on greater significance. Therefore, it is important for forest owners to minimize the portion of random yield loss. As a result of large-scale changes in meteorological conditions, model calculations provide evidence that in large parts of Europe the number of storms could increase during the next decades (KNIPPERTZ et al., 2000; BENGTSSON et al. 2006; LAMBERT and FYFE 2006; LECKEBUSCH et al., 2006; LECKEBUSCH et al., 2007; PINTO et al. 2007a,b). It is against this background that the importance to take as many measures as possible to reduce the vulnerability of forests against storms will increase. 40 Storm damages in forests do not only occur when trees are exposed to high mean wind speeds (high static wind loads), but also because of the responses of trees to dynamic wind loading. Due to dynamic wind loading trees start to sway. Bending sways at their natural sway frequency are the principal reaction to dynamic wind loading (MAYER, 1987; GARDINER, 1992, 1994; PELTOLA, 1996; FLESCH and WILSON, 1999). Resonance between dynamic wind loads and tree reactions can cause wind damages at wind speeds far below the maximum static wind load trees can resist (OLIVER and MAYHEAD, 1974; MAYER, 1987; Blackburn et al., 1988; GARDINER, 1992). The risk of the occurrence of storm damages in forests can be divided into a permanent (due to soil characteristics, topography, weather/climate) and a temporary (due to stand and tree characteristics) part (LEKES and DANDUL, 2000). The permanent part is hard to change by economic reasonably silvicultural treatment. On the other hand, the temporary part might be, at least in parts, reduced by suitable silvicultural treatment. For example, MAYER (1987) proposed to increase the natural sway frequency of trees to minimize the wind energy transfer into tree movement. The utilisation of the silvicultural potentials to reduce the risk of wind damages in forests requires a profound knowledge and understanding of (i) tree reactions to dynamic wind loading, (ii) the damping of wind-induced tree sways. Total tree damping results from the interaction of crown collisions (BLACKBURN et al., 1988; MILNE, 1991; RUDNICKI et al., 2001, 2003), aerodynamic drag on the tree crowns (MAYHEAD, 1973; RUDNICKI et al., 2004), structural damping (NIKLAS, 1992; JAMES, 2003; JAMES et al., 2006), and viscous damping of the wood (MILNE, 1991). These components of total tree damping contribute to the dissipation of wind energy transferred into forests. In order to understand wind forces acting on forest trees, tree-pulling experiments are carried out (FRASER, 1962; SMITH et al., 1987; FREDERICKSON et al., 1993; BRÜCHERT et al., 2000; MOORE, 2000; PELTOLA et al., 2000; NICOLL et al., 2005; PELTOLA, 2006). In the course of the tree-pulling experiments, trees are displaced with the help of a winch system to study their biomechanical properties. Based on the results of tree-pulling experiments, tree stability against wind as well as sway behaviour of trees can be analysed. The results of tree-pulling experiments are an important basis for mechanistic storm damage risk models like HWIND (PELTOLA et al., 1999), ForestGALES (GARDINER et al., 2000), and FOREOLE (ANCELIN et al., 2004). MOORE and MAGUIRE (2004) summarized studies that measured the natural sway frequency and damping ratios of conifer trees. The purpose of this study was to analyse static and dynamic biomechanical properties of Scots pine trees at the forest meteorological experimental site Hartheim. Here, the experimental setup and first results of static and dynamic tree pulling experiments are presented. 2. Methods 2.1 Measurement site The tree-pulling experiments were carried out at the forest meteorological experimental site Hartheim of the Meteorological Institute (MAYER et al., 2008) over the period June to September 2006. The experimental site is located in a planted Scots pine (Pinus sylvestris L.) forest in the southern Upper Rhine Valley in SW Germany (47°56´04´´N, 41 7°36´02´´E, 201 m a.s.l.). In the year 2006 its mean stand density was 800 trees/ha, the mean stand height sh was 14.5 m, the mean trunk diameter at breast height (dbh) was 17.5 cm, and the plant area index (PAI) was 1.9. The upper soil at the measurement site consists of sandy loam without stone fraction. Its mean depth is 0.4 m (HÄDRICH, 1979). The field capacity and the permanent wilting point amount to 33 vol.% and 12 vol.% (WELLPOTT et al., 2005). The lower soil is sandy gravel and is not penetrated by the roots of the Scots pine trees. In order to determine biomechanical properties of Scots pine at this site, 25 (T1-T25) sample trees (Table 1) were pulled by a hand winch system. Table 1: Characteristics of 25 Scots pine trees (T1-T25) assigned to five groups (G1G5) at the forest meteorological experimental site Hartheim Group Tree Tree height h (m) Diameter at breast height dbh (cm) h/dbh G1 T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 T11 T12 T13 T14 T15 T16 T17 T18 T19 T20 T21 T22 T23 T24 T25 14.6 15.1 14.2 14.2 14.7 15.4 13.6 14.1 13.3 12.9 13.6 16.1 14.3 15.1 14.1 13.0 12.5 12.6 14.9 14.0 13.8 13.7 12.7 13.3 12.3 15.4 18.1 16.0 14.6 17.8 21.0 14.5 19.8 18.0 14.0 20.5 21.8 16.3 19.5 14.5 14.8 15.9 12.9 15.7 16.5 15.4 19.2 12.7 18.8 13.7 95 83 89 97 83 73 94 71 74 92 66 74 88 77 97 88 79 98 95 85 90 71 100 71 90 G2 G3 G4 G5 The vertical variation in the lateral extent of the sample tree crowns was measured by a method similar to the method presented by RAUTIAINEN and STENBERG (2005). Cubic spline interpolation was used to determine the mean crown profiles. The mean outer shape of the tree crowns was obtained by rotating the mean crown profile around the vertical stem axis. On the days when tree-pulling experiments were carried out the daily mean values of the volumetric soil moisture content Θ ranged between 15 and 25 vol.%. The Θ-values were in 42 a close range and clearly below the field capacity. MAYER (1985, 1987) has shown that very moist soil might increase the sway period and decrease tree damping due to lower friction resistance of the roots in the soil. 2.2 Tree-pulling and tree displacement measurements In a first step, a hand winch was used to pull single trees. The pulling force was measured by digital scale (HCB 200K500, Kern, Germany) in kilograms. For all trees, the pull direction was 225° and corresponds to the sector of the most frequent wind direction observed at the experimental site. The attachment point of the rope was directly below the crown of the Scots pine trees representing 56 to 81% of the tree height. According to WOOD (1995) an attachment height of about 80% of the tree height yields a uniform stress distribution in the outer wood fibres. In other studies (SMITH et al., 1987; FREDERICKSON et al., 1993; PELTOLA et al., 1999) the rope was attached between onethird and one-half of the tree height. Each sample tree was pulled with a pulling force that started at 100 N. The pulling force was then gradually increased to at least 500 N by increments of 100 N. For a few trees, maximum applied pulling force was 1000 N. In a second step, five trees pooled in five groups (G1-G5, Table 1) were pulled simultaneously. In order to be able to pull G1-G5, five lashing straps as well as five ropes were used. One lashing strap was attached to one rope used to pull a tree. The pulling force per tree was measured by digital scales (HCB 200K500, Kern, Germany). First, the tree closest to the anchorage point of the lashing straps was pulled. Then, the second closest tree was pulled and so on. After all trees were pulled with the defined force (250 N, 500 N, and 750 N), a snap shackle was used to release all trees at once. Guide pulleys ensured that all trees were pulled into the same direction (225°). In order to prevent crown interactions between the sample trees and the trees surrounding them, the surrounding trees were pulled away. Tree displacement – as a result of release from the pulling ropes – was measured by biaxial clinometers (model 902-45, Applied Geomechanics, USA). Three clinometers were mounted on each of the sample trees. In order to get comparable stem displacement measurement heights among the sample trees, the clinometers were mounted at the relative heights hc1 = 1/7h, hc2 = 3/7h, and hc3 = 5/7h, h being the sample tree height. The biaxial clinometers measured the two horizontal stem displacement components X (E-W) and Y (N-S) and were sampled at a frequency of 10 Hz by a CR5000 micrologger (Campbell Scientific, USA). In order to standardize stem displacement measurements and to minimize sensor heating, all clinometers were pointed northward. 3. Results and discussion 3.1 Tree movement Fig. 1 shows the movement pattern of tree 5 (T5) at the zero plane displacement height zd = 0.78sh – which is the mean height at which momentum is absorbed by individual trees (THOM, 1975) – after the release from the towing rope (pulling force FP: 500 N). As reported from numerous studies (e.g. OLIVIER and MAYHEAD, 1974; MAYER, 1987; GARDINER, 1995) before, it can be seen that bending sways in the pull direction dominate the movement pattern. Lateral displacement is small compared to the displacement in pull direction. 43 Fig. 1: Movement pattern of tree 5 (T5) at the zero-plane displacement height zd = 0.78sh after the release from the towing rope (pulling force: 500 N) 3.2 Biomechanical tree properties In Fig. 2, the displacement in pull direction dp of T5 at zd is shown as a function of applied pulling force FP. T5 was pulled in ten steps from 100 N to 1000 N (increment: 100 N). With increasing FP, dp increases linearly (X = 0.002FP – 0.09, R2 = 0.99). Fig. 2: Displacement in pull direction dp of tree 5 (T5) at the zero-plane displacement height zd = 0.78sh as a function of pulling force FP 44 Similar relationships between FP and dp were determined for all other sample trees. As soon as a wind load displaces a tree, bending moments due to the wind load and due to the tree weight are induced (GARDINER et al., 2000; MATTHECK, 2002). With increasing tree displacement these bending moments increase (data not shown). Due to their dimensions and biomechanical properties, not all sample trees could be pulled with forces up to 1000 N. A common parameter used to analyse tree stability against wind loading is h/dbh. Fig. 3 shows the fit of the linear relationship between dp at zd and h/dbh (Fp: 500 N). It yields dp = 0.10×h/dbh – 5.99 (R2 = 0.76) and shows that trees with larger h/dbh react to static wind loading with larger tree displacement than trees with smaller h/dbh. Owing to technical problems with the datalogger while pulling T18, data of only 24 Scots pine trees are presented. Fig. 3: Displacement in pull direction dp of 24 Scots pine trees at the zero-plane displacement height zd = 0.78sh as a function of h/dbh (pulling force FP: 500 N) An important parameter used to describe the resistance of a tree to bending is the bending stiffness or flexural rigidity EI (SILK KUHN et al., 1982; SPECK, 1990; BRÜCHERT , 1998; BRÜCHERT and BECKER, 2000; BRÜCHERT et al., 2000). EI is the product of the second moment of area I and the modulus of elasticity E and was calculated according to SPECK (1990). Since I is a function of only the stem diameter, it decreases rapidly with increasing height along the stem (BRÜCHERT and BECKER, 2000). The EI-curves of several Scots pine trees presented in Fig. 4 illustrate the range of EI-values to be expected among the sample trees. Up to 0.3hi (i = 1, …, 24), EI shows marked differences among the sample trees. Further up, EI of all Scots pine trees decreases to close to zero. In Fig. 5, dp at 0.3hi is shown as a function of EI at 0.3hi (i = 1, …, 24). At this height, dp decreases clearly with increasing EI. Accordingly, with increasing EI the resistance of the sample trees to bending increases. 45 The natural sway frequency f of the sample trees was determined from consecutive maxima and minima (MOORE and MAGUIRE, 2004) in the time series of tree displacement in x-direction dx after release from the pulling rope. Since groups of five trees were released at the same time, crown interactions might have affected f. Fig. 4: Decrease of bending stiffness EI (MNm2) with increasing relative tree height z/h of the sample trees T6, T7, T8, T9, and T12 (pulling force FP: 500 N) Thus, extreme values of dx were chosen only after crown interactions had no longer an effect on tree displacement. The natural sway frequencies of the sample trees range between 0.23 and 0.39 Hz. Fig. 5: Tree displacement in pull direction dp (m) of 24 Scots pine trees at z/hi = 0.3 (i = 1, …, 24; pulling force FP: 500 N) as a function of bending stiffness EI (MNm2) 46 The natural sway frequencies were in the magnitude of results observed in previous studies (MAYHEAD, 1973; BLACKBURN et al., 1988; MILNE, 1991; FLESCH and WILSON, 1999; BRÜCHERT and GARDINER, 2006). The relationship between dbh/h2 and f (Fig. 9) yields similar results like those reported by GARDINER (1992), MOORE and MAGUIRE (2004) and BRÜCHERT and GARDINER (2006): f increases with increasing dbh/h2. Since damping of tree displacement is not only a function of internal tree properties but also of aerodynamic drag on the tree crowns and structural damping, and since the diameter range of the sample trees is comparatively narrow, the coefficient of determination R2 is lower than for the results obtained by the static tree pulling tests. Fig. 6: Natural sway frequency f of 25 Scots pine trees as a function of dbh/h2 4. Conclusions The results of the tree pulling experiments show that biomechanical properties of the Scots pine trees have an important effect on the static and dynamic response of the trees to the applied pulling force. Based on the results it can be concluded that silvicultural practices that aim to increase dbh (and with it EI) are an option to increase the resistance of individual trees to stem bending. This can be achieved by early reducing the stem number per unit area (BURSCHEL and HUSS, 1987). Greater EI values along the stem reduce the bending moment due to tree mass and the leeward relocation of the mass centre. This minimises the risk of damage as a result of the trees own weight (MITSCHERLICH, 1973). Trees with larger h/dbh might rely more on the support of neighbouring trees to damp their displacement. References ANCELIN, P., B. COURBAUD, T. FOURCAUD, 2004: Development of an individual tree-based mechanical model to predict wind damage within forest stands. – For. Ecol. Manag. 203, 101-121. 47 BENGTSSON, L., K.I. HODGES, E. ROECKNER, 2006: Storm tracks and climate change. – J. Clim. 19, 3518-3543. BLACKBURN, P., J.A. PETTY, K.F. MILLER, 1988: An assessment of the static and dynamic factors involved in wind-throw. – Forestry 61, 29-43. BRÜCHERT, F., 1998: Die biegemechanischen Eigenschaften von Fichten (Picea abies (L.) Karst.) bei unterschiedlichen Wuchsbedingungen. Diss. Univ. Freiburg, 201 pp. BRÜCHERT, F., G. BECKER, 2000: Biegemechanische Eigenschaften von Fichten (Picea abies (L.) Karst.) bei unterschiedlichen Wuchsbedingungen – Grundlagen zur Abschätzung der Stabilität von Bäumen gegenüber mechanischen Belastungen durch Sturm und Schnee. Forstarchiv 71, 102-111. BRÜCHERT, F., B. GARDINER, 2006: The effect of wind exposure on the tree aerial architecture and biomechanics of Sitka spruce (Picea sitchensis). – Amer. J. Bot. 93, 1512-1521. BRÜCHERT, F., G. BECKER, T. SPECK, 2000: The mechanics of Norway spruce [Picea abies (L.) Karst]: mechanical properties of standing trees from different thinning regimes. – For. Ecol. Manag. 135, 45-62. BURSCHEL, P., J. HUSS, 1987: Grundriss des Waldbaus. Paul Parey Verlag, Hamburg, 487 pp. FLESCH, T.K., J.D. Wilson, 1999: Wind and remnant tree sway in forest cutblocks. II. Relating measured tree sway to wind statistics. – Agric. For. Meteorol. 93, 243-258. FRASER, A.I., 1962: The soil and roots as factors in tree stability. – Forestry 35, 117-127. FREDERICKSON, T.S., R.L. HEDDEN, S.A. WILLIAMS, 1993: Testing loblolly pine wind firmness with a simulated wind stress. – Can. J. For. Res. 23, 1760-1765. FVA, 2006: Waldzustandsbericht 2006 der Forstlichen Versuchs- und Forschungsanstalt Baden-Württemberg, 57 pp. GARDINER, B.A., 1992: Mathematical modelling of the static and dynamic characteristics of plantation trees. In: J. Franke, A. Roeder (Eds.): Mathematical modelling of Forest Ecosystems. Sauerländer’s Verlag, Frankfurt am Main, 40-61. GARDINER, B.A., 1994: Wind and wind forces in a plantation spruce forest. – Boundary-Layer Meteorol. 67, 161-184. GARDINER, B.A., 1995: The interactions of wind and tree movement in forest canopies. In: Wind and Trees, ed. M.P. Coutts, J. Grace. Cambridge Univ. Press, Cambridge, 41-59. GARDINER, B., H. PELTOLA, S. KELLOMÄKI, 2000: Comparison of two models for predicting the critical wind speeds required to damage coniferous trees. – Ecol. Model. 129, 1-23. GOYETTE, S., M. BENISTON, D. CAYA, R. LAPRISE, P. JUNGO, 2001: Numerical investigation of an extreme storm with the Canadian Regional Climate Model: the case study of windstorm VIVIAN, Switzerland, February 27, 1990. – Clim. Dyn. 18, 145-168. HÄDRICH, F., 1979: Der Wasserhaushalt einer Zweischicht-Pararendzina unter Kiefernjungbeständen im Trockengebiet am südlichen Oberrhein. – Mitt. Dtsch. Bodenk. Gesellsch. 29, 149-158. JAMES, K., 2003: Dynamic loading of trees. – J. Arboriculture 29, 165-171. JAMES, K.R., N. HARITOS, P.K. ADES, 2006: Mechanical stability of trees under dynamic loads. – Amer. J. Bot. 93, 1522-1530. KNIPPERTZ, P., U. ULBRICH, P. SPETH, 2000: Changing cyclones and surface wind speeds over the North Atlantic and Europe in a transient GHG experiment. – Clim. Res. 15, 109-122. LAMBERT, S.J., J.C. FYFE, 2006: Changes in winter cyclone frequencies and strengths simulated in enhanced greenhouse warming experiments: results from the models participating in the IPCC diagnostic exercise. – Clim. Dynam. 26, 713-728 48 LECKEBUSCH, G.C., U. ULBRICH, L. FRÖHLICH, J.G. PINTO, 2007: Property loss potentials for European midlatitude storms in changing climate. – Geophys. Res. Lett. 34, L05703. LECKEBUSCH, G.C., B. KOFFI, U. ULBRICH, J.G. PINTO, T. SPANGEHL, S. ZACHARIAS, 2006: Analysis of frequency and intensity of European storm events from a multi-model perspective at synoptic and regional scales. – Clim. Res. 31, 59-74. LEKES, V., I. DANDUL, 2000: Using airflow modelling and spatial analysis for defining wind damage risk classification. – For. Ecol. Manag. 135, 331-344. MATTHECK, C., 2002: Mechanik am Baum. Verlag Forschungszentrum Karlsruhe GmbH, Karlsruhe, 139 pp. MAYER, H., 1985: Baumschwingungen und Sturmgefährdung des Waldes. Wiss. Mitt. Meteorol. Inst. Univ. München, No. 51, 244 pp. MAYER, H., 1987: Wind-induced tree sways. – Trees 1, 195-206. MAYER, H., D. SCHINDLER, 2002: Forstmeteorologische Grundlagen zur Auslösung von Sturmschäden im Wald im Zusammenhang mit dem Orkan "Lothar". – Allg. Forst- u. J.-Ztg., 173, 200-208. MAYER, H., D. SCHINDLER, J. HOLST, D. REDEPENNING, G. FERNBACH, 2008: The Forest Meteorological Experimental Site Hartheim of the Meteorological Institute, Albert-LudwigsUniversity of Freiburg. – Ber. Meteor. Inst. Univ. Freiburg Nr. 17, 17-38. MAYHEAD, G.J., 1973: Some drag coefficients for British forest trees derived from wind tunnel studies. – Agric. For. Meteorol. 12, 123-130. MILNE, R., 1991: Dynamics of swaying of Picea sitchensis. – Tree Physiol. 9, 383-399. MITSCHERLICH, G., 1973: Wald und Wind. – Allg. Forst- u. J.-Ztg. 144, 76-81. MOORE, J.R., 2000: Differences in maximum resistive bending moments of Pinus radiata trees grown on a range of soil types. – For. Ecol. Manag. 135, 63-71. MOORE, J.R., D.A. MAGUIRE, 2004: Natural sway frequencies and damping ratios of trees: Concepts, review and synthesis of previous studies. – Trees 18, 195-203. NICOLL, B.C., A. ACHIM, S. MOCHAN, B.A. GARDINER, 2005: Does steep terrain influence tree stability? A field investigation. – Can. J. For. Res. 35, 2360-2367. NIKLAS, K.J., 1992: Plant Biomechanics: an engineering approach to plant form and function. The University of Chicago Press, Chicago, 607 pp. OLIVER, H.R., G.J. MAYHEAD, 1974: Wind measurements in a pine forest during a destructive gale. – Forestry 47, 185-195. PELTOLA, H., 1996: Swaying of trees in response to wind and thinning in a stand of Scots pine. – Boundary-Layer Meteorol. 77, 285-304. PELTOLA, H., 2006: Mechanical stability of trees under static loads. – Amer. J. Bot. 93, 15011511. PELTOLA, H., S. KELLOMÄKI, H. VÄISÄNEN, V.-P. IKONEN, 1999: A mechanistic model for assessing the risk of wind and snow damage to single trees and stands of Scots pine, Norway spruce and birch. – Can. J. For. Res. 29, 647-661. PELTOLA, H., S. KELLOMÄKI, A. HASSINEN, M. GRANANDER, 2000: Mechanical stability of Scots pine, Norway spruce and birch: an analysis of tree-pulling experiments in Finland. – For. Ecol. Manag. 135, 143-153. PINTO, J.G., L. FRÖHLICH, G.C. LECKEBUSCH, U. ULBRICH, 2007a: Changing European storm loss potentials under modified climate conditions according to ensemble simulations of the ECHAM5/MPI-OM1 GCM. – Nat. Haz. Earth Sys. Sci. 7, 165-175. 49 PINTO, J.G., U. ULBRICH, G.C. LECKEBUSCH, T. SPANGEHL, M. REYERS, S. ZACHARIAS, 2007b: Changes in storm track and cyclone activity in three SRES ensemble experiments with the ECHAM5/MPI-OM1 GCM. – Clim. Dynam. 29, 195-210. RAUTIAINEN, M., P. STENBERG, 2005: Simplified tree crown model using standard forest mensuration data for Scots pine. – Agric. For. Meteorol. 128, 123-129. RUDNICKI, M., V.J. LIEFFERS, U. SILINS, 2003: Stand structure governs the crown collisions of lodgepole pine. – Can. J. For. Res. 33, 1238-1244. RUDNICKI, M., S.J. MITCHELL, M.D. NOVAK, 2004: Wind tunnel measurements of crown streamlining and drag relationships for three conifer species. – Can. J. For. Res. 34, 666-676. RUDNICKI, M., U. SILINS, V.J. LIEFFERS, G. JOSI, 2001: Measure of simultaneous tree sways and estimation of crown interactions among a group of trees. – Trees 15, 83-90. SCHELHAAS, M.-J., G.-J. NABUURS, A. SCHUCK, 2003: Natural disturbances in the European forests in the 19th and 20th centuries. – Global Change Biol. 9, 1620-1633. SCHÜEPP, M., H.H. SCHIESSER, H. HUNTRIESER, H.U. SCHERRER, H. SCHMIDTKE, 1994: The Winterstorm "Vivian" of 27 February 1990: About the Meteorological Development, Wind Forces and Damage Situation in the Forests of Switzerland. – Theor. Appl. Climatol. 49, 183-200. SILK KUHN, W., WANG, L.L., CLELAND, R.E., 1982: Mechanical properties of the rice panicle. – Plant Physiol. 70, 460-464. SMITH, V.G., M. WATTS, D.F. JAMES, 1987: Mechanical stability of black spruce in the clay belt region of northern Ontario. – Can. J. For. Res. 17, 1080-1091. SPECK, T., 1990: Biomechanische Untersuchungen an aufrechten Pflanzenachsen unter besonderer Berücksichtigung der frühen „Gefäß“-Landpflanzen sowie selbsttragender und nicht selbsttragender Lebensformtypen. Diss. Univ. Freiburg, 480 pp. THOM, A.S., 1975: Momentum, mass and heat exchange of plant communities. In: J.L. Monteith (Ed.): Vegetation and the Atmosphere, Vol.1 Principles. Academic Press, London, pp. 57109. WELLPOTT, A., F. IMBERY, D. SCHINDLER, H. MAYER, 2005: Simulation of drought for a Scots pine forest (Pinus sylvestris L.) in the southern upper Rhine plain. – Met. Z. 14, 143-150. WOOD, C.J., 1995: Understanding wind forces on trees. In: Wind and Trees, M.P. Coutts, J. Grace (Eds.), Cambridge Univ. Press, Cambridge, 133-164. Authors’ address: Dr. Dirk Schindler ([email protected]) Dipl.-Forstw. Jan-Philipp Egner Dipl.-Forstw. Jochen Schönborn ([email protected]), Prof. Dr. Helmut Mayer ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 50 51 Using GIS for mapping storm-endangered forests in Baden-Württemberg Karin Grebhan, Dirk Schindler and Helmut Mayer Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Abstract Storm is a significant natural hazard in Baden-Württemberg. The work presented in this paper is aimed at reducing the risk of wind damages in forests in Baden-Württemberg. Therefore, two hazard maps will be generated showing the forest vulnerability to wind damage under current (1971-2000) and future (2021-2050) climate conditions. The paper presents the collected data, the GIS-based approach for mapping storm-endangered forests and first results. GIS-gestützte Erstellung von Sturmgefährdungskarten für die Wälder in Baden-Württemberg Zusammenfassung Sturm stellt in Baden-Württemberg eine bedeutende Naturgefahr dar. Ziel dieser Arbeit ist es, das Sturmschadensrisiko der Wälder Baden-Württembergs zu reduzieren. Hierfür sollen zwei Gefährdungskarten erstellt werden, die die potenzielle Sturmgefährdung der Wälder unter derzeitigen (Referenzzeitraum 1971-2000) und zukünftigen (Referenzzeitraum 2021-2050) Klimabedingungen aufzeigen. In diesem Beitrag werden die bisher vorhandenen Daten, der GISbasierte Ansatz für die Kartenerstellung und erste Ergebnisse vorgestellt. 1. Introduction The latest results of climate research show that the present climate change can lead to an enhanced number and increased intensity of severe storms in the mid-latitudes (LECKEBUSCH et al., 2006; IPCC, 2007; PINTO et al., 2007). Forest damages belong to the catastrophic results that can be linked with such natural events. For example, the hurricane “Lothar” of December 1999 blow down 25 millions m³ only in BadenWürttemberg. This is approximately the threefold annual logging in this Federal State and in accordance with an economic loss from approximately 770 millions € (MAYER and SCHINDLER, 2002). In this study the Geographical Information System (GIS) ArcGIS 9.2 will be used for map creation. Based on the hazard maps, recommendations and measures for reducing the risk of wind damages in Baden-Württemberg forests should be derived. The effect of wind on forests is extensively reviewed in many studies (HÜTTE, 1967; BAKER et al., 2002; DOBBERTIN, 2002; SCHMOECKEL, 2006; EVANS et al., 2007). It is determined by complex interactions between numerous factors. The key biotic and abiotic factors that influence the wind risk on forests are shown in Fig.1. Based on the past studies, it is suspected that the forest susceptibility to wind damage in Baden-Württemberg will increase with increasing 1. wind load, 2. tree height, 52 3. terrain complexity. meteorological conditions air flow, precipitation, air temperature human influences logging, road construction forest forest type, stand age and structure, h/d ratio, vitality topography elevation, exposure, slope angle, terrain complexity soil soil type, depth, moisture, pH Fig. 1: Main biotic and abiotic factors affecting the degree of wind damage in forests It is also hypothesized that wind disturbance in forests will be more likely to be found 1. in exposed areas, on ridges, in valleys, 2. on moist, thin or acid soils, 3. in forests dominated by spruces, 4. in areas with forest edges. 2. Materials and methods 2.1 Description of the study area Forest ownership in Baden-Württemberg is a patchwork of corporation (39%), public (24%) and private (37%) forests (MLR, 2008). Based on data of the CORINE Land Cover 2000 project, the forests in Baden-Württemberg cover an area of approximately 13.700 km2, i.e. 38% of the land surface (Fig. 2). Therefore, Baden-Württemberg belongs to the well forested Federal States in Germany. Only Rhineland-Palatinate and Hesse have a higher amount of forested land (KÄNDLER et al., 2005). The most common forest type in the study area is coniferous forest (45%) followed by mixed (35%) and broad-leaved forests (20%). 53 Fig. 2: Map of the Federal State of Baden-Württemberg showing the spatial distribution of forests (according to State Agency for Environment, Measurements and Nature Conservation Baden-Württemberg (LUBW)) 54 The topography of Baden-Württemberg is complex. Striking elements are the Black Forest, whose highest mountain is the Feldberg (1.493 m a.s.l.), the Swabian Jura forming the other range up to 1.000 m a.s.l. and the southern upper Rhine plain adjacent to the west of the Black Forest. 2.2 Data and methods The key procedures used for mapping storm-endangered forests in the Federal State of Baden-Württemberg are summarized in Fig. 3. literature review derivation of further information from the data levels evaluation and weighting of the data levels hazard map under current climate conditions hazard map under future climate conditions simulated air flow 1971 - 2000 workflow development of hazard maps data level arrangement and processing simulated air flow 2021 - 2050 Fig. 3: Flow chart for mapping storm-endangered forests in Baden-Württemberg Data from different sources and derived data sets are used in this study. DTM50 (digital terrain model on a 50 m scale grid) and mean air temperature, mean precipitation, soil types, groundwater as well as backwater affected soils, geology and land cover shapefiles were provided by the State Agency for Environment, Measurements and Nature Conservation Baden-Württemberg (LUBW). Rivers and Topex-to-distance-values in form of shapefiles were supplied by the Forest Research Institute of BadenWürttemberg (Germany) and a street shapefile was extracted from ESRI data (DVD for ArcGIS 9). The collected data was stored in the GIS. Due to the various data sources, it was necessary to define one coordinate system for all data sets. Therefore the Transverse Mercator projection (Bessel 1841 spheroid) was chosen. In addition, all final shapefiles were converted to a 50 m resolution grid for spatial analyses. The second fundamental step will be the development of hazard maps. This process is based on the assumption that forests in certain environments, which are comparable 55 with those where severe damages by storm already occurred, should show a similar high risk. These prone environments are represented by factors affecting the wind risk of forests (Fig. 1). To identify the major affecting factors, the data levels must be evaluated and weighted according to their importance. Taking into account the simulated air flow for the period 1971-2000, which is constructed by the Institute for Meteorology and Climate Research, University of Karlsruhe (Germany) the hazard map for the current climate conditions can be produced. Finally, the hazard map for future climate conditions will be constructed by using the simulated air flow for 2021-2050. 3. Preliminary results The DTM50 was used for extracting (i) elevation, (ii) slope angle, (iii) slope aspect and (iv) curvature. This task has been implemented in ArcGIS 9.2 by using the Spatial Analyst. The values of elevation are divided into fifteen classes with 100 m intervals. The values of slope angle are classified into eight classes: 0°-5°, >5°-10°, >10°-15°, >15°20°, >20°-25°, >25°-30°, >30°-35° and >35°. The values of slope aspect are divided into eight classes according to 45° intervals and one class, which represents flat terrain. The curvature values were classified into three classes: convex, concave and flat. The soil map provided for this investigation has polygons representing 396 soil classes. These classes were merged into 19 classes to simplify the spatial analysis. Further information to create maps of soil depth (5 classes), substrate (17 classes) and soil acidification (13 classes) are included in the table of attributes of the soil map. The land cover map of the CORINE project was used to identify forested areas in Baden-Württemberg. It was the basis to construct a map, which shows the distance of forests to their edge. As an example, Fig. 4 contains a map extract for the region around Sindelfingen. Areas with storm damages within forests areas caused by the hurricane “Lothar” (26 December 1999), which could be derived from the CORINE data base (KEIL et al., 2005), are marked in this map. This information enables to identify the importance of the distance of forest edges with respect to forest vulnerability to strong air flow. Fig. 5 shows the distribution of the forest patches impacted by the hurricane “Lothar” in relation to their minimum distance to the edges of the forests. Two-thirds of all affected patches are located no more than 300 m away from the forest edges. About 45% of the storm damaged areas are directly located at the forest edges. Due to their significance, maps of roads and rivers were also used to generate maps showing their distances to forest stands. The Topex2000 to 2000 m was used to represent the topographic exposure for each grid cell. For a given location, this index is calculated by summing each of the vertical angles to the skyline for the eight main compass directions (QUINE and WHITE, 1998). The used Topex is distance limited to 2000 m. 56 Fig. 4: Map extract of the distance to the forest edge for the region around Sindelfingen, framed by red lines: areas with storm damages caused by the hurricane “Lothar” on 26 December 1999 57 300 number of affected patches 250 200 150 100 50 0 0 50 100 200 300 400 500 1000 1500 2000 2500 3000 minimum distance to forest edge (m) Fig. 5: Histogram of minimum distance to forest edge, columns are all stands affected by the hurricane “Lothar” (26 December 1999), which were mapped in the CORINE Land Cover 2000 project (KEIL et al., 2005) Table 1 summarizes preliminary results based on overlay analyses. They are based on disturbance areas caused by the hurricane “Lothar”. They were taken from the CORINE Land Cover 2000 project. The hurricane hit mainly sites at elevations between 400 and 650 m a.s.l. Generally a continuous increase in the number of disturbance areas was detected in an elevation range from 200 to 550 m a.s.l. About 50% of the windthrow patches affected by “Lothar” were found on gentle slopes with a slope less than 6°. The reason for this could be that soils on low slopes are more moistened in comparison with soils on steep slopes, so that trees are less stable (MAYER et al. 2005). The results also show that especially stands on southern slopes were blown down. Table 1: Summary of the previous data analysis, the percentage of disturbance areas vs. elevation, slope angle, slope aspect and Topex2000 to 2000 classes is represented in brackets elevation slope angle slope aspect topex2000 to 2000 m - 500-550 m a.s.l. (14%) - 2°-4° (24%) - south (28%) - 5-10 (15%) - 450-500 m a.s.l. (11%) - 0°-2° (15%) - southwest (24%) - 10-15 (15%) - 600-650 m a.s.l. (10%) - 4°-6° (12%) - southeast (20%) - 0-5 (12%) - 400-450 m a.s.l. (9%) - 15-20 (11%) - 550-600 m a.s.l. (9%) - 20-25 (7%) - 25-30 (7%) 58 Around two-thirds of the storm affected patches have low Topex scores. After WILSON (1984) a Topex score less than 30 signifies a very high exposure with no or only little local shelter. 77% of the damaged sites were located on soils with low pH value (pH < 3.8 to 4.2). MAYER et al. (2005) also found that the pH value of the soil was one of the most significant factors on storm damaged plots. 4. Conclusions In view of the predicted increase in frequency and intensity of severe storms in the midlatitudes (LECKEBUSCH et al, 2006), wind risk assessment of is very important. The aim of this study is the generation of two hazard maps showing the vulnerability of forests in the Federal State of Baden-Württemberg (SW Germany) to storm damages under current (1971-2000) and future (2021-2050) climate conditions. To produce such hazard maps, the relationship between storm damages and influencing factors must be analyzed. Because of the complex interactions between storm damages in forests and various factors, which are included in the causal chain of storm damages in forests (MAYER and SCHINDLER, 2002), a range of data sets is necessary to meet the objectives. Acknowledgement The authors are indebted to the State Agency for Environment, Measurements and Nature Conservation Baden-Württemberg for funding this study within the joint research project RESTER (strategies to reduce the storm risks of forests in Baden-Württemberg), which is part of the research program “challenge climate change” of the Ministry of the Environment of the Federal State of Baden-Württemberg. References BAKER, W.L., P.H. FLAHERTY, J.D. LINDEMANN, T.T. VEBLEN, K.S. EISENHART, D.W. KULAKOWSKI, 2002: Effect of vegetation on the impact of a severe blowdown in the southern Rocky Mountains. - Forest Ecology and Management 168, 63-75. DOBBERTIN, M., 2002: Influence of stand structure and site factors on wind damage comparing the storms Vivian and Lothar. - Forest, Snow and Landscape Research 77, 187-205. EVANS, A.M., A.E. CAMP, M.L. TYRELL, C.C. RIELY, 2007: Biotic and abiotic influences on the wind disturbance in forests of NW Pennsylvania, USA. - Forest Ecology and Management 245, 44-53. HÜTTE, P., 1967: Die standörtlichen Voraussetzungen der Sturmschäden. - Forstwissenschaftliches Centralblatt 86, 276-295. IPCC, 2007: Working Group I: The physical science basis of climate change. - http://ipccwg1.ucar.edu/wg1/wg1-report.html. KÄNDLER, G., B. BÖSCH, M. SCHMIDT, 2005: Wesentliche Ergebnisse der zweiten Bundeswaldinventur in Baden-Württemberg – Rückblick und Ausblick. - Forstarchiv 60, 45-49. KEIL, M., R. KIEFL, G. STRUNZ, 2005: CORINE Land Cover 2000 – Germany. Final Report. Wessling. LECKEBUSCH, G.C., B. KOFFI, U. ULBRICH, J.G. PINTO, T. SPANGEHL, S. ZACHARIAS, 2006: Analysis of frequency and intensity of winter storm events in Europe on synoptic and regional scales from a multi-model Perspective. - Clim. Res. 31, 59-74. 59 MAYER, H., D. SCHINDLER, 2002: Forstmeteorologische Grundlagen zur Auslösung von Sturmschäden im Wald in Zusammenhang mit dem Orkan „Lothar“. - Allg. Forst- u. J.-Ztg. 173, 200-208. MAYER, P., P. BRANG, M. DOBBERTIN, D. HALLENBARTER, J.-P. RENAUD, L. WALTHERT, S. ZIMMERMANN, 2005: Forest storm damage is more frequent on acidic soils. - Ann. For. Sci. 62, 303-311. MLR, 2008: Wald in Zahlen. Wem gehört der Wald? - http://www.wald-onlinebw.de/rubriken/wald/wald-in-zahlen/wem-gehoert-der-wald/ PINTO, J.G., U. ULBRICH, G.C. LECKEBUSCH, T. SPANGEHL, M. REYERS, S. ZACHARIAS, 2007: Changing European storm loss potentials under modified climate conditions according to ensemble simulations of the ECHAM5/MPI-OM1GCM. - Nat. Hazards Earth Syst. Sci. 7, 165175. QUINE, C.P., I.M.S. WHITE, 1998: The potential of distance-limited topex in the prediction of site windiness. - Forestry 71, 325-332 SCHMOECKEL, J., 2006: Orographischer Einfluss auf die Strömung abgeleitet aus den Sturmschäden im Schwarzwald während des Orkans „Lothar“. - Dissertation, Universität Karlsruhe, 137 S. WILSON, J. D., 1984: Determining a topex score. - Scottish Forestry 38, 251-256. Authors' address: Dipl.-Geogr. Karin Grebhan ([email protected]) Dr. Dirk Schindler (dirk.schindler@meteo,uni-freiburg.de) Prof. Dr. Helmut Mayer ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 60 61 Analyses of water balance components of beech stands in south-western Germany using BROOK90 Jutta Holst1, Thomas Holst1,2 and Helmut Mayer1 1 2 Meteorological Institute, Albert-Ludwigs University of Freiburg, Germany Department of Physical Geography and Ecosystems Analysis, Lund University, Sweden Abstract In 1999, forest research sites were established in 80 yrs old beech stands at the Swabian Jura mountain range in south-western Germany. These research sites are located on the opposite slopes of a narrow valley near Tuttlingen, which causes a gradient of climate due to the different orientation of the slopes. Additionally, the microclimatic conditions below the canopy of the beech stands were modified by establishing research plots with different canopy densities on each of the slopes. To analyse the combined effects of the climate gradient and the different canopy densities, the water balances of the beech stands at the research plots were investigated using long-term forest-meteorological measurements and simulation results with the water balance model BROOK90. In this study, results for interception, soil moisture, soil water content, total plant available soil water and transpiration are discussed. Analysen der Wasserhaushaltskomponenten von Buchenbeständen in Südwestdeutschland unter Verwendung von BROOK90 Zusammenfassung 1999 wurden forstliche Untersuchungsflächen in einem 80-jährigen Buchenbestand auf der Schwäbischen Alb in SW Deutschland in Betrieb genommen. Die Untersuchungsflächen liegen auf den gegenüberliegenden Hängen eines engen Tals nahe Tuttlingen. Durch die entgegengesetzte Orientierung der Hänge wird ein unterschiedliches Mikroklima verursacht. Zusätzlich wird das Mikroklima unterhalb des Kronenschirms modifiziert, indem Flächen unterschiedlicher Bestandesdichte auf beiden Hängen etabliert wurden. Um die kombinierten Effekte des Klimagradienten und der unterschiedlichen Bestandesdichte zu analysieren, wurde der Wasserhaushalt der Buchenbestände auf den Untersuchungsflächen auf Basis der langjährigen forstmeteorologischen Messungen mit Hilfe des Wasserhaushaltmodells BROOK 90 untersucht. In diesem Artikel werden Ergebnisse zu Interzeption, Bodenfeuchte, Bodenwassergehalt, pflanzenverfügbarem Bodenwasser sowie Transpiration diskutiert. 1. Introduction Beech forests (Fagus sylvatica L.) are of great importance for Central Europe as beech species belong to the dominating, natural vegetation (ELLENBERG, 1996). Prior to anthropogenic impacts 75% of Central Europe were covered with beech. Nevertheless, the knowledge about ecological and eco-physiological processes within a beech ecosystem is still limited (ADAMS and GRIERSON, 2001). Differences between the expected and the actual growth rates indicate long-term changes of the ecosystem (SPIECKER et al., 1996). 62 Results from regional climate models predict more frequent and longer lasting dry periods in Central European summers with reduced precipitation and higher air temperature for the next decades (GERSTENGARBE et al., 2003; IPCC, 2007). These relatively fast changes in climate exceed the adaptation time of beech forests. Beech forests are adapted to a variety of climate conditions. But their competitiveness is reduced on very wet or very dry soils (RENNENBERG et al., 2004; AMMER et al., 2005). Changes in regional and local climate conditions have an impact on the productivity of beech forests. Therefore, climate change poses a challenge for forest management. In this context methods of forest management can be applied to mitigate negative consequences of climate change on beech forests. In 1999, an extensive joint research project was established in SW Germany to analyse the different growing conditions of beech ecosystems under changing climate conditions and silvicultural management strategies. The experimental design is characterized by a set of three different stand densities on each of two opposite slopes (NE and SW) of a narrow valley (HOLST, 2004). The investigation of the water balance of the beech stands presented here is part of extensive forest meteorological investigations within the joint interdisciplinary research project (e.g. GESSLER et al., 2001; FOTELLI et al., 2003; HAUSER, 2003; DANNENMANN et al., 2007). 2. Methods The research sites are located in south-western Germany near Tuttlingen (47° 59’ N, 8° 45’ E, about 800 m above sea level) on the two opposite slopes of a narrow valley. One slope is facing towards SW direction (defined as SW-slope), while the aspect of the second slope is NE (NE-slope). Both slopes have an inclination of 23° to 30° and the horizontal distance between both sites is about 800 m. Both hillsides are covered with 70 to 80-year-old single-layer, beech-dominated (> 90%, Fagus sylvatica L.) forest stands. The mean stand height hg varied between 21.1 m and 31.9 m (Table 1). The location of the beech stands with similar site and soil conditions on the opposite NE and SW slopes of the valley enables the simulation of a climate gradient (HOLST et al., 2005). As the objectives of the joint research project include the analysis of ecological effects of different stand densities, shelterwood fellings were carried out in research plots on both slopes (HOLST, 2004). Table 1: Stand characteristics of the beech stands at the research sites near Tuttlingen (SW Germany; HOLST and MAYER, 2005). BA: basal area, hg: mean tree height, dbh: breast height diameter, PAI: plant area index (2003), NE: slope with NE aspect, SW: slope with SW aspect control plot BA trees/ha hg dbh PAI NE 27 m² ha-1 526 26.5 m 25.6 cm 5.8 m² m-2 SW 21 m² ha-1 576 21.1 m 21.5 cm 5.8 m² m-2 weak shelterwood felling strong shelterwood felling NE SW NE SW -1 -1 -1 14.7 m² ha 15 m² ha 11.4 m² ha 10.4 m² ha-1 254 312 123 199 27.1 m 23.5 m 31.9 m 22.8 m 27.1 cm 24.7 cm 34.4 cm 25.8 cm 4.0 m² m-2 4.0 m² m-2 2.7 m² m-2 2.5 m² m-2 63 With respect to the water balance components, gross precipitation is continuously recorded by the use of tipping buckets (ARG100, Vaisala, Helsinki, FIN) on the top of two forest-meteorological research towers (1.5 hg) installed at the control plots on both slopes. At each of the differently managed research plots, throughfall is measured by the use of a combination of standardized precipitation gutters with tipping buckets at their end. Soil moisture is continuously determined using probes according to the time domain reflectometry method (TDR, CS615, Campbell Sci., Shepshed, GB) also at each of the research plots. The TDR soil moisture recordings were compared against gravimetrically taken soil moisture measurements to allow for a soil-specific calibration. To obtain a more extensive information on the effects of aspect and canopy density on all components of the water balance of the beech stands (e.g. evapotranspiration, transpiration, available soil water, total soil water) simulations were performed using the forest-hydrological water balance model BROOK90 (HAMMEL and KENNEL, 2001; FEDERER et al., 2003; WELLPOTT et al., 2005). 3. Results 3.1 Interception The density of the forest canopy, which is quantified in this project by the plant area index PAI, influences the water fluxes into the soil through the interception I. In this investigation, I was calculated as the ratio of throughfall and gross precipitation. Since stemflow was measured only discontinuously, it was not considered here. Due to the specific site conditions (e.g. complex terrain, uncentered crowns of the beech trees), the stemflow of the beech stands at the research sites was considerably lower than comparative values given in the literature for beech stands in a relatively flat terrain (GEIGER et al., 1995; PECK, 2004). The exemplary results for I of the beech stands on the NE slope (Table 2) indicate that I at the control plot (NE-C) is about 20% higher than mean I measured at the strong shelterwood felling plot (NE-2) in the growing season. Table 2: Monthly means of interception I at the control plot (-C) and strong shelterwood felling plot (-2) on the NE exposed slope near Tuttlingen in the growing seasons 2001 to 2006 NE-C NE-2 Apr 25% 23% May 42% 25% Jun 39% 28% Jul 42% 26% Aug 37% 24% Sep 41% 26% Oct 35% 19% 3.2 Soil moisture Soil moisture determines the hydrological state of the soil which results from infiltrated water, soil properties, and interactions between soil and plant physiological processes. Due to the higher irradiation on the SW oriented slope, the reduction of soil moisture during the summer months was higher on the SW orientated slope than on the NE orientated slope in the period 1999 to 2006 (Table 3). In addition the reduction was higher at the control plots (-C) compared to the plots with reduced stand density (-2). The reason can be found in the larger amount of biomass which caused a higher transpiration at stand scale at the control plots compared to the shelterwood plots. 64 Table 3: Monthly means of normalized volumetric soil moisture content SWC/SWCmax at the control plots (-C) and strong shelterwood felling plots (-2) on the two opposite slopes (NE, SW aspect) of a narrow valley near Tuttlingen; period: Nov 1999 to Dec 2006 NE-C NE-2 SW-C SW-2 Jan 0.95 0.88 0.94 0.96 Feb 0.96 0.89 0.96 0.97 Mar 0.98 0.91 0.99 0.99 Apr 1.00 1.00 1.00 1.00 May 0.96 0.93 0.89 0.95 Jun 0.78 0.78 0.67 0.72 Jul 0.69 0.71 0.54 0.64 Aug 0.73 0.74 0.56 0.66 Sep 0.73 0.74 0.55 0.63 Oct 0.86 0.83 0.68 0.76 Nov 0.91 0.88 0.81 0.87 Dec 0.94 0.89 0.92 0.95 3.3 Model results The general course of the simulated soil water in the upper most 20 cm (SW) corresponds with the measurements in the period 2001 to 2006 (Fig. 1). During winter, the simulated SW is generally lower than the measured SW. Possible reasons are effects of freezing or melting snow. However, as this study is focussed on the effects of exposition and stand density on the microclimate and growing conditions for beech ecosystems, it concentrates on the growing season. Especially during this period, the model results agree reasonably well with the measurements. Thus, it is assumed, that BROOK90 is suitable to simulate the water balance of the forest stands at the research sites near Tuttlingen. This allows the analysis of further parameters of the water budget, which were not measured. 140 120 80 60 40 2001 2002 2003 2004 time (doy) 2005 335 245 65 155 340 250 160 70 346 256 166 76 351 261 171 81 356 266 176 86 361 1 91 0 271 SW(measurements) SW(BROOK90) AW(BROOK90) 20 181 SW 20cm, AW (mm) 100 2006 Fig. 1: Daily means of simulated (BROOK90) and measured soil water (SW) in the upper 0.2 m of the soil as well as total plant available soil water content (AW) in layers with roots at the control plot NE-C near Tuttlingen in the period 2001 to 2006 65 The total plant available water in layers with roots (AW) shows a large annual course. During winter, it reaches values up to 130 mm and during summer, it decreases to values between 40 mm and 60 mm. In summer 2003, which was characterized by extremely hot and dry conditions during summer in whole Central Europe (SCHÖNWIESE et al., 2004; REBETEZ et al., 2006), available soil water decreased to nearly 0 mm. This strong reduction can also be seen in less intense and shorter dry period in July 2006, indicating that plant available water is very limited even during less severe situations. 40 1500 Ta P 2001 2002 2003 2004 2005 335 245 155 65 340 250 160 70 346 256 76 166 351 261 0 81 -20 171 250 356 -10 266 500 86 0 176 750 361 10 271 1000 181 20 1 1250 91 30 ΣP (mm) G (MJ/m²d), T a (°C) G 2006 time (doy) Fig. 2: Daily totals of incident radiation G, daily mean air temperature Ta and cumulated precipitation P above the canopy on the NE slope near Tuttlingen in the period 2001 to 2006 One of the main steering factors for transpiration is the available energy AE. As an indicator for AE, Fig. 2 shows daily totals of incident solar radiation G on the NE oriented slope. Furthermore, daily means of air temperature Ta and cumulated totals of precipitation P above the canopy are shown. While the years 2004 to 2006 were characterised by average annual totals of precipitation (approx. 800 mm), the years 2001 and 2002 showed distinctly higher annual precipitation totals. The precipitation pattern in 2006 was distributed irregularly with high P in spring and late summer and rather low P during June and July. During 2002, air temperature Ta was rather low throughout the summer. On contrary, 2003 was characterised by extremely low precipitation totals (610 mm), high air temperature and high incident radiation. Besides the available energy, the actual transpiration T depends considerably on the available soil water. Therefore, T was reduced in summer 2003 (Fig. 3). During years with sufficient water availability, the ratio of the actual transpiration at the shelterwood plot (NE-2) to that at the control plot (NE-C) was constant throughout the vegetation period. In periods with limited plant available water, the ration increased due to a decrease of T at the control plot. Due to the higher PAI and the therewith larger area of 66 transpiring canopy, this decrease of T was stronger at the control plot than at the shelterwood plot. 120 NE-C NE-2 NE-2/NE-C 2001 2002 2003 2004 2005 335 245 155 65 340 250 70 160 346 256 76 166 351 261 0 81 0 171 20 356 50 266 40 86 100 176 60 361 150 271 80 181 200 1 100 91 250 ΣTNE-2/ΣTNE-C (%) ΣT (mm) 300 2006 time (doy) Fig. 3: Simulated cumulated transpiration ΣT of two beech stands with different canopy density (control plot -C; strong shelterwood felling -2) on the NE slope near Tuttlingen in the period 2001 to 2006 4. Conclusions The results show how forest management influences the water balance components of beech stands (e.g. interception, transpiration, soil moisture). With respect to regional climate change in Central Europe, the simulation of the water balance of the beech stands enables an estimation, how the drought risk for beech stands, which will be more frequent in the future, can be reduced by silvicultural management leading e.g. to lower canopy densities. Acknowledgement Financial support was provided by the grants SFB433 - TP1, MA 749/17-1, MA 749/17-2 and DFG FG788 – MA 749/21-1 of the German Research Foundation DFG (Deutsche Forschungsgemeinschaft). Thanks go to D. Redepenning and G. Fernbach for their field assistance. References ADAMS, M.A., P.F. GRIERSON, 2001: Stable isotopes at natural abundance in terrestrial plant ecology and ecophysiology: an update. – Plant Biol. 3, 299-310. AMMER, CH., L. ALBRECHT, H. BORCHERT, F. BROSINGER, C. DITTMAR, W. ELLING, J. EWALD, B. FELBERMEIER, H. VON GILSA, J. HUSS, G. KENK, C. KÖLLING, U. KOHNLE, P. MEYER, R. MOSANDL, H.-U. MOOSMAYER, S. PALMER, A. REIF, K.-E. REHFUESS, B. STIMM, 2005: Zur 67 Zukunft der Buche (Fagus sylvatica L.) in Mitteleuropa - kritische Anmerkungen zu einem Beitrag von Rennenberg et al. (2004). – Allg. Forst- u. J.-Ztg. 176, 60-67. DANNENMANN, M., R. GASCHE, A. LEDEBUHR, T. HOLST, H. MAYER, H. PAPEN, 2007: The effect of forest management on trace gas exchange at the pedosphere-atmosphere interface in beech (Fagus sylvatica) forests stocking on calcareous soils. – Eur. J. For. Res. 126, 331346. ELLENBERG, H., 1996: Vegetation Mitteleuropas mit den Alpen. – Ulmer Verlag, Stuttgart (in German). FEDERER, C.A., C. VÖRÖSMARTY, B. FEKETE, 2003: Sensitivity of annual evaporation to soil and root properties in two models of contrasting complexity. – J. Hydrometeorol. 4, 1276-1290. FOTELLI, M.N., H. RENNENBERG, T. HOLST, H. MAYER, A. GESSLER, 2003: Carbon isotope composition of various tissues of beech (Fagus sylvatica) regeneration is indicative of recent environmental conditions within a forest understorey. – New Phytol. 159, 229244. GEIGER, R., R.H. AARON, P. TODHUNTER, 1995: The climate near the ground. – Vieweg, Braunschweig. GERSTENGARBE, F.-W., F. BADECK, F. HATTERMANN, V. KRYSNOVA, W. LAHMER, P. LASCH, M. STOCK, F. SUCKOW, F. WECHSUNG, P. WERNER, 2003: Studie zur klimatischen Entwicklung im Land Brandenburg bis 2055 und deren Auswirkungen auf den Wasserhaushalt, die Forst- und Landwirtschaft sowie die Ableitung erster Perspektiven. – PIK Report No. 83. Potsdam-Institut für Klimafolgenforschung, Potsdam (in German). GESSLER, A., S. SCHREMPP, A. MATZARAKIS, H. MAYER, H. RENNENBERG, M.A. ADAMS, 2001: Radiation modifies the effect of water availability on the carbon isotope composition of bech (Fagus sylvatica L.). – New Phytol. 150, 653-664. HAMMEL, K., M. KENNEL, 2001: Characterisation and analysis of the water availability and the water balance of forest sites in Bavaria using the simulation model BROOK90. – Forstliche Forschungsberichte München No. 185, 148 pp (in German). HAUSER, S., 2003: Dynamik hochaufgelöster radialer Schaftveränderungen und des Dickenwachstums bei Buchen (Fagus sylvatica L.) der Schwäbischen Alb unter dem Einfluß von Witterung und Bewirtschaftung. – PhD-thesis, Faculty of Forest and Environmental Sciences, Albert-Ludwigs University of Freiburg. HOLST, T., 2004: Influence of exposition and stand density on the microclimate in beech forests. – PhD-thesis, Faculty of Forest and Environmental Sciences, Albert-LudwigsUniversity of Freiburg (in German). HOLST, T., J. ROST, H. MAYER, 2005: Net radiation balance for two forested slopes on opposite sides of a valley. – Int. J. Biometeorol. 49, 275-284. IPCC (Intergovernmental Panel on Climate Change), 2007: Climate Change 2007: The physical science basis. Contribution of Working Group I to the fourth Assessment Report of the Intergovernmental Panel on Climate Change. Summary for Policymakers. – IPCC secretariat, c/o WMO, Switzerland. (http://www.ipcc.ch, 12.03.2007). PECK, A.K., 2004: Hydrometeorologische und mikroklimatische Kennzeichen von Buchenwäldern. – Ber. Meteor. Inst. Univ. Freiburg Nr. 10. REBETEZ, M., H. MAYER, D. SCHINDLER, O. DUPONT, K. GARTNER, J. KROPP, A. MENZEL, 2006: Caractéristiques climatiques de l'été 2003. – RDV techniques 11, 14-18. 68 RENNENBERG, H., W. SEILER, R. MATYSSEK, A. GEßLER, J. KREUZWIESER, 2004: Die Buche (Fagus sylvatica L.) - ein Waldbaum ohne Zukunft im südlichen Mitteleuropa? – Allg. Forstu. J.-Ztg. 175, 210-244. SCHÖNWIESE, C.-D., S. STAEGER, S. TRÖMEL, 2004: The hot summer 2003 in Germany. Some preliminary results of a statistical time series analysis. – Meteorol. Z. 13, 323-327. SPIECKER, H., K. MIELIKAINEN, M. KÖHL, J.P. SKOVSGAARD, [eds.], 1996: Growth trends in European forests. – European Forest Research Report No. 5, Springer Verlag, Berlin. WELLPOTT, A., F. IMBERY, D. SCHINDLER, H. MAYER, 2005: Simulation of drought for a Scots pine forest (Pinus sylvestris L.) in the southern upper Rhine plain. – Meteorol. Z. 14, 143-150. Authors' address: Dr. Jutta Holst ([email protected]), Prof. Dr. Helmut Mayer ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany Dr. Thomas Holst ([email protected]) Department of Physical Geography and Ecosystems Analysis, Lund University Sölvegatan 12, SE-22362, Lund, Sweden 69 Ecosystem flux measurements of BVOC at a sub-arctic wetland site Thomas Holst, Almut Arneth, Sean Hayward and Anna Ekberg Department of Physical Geography and Ecosystems Analysis, Lund University, Sweden Abstract In this contribution, measurements of biogenic volatile organic compounds (BVOC) are presented, which have been conducted at a sub-arctic wetland in northern Sweden, about 250 km north of the Arctic Circle, by use of a disjunct eddy-covariance (DEC) system. The DEC system consisted of a Proton Transfer Reaction-Mass Spectrometer (PTR-MS) and a sonic anemometer and was used to observe the ecosystem exchange of a set of BVOCs above a wetland site with vegetation dominated by Eriophorum and Carex species over a period of several weeks. Here, flux measurements of isoprene and methanol for a 15 day-period during summer are discussed and compared to air temperature (Ta) and photosynthetically active radiation (PAR). As expected, the emissions revealed a daily cycle dependent on Ta and PAR, but isoprene fluxes were significantly higher compared to emissions parameterised by the standard Guenther algorithms at high Ta. Data on BVOC fluxes measured on ecosystem scale are still rare especially for regions at high latitudes. Therefore, the results presented will be valuable for the understanding of the processes involved and for the development and validation of biogeochemical models. Ökosystem-Fluss-Messungen von BVOC in einem sub-arktischen Moor Zusammenfassung In diesem Beitrag werden Messungen von flüchtigen Kohlenstoffverbindungen (biogenic volatile organic compounds, BVOC) vorgestellt, die in einem sub-arktischem Moorgebiet in NordSchweden, gelegen etwa 250 km nördlich des Polarkreises, durchgeführt worden sind. Für die Messung der turbulenten Flüsse der BVOC wurde eine Variante der Eddy-Kovarianz-Methode (disjunct eddy-covariance, DEC) verwendet, die auf schnellen Konzentrationsmessungen mit Hilfe eines Massenspektrometers (Proton Transfer Reaction-Mass Spectrometer, PTR-MS) und der Erfassung der Windfluktuationen durch ein Ultraschall-Anemometer beruht. Die Vegetation der Untersuchungsfläche war von Eriophorum and Carex-Arten dominiert und der turbulente Austausch einiger VOC wurde über einen Zeitraum von mehreren Wochen erfasst. Dieser Beitrag konzentriert sich auf die Messungen des Isopren- und Methanolaustausches über einen Zeitraum von 15 Tagen im Sommer (Anfang August) und den Vergleich mit Lufttemperatur (Ta) und photosynthetisch aktiver Strahlung (PAR). Erwartungsgemäß waren die Emissionen von Isopren und Methanol eng mit Ta und PAR verknüpft und zeigten einen teils deutlichen Tagesgang. Allerdings wurde für Isopren ein deutlich stärkerer Anstieg der Emissionen bei höher Ta beobachtet, als mit dem weit verbreitetem Guenther-Algorithmus vorhergesagt. Messungen auf Ökosystem-Ebene zur Emission von BVOC sind vergleichsweise selten. Dies gilt besonders für Ökosysteme in den hohen Breiten. Daher liefern die vorgestellten Messungen zum turbulenten Austausch von BVOC auf Ökosystem-Ebene einen wichtigen Beitrag zum Verständnis der biochemischen Prozesse. Die Resultate können auch für die Entwicklung und Validierung von biogeochemischen Modellen verwendet werden. 70 1. Introduction On a global scale, terrestrial ecosystems add c. 1000 Tg C y-1 of biogenic volatile organic compounds (BVOC) to the atmosphere, with isoprene contributing to roughly about half of the emissions (GUENTHER et al., 1995; KESSELMEIER and STAUDT, 1999). BVOCs have a high impact on atmospheric chemistry, since they react with OH on very short time scales and influence e.g. levels of atmospheric methane and ozone (e.g., DERWENT et al., 1995; ATKINSON, 2000; DI CARLO et al., 2004). Furthermore, VOCs from terrestrial ecosystems are known to participate in the production and growth of secondary organic aerosols, which alter the radiation balance of the atmosphere and may act as cloud condensation nuclei (e.g., HOFFMANN et al., 1997; KULMALA et al., 2000; KULMALA et al., 2004). In contrast with the high importance of BVOCs for atmospheric chemistry, many of the physiological and biochemical processes leading to the emission of BVOCs remain poorly understood, and some atmospheric processes which involve BVOC reactions on short and long time scales are still unknown (DI CARLO et al., 2004; NIINEMETS et al., 2004). BVOC emission data that can be used to develop and validate biogeochemical models has been measured for a wide range of different plant species, but was conducted mostly with chamber and cuvette techniques and often restricted to lab studies or to short campaigns. During the last few years, the number of studies conducted on the ecosystem scale has increased (KARL et al., 2001; RINNE et al., 2002; GRABMER et al., 2004; SPIRIG et al., 2005), but still is mostly limited to periods of days or weeks. For ecosystems in the higher latitudes, only a very few studies have been published (RINNE et al., 2000; HAAPANALA et al., 2006) despite of the fact that ecosystems in boreal and sub-arctic regions will be strongly affected by climate change. 2. Methods 2.1. Field site The field site for the measurement of ecosystem-scale fluxes of different biogenic VOCs was located in northern sub-arctic Sweden, 250 km north of the Arctic Circle. Stordalen mire (68° 20’ N, 19° 03’ E, 351 m asl) is situated about 10 km east of Abisko and Abisko Naturvetenskabliga Station (Abisko Scientific Research Station, ANS) where long-term measurements of climate and environmental conditions have been conducted and a long record of various environmental research for this region has been published (e.g., SVENSSON et al. 1999; JOHANSSON et al., 2006; STRÖM and CHRISTENSEN, 2007). At this high-latitude region, the growing season usually starts at the end of May, while early frost and snow might occur as early as in late August or in the beginning of September. The annual mean air temperature (1913 to 2003) at the ANS was – 0.7°C, and the mean air temperature during July was 11°C for the same period. Annual precipitation is low (300 mm), which reflects the influence of the mountain ranges the area is located in. The region is characterized by discontinuous permafrost conditions. Large parts of the soil at Stordalen mire is permanently frozen, and a slightly elevated, dry active layer of a few dm depth of melted organic soil evolves throughout the arctic summer at the surface layer, before starting to freeze again in autumn. Nevertheless, in some parts of the 71 mire, permafrost is lacking and fen-like parts developed, which consist of less elevated semi-wet and wet patches. According to their growing conditions, the dry parts, palsas and the semi-wet areas are mostly dominated by Eriophorum vaginatum, Carex rotrundata and Sphagnum ssp., while in the wet parts, the vegetation is dominated by Eriophorum angustifolium and Carex rostrata. 2.2. Instrumentation For the determination of the emission and ecosystem exchange of different biogenic VOCs, a measuring system consisting of a sonic anemometer (Metek, Germany) and a Proton Transfer Reaction-Mass Spectrometer (PTR-MS; Ionicon, Austria) has been used. The sonic anemometer was mounted on a small mast at a height of 2.95 m agl in the centre of the wetland site. An inlet was attached right beneath the sonic anemometer head, and a Teflon inlet line (8 mm i.d.) was connecting the mast with a small hut located about 12 m NE from the mast. A diaphragm pump was continuously pulling 20 l min-1 through the main inlet line, and a sub-sample of 0.3 l min-1 of the sample air was taken from this flow and analysed by the PTR-MS located in the hut. To avoid condensation, the main inlet line was heated. The PTR-MS is an instrument allowing the measurement of a variety of VOCs at a very high sampling rate and with high precision (e.g., LINDINGER et al., 1998; DE GOUW and WARNECKE, 2007). As it is sensible only to compounds with a higher proton affinity than water, compounds with high mixing ratios in ambient air like CO2, N2 or O2 cannot be detected, and the instrument is very sensitive to VOC even at low concentrations in the pptv-range. Within the drift tube of the PTR-MS, VOCs are protonated by H3O+ions (‘soft ionisation’-technique), and are subsequently detected by a quadrupole mass spectrometer. With this technique, the sample gas needs no preparation or separation into different constituents before analysis, and different compounds are detected sequentially. In this study, the PTR-MS was used to monitor a set of BVOC, including methanol (protonated mass 33, CH3OH + H+) and isoprene (mass 69, C5H8 + H+). Due to the low ambient concentration at the sub-arctic wetland site, each of the compounds of the scanning sequence was sampled for 0.5 s, and the complete sequence took about 3 s. Instrument background to calculate the uncertainty of concentrations and fluxes obtained was checked regularly by use of a catalytic converter (Zero Air Generator, Parker Balston, USA) which provided VOC-free air. 2.3. Flux determination To determine the turbulent exchange of BVOCs between the atmosphere and the vegetation at the sub-arctic mire, the high-frequency (20 Hz) measurements of the 3D wind vector obtained by the sonic anemometer and the fast measurements of BVOC concentrations taken with the PTR-MS were combined by application of the disjunct eddycovariance technique (DEC) which has been used for BVOC flux measurements with a PTR-MS before (RINNE et al., 2001; KARL et al., 2002; GRABMER et al., 2004). In contrast to standard eddy-covariance (EC) techniques, where wind vector and concentration measurements are sampled as synchronized high-frequency data with the same temporal resolution (e.g., AUBINET et al., 2000), the data set obtained from the PTR-MS was i) at a lower temporal resolution compared with the sonic data due to the increased integration time of 0.5 s for each of the BVOC, and ii) it was not continuous (but ‘disjunct’), as 72 the PTR-MS provided data samples for a specific compound about every 3 s due to the sequential sampling. With DEC-techniques, the turbulent exchange is calculated from the covariance of the deviations of vertical wind velocity (w’) and concentration (c’) from their mean values only for N disjunct data samples within the averaging period (30 min in this case), determined by the available measurements of the PTR-MS (RINNE et al., 2001; KARL et al., 2002; GRABMER et al., 2004). With one sequential sampling of the PTR-MS taking ~3 s, the number of samples of w’ and c’ to calculate the turbulent flux was reduced to a fraction of ~16 % of a continuously recorded data set. It has been shown, that the turbulent exchange calculated using a reduced, disjunct data set has a negligible influence on the result as long as the data set is not reduced too much (KARL et al, 2002; GRABMER et al., 2004). In the experimental set up applied in this study, a significant error based on the disjunct sampling would have occurred with a sub-sample of only 10 % and less (HOLST et al, 2008). The temporal resolution of the PTR-MS in this study was only 2 Hz due to a comparatively long period needed for integration caused by the low ambient concentrations. Resulting from the 2 Hz temporal resolution, the turbulent energy contained in the highfrequency part of the spectrum was missed and thus the turbulent exchange underestimated. Nevertheless, as the major part of the energy is transported in the lower frequency domain (e.g., KAIMAL and FINNIGAN, 1994; FOKEN, 2003), most of the turbulent flux was still captured. An estimate of the flux loss caused by the low temporal resolution based on spectral analyses and a comparison of sensible heat flux H estimates resulted in an underestimation of 10% of the flux (HOLST et al., 2007; HOLST et al., 2008), which has been found for comparable flux measurements with a PTR-MS system before (LEE et al., 2005; SPIRIG et al., 2005). 3. Results and discussion Flux measurements of BVOC have been conducted during a period approximately covering half of the growing season at this sub-arctic field site. Exemplarily, results from the beginning of August (01-15 August 2006) will be presented, which can be regarded as representative for the mature vegetation. In Fig. 1 measurements of air temperature (Ta) and photosynthetically active radiation (PAR) are shown. While at this high latitude site daily maximum Ta was frequently exceeding 17°C during the first half of August and maximum Ta reached 23.2 °C (01 August 2006), low Ta was observed during nights on cloudless days with a minimum Ta of 1.4°C recorded for this period on 10 August. High values for incoming shortwave radiation (e.g., PAR) were observed at the field site in Stordalen during the first half of August except for a short 3-day period (06 to 08 August). As the emission of isoprene is known to be highly correlated to leaf temperature or Ta and PAR (GUENTHER et al., 1993; GUENTHER et al., 1995), high emissions of isoprene were expected during warm and sunny days. In Fig. 2, measurements of the turbulent exchange of isoprene by use of a DEC system including a PTR-MS are presented. Gaps in the time series were caused mostly by data rejected during flux quality analyses (friction velocity threshold of 0.25 m s-1, despiking) or data recorded while testing instrument performance (background noise measurements). 73 2500 25 Ta 2250 PAR 2000 20 Ta (°C) 1500 15 1250 1000 10 750 PAR (µmol m-2 s-1) 1750 500 5 250 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 day (August 2006) Fig. 1: Air temperature (Ta) and photosynthetically active radiation (PAR) measured at the Stordalen field site in sub-arctic Sweden during the first half of August 2006 250 -2 -1 Fi sopr ene (µg C m h ) 200 150 100 50 0 -50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 day (August 2006) Fig. 2: Turbulent fluxes of isoprene (C5H8, Fisoprene) measured with a disjunct eddycovariance system above a sub-arctic wetland in northern Sweden during the first half of August 2006 An estimation of flux determination error lead to uncertainties of 6 µg C m-2 h-1 (based on 2 times standard deviation (σ) of the noise of the flux covariance). Isoprene fluxes up to 224 µg C m-2 h-1 have been observed during daytime at high PAR-level, but emis- 74 sions were low under cloud cover (06 to 08 August) and fluxes were virtually zero during nights. This reflects that isoprene is only produced in plants dependent on the availability of light and not stored in the leaves (e.g., KESSELMEIER and STAUDT, 1999). The turbulent fluxes measured for methanol are shown in Fig. 3. Compared to isoprene, patterns obtained for methanol were more variable, caused by higher background noise for methanol leading to a flux determination error of 13 µg C m-2 h-1. Nevertheless, as for isoprene, daily patterns were obtained during the warm and sunny days at the beginning of August. Methanol emission is known to be related to stomatal conductance (NEMECEK-MARSHALL et al., 1995; HÜVE et al., 2007), but methanol has also been proven to be stored within the leaves and to be emitted during growth and decay (NEMECEK-MARSHALL et al., 1995; HÜVE et al., 2007; SECO et al., 2007) as well. This might explain the unclear behaviour of the measured fluxes in the latter part of the period shown. 100 -2 -1 Fmethanol (µg C m h ) 75 50 25 0 -25 -50 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 day (August 2006) Fig. 3: Turbulent fluxes of methanol (CH3OH, Fmethanol) measured with a disjunct eddy-covariance system above a sub-arctic wetland in northern Sweden during the first half of August 2006 To analyse the dependence of the isoprene emission on the meteorological conditions, the measured fluxes were compared to Ta and PAR (Figs. 4, 5). As expected from the Guenther algorithms (GUENTHER et al., 1993; GUENTHER et al., 1995), a strong nonlinear relationship could be observed both with Ta and PAR. However, if measured Ta and PAR were used to model the isoprene emission based on the standard Guenther algorithm (GUENTHER et al., 1993; GUENTHER et al., 1995) with an emission capacity of 224 µg m-2 h-1 (equivalent to 198 µg C m-2 h-1) as reported by HELLÉN et al. (2006) from chamber studies at a boreal wetland in south-west Finland (Fig. 5), the measured fluxes agreed well up to ~ 13°C, but showed a steeper increase with Ta than the algorithm at higher Ta. 75 250 -2 -1 Fisopr ene (µg C m h ) 200 150 100 50 0 -50 0 250 500 750 1000 -2 1250 1500 -1 PAR (µmol m s ) Fig. 4: Turbulent fluxes of isoprene (Fisoprene) measured with a disjunct eddycovariance system above a sub-arctic wetland in Sweden compared to photosynthetically active radiation (PAR) measured at the site during 01-15 August 2006 250 Guenther algorithm measurements -2 -1 Fisoprene (µg C m h ) 200 150 100 50 0 -50 0 5 10 15 20 25 Ta (°C) Fig. 5: Turbulent fluxes of isoprene (Fisoprene) measured with a disjunct eddycovariance system above a sub-arctic wetland in Sweden compared to air temperature (Ta) and isoprene emission modelled using the standard Guenther algorithm with a emission capacity of 224 µg m-2 h-1 and based on Ta and PAR data (period: 01-15 August 2006) 76 With higher values for the emission capacity (i.e., 600 µg C m-2 h-1 as found by HAAPANALA et al., 2006), measured emission was clearly overestimated at lower Ta, but still slightly underestimated at higher Ta. This steep increase of isoprene emission with Ta exceeding the estimates of the Guenther algorithm was also found by EKBERG et al. (2008) with chamber studies conducted at Stordalen. 4. Summary and conclusion In this contribution, measurements of the turbulent exchange of isoprene and methanol above a sub-arctic wetland in northern Sweden have been presented. The data was obtained with a PTR-MS system using DEC techniques, and emissions were comparable to results published from other boreal ecosystems, which were mostly based on chamber and leaf-scale measurements. However, a much steeper increase of the isoprene emission with air temperature than expected from the Guenther algorithm was found. The results point out that more data on the BVOC emission characteristics is needed to fully understand and model regional and global emission of BVOC, especially for ecosystems in sub-arctic environments which will be most affected by climate change. Furthermore, BVOCs have a strong impact on atmospheric chemistry, oxidation products and have been shown to foster the production of secondary organic aerosol. Therefore, long-term measurements on the ecosystem scale of BVOC exchange are important to include different seasonal behaviour like plant growth and decay, but until now, only a few studies have been published and almost no data exists for the high northern latitudes. Acknowledgements The authors would like to thank T. Friborg (University of Copenhagen), P.M. Crill (Stockholm University) and M. Jackowicz-Korczynski (Lund University) for meteorological data from Stordalen as well as the staff of Abisko Naturvetenskabliga Station (ANS) for their help. References AUBINET, M., A. GRELLE, A. IBROM, Ü. RANNIK, J. MONCRIEFF, T. FOKEN, A.S. KOWALSKI, P.H. MARTIN, P. BERBIGIER, C. BERNHOFER, R. CLEMENT, J. ELBERS, A. GRANIER, T. GRÜNWALD, K. MORGENSTERN, K. PILEGAARD, C. REBMANN, W. SNIJDERS, R. VALENTINI, T. VESALA, 2000: Estimates of the annual net carbon and water exchange of forests: the EUROFLUX methodology. – Advances in Ecological Research 30, 113-175. ATKINSON, R., 2000; Atmospheric chemistry of VOCs and Nox. – Atmospheric Environment 34, 2063-2101. DE GOUW, J., C. WARNEKE, 2007: Measurements of Volatile Organic Compounds in the Earth’s atmosphere using Proton-Transfer-Reaction Mass Spectrometry. – Mass Spectrometry Reviews 26, 223-257. DERWENT, R.G, D.R. MIDDLETON, R.A. FIELD, M.E. GOLDSTONE, J.N. LESTER, R. PERRY, 1995: Analysis and interpretation of air quality data from an urban roadside location in central London over the period from July 1991 to July 1992. – Atmospheric Environment 29, 923-946. DI CARLO, P., W.H. BRUNE, M. MARTINEZ, H. HARDER, R. LESHER, X. REN, T. THORNBERRY, M.A. CARROLL, V. YOUNG, P.B. SHEPSON, D. RIEMER, E. APEL, C. CAMPBELL, 2004: Missing OH reactivity in a forest: Evidence for unknown reactive biogenic VOCs. – Science 304, 722-725. 77 EKBERG, A., A. ARNETH, H. HAKOLA, S. HAYWARD, T. HOLST, 2008: Leaf isoprene emission in a subarctic wetland sedge community. – Oecologia, submitted. FOKEN, T., 2003: Angewandte Meteorologie. – Springer Verlag, Berlin, pp. 289. GRABMER, W., M. GRAUS, C. LINDINGER, A. WISTHALER, B. RAPPENGLÜCK, R. STEINBRECHER, A. HANSEL, 2004: Disjunct eddy covariance measurements of monoterpene fluxes from a Norway spruce forest using PTR-MS. – International Journal of Mass Spectrometry 239, 111-115. GUENTHER, A.B., P.R. ZIMMERMAN, P.C. HARLEY, R.K. MONSON, R. FALL, 1993: Isoprene and monoterpene emission rate variability - Model evaluations and sensitivity analyses. – Journal of Geophysical Research 98, 12609-12617. GUENTHER, A., C.N. HEWITT, D. ERICKSON, R. FALL, C. GERON, T. GRAEDEL, P. HARLEY, L. KLINGER, M. LERDAU, W.A. MCKAY, T. PIERCE, B. SCHOLES, R. STEINBRECHER, R. TALLAMRAJU, J. TAYLOR, P. ZIMMERMAN, 1995: A global model of natural organic compound emissions. – Journal of Geophysical Research-Atmospheres 100, 8873-8892. HAAPANALA, S., J. RINNE, K.-H. PYSTYNEN, H. HELLÉN, H. HAKOLA, T. RIUTTA, 2006: Measurements of hydrocarbon emissions from a boreal fen using the REA technique. – Biogeosciences 3, 103-112. HELLÉN, H., H. HAKOLA, K.-H. PRYSTYNEN, J. RINNE, S HAAPANALA, 2006: C2-C10 hydrocarbon emissions from a boreal wetland and forest floor. – Biogeosciences 3, 167-174. HOFFMANN, T., J.R. ODUM, F. BOWMAN, D. COLLINS, D. KLOCKOW, R.C. FLAGAN, J.H. SEINFELD, 1997: Formation of organic aerosols from the oxidation of biogenic hydrocarbons. – Journal of Atmospheric Chemistry 26, 189-222. HOLST, T., S. HAYWARD, A. EKBERG, A. ARNETH, 2007: Measurements of biogenic volatile organic compounds above a sub-arctic wetland in northern Sweden. – In: HANSEL, A., T.D. MÄRK (Eds.): Contributions 3rd International Conference on Proton Transfer Reaction Mass Spectrometry and Its Applications. – Innsbruck University Press, 174-177. HOLST, T., A. ARNETH, S. HAYWARD, A. EKBERG, M. MASTEPANOV, M. JACKOWICZKORCZYNSKI, T. FRIBORG, P.M. CRILL, 2008: BVOC ecosystem flux measurements at a high latitude wetland site. – Atmospheric Chemistry and Physics Discussion, in preparation. HÜVE, K., M.M. CHRIST, E. KLEIST, R. UERLINGS, Ü. NIINEMETS, A. WALTER, J. WILDT, 2007: Simultaneous growth and emission measurements demonstrate an interactive control of methanol release by leaf expansion and stomata. – Journal of Experimental Botany 58, 17831793. JOHANSSON, T., N. MALMER, P.M. CRILL, T. FRIBORG, J.H. ǺKERMAN, M. MASTEPANOV, T.R. CHRISTENSEN, 2006: Decadal vegetation changes in a northern peatland, greenhouse gas fluxes and net radiative forcing. – Global Change Biology 12, 1-18. KAIMAL, J.C., J.J. FINNIGAN, 1994: Atmospheric boundary layer flows: Their structure and measurement. – Oxford University Press, New York, pp. 289. KARL, T., A. GUENTHER, A. JORDAN, R. FALL, W. LINDINGER, 2001: Eddy covariance measurement of biogenic oxygenated VOC emissions from hay harvesting. – Atmospheric Environment 35, 491-495. KARL, T.G., C. SPIRIG, J. RINNE, C. STROUD, P. PREVOST, J. GREENBERG, R. FALL, A. GUENTHER, 2002: Virtual disjunct eddy covariance measurements of organic compound fluxes from a subalpine forest using proton transfer reaction mass spectrometry. – Atmospheric Chemistry and Physics 2, 279-291. 78 KESSELMEIER, J., M. STAUDT, 1999: Biogenic volatile organic compounds (VOC): An overview on emission, physiology and ecology. – Journal of Atmospheric Chemistry 33, 23-88. KULMALA, M., K. HÄMERI, J.M. MÄKELÄ, P.P. AALTO, L. PIRJOLA, M. VÄKEVA, E.D. NILSSON, I.K. KOPONEN, G. BUZORIUS, P. KERONEN, Ü. RANNIK, L. LAAKSO, T. VESALA, K. BIGG, W. SEIDL, R. FORKEL, T. HOFFMANN, J. SPANKE, R. JANSON, M. SHIMMO, H.C. HANSSON, C. O’DOWD, E. BECKER, J. PAATERO, K. TEINILÄ, R. HILLAMO, Y. VIISANEN, A. LAAKSONEN, E. SWIETLICKI, J. SALM, P. HARI, N. ALTIMIR, R. WEBER, 2000: Biogenic aerosol formation in the boreal forest. – Boreal Environment Research 5, 281-297. KULMALA, M., T. SUNI, K.E.J. LEHTINEN, M. DAL MASO, M. BOY, A. REISSELL, Ü. RANNIK, P. AALTO, P. KERONEN, H. HAKOLA, J. BÄCK, T. HOFFMANN, T. VESALA, P. HARI, 2004: A new feedback mechanism linking forests, aerosols, and climate. – Atmospheric Chemistry and Physics 4, 557-562. LEE, A., G.W. SCHADE, R. HOLZINGER, A.H. GOLDSTEIN, 2005: A comparison of new measurements of total monoterpene flux with improved measurements of speciated monoterpene flux. – Atmospheric Chemistry and Physics 5, 505-513. LINDINGER, W., A. HANSEL, A. JORDAN, 1998: On-line monitoring of volatile organic compounds at pptv levels by means of proton-transfer-reaction mass spectrometry (PTR-MS) medical applications, food control and environmental research. – International Journal of Mass Spectrometry and Ion Processes 173, 191-241. NEMECEK-MARSHALL, M., R.C. MACDONALD, J.J. FRANZEN, C.L. WOJCIECHOWSKI, R. FALL, 1995: Methanol emissions from leaves. – Plant Physiology 108, 1359-1368. SECO, R., J. PEÑUELAS, I. FILELLA, 2007: Short-chain oxygenated VOCs: Emission and uptake by plants and atmospheric sources, sinks, and concentrations. – Atmospheric Environment 41, 2477-2499. NIINEMETS, Ü., F. LORETO, M. REICHSTEIN, 2004 : Physiological and physicochemical controls on foliar volatile organic compound emission. – Trends in Plant Science 9, 180—186. RINNE, J., J.-P. TUOVINEN, T. LAURILA, H. HAKOLA, M. AURELA, H. HYPÉN, 2000: Measurements of hydrocarbon fluxes by a gradient method above a northern boreal forest. – Agricultural and Forest Meteorology 102, 25-37. RINNE, H.J.I., A. GUENTHER, C. WARNECKE, J.A. DE GOUW, S.L. LUXEMBOURG, 2001: Disjunct eddy covariance technique for trace gas measurements. – Geophysical Research Letters 28, 3139. RINNE, H.J.I., A.B. GUENTHER, J.P. GREENBERG, P.C. HARLEY, 2002: Isoprene and monoterpene fluxes measured above Amazonian rainforest and their dependence on light and temperature. – Atmospheric Environment 36, 2421-2426. SPIRIG, C., A. NEFTEL, C. AMMANN, J. DOMMEN, W. GRABMER, A. THIELMANN, A. SCHAUB, J. BEAUCHAMP, A. WISTHALER, A. HANSEL, 2005: Eddy covariance flux measurements of biogenic VOCs during ECHO 2003 using proton transfer reaction mass spectrometry. – Atmospheric Chemistry and Physics 5, 465-481. STRÖM, L., T.R. CHRISTENSEN, 2007: Below ground carbon turnover and greenhouse gas exchanges in a sub-arctic wetland. – Soil Biology & Biochemistry 39, 1689-1698. SVENSSON, B.H., T.R. CHRISTENSEN, E. JOHANSSON, M. ÖQUIST, 1999: Interdecadal Changes in CO2 and CH4 Fluxes of a Subarctic Mire: Stordalen Revisited after 20 Years. – Oikos 85, 22-30. 79 Authors' address: Dr. Thomas Holst ([email protected]) Dr. Almuth Arneth ([email protected]) Dr. Sean Hayward ([email protected]) Dr. Anna Ekberg ([email protected]) Department of Physical Geography and Ecosystems Analysis, Lund University Sölvegatan 12, SE-22362, Lund, Sweden 80 81 Eigenschaften des Klimas in der Überflutungsaue des Oberrheins Dieter Ahrens und Werner Möhle Landesanstalt für Umwelt, Messungen und Naturschutz Baden-Württemberg Zusammenfassung Die Veränderungen der Landnutzung in Europa hin zu trockeneren Oberflächen durch Bebauung und landwirtschaftliche Nutzung in den letzten Jahrzehnten sind ein Ursachenfaktor für die beobachteten regionalen Klimaveränderungen. Die verbliebenen Reste der feuchten Überflutungsauen am Oberrhein sind in diesem Zusammenhang ein „Klimadenkmal“. Anhand von 10 jährigen kontinuierlichen Messungen im Naturschutzgebiet „Rastatter Rheinaue“ können vergleichend zu urban und anthropogen geprägten Flächen die spezifischen Unterschiede von Lufttemperatur, Luftfeuchte, Albedo und langwelligen Strahlungskomponenten dargestellt werden. Durch die besondere, hochwassersichere Messanordnung konnten diese Parameter auch während Überflutungen des Untergrundes erfasst werden. Während der Messungen traten der höchste Pegelstand und die längste Überflutungsperiode auf. Die Vergleiche umfassen neben Jahres-, Monats- und Tagesmitteln auch solche Extremsituationen. Climatic aspects of the floodplains of the upper Rhine basin Abstract The recent changes in land cover in Europe during the last decades towards to dryer surfaces caused by urbanization and farming are one of the influences for regional climate change. The remaining rests of the wet flood plains of the Upper Rhine are therefore a “Climatic Memorial”. Referring 10 years of continues monitoring within the preserved area “Rastatt flood plains”, it is possible to compare the influence of this wet surface on air temperature, air moisture, albedo and long wave radiation with urban and lawn surfaces. Due to the specific installation of the measuring systems, which were protected against flood, the recordings could be continued even during these events. The highest water mark ever observed and the longest temporal duration of inundation happened in the monitoring period. Besides annual, monthly and daily means, the comparison of meteorological parameters between the selected land use types even includes such extreme events. 1. Einleitung Die zunehmende Häufigkeit extremer Hochwasserereignisse am Rhein in den vergangenen Jahren ist mitverantwortlich dafür, dass die Problematik von „schleichenden Veränderungen der Landnutzung und der Oberflächenstruktur und ihrer Wirkung auf das Klima“ (REKLIP, 1987) zunehmend ins öffentliche Interesse gerückt ist. Die aktuelle Diskussion um extremes Hochwasser im August 2007 - am 10. August 2007 wurde in Plittersdorf der zweithöchste (7.58 m) jemals gemessene Pegel (nach dem 13. Mai 1999 mit 7.68 m) registriert - und Polderbauten am Oberrhein belegt dies erneut (BNN, 2007; LUBW, 2007). Der Verlust an Überflutungsflächen entlang des Rheins ist allerdings nicht als „schleichend“ zu bewerten. Innerhalb nur weniger Jahrzehnte ist eine einschneidende Abnahme der Überschwemmungsfläche erfolgt. Von ehedem einigen hundert sind heute nur 82 noch etwa 50 Quadratkilometer Auenfläche beiderseits des Oberrheins zwischen Basel und Bingen überflutet. Die Veränderung der Landschaft ist untrennbar mit der historischen Entwicklung des Oberrheinausbaus verknüpft. Dieser Eingriff in den Flusslauf, welcher die Phasen Rheinkorrektion im 19. Jahrhundert, Rheinregulierung und zuletzt Staustufenausbau im 20.Jahrhundert umfasst, ist in der Literatur ausführlich dokumentiert (DISTER, 1985a; LEHLE, 1985). Die verbliebenen Rheinauen, welche durch die Dynamik des Flusses mit den kurzzeitigen Wasserstandsschwankungen und alljährlichen Überschwemmungen gekennzeichnet sind, stellen eine einzigartige Naturlandschaft dar. „Kein anderes Ökosystem in den gemäßigten Breiten ist so strukturreich wie die amphibische Überschneidungszone zwischen Land und Wasser. Am Oberrhein sorgen außerdem das milde Klima, die nährstoffreichen Überschwemmungen und die hohen Grundwasserstände für nahezu tropische Bedingungen mit entsprechend üppigem Pflanzenwuchs“ (KUTTLER und SPÄTH, 1993). Die klimatisch wichtigste Eigenschaft der Aue ist ihre Rolle als eine bedeutende Quelle für Wasserdampf, da auch im Sommer, anders als in Wäldern an anderen Standorten, meist ausreichend Wasser für die Verdunstung vorhanden ist. Weiter sind charakteristisch: große aerodynamische Rauigkeit durch wechselnde Bestandshöhen und dazwischen liegende offene Wasserflächen. Überschwemmungen, die zu jeder Jahreszeit kurzfristig auftreten, verändern die Energiebilanz der Auewälder innerhalb von Stunden gravierend. Sie verbleiben in diesem Extremzustand oft über Tage und Wochen. Die bodennahe Krautschicht unterliegt dabei je nach Entwicklung Zerstörungen oder Wachstumsförderung. Auch hierin unterscheidet sich die Überflutungsaue grundsätzlich von anderen Waldlandschaften Mitteleuropas. Während die Auen botanisch, zoologisch und hydrologisch (DISTER, 1985b, 1988; HOMAGK, 1992) intensiv erforscht wurden, gibt es zum Klima der Auen kaum Untersuchungen (UHRECKÝ et. al., 1985; MAYER und AHRENS, 1997). Im Rahmen des Deutsch-Französisch-Schweizerischen REGIO-Klima-Projektes (REKLIP, 1987) erfolgten die hier beschriebenen fast 10 jährigen Messungen im Zeitraum von 1991 bis 2001. Sie ermöglichen es erstmalig, die besonderen Kennzeichen des Klimas der Rheinaue in Details darzustellen. 2. Untersuchungsgebiet Das Untersuchungsgebiet „Rastatter Rheinaue“ gehört morphologisch noch zu der im südlichen Teil der Oberrheinebene gelegenen Furkationszone des Stromes bzw. markiert den Übergang zur sich unmittelbar nördlich anschließenden Mäanderzone. 2.1 Messstandort Der Messstandort befindet sich rechtsrheinisch auf der Gemarkung der Stadt Rastatt, Ortsteil Plittersdorf, und gehört zur sog. Jungaue, im Folgenden als „Überflutungsaue“ bezeichnet. Die Trennungslinie zur Altaue wird heute durch einen Hochwasserdamm markiert, der hier bis an die Siedlungsgrenze des Ortsteiles Plittersdorf reicht. 83 Die Überflutungsaue hat in diesem Bereich auf deutscher Seite eine Breite von bis zu 2 km. Sie ist Teil der 1984 grenzüberschreitend als Naturschutzgebiet ausgewiesenen „Rastatter Rheinaue“ (WWF, 1995) mit einer Fläche von 8,45 km². Der Schutzzweck ist die naturnahe Erhaltung der verbliebenen Reste der ursprünglichen Überflutungsaue vor der Rheinregulierung. Es bestehen hier noch zahlreiche offene Verbindungen zu Altund Seitengewässern welche auch bei Normalwasserstand regelmäßig durchflutet werden. Als Untersuchungsstandort für eine ortsfeste Station wurde eine ständig feuchte, weitgehend Natur belassene und für die unbewaldeten Teile der Aue damit typische Überflutungswiese gewählt (MAYER und AHRENS, 1997). Diese befindet sich nahe am Flussufer bei Rhein-Kilometer 340 (108 m ü. NN). Ein wesentliches Charakteristikum des Messortes ist die zeitweise vollständige oder teilweise Überschwemmung des Untergrundes. Bereits bevor die gesamte Aue flächenhaft überflutet wird, beginnt das Wasser über die Altrheinarme hereinzuströmen und führt zu Teilüberflutungen. Auch bei starken Niederschlägen wird die Messwiese und Umgebung zeitweise mit großen Wasserflächen durchsetzt. Die überwiegenden Flächen der Aue sind Waldbestand, Weiden (Salix) und Pappeln (Populus); letztere erreichen Höhen von über 30 m. In der solchermaßen „gestörten“ Umgebung wäre die Messung vollständiger Energiebilanzen der Erdoberfläche nur mit extrem großem Aufwand möglich gewesen. Sie hätte die Errichtung eines Messturmes von mindestens 50 m Höhe erfordert. Die Instrumentierung der Station „Rheinaue Plittersdorf“ war somit zweckgebunden gegenüber einer voll ausgerüsteten Energiehaushaltsstation eingeschränkt, gestattete aber trotzdem die Ableitung charakteristischer Klimaparameter der Überflutungsaue. Durch Vergleiche einzelner Parameter mit anderen REKLIP Stationen wird die Aussage ergänzt. Tab. 1: Ausstattung der Station „Rheinaue Plittersdorf“ mit meteorologischen Messwertgebern Parameter Symbol (Einheit) Messwertgeber Windgeschwindigkeit v (m/s) Propeller Anemometer Windrichtung (Grad) nach Gill Lufttemperatur Ta (°C) Pt 100 Widerstandstherm. Luftfeuchte (Taupunkt) Td (°C) LiCl Kammertemperatur langwellige atmosphärische Gegenstrahlung L↓ (W/m2) langwellige Ausstrahlung des Erdbodens L↑ (W/m2) Globalstrahlung K↓ /W/m2) kurzwellige Reflexstrahlung 2 Pyrradiometer 2 Pyranometer 2 K↑ (W/m ) Ein besonderes Problem stellten die häufigen Überflutungen des Messorts dar. Die Datenerfassung erfolgte deshalb in einem Container, welcher auf einem 1.6 m hohen, extra 84 konstruierten Untergestell, sicher für mehr als 100 jähriges Hochwasser, untergebracht werden musste. Während der Überflutungszeiten konnte wegen Unpassierbarkeit der Zufahrtsstraßen die regelmäßige Wartung der Anlagen nicht durchgeführt werden, was teilweise zu Datenausfällen führte. Die kontinuierlich gemessenen meteorologischen Parameter sind in der Tab. 1 zusammengestellt. Die Abtastrate aller Geräte betrug 2 Sekunden. Daraus erfolgte die Bildung von 30 Minuten Mittelwerten für alle Komponenten über einen zentralen Stationsrechner. Diese Werte wurden in einer Datenbank gespeichert. Die Windmessung erfolgte in 17 m Höhe über Grund an einem Mast, der etwa 50 m vom Ufer entfernt im Container eingebaut war. Störungen durch den ca. 50 m östlich gelegenen Pappelhochwald bezüglich Windrichtung und Windgeschwindigkeit konnten wegen dieser Messhöhe verringert, aber nicht ganz ausgeschaltet werden. In der Hauptwindrichtung im Südwesten waren kaum Hindernisse vorhanden, in der zweithäufigsten Windrichtung Nordosten einige hohe Baumreihen. Daher war die Windmessung aus den nordöstlichen Sektoren am stärksten beeinträchtigt. Ein einschneidendes Ereignis waren die Sturmschäden durch den Orkan „Lothar“ am 26. Dezember 1999. Das Untersuchungsgebiet befand sich regional im Hauptschadensbereich, wobei der Auewald über mehrere Quadratkilometer vollständig entwurzelt wurde. Es entstanden weite Kahlflächen, welche das typische Klima der Aue, auch am Messort, auf Jahre hinaus veränderten. Folglich erhöhte sich dadurch die Windgeschwindigkeit am Standort im Jahr 2000. Die Zerstörung der Stromversorgung im gesamten Gebiet führte zu einer mehrtägigen Unterbrechung der Messungen. 2.2 Vergleichsstationen Zum regionalen Vergleich wurden für die Komponenten Lufttemperatur und Luftfeuchte die 7 km östlich gelegene Luftqualitäts-Messstation „Rastatt“ (117 m ü. NN) herangezogen, welche sich auf der trockenen und sandigen sog. „Niederterrasse“ in einem locker bebauten Stadtrandgebiet von Rastatt befindet. Diese Station war für die vergleichsweise benutzten meteorologischen Parameter mit identischen Messfühlern ausgestattet. Messungen der kurz- und langwelligen Strahlungskomponenten im näheren Bereich und über einen längeren Zeitraum außerhalb der Aue sind nicht verfügbar. Zum Vergleich musste deshalb auf die REKLIP Station „Bremgarten“ zurückgegriffen werden (IZIOMON et al., 2001), die sich etwa 90 km südlich von Plittersdorf ebenfalls in der Oberrheinebene befindet. Dort fanden die Messungen zeitgleich vom Juli 1991 bis zum September 1996 auf einem ehemaligen Militärflugplatz (212 m ü. NN) mit kurzer Grasoberfläche statt. Der Messort lag etwa 3 km vom Rhein entfernt in der südbadischen Trockenzone (REKLIP, 1995). 3. Allgemeine klimatische Bedingungen Obwohl die Oberrheinebene über große Entfernungen hinweg eben ist, mit Höhendifferenzen von nur wenigen Metern, können sich dennoch signifikante lokale Unterschiede der wesentlichen Klimaelemente wie Lufttemperatur, Luftfeuchte und Wind herausbilden. Die Ursache hierfür sind Unterschiede in Art und Höhe des Bewuchses, der Bodenfeuchte und der Windexposition. 85 3.1 Lufttemperatur Bei den Extremwerten der Lufttemperatur Ta (Tab. 2) und den Ta Monatsmitteln (Abb. 1) an der Station „Rheinaue Plittersdorf“ kommt das insgesamt milde Makroklima der mittleren Oberrheinebene zum Ausdruck. Durch die südliche Lage in Mitteleuropa und die geringe Meereshöhe von nur etwa 100 m ü. NN liegt im Winter die Lufttemperatur meist deutlich über dem Gefrierpunkt, im Sommer können Ta Monatsmittel von über 20 °C auftreten. In dieser Jahreszeit werden häufig sog. „heiße Tage“ (Ta,max > 30 °C) erreicht. Tab. 2: Extreme der Lufttemperatur Ta an der Station „Rheinaue Plittersdorf“ im Zeitraum 1991 bis 2000 Mittelungszeit Jahr Monat Tag Stunde Maximum (°C) 11.6 (2000) 22.3 (Juli 1994) 27.7 (9.8.1992) 36.7 (8./9.8.1992) Minimum (°C) 9.7 (1997) -2.8 (Jan 1997) -11.4 (1.1.1997) -15.5 (2.1.1997) 25 20 10 Ta (°C) 15 5 0 1991 1996 mean 1992 1997 1993 1998 1994 1999 1995 2000 -5 1 2 3 4 5 6 7 month 8 9 10 11 12 Abb. 1: Jährliche Monatsmittel der Lufttemperatur Ta an der Station „Rheinaue Plittersdorf“ im Zeitraum 1991 bis 2000 Die Lufttemperatur wird in der bodennahen Atmosphäre durch die Strahlungsverhältnisse, den Energieumsatz an der Oberfläche, die Art der Landnutzung (z.B. Vegetation oder Bebauung), die Verdunstung und den Luftmassenaustausch bestimmt (GEIGER, 1961). Der Ta Vergleich zwischen der feuchten Überflutungswiese und der 7 km östlich 86 gelegenen Luftqualitätsstation „Rastatt“ belegt ein unterschiedliches thermisches Mikroklima innerhalb nur kurzer Entfernungen, aber bei gleicher Höhenlage. Ta,Plittersdorf - Ta,Rastatt (°C) 2 1 0 -1 -2 -3 0 5 10 Ta,Rastatt 15 20 25 (°C) Abb. 2: Differenzen der Ta Monatsmittel zwischen den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ in Abhängigkeit von den Ta Monatsmitteln an der Station „Rastatt“ im Zeitraum Mai 1993 bis August 2000 Aus der Gegenüberstellung der Ta Monatsmittel an beiden Stationen (Abb. 2) im verfügbaren Vergleichszeitraum von Mai 1993 bis August 2000 erkennt man, dass es in der Rheinaue fast immer kühler war als in Rastatt. Die wenigen Ausnahmen beschränken sich auf die Übergangsjahreszeiten und den Winter. Bei niedriger Lufttemperatur, also im Winter, sind die Differenzen zwischen beiden Stationen gering und unterscheiden sich nur um wenige Zehntel Grad. Dagegen ist es in den Sommermonaten bei hoher Lufttemperatur in der Aue im Mittel um bis zu 1.5 °C kühler. Der Grund hierfür sind die Energieumsätze über den verschiedenartigen Oberflächen, wobei in der Rheinaue die hohe Verdunstungsrate über der feuchten Wiese abkühlend wirkt, während auf der vom Untergrund her trockenen, zusätzlich durch Flächenversiegelung urban beeinflussten Umgebung der Station Rastatt häufig bereits der Welkepunkt der Vegetation erreicht wird und damit die Verdunstung stark eingeschränkt ist. Eine monatliche Differenz der Lufttemperatur von 1.5° C entspricht im Sommer klimatisch immerhin einer vertikalen Höhenstufe von rund 200 m am Schwarzwaldrand (REKLIP, 1995). 3.2 Luftfeuchte Bei der Luftfeuchte, welche als Wasserdampfgehalt der Atmosphäre durch den Dampfdruck VP (hPa) oder als relative Feuchte RH (%) angegeben werden kann, ist der signifikant höhere Feuchtegehalt in der bodennahen Luftschicht über der Überflutungswiese 87 charakteristisch. Bei den VP Monatsmitteln (Abb. 3) kommt das besonders im Sommer feuchte Mikroklima der Auelandschaft zum Ausdruck. VP Monatsmittel um oder über 20 hPa sind auch in der Oberrheinebene als außergewöhnlich hohe Luftfeuchte anzusehen (REKLIP, 1995). Gerade im Juli zeigt die große Schwankungsbreite des monatlichen Wasserdampfgehaltes von VP zwischen 15 und 21 hPa den Einfluss wechselnder Wasserstände und damit Verdunstungsraten auf die bodennahe Luftschicht in der Überflutungsaue. 20 10 VP (hPa) 15 5 1991 1996 mean 1992 1997 1993 1998 1994 1999 1995 2000 0 1 2 3 4 5 6 7 month 8 9 10 11 12 Abb. 3: Jährliche Monatsmittel des Dampfdruckes VP an der Station „Rheinaue Plittersdorf“ im Zeitraum 1991 bis 2000 Aus der Gegenüberstellung der VP Monatsmittel an beiden Stationen (Abb. 4) im verfügbaren Vergleichszeitraum werden die in allen Jahreszeiten stets sehr viel höheren Feuchtewerte in der Aue deutlich. Der Unterschied zu trockenen Gebieten der Oberrheinebene tritt bei dieser Klimakomponente noch stärker als bei der Lufttemperatur hervor. Während bei absolut kleinen Feuchtegehalten im Winter auch die Differenzen gering sind, da in dieser Jahreszeit die Verdunstung der Vegetation stark eingeschränkt ist, nehmen diese mit steigenden Dampfdruckwerten zu und erreichen in manchen Sommermonaten Differenzwerte von mehr als 5 hPa. Durch das Wasserangebot in der Aue wird im Sommer die stets mögliche Verdunstung zur bestimmenden Komponente des Energiehaushaltes an der Erdoberfläche und zur typischen Komponente des Aueklimas. Aus der in der Tab. 3 enthaltenen Häufigkeit von Überschreitungen hoher VP Schwellenwerte an den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ lässt sich die überaus große Häufigkeit hoher Feuchtewerte in der Aue eindrucksvoll erkennen. VP Werte > 25 hPa sind in „Rastatt“ während der gesamten 7 jährigen Vergleichsperiode überhaupt nicht aufgetreten und VP Werte > 20 hPa nur sehr selten (0.25% aller Stunden). In der Rheinaue lag dagegen VP an immerhin 3.35% aller Stunden im Vergleichszeitraum 88 über 20 hPa. Der höchste Einzelwert des Dampfdruckes (1 Std.-Mittel) wurde am 21. Juli 1995 mit VP=34 hPa gemessen (vgl. auch Abb. 16). VP,Plitterdorf - VP,Rastatt (hPa) 5 4 3 2 1 0 4 6 8 10 VP Rastatt 12 14 16 (hPa) Abb. 4: Differenzen der VP Monatsmittel zwischen den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ in Abhängigkeit von den VP Monatsmitteln an der Station „Rastatt“ im Zeitraum Mai 1993 bis August 2000 Tab. 3: Anzahl von Stundenmittelwerten hohen Dampfdruckes VP (Mai 1993 bis August 2000) Station VP ≥ 20 hPa VP ≥ 25 hPa VP ≥ 30 hPa „Rheinaue Plittersdorf“ 2101 151 3 „Rastatt“ 157 - - 3.3 Wind Die an einem Standort gemessene Windgeschwindigkeit v ist in einem ebenen Gelände stark von den lokalen Verhältnissen, im Wesentlichen von Hindernisstrukturen wie Bebauung und Bewuchs, abhängig. In der Rheinaue betrug der v Mittelwert 1.9 m/s. In der Stadtrandlage von Rastatt lag er bei 2.3 m/s. In freien Feldlagen der mittleren Oberrheinebene sind v Mittelwerte um 3.0 m/s als charakteristisch anzusehen (REKLIP, 1995). Am Auenstandort „Rheinaue Plittersdorf wurde v im Sommer infolge des Einflusses der Belaubung im Mittel auf 1.5 m/s reduziert. Diese Verminderung von v im Sommer ist von großem Einfluss auf das Feuchtemilieu, weil damit die horizontale Diffusion des aus der Verdunstung stammenden Wasserdampfes stark eingeschränkt ist. Im Winter erfolgte zwar keine Angleichung der Windgeschwindigkeit an die Freilandverhältnisse, 89 aber doch eine Erhöhung auf 2.5 m/s, da in der Aue nur Laub abwerfende Bäume wachsen. Durch die großflächigen Baumwürfe des Orkans „Lothar” vom 26. Dezember 1999 wurde der Auenstandort wesentlich offener in alle Richtungen. Eine Erhöhung der mittleren Windgeschwindigkeit im Jahr 2000 auf v = 2.8 m/s war die Folge. 3.4 Strahlung Die kurz- und langwelligen Strahlungsflussdichten in der bodennahen Atmosphäre setzen sich aus folgenden Komponenten zusammen: K↓ : Globalstrahlung K↑: kurzwellige Reflexstrahlung der Erdoberfläche L↓: langwellige atmosphärische Gegenstrahlung L↑: langwellige Ausstrahlung des Erdbodens Die Bilanz dieser Komponenten besitzt eine Schlüsselfunktion für die regionale Klimadifferenzierung. Durch die Strahlungsbilanzgleichung Q* = K↓ - K↑ + L↓ - L↑ (1) sind die einzelnen Strahlungsflussdichten zur Strahlungsbilanz Q* verknüpft. Strahlungsbilanz Die Monatsmittel der Strahlungsbilanz Q* über der Überflutungswiese am Standort „Rheinaue Plittersdorf“ (Abb. 5) nehmen im Jahresgang Werte zwischen –15 W/m2 im Januar und +175 W/m2 im Juli an. Zum Vergleich sind die minimalen und maximalen Werte von der trockenen Rasenfläche in Bremgarten (ROST, 2004) als Linien eingetragen. Im Winter sind die Unterschiede zwischen beiden Oberflächen absolut gesehen gering. In den anderen Jahreszeiten gehen die unterschiedlichen Oberflächenstrukturen ganz entscheidend in die Strahlungsbilanz ein. Tendenziell ist die Oberfläche in der Aue feuchter und damit durch Verdunstung auch kühler, d.h. die langwellige Ausstrahlung L↑ ist kleiner. Unter der nahe liegenden Voraussetzung, dass bei beiden Stationen die von oben einfallende Globalstrahlung K↓ und die langwellige Gegenstrahlung der Atmosphäre L↓ in etwa die gleichen Werte annehmen, ergibt sich daraus eine größere positive Strahlungsbilanz Q* in der Aue. Besonders deutlich wirken sich Überflutungen in dieser Richtung aus, wie die Werte von Mai bis Juli 1999 belegen. Die jährlichen Unterschiede sind in der Aue im selben Monat, bedingt durch die Veränderungen des Untergrundes, sehr viel größer als in Bremgarten. Daher weisen die Energieumsätze in der bodennahen Atmosphäre über der Überflutungsfläche eine stärkere zeitliche Variabilität gegenüber einer Kulturwiese auf, was damit ein besonderes klimatisches Merkmal dieses Naturraumes ist. Auf Grund der Strahlungsbilanzgleichung sind bei gleichen Werten für die atmosphärischen Komponenten an beiden Standorten die kurzwellige Reflexstrahlung K↑ des Erdbodens und die langwellige Ausstrahlung L↑ für die Differenzen der Strahlungsbilanz verantwortlich. 90 150 50 Q* (W/m²) 100 0 1991 1992 1996 1997 min Bremgarten mean Plittersdorf -50 1 2 3 4 5 1993 1994 1998 1999 max Bremgarten 6 7 month 8 1995 2000 9 10 11 12 Abb. 5 Monatsmittel der Strahlungsbilanz Q* an der Station „Rheinaue Plittersdorf“ im Zeitraum Juli 1991 bis Dezember 2000 sowie mittlere maximale und minimale monatliche Strahlungsbilanz Q* an der REKLIP Station „Bremgarten“ im Zeitraum Juli 1991 bis September 1996 Albedo Normiert man die kurzwellige Reflexstrahlung K↑ auf die einfallende Globalstrahlung K↓, ergibt sich die Albedo a (in %) als Maß für die kurzwellige Reflektivität der Oberfläche: a=100%* K↑ / K↓ (2) Die häufigen Oberflächenveränderungen der Überflutungswiese wirken sich wie bei der Strahlungsbilanz sehr stark auf den zeitlichen Verlauf der Albedo aus. Ganzjährig hat die Kulturwiese am Standort „Bremgarten“ (Abb. 6) eine deutlich höhere Albedo (ROST, 2004) als die Natur belassene Oberfläche am Standort „Rheinaue Plittersdorf“. Einigermaßen vergleichbar sind nur die Albedowerte im Sommer. Besonders große Differenzen treten im Winter auf. Hohe winterliche Albedowerte auf der Kulturwiese entstehen durch Schneebedeckung oder in milden Wintern durch die Erhaltung der grünen Grasoberfläche. Auf der Überflutungswiese werden die niedrigsten Albedowerte in den Wintermonaten erreicht, wenn der dichte Kraut- und Grasbewuchs vertrocknet oder verrottet ist und eine dunkelbraune Färbung aufweist. Das Maximum liegt im Frühsommer, meist im Mai, wenn die Überflutungswiese vorwiegend durch die aufwachsenden Goldruten (Solidago canadensis/gigantea) und das Indische Springkraut (Impatiens glandulifera) 91 hellgrün bis hellgelb gefärbt ist. Die Vegetationsentwicklung formt zu diesem Zeitpunkt mit starkem Wachstum, Verdichtung und Blühvorgängen eine dichte Reflexionsschicht für die kurzwellige Sonneneinstrahlung, außer der Boden ist teilweise oder ganz durch Hochwasser überflutet wie z.B. im Frühjahr 1999. Deshalb sind die Unterschiede von Jahr zu Jahr im Mai und Juni besonders groß. Der folgende höhere Albedowert bis September ist auf die Ausbildung einer dichten, nicht welkenden Vegetationsschicht zurückzuführen. 35 30 25 a (%) 20 15 10 1991 1992 1997 1998 mIn Bremgarten mean Plittersdorf 5 0 1 2 3 4 5 1993 1994 1995 1999 2000 max Bremgarten 6 7 month 8 9 10 11 12 Abb. 6: Monatsmittel der Albedo a an der Station „Rheinaue Plittersdorf“ im Zeitraum Juli 1991 bis Dezember 2000 sowie mittlere maximale und minimale monatliche Albedowerte an der REKLIP Station „Bremgarten“ im Zeitraum Juli 1991 bis September 1996 Eine Schneebedeckung im Winter ist ein seltenes Ereignis in der Rheinaue und führte nur an wenigen Tagen zu einer erhöhten Albedo, wie im November und Dezember 1999. Eine typische Überflutung im Winter, sog. „Weihnachtshochwasser“, kann andererseits wiederum zu niedrigeren Albedowerten führen, wie im Winter 1991/92. Zusammenfassend ergibt sich, dass die Unterschiede in der Albedo in den einzelnen Monaten von Jahr zu Jahr, bedingt durch die Wasserstandsdynamik und die Folgen für die Vegetation, sehr groß sind. Auch der Jahresgang selbst ist sehr viel stärker ausgeprägt als über einer Kulturwiese mit kurzem, gemähtem Gras, wo er nur wenige Prozent beträgt, abgesehen von einer winterlichen Schneebedeckung. 92 Spezifische Strahlungsbilanz Eine weitere terrestrisch beeinflusste Komponente ist die langwellige „spezifische Strahlungsbilanz“ L*L. Der Einfluss der Ausstrahlung der Erdoberfläche (L↑)auf die gesamte Strahlungsbilanz lässt sich nach der Gleichung L*L =(L↓-L↑/K↓)*100% (3) analysieren. 100 1991 1992 1997 1998 min Bremgarten mean Plittersdorf 80 1993 1994 1999 2000 max Bremgarten 1995 L*L (%) 60 40 20 0 1 2 3 4 5 6 7 month 8 9 10 11 12 Abb. 7: Monatsmittel der spezifischen langwelligen Strahlungsbilanz L*L an der Station „Rheinaue Plittersdorf“ im Zeitraum Juli 1991 bis Dezember 2000 sowie mittlere maximale und minimale monatliche Monatsmittel der spezifischen langwelligen Strahlungsbilanz L*L an der REKLIP Station „Bremgarten“ im Zeitraum Juli 1991 bis September 1996 Die spezifische langwellige Strahlungsbilanz L*L hängt über L↑ von den turbulenten Strömen latenter und fühlbarer Wärme sowie dem Boden- und Bestandswärmestrom einer Oberfläche ab (KESSLER, 1983). Da bei der Verdunstung viel Energie verbraucht wird, und damit L↑ kleiner wird, hat eine trockene Oberfläche einen größeren Wert von L*L als eine feuchte. So weist bei der Überflutungswiese L*L auf der Basis von Monatsmittelwerten einen ausgeprägten Jahresgang mit einem weiten monatlichen Streubereich auf (Abb. 7). Die niedrigsten Werte im Sommer, bei starker Verdunstung der Pflanzen und bei Überflutung, liegen unter 10% und steigen mit zunehmender Abtrocknung im Herbst und Winter bis auf 50% an, sofern kein Hochwasser die Messwiese überflutet (z.B. 1997 und 1999). Neben der jahreszeitlichen Dynamik fallen wiederum die von Jahr zu Jahr in den einzelnen Monaten über einen weiten Bereich schwankenden L*L Werte auf, welche noch stärker als die Albedo durch die jeweiligen Niederschlags- und Wasserstände 93 geprägt sind. Besonders groß ist hier der Einfluss der Verdunstung, wodurch L*L weitgehend durch die Wasserverfügbarkeit in der Aue beeinflusst wird. Für trockenere Oberflächen, wie sie heute fast überwiegend in Mitteleuropa vorherrschen, werden L*L Sommerwerte von 20 bis über 30% als typisch angegeben, bei nur geringen jahreszeitlichen Schwankungen (KESSLER, 1983). Dagegen wurden über der Überflutungswiese im Sommer meist Werte von L*L < 20% ermittelt, ein weiteres Indiz für eine ständig sehr viel feuchtere Oberfläche. Der Herbst und Winter ist in manchen Jahren in der Rheinaue durch lang anhaltende Niedrigwasserstände gekennzeichnet. Dichte, ausgetrocknete und deshalb weitgehend undurchlässige Schlammoberflächen aus vorangegangenen Überflutungen behindern dann die Verdunstung und ziehen ein Austrocknen des Kraut- und Grasbestandes nach sich. Daraus resultieren dann höhere Werte von L*L, wie z.B. im Herbst 1993. Dennoch werden die Winter- und Herbstwerte der „trockenen“ Kulturwiese am Standort „Bremgarten“ nur in wenigen Monaten erreicht. Die Auswertung der spezifischen Strahlungsbilanz L*L zeigt eindrucksvoll die Bedeutung der Verdunstung für den Energiehaushalt auf und belegt damit das feuchtere Klima in der Überflutungsaue auch anhand dieser Größe der theoretischen Klimatologie. 4. Ausgewählte meteorologische Bedingungen Eine Besonderheit des Untersuchungsortes sind die zeitweilig vollständigen oder teilweisen Überschwemmungen des Untergrundes. Alle mikrometeorologischen Parameter, erfahren dadurch eine vergleichsweise starke Veränderung; am raschesten wirkt sich diese auf die Albedo aus. 4.1 Überflutungen Im Frühjahr 1999 ereignete sich eine der am längsten andauernden Hochwassersituationen des letzten Jahrhunderts mit dem höchsten jemals gemessenen Wasserstand. Vom 13. Mai bis zum 15. Juni 1999, also 34 Tage lang, war der Messstandort ohne Unterbrechung mit Wasser bedeckt. Für den Zeitraum der ständigen Überschwemmung können neben dem Verlauf der Albedo auch Unterschiede von Lufttemperatur und Luftfeuchte zur Station „Rastatt“ auf der trockenen Niederterrasse dargestellt werden (Tab. 4). Tab. 4: Meteorologische Kenngrößen während der Überflutung vom 13. Mai bis zum 15. Juni 1999 Mittelwerte von: „Rheinaue Plittersdorf“ „Rastatt“ Lufttemperatur Ta (°C) 15,8 17,8 Dampfdruck VP (hPa) 14,8 11,2 79 60 Relative Feuchte RH (%) Danach war die Luft im Überflutungsbereich im Mittel um 2,0 °C kälter, die Luftfeuchte, ausgedrückt durch den Dampfdruck VP, dagegen um 3,6 hPa bzw. in der relativen Luftfeuchte um 19% höher. 94 Zeitlich gesehen reagierte die Albedo a am schnellsten auf die Veränderung der Oberfläche von Land zu Wasser. Sobald das Wasser die Messwiese bedeckte, wurde die Reflexion kurzwelliger Strahlung drastisch herabgesetzt und die Albedowerte sanken innerhalb eines Tages von a = 24% auf a = 8% ab (Abb. 8). 25 a (%) 20 15 10 5 0 1 4 7 1013 16 19 22 25 28 31 3 6 9 12 15 18 21 24 27 30 May - June 1999 Abb. 8: Tagesmittel der Albedo a am Standort „Rheinaue Plittersdorf“ während der Überflutungszeit vom 13. Mai bis zum 15. Juni 1999 (markiert) Die in den folgenden Tagen noch auftretenden Anstiege und Schwankungen der Albedo sind darauf zurückzuführen, dass Treibholz und schwimmende Pflanzenreste oft große Flächen bei Strömungshindernissen wie Bäume und Sträucher bedecken. Die Albedo blieb bis zum 15. Juni 1999 bei etwa 10%. Erst danach stiegen die a Werte mit zunehmender Austrocknung bei fallenden Wasserständen langsam wieder auf über 15% an. Der zeitliche Verlauf der Tagesmittel von Lufttemperatur Ta und Dampfdruck VP ist in den Abb. 9 und 10 dargestellt. Die Messwerte zeigen, dass die wasserbedeckte Aue bei Plittersdorf stets kühler als „Rastatt“ war und weiterhin eine Tendenz zu wachsenden Differenzen bei höherer Lufttemperatur bestand. Die für die Jahreszeit schon ungewöhnlich hohe Lufttemperatur Ende Mai 1999 war großräumig durch die Wetterlage hervorgerufen, wurde aber lokal durch den Untergrund stark modifiziert. Dadurch waren die Ta Tagesmittel in der Aue um ca. 3 °C niedriger, was durch den relativ kalten “Untergrund” verursacht wurde, d.h. die Wasseroberfläche verhinderte tagsüber eine Erwärmung des Erdbodens. Die auch außerhalb von Überflutungszeiten schon großen Unterschiede in der Luftfeuchte zwischen beiden Messorten steigerten sich an heißen Tagen während der Überschwemmung weiter auf eine Differenzen von bis zu 8 hPa im Tagesmittel (Abb. 10). Die über die Tagesmittel erkennbaren Besonderheiten der Aue treten bei einer Analyse der täglichen Maxima von Temperatur Ta und Dampfdruck VP noch deutlicher hervor. In der Abb. 11 sind die täglichen Ta Maxima (Halbstundenmittelwerte) für den genannten Zeitraum gegenübergestellt. Die Differenz bei niedrigerer Lufttemperatur, d.h. an 95 bewölkten und windigen Tagen, betrug etwa -1.8 °C. Sie erreichte ca. -5.0 °C an heißen Tagen. 24 Rastatt Plittersdorf 22 Ta (°C) 20 18 16 14 12 13 16 19 22 25 28 31 3 6 running days May - June 1999 9 12 15 Abb. 9: Tagesmittel der Lufttemperatur Ta während der Überflutungszeit (13. Mai bis zum 15. Juni 1999) an den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ 22 Plittersdorf Rastatt 20 VP (hPa) 18 16 14 12 10 8 6 13 16 19 22 25 28 31 3 6 9 running days May - June 1999 12 15 Abb. 10: Tagesmittel des Dampfdrucks während der Überflutungszeit (13. Mai bis zum 15. Juni 1999) an den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ Ebenso bedeutend waren die Unterschiede beim maximalen Dampfdruck VP, welcher am Standort „Rheinaue Plittersdorf“ ständig höher war als in Rastatt (Abb. 12). An einzelnen Tagen stieg diese Differenz bis auf 9 hPa an. Tatsächlich herrschte in der Aue bereits ein “Seenklima”, denn die überschwemmte Wasserfläche erreichte mit mehr als 96 3 km Breite und etwa 25 km Länge die Dimensionen mancher mitteleuropäischer Binnenseen. Ta, Plittersdorf - Ta,Rastatt (°C) 0 -1 -2 -3 -4 -5 -6 16 18 20 22 24 26 28 Ta, Rastatt (°C) 30 32 34 Abb. 11: Differenzen der täglichen Ta Maxima (Basis: Halbstundenmittelwerte) zwischen den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ in Abhängigkeit von den täglichen Ta Maxima an der Station „Rastatt“ während der Überflutungszeit (13. Mai bis zum 15. Juni 1999) an den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ VP,Plittersdorf - VP,Rastatt (hPa) 10 8 6 4 2 0 10 11 12 13 14 15 VP, Rastatt (hPa) 16 17 Abb. 12: Differenzen der täglichen VP Maxima (Basis: Halbstundenmittelwerte) zwischen den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ in Abhängigkeit von den täglichen VP Maxima an der Station „Rastatt“ während der Überflutungszeit (13. Mai bis zum 15. Juni 1999) an den Stationen „Rheinaue Plittersdorf“ und „Rastatt“ 97 Die skizzierten Besonderheiten des Klimas in der überfluteten Aue traten an einzelnen warmen und windschwachen Strahlungstagen noch verstärkt hervor, exemplarisch dargestellt für den 29. Mai 1999. Besonders auffällig ist, dass Ta in der Aue ein stark gekapptes nachmittägliches Maximum aufweist (Abb. 13), welches um fast 5 °C unter der Höchsttemperatur von Rastatt bleibt, und die Erwärmung am Vormittag stark verzögert stattfand, mit 1.6 °C/Stunde in der Aue gegenüber 2.2 °C/Stunde in „Rastatt“. 34 Rastatt Plittersdorf 32 30 28 (°C) 24 Ta 26 22 20 18 16 14 0 3 6 9 12 15 CET (hrs) 18 21 24 Abb. 13: Tagesgänge der Lufttemperatur Ta bei vollständiger Überflutung am Standort „Rheinaue Plittersdorf“ und am nicht überfluteten Vergleichsstandort „Rastatt“ am 29. Mai 1999 24 Plittersdorf Rastatt 22 (hPa) 18 VP 20 16 14 12 10 0 3 6 9 12 15 18 21 24 CET (hrs) Abb. 14: Tagesgänge der Lufttemperatur Ta bei vollständiger Überflutung am Standort „Rheinaue Plittersdorf“ und am nicht überfluteten Vergleichsstandort „Rastatt“ am 29. Mai 1999 98 Der Dampfdruck VP war am Standort „Rheinaue Plittersdorf“ bis zu 6 hPa höher als in „Rastatt“ (Abb. 14). Die hohe Verdunstungsrate führte nachts in der Aue bis zur Feuchtesättigung (RH = 100%), was für die Jahreszeit und Wetterlage außergewöhnlich ist. Überflutungen, die zu jeder Jahreszeit auftreten können, haben im Früh- und Hochsommer stärkere mikroklimatische Auswirkungen als im Winter. Sie verändern die Eigenschaften der Oberfläche, welche den Energieumsatz der bodennahen Atmosphäre steuert, im Sommerhalbjahr vergleichsweise stärker gegenüber einem normalen, trockenen Untergrund oder Vegetationsbestand. Eine Folge davon sind eine geringere Albedo, eine Absenkung der Oberflächentemperatur und damit der Lufttemperatur sowie eine stark erhöhte Luftfeuchte wegen höherer Verdunstung. 4.2 Sommertage ohne Überflutung Im Sommer sind heiße Tage (Ta,max > 30 °C) in der Oberrheinebene, bedingt durch die südliche und tiefe Lage, häufiger als sonst in Mitteleuropa. Als ein typischer Hochsommertag wurde der 21. Juli 1995 ausgewählt. Durch den Vergleich mit „Rastatt“ lässt sich aufzeigen, wie bei fast gleicher Stationslage im Zentrum des Rheingrabens, also mit identischem Mesoklima, durch verschiedene Oberflächenstrukturen erhebliche lokale Unterschiede in den einzelnen Klimaelementen entstehen. Während am Vormittag die Lufttemperatur Ta an beiden Standorten (Abb. 15) bis gegen 12 Uhr MEZ fast identische Werte annahm, erwärmte sich das trockenere Vorstadtgebiet am Nachmittag weiter auf 35 °C gegenüber 34 °C in der Aue. Bioklimatisch bedeutender ist nicht die höhere Maximaltemperatur, sondern vielmehr das weit in den Abend hinein andauernde höhere thermische Niveau, welches sich zwischen 18 und 21 Uhr MEZ in einer um mehr als 3 °C höheren Lufttemperatur in „Rastatt“ gegenüber der Aue ausdrückte. 36 Plittersdorf Rastatt 28 Ta (°C) 32 24 20 0 3 6 9 12 15 18 21 24 CET (hrs) Abb. 15: Tagesgänge der Lufttemperatur Ta an den Standorten „Rheinaue Plittersdorf“ und „Rastatt“ 21. Juli 1995 (heißer Tag) 99 35 Plittersdorf Rastatt 25 VP (hPa) 30 20 15 0 3 6 9 12 15 18 21 24 CET (hrs) Abb. 16: Tagesgänge des Dampfdrucks VP an den Standorten „Rheinaue Plittersdorf“ und „Rastatt“ 21. Juli 1995 (heißer Tag) Bei der Luftfeuchtigkeit (Abb. 16) wirkte sich dagegen das höhere Wasserdampfangebot in der bodennahen Luftschicht in der Aue aus. Besonders herausragend waren die Unterschiede beim Dampfdruck am Abend, wobei am Standort „Rheinaue Plittersdorf“ der höchste Wert während der gesamten Messzeit mit VP = 34 hPa erreicht wurde. Eine solche Luftfeuchte ist für Mitteleuropa außergewöhnlich. Die Ursachen sind das reichliche Wasserangebot zur Verdunstung von der feuchten Wiese und zusätzlich aus dem nahe gelegenen Auenwald mit seiner enormen Biomasse. In Rastatt wurden dagegen nur VP = 23 hPa gemessen, ein ebenfalls hoher Wert, aber dennoch ganz deutlich unter den “tropischen” Verhältnissen in der Rheinaue. Mesoklimatisch sind Werte von VP = 25 hPa als extreme Luftfeuchte anzusehen (REKLIP, 1995). Die hier beispielhaft gezeigte überaus starke Feuchtezunahme an einem Sommerabend ist charakteristisch für das Klima der feuchten Rheinaue und konnte in dieser Form an vielen Tagen im Sommer nachgewiesen werden. 5. Schlussfolgerung Die für grundlegende Klimavariable präsentierten Ergebnisse zeigen die besonderen Eigenschaften des Klimas in der Überflutungsaue des Oberrheins auf. Sie bilden die Randbedingung für biologische Prozesse und daraus resultierende Erscheinungsformen in Flora und Fauna, die sich in der Bezeichnung „strukturreiche Naturlandschaft“ widerspiegeln. Literatur BNN, 2007: Tageszeitung Badische Neueste Nachrichten vom 8. bis 11.08.2007, Karlsruhe. DISTER, E., 1985a: Auen-Land unter Wasser. - WWF Journal 2/85, 4-10. DISTER, E., 1985b: Taschenpolder als Hochwasserschutzmaßnahme am Oberrhein. - Geographische Rundschau 37, 241-247. 100 DISTER, E., 1988: Ökologie der mitteleuropäischen Auenwälder. - Die Auenwälder gestern und heute und morgen? Wilhelm-Münker-Stiftung, Heft 19, 6-30. GEIGER, R., 1961: Das Klima der bodennahen Luftschicht. - Vieweg Verlag, Braunschweig. Homagk, P., 1992: Zielkonflikt zwischen Hochwasserschutz und Ökologie? - Wasserwirtschaft 82, 32-36. Iziomon M.G., H. Mayer, W. Wicke, A. Matzarakis, 2001: Radiation balance over low-lying and mountainous areas in south-west Germany. - Theor. Appl. Climatol. 68, 219-231. KESSLER, A., 1983: Über die spezifischen Strahlungsumsätze verschiedener Oberflächen in Mitteleuropa, klimatologisch betrachtet. - Wiss. Ber. Meteor. Inst. Univ. Karlsruhe Nr.4, 101-112. KUTTER, S., V. SPÄTH, 1993: Rheinauen - Bedrohtes Paradies am Oberrhein. - Verlag G. Braun, Karlsruhe. LEHLE, M., 1985: Hochwasserschutz am Rhein für den Raum Mannheim. - Wasserwirtschaft 75, 11-14 LUBW, 2007: Landesanstalt für Umwelt, Messungen und Naturschutz, Baden-Württemberg, Hochwasser Vorhersage Zentrale (HVZ ) http://www.lubw.baden-wuerttemberg.de/public /hvz/ MAYER, H., D. AHRENS, 1997: Kennzeichen des Klimas in der Überflutungsaue des Rheins. Wetter und Leben 49, 89-105. REKLIP, 1987: Wissenschaftlicher Plan der Programmkommission, Karlsruhe. REKLIP, 1995: Klimaatlas, Oberrhein Mitte-Süd. - Verlag Corpür, Offenbach, Zürich, Strasbourg. ROST, J., 2004: Vergleichende Analyse der Energiebilanz zweier Untersuchungsflächen der Landnutzungen „Grasland“ und „Wald“ in der südlichen Oberrheinebene. – Ber. Meteor. Inst. Univ. Freiburg Nr. 14. UHRECKÝ, I., Z. SMOLIK, V. HAVLICEK, R. MRKVA, 1985: Radiation, temperature and rainfall regimes of the floodplain forest ecosystem. In: M. Penka, M. Vyscot, E. Klimo and F. Vasicek (eds.), Floodplain Forest Ecosystem I, Elsevier, New York, 33-51. WWF, 1995: Gebietskarte der Rastatter Rheinaue 1:25000.- Auen-Institut Rastatt. Anschrift der Autoren: Dr. Dieter Ahrens ([email protected]) Ing. (grad.) Werner Möhle Landesanstalt für Umwelt, Messungen und Naturschutz Baden-Württemberg Postfach 10 01 63, D-76231 Karlsruhe, Deutschland 101 KLIMES - a joint research project on human thermal comfort in cities Helmut Mayer Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Abstract Due to results of regional climate simulations for Central Europe, the likelihood is very high that not only the near-surface air temperature is increasing but also an intensification of extreme heat waves in summer will occur. The change of the large-scale thermal background conditions is strengthened in cities by their specific urban climate. Against this background, the demand for human-biometeorologically based concepts is continuously increasing in town planning, by which the stronger impairments of human thermal comfort for citizens in the future can be minimised. A comprehensive analysis of human thermal comfort within cities during large-scale heat conditions requires a coordinated combination of different methods: (i) experimental investigations and numerical simulations to calculate comfort-relevant thermal assessment indices, (ii) questionnaires with respect to the individual perception of the local thermal environment and (iii) monitoring of behaviour patterns of citizens in urban open spaces. The general objective of the BMBF joint research project KLIMES is the development of urbanistic concepts to mitigate the impacts of extreme heat on citizens. The fundamentals are worked out in four KLIMES subprojects. Within the scope of the subproject KLIMES ALUF-2, experimental investigations on human thermal comfort in different urban quarters in Freiburg, the warmest city in Germany, are conducted on typical summer days. The internationally used physiologically equivalent temperature is applied as thermo-physiologically significant assessment index. The investigation design as well as exemplary results obtained for the selected site “Rieselfeld” are presented and discussed. KLIMES – ein Verbundprojekt zum thermischen Komfort für Menschen in Städten Zusammenfassung Aufgrund von Ergebnissen aus regionalen Klimasimulationen ist die Wahrscheinlichkeit sehr hoch, dass in Mitteleuropa nicht nur die bodennahe Lufttemperatur ansteigt, sondern dass es auch zu einer Intensivierung von extremen Hitzeperioden im Sommer kommt. Die Veränderungen der großräumigen thermischen Hintergrundbedingungen werden in Städten durch ihr eigenes Stadtklima noch verschärft. Vor diesem Hintergrund steigt in der Stadtplanung die Anforderung nach human-biometeorologisch basierten Strategien stetig an, mit denen die zukünftig stärkeren Beeinträchtigungen des thermischen Komforts für Menschen in der Stadt möglichst gering gehalten werden können. Eine umfassende Analyse des thermischen Komforts bei großräumiger Hitze erfordert eine aufeinander abgestimmte Kombination von verschiedenen Methoden: (i) experimentelle Untersuchungen und numerische Simulationen zur Berechnung von komfortrelevanten thermischen Bewertungsindizes, (ii) Umfragen zur individuellen Wahrnehmung der lokalen thermischen Umgebungsbedingungen und (iii) Beobachtungen von Verhaltensmustern von Menschen im städtischen Freiraum. Das BMBF Verbundprojekt KLIMES hat als Ziel, städtebauliche Entwurfsbausteine zur Abpufferung der Auswirkungen von extremer Hitze auf Menschen in der Stadt zu entwickeln. Die Grundlagen dafür werden in vier KLIMES Teilprojekten erarbeitet. Im Teilprojekt KLIMES ALUF-2 werden an typischen Sommertagen experimentelle Untersuchungen zum thermischen Komfort in verschiedenen Stadtquartieren in Freiburg, der wärmsten Stadt Deutschlands, durchgeführt. Als thermischer Bewertungsindex wird die international angewandte physiologisch 102 äquivalente Temperatur verwendet. Das Untersuchungsdesign sowie exemplarische Ergebnisse für den ausgewählten Standort „Rieselfeld“ werden präsentiert und diskutiert. 1. Introduction Due to their specific features, cities modify the large-scale weather conditions and thus form a distinct urban climate (e.g. ARNFIELD, 2003; MILLS, 2007). Its most well-known phenomenon is represented by the urban heat island (e.g. GRIMMOND, 2006; OKE, 2006) or urban heat archipelago, if the intra-urban thermal conditions are of interest (MAYER, 1988). Large-scale heat in summer is intensified within cities and, therefore, affects efficiency, well-being and health of citizens. The two heat waves in Central Europe in 2003, which are presented in Fig. 1 for the city of Freiburg (SW Germany), are examples for extreme heat in summer. The heat waves covered almost the whole June 2003 and the first half of August 2003. Freiburg 35 heat waves 2003 30 25 Ta (°C) 20 Ta,1961-1990 + σ Ta , 2003 15 10 Ta,1961-1990 - σ 5 0 Ta, 1961-1990 1961-1990: T a,mean = 10.7 °C; ♦ = 6.5 °C 2003: T a,mean = 12.7 °C; ♦ = 9.0 °C -5 -10 0 30 60 90 120 150 180 DOY 210 240 270 300 330 360 Fig. 1: Daily mean values of air temperature Ta in Freiburg (SW Germany) in 2003 and averaged over the climate standard period 1961-1990, σ: standard deviation, data source: Deutscher Wetterdienst DWD Regional climate models predict the likelihood that heat waves will be more frequent, more intense and longer lasting in the future (e.g. MEEHL and TEBALDI, 2004). Therefore, methods of town planning, which are aimed for the optimisation of human thermal comfort within cities, become more and more important (e.g. ELIASSON, 2000). They must consider the limited area of action due to existing urban structures in central European cities. The joint research project KLIMES, which is introduced in this article, meets this demand. 103 The main objectives of this study are (i) to explain human thermal comfort in cities in general, (ii) to describe the coordinated design of different methodological approaches applied in KLIMES and (iii) to discuss preliminary results on the micro-scale variability of human thermal comfort within an urban street canyon. 2. Human thermal comfort Defined by the ASHRAE standard 55, which is equivalent to the British standard BS EN ISO 7730, human thermal comfort is a state of mind that expresses satisfaction of people with the thermal surroundings (HSE, 2008). Human thermal comfort describes a person's psychological state of mind and is usually referred to in terms of whether someone is feeling too hot or too cold. Human thermal comfort is achieved, when the heat generated by human metabolism is allowed to dissipate. This leads to a thermal equilibrium with the thermal surroundings. People react to the atmospheric environment through heat exchange, which is described in the human energy balance (e.g. HÖPPE, 1993). It takes into account all mechanisms of the heat exchange between the human body and its environment. The mechanisms are important in maintaining human thermal comfort. In order to ensure a body core temperature at around 37 °C, the mechanisms must keep heat production and heat loss in a state of equilibrium. Any heat gain or loss beyond this generates the sensation of thermal discomfort. As human thermal comfort depends on different factors, it is almost impossible to predict thermal comfort for individuals. Therefore, human thermal comfort is mostly related to a collective of people. According to the British Health & Safety Executive (HSE, 2008), 80% is a reasonable limit for the minimum number of people, who should be thermally comfortable in an environment. Based on the human energy balance, six physical basic factors determine human thermal comfort. They cooperate in a complex way, which is schematically shown in Fig. 2. Fig. 2: Model of the cooperation of six physical basic factors governing the human energy balance to determine human thermal comfort (according to HSE, 2008) Besides the physical factors, human thermal comfort depends on different personal factors like psychological or socio-cultural factors. They influence the adaptation and acclimatisation of people to climate and weather conditions. 104 3. Importance of human thermal comfort in applied urban climatology In general, a strong public interest in the quality of open spaces can be observed in cities. Therefore, the importance of making urban spaces attractive and accessible is increasing. Besides other features (e.g. air pollution, noise, …), human thermal comfort is a critical parameter for the use of open spaces in the urban environment. The use of urban open spaces for different human activities is more likely to increase if the outdoor environment is perceived as thermo-physiologically comfortable. In addition, a thermally comfortable outdoor environment has a positive impact on the indoor climate, which is particularly significant during extreme weather like heat waves. The maintenance of human thermal comfort represents an essential requirement for efficiency, well-being and health of citizens, i.e. human thermal comfort enhances the quality of life within cities (e.g. MAYER et al., 2008). 4. Methodology to analyse human thermal comfort The investigation of human thermal comfort represents a major task in human-biometeorology (e.g. HÖPPE, 1993, 2002). The analysis of human thermal comfort includes a coordinated design of - experimental investigations and numerical simulations, - field surveys (questionnaires and monitoring). human energy balance atmospheric environment ¾ ¾ ¾ ¾ ¾ comfort equation after FANGER ¾ MEMI model air temperature vapour pressure wind speed short- and long-wave radiation flux densities from the three-dimensional surroundings (Î mean radiant temperature) human beings ¾ ¾ ¾ ¾ activity heat transfer resistance of the clothing albedo, emissivity … questionnaires ¾ ¾ ¾ ¾ ENVI-met model BOTworld model IMEM model ... thermophysiological assessment indices ¾ ¾ ¾ ¾ ¾ ¾ ¾ predicted mean vote PMV physiologically equivalent temperature PET perceived temperature pt OUT_SET* (UTCI) thermophysiological variables (e.g. Tsk) … ¾ graded assessment scale ¾ … results (maps, frequency distributions, temporal variability, ...) Fig. 3: Scheme for the human-biometeorological analysis of the thermal environment; the universal thermal climate index UTCI is put in brackets, as currently it is still under development (according to MAYER, 2006) The general method to analyse the thermal environment in a humanbiometeorologically significant way is schematically presented in Fig. 3. To quantify the level of thermal comfort, thermo-physiological assessment indices are used, which are derived from human energy balance models developed in the current human- 105 biometeorology (e.g. MAYER, 1993). Various thermo-physiological assessment indices, which are based on different formulations for the fluxes of the human energy balance, are available worldwide for stationary and unsteady conditions (Fig. 3). Among these indices, the physiologically equivalent temperature PET (MAYER and HÖPPE, 1987) has turned out to be a widely used thermal index (THORSSON et al., 2004; JOHANSSON and EMMANUEL, 2006; OLIVEIRA and ANDRADE, 2007, LIN and MATZARAKIS, 2008). To determine PET, the meteorological variables air temperature, vapour pressure, wind speed and mean radiant temperature as a measure of the absorbed radiation heat must be available. Characteristics of human beings are often set constant, if thermal stress level at different sites should be compared. The procedure to determine PET as output of the MEMI model is described by MAYER and HÖPPE (1987) as well as HÖPPE (1999) in detail. PET provides a measure for the perception of heat by a collective of citizens, which is represented by standardised standing person. According to the methods in human-biometeorology, the reference height for PET is 1.1 m a.g.l. Therefore, the meteorological variables necessary to determine PET should be measured or simulated in the same height or related to it, if meteorological variables are available only in screen level (2 m a.g.l.). Field surveys are necessary as - real thermal sensations of citizens may slightly differ from those predicted by thermo-physiological indices, - about only 50% of the variance in subjective comfort evaluation can be explained by physical parameters (NIKOLOPOULOU and STEEMERS, 2003), - perception of the thermal environment strongly influences usage patterns, - effects of different geographical and climatic zones as well as different cultures on the perception of the thermal environment can be observed, - effects of adaptation and acclimatisation should be taken into account. Hence, the necessity follows that psychological parameters must be involved in the thermal assessment of the outdoor environment, as the psychological adaptation has become important. The main objectives of questionnaires are to get information on subjective human variables in the outdoor context, e.g. behaviour, thermal history, perception of the current thermal situation and usage of outdoor spaces by individuals. It should be noticed that many individuals form a collective of citizens. Results of questionnaires are especially necessary to take into account effects of adaptation and acclimatization, if the numerical values calculated for a thermo-physiological assessment index should be classified according to a graded verbal assessment scale. The main objective of monitoring is to get information on subjective human variables in the outdoor context, e.g. current behaviour or current use of different urban open spaces. 106 5. Joint research project KLIMES The joint research project “Development of strategies to mitigate enhanced heat stress in urban quarters due to regional climate change in Central Europe” - abbreviated by KLIMES (MAYER et al., 2008) - is conducted out by four German research groups. KLIMES is part of the research initiative “klimazwei” of the German Federal Ministry of Education and Research (BMBF). Based on an overview on the state-of-the-art in the planning-related urban human-biometeorology and after an identification of deficits, working hypotheses were derived for KLIMES, which led to its general objectives (KATZSCHNER et al., 2007a, b; MAYER et al., 2008): - update of human-biometeorological methods available to quantify the perception of heat by citizens under current and future climate conditions, - quantification of the perception of human thermal comfort (discomfort) in different urban quarters (outdoors and indoors) during extreme summer heat, - development and verification of urbanistic strategies based on human-biometeorological results to mitigate the negative impacts of climate trends and extreme weather on citizens in different urban quarters (optimisation of human thermal comfort under consideration of objectives of environmental protection, e.g. abandonment of electric air conditioning), - synthesis of all results in a guideline for town planning orientated to the challenges due to regional climate change in Central Europe. To achieve the objectives, a coordinated design of different methods is applied in KLIMES (Fig. 4): - experimental investigations on the perception of heat by citizens in different urban quarters in Freiburg (SW Germany), which is the warmest city in Germany, - questionnaires about citizens' current perception of heat under consideration of their thermal history and their use of open spaces (see also KNEZ and THORSSON, 2006; KATZSCHNER et al. 2007a, b), - model-based simulations of human thermal comfort in different urban quarters (outdoors and indoors) under current and future thermal conditions using the stationary model ENVI-met (BRUSE and FLEER, 1998) and the unsteady model BOTworld (BRUSE, 2007), - development of human-biometeorologically based strategies for town planning to optimise human thermal comfort outdoors and indoors against the background of predictions on heat in the future, - permanent dialogue with the planning practice and the public. The joint research project KLIMES is coordinated by the Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany (KLIMES ALUF-1). The KLIMES subprojects have different coordinated aims (Fig. 4). They are conducted by: - Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany (KLIMES ALUF-2), - Department of Environmental Meteorology, University of Kassel, Germany (KLIMES KAS-1), 107 - Environmental Modelling Group of the Institute for Geography, University of Mainz, Germany (KLIMES JGUM), - Department of Urban Development, University of Kassel, Germany (KLIMES KAS2). Funded by BMBF Joint research project KLIMES Coordination: KLIMES ALUF-1 Hypotheses Objectives Working plan Dialogue with experts and the public Design of thermal index, experimental investigations on human thermal comfort: Experimental investigations and interviews with people on human thermal comfort: KLIMES ALUF2 KLIMES KAS-1 Transformation of results on human thermal comfort into urbanistic concepts: KLIMES KAS-2 Simulations on the human thermal comfort under current and future climate conditions: KLIMES JGUM Concept modules for town planning aimed at regional climate change Fig. 4: Scheme for the operational procedures in the joint research project KLIMES containing the different KLIMES subprojects 6. Experimental design of the subproject KLIMES ALUF-2 Regarding the demands of town planning, investigation sites in Freiburg have been carefully selected by all KLIMES research groups. For each KLIMES site, the experimental design of the subproject KLIMES ALUF-2 consists of a stationary humanbiometeoro-logical station (Fig. 5) and a mobile human-biometeorological station (Fig. 6). As the mean radiant temperature Tmrt was calculated on the basis of measured shortand long-wave radiation flux densities from the three-dimensional surroundings of a standing person according to Höppe (1992) and THORSSON et al. (2007), both types of human-biometeorological stations include appropriate systems to measure the incoming and outgoing, short- and long-wave radiation flux densities in vertical directions as well as the short- and long-wave radiation flux densities from the four main horizontal directions. Technical details of the experimental equipment are compiled in Table 1. 108 Fig. 5: Stationary human-biometeorological station used in experimental investigations on human thermal comfort in Freiburg (SW Germany) within the scope of the subproject KLIMES ALUF-2 Fig. 6: Mobile human-biometeorological station (version 2008) used in experimental investigations on human thermal comfort in Freiburg (SW Germany) within the scope of the subproject KLIMES ALUF-2 109 Table 1: Instrumentation of the stationary and mobile human-biometeorological measurement stations used in 2007 in the subproject KLIMES ALUF-2 (according to MAYER et al., 2008) meteorological variable stationary mobile air temperature Humicap HMP45D, Vaisala Comp., 1.1 m a.g.l. via psychrometer principle containing Pt 100, Friedrichs Comp., 2 m a.g.l. (since 2008: 1.2 m a.g.l.) vapour pressure Humicap HMP45D, Vaisala Comp., 1.1. m a.g.l. via psychrometer principle, Friedrichs Comp., 2 m a.g.l. (since 2008: 1.2 m a.g.l.) wind speed 3-D sonic anemometer 81000 VRE, Fa. Young, 1.2 m a.g.l. hot-wire anemeometer, Dantec Comp., 2 m a.g.l. (since 2008: 1.2 m a.g.l.) short-wave radiation CM3 (as part of CNR1), Kipp & Zonen Comp., 1.1 m a.g.l. CM21, Kipp & Zonen Comp., 1.1 m a.g.l. long-wave radiation CG3 (as part of CNR1), Kipp & Zonen Comp., 1.1 m a.g.l. CG1, Kipp & Zonen Comp., 1.1 m a.g.l. data recording CR3000 datalogger + AM16/32 multiplexer, Campbell Comp. manually scan interval averaging period 5s instantaneous 1 min 1 min KLIMES ALUF-2 sites in different urban quarters in Freiburg Rieselfeld Fig: 7: Sites with experimental investigations on human thermal comfort in summer 2007 in different urban quarters in Freiburg (KLIMES ALUF-2) 110 The map in Fig. 7 shows an overview on the different urban quarters in Freiburg, where experimental investigations on human thermal comfort were carried out in summer 2007. The stationary human-biometeorological station and the points of the mobile human-biometeorological measurements were always located in urban street canyons. Fish-eye photos were used to calculate the sky view factor SVF at each site. In addition, the aspect ratio H/W (H: building height, W: street width) and the orientation of the street canyons to the sun were determined for each site. 7. Exemplary results of experimental investigations on human thermal comfort Within the subproject KLIMES ALUF-2 experimental investigations on human thermal comfort were conducted in the urban quarter “Rieselfeld” in Freiburg. “Rieselfeld” is a neigh-bourhood in the western part of Freiburg planned and built since the nineties of the last century. It is characterised by modern, three- to four-storeyed block buildings. Some streets are tree-lined and have small grass-covered front gardens. To get a picture of the urban structure of Rieselfeld, Fig. 8 shows the urban street canyon, where humanbiometeorological measurements were carried out on 19 June 2007, which was a typical summer day related to the climate conditions in Freiburg. Fig: 8: Stationary human-biometeorological station in an urban street canyon at the “Rieselfeld” site in Freiburg on 19 June 2007 In the following, selected results based on human-biometeorological measurements at the stationary station and at the measurement point no. 1 (= MP1) of the mobile investigations are discussed as examples for the spatial variability of human thermal comfort within urban street canyons. Both micro-sites were located within the same NW-SE oriented street canyon, but on the opposite sidewalks (Table 2). 111 Table 2: Characteristics of the stationary micro-site and measuring point no. 1 (MP1) in the “Rieselfeld” quarter in Freiburg characteristics stationary micro-site MP1 NW-SE NW-SE SW oriented sidewalk NE oriented sidewalk street canyon H/W 0.49 0.49 SVF at the specific micro-site 0.51 0.47 front garden yes yes ≈3m ≈3 m orientation of the urban street canyon specific location horizontal distance to the nearest house wall Due to the investigation design, the results in Figs. 9 to 11 contain 1-hr mean values for the stationary micro-site and 1-min mean values for MP1. Despite the different temporal resolution, a typical behaviour of the presented variables can be recognised, which reflects the influence of the orientation of both micro-sites. The values for the near-surface air temperature Ta at both micro-sites (Fig. 9) were similar in the morning. As MP1 was shaded while the stationary micro-site was still in the sun, the patterns of Ta dispersed and Ta was higher at the stationary micro-site. The peak Ta difference between both opposite micro-sites was approximately 2 °C. Freiburg, Rieselfeld, 19-06-2007 32 30 Ta (°C) 28 26 24 22 20 8:00 stationary, NW-SE street canyon, H/W=0.49, SW exp. sidewalk, SVF=0.51 mobile, MP1, NW-SE street canyon, H/W=0.49, NE exp. sidewalk, SVF=0.47 10:00 12:00 14:00 16:00 18:00 20:00 22:00 0:00 CET (hrs) Fig: 9: Near-surface air temperature Ta at the stationary micro-site and at MP1 of mobile investigations within the “Rieselfeld” site in Freiburg on 19 June 2007 112 Freiburg, Rieselfeld, 19-06-2007 70 60 Tmrt (°C) 50 40 30 20 stationary, NW-SE street canyon, H/W=0.49, SW exp. sidewalk, SVF=0.51 mobile, MP1, NW-SE street canyon, H/W=0.49, NE exp. sidewalk, SVF=0.47 10 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 0:00 CET (hrs) Fig: 10: Mean radiant temperature Tmrt at the stationary micro-site and at MP1 of mobile investigations within the “Rieselfeld” site in Freiburg on 19 June 2007 Freiburg, Rieselfeld, 19-06-2007 50 PET (°C) 40 30 20 stationary, NW-SE street canyon, H/W=0.49, SW exp. sidewalk, SVF=0.51 mobile, MP1, NW-SE street canyon, H/W=0.49, NE exp. sidewalk, SVF=0.47 10 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 0:00 CET (hrs) Fig: 11: Physiologically equivalent temperature PET at the stationary micro-site and at MP1 of mobile investigations within the “Rieselfeld” site in Freiburg on 19 June 2007 113 The difference between the daily patterns of the mean radiant temperature Tmrt at both micro-sites (Fig. 10) was more pronounced than for Ta. As soon as MP1 was shaded, Tmrt decreased directly at this micro-site. When Tmrt reached its peak value (63.4 °C) at the stationary micro-site in the afternoon, the maximum Tmrt difference (≈ 33 °C) between the sunny stationary micro-site and the shaded micro-site MP1 was observed. As Tmrt was calculated for a standardised standing person, the absorbed short- and longwave radiation flux densities from the four main horizontal directions basically determined Tmrt (ALI-TOUDERT and MAYER, 2007a; MAYER et al., 2008). However, they are influenced on sunny days by the solar radiation. Therefore, the shading of the solar radiation caused a direct decrease of Tmrt. From regression analyses performed for sites, which are not influenced by sea breezes, it is known (e.g. MAYER et al., 2008) that on typical summer days the physiologically equivalent temperature PET has the strongest correlation with Tmrt. Therefore, the similarity of the patterns for Tmrt and PET (Fig. 11) is not surprising. PET exceeded the 40 °C level during the whole afternoon at the sunny stationary micro-site, which can be interpreted as strong heat stress. In contrast, at the shaded MP1 PET was in the same period between 25 and 30 °C, which indicates a mean thermal sensation of “slightly warm”. This distinct micro-scale reduction of the perceived heat level by citizens resulted mainly from the lower radiation heat, which characterised the shaded situation. Height-to-width ratio H/W Orientation E-W H W 0.5 1 2 3 4 N-S NE-SW 1.2 m 2m street: 8 x 1 m NW-SE Fig: 12: Scenarios for the design of urban street canyons as basis for simulations of human thermal comfort by use of the ENVI-met model (ALI-TOUDERT, 2005), Fig. made by ALI-TOUDERT (pers. comm., 2007) 114 8. Simulations of human thermal comfort The results of simulations of human thermal comfort in different urban street canyons carried out by ALI-TOUDERT (2005) with the micro-scale ENVI-met model (BRUSE and FLEER, 1998) and partly discussed in ALI-TOUDERT and MAYER (2006, 2007b) represented important background conditions for the design and the objectives of the joint research project KLIMES. The simulations are related to various scenarios for the design of urban street canyons, which are described in Fig. 12. The simulations were conducted for hot and dry climate conditions, which are represented by the city of Ghardaia, Algeria, on 1 August. To identify the level of human thermal comfort, the thermal index PET was applied. As results tow-dimensional patterns of PET are available showing the behaviour of PET in 1.2 m a.g.l. during the daylight hours (vertical axis) in dependence on the width of the street canyon (horizontal axis). Fig: 13: Patterns of simulated physiologically equivalent temperature PET in 1.2 m a.g.l. during the daylight hours in urban street canyons dependent on different orientations and aspect ratios H/W, simulations carried out for hot and dry climate (Ghardaia, Algeria, 1 August) by use of the ENVI-met model (ALITOUDERT, 2005), Fig. made by ALI-TOUDERT (pers. comm., 2007) As example of the comprehensive results by ALI-TOUDERT (2005), Fig. 13 contains the PET patterns for an East-West oriented street canyon (upper row) and a North-South oriented street canyon (lower row) dependent on the aspect ratio H/W. It varies from H/W=0.5 (left column) to H/W=4 (right column). 115 Based on the areas with a dark-red colour, which indicates extreme heat stress, the results in Fig. 13 show, that extreme heat stress during the daylight hours characterises the thermal conditions in an urban street canyon, which is East-West oriented and has an aspect ratio of H/W=0.5. In contrast, heat stress is strongly reduced even under hot and dry climate background conditions, if the street canyon is North-South oriented and very deep, i.e. the aspect ratio is relatively high (e.g. H/W=4 in Fig. 13). 9. Conclusions Thermal comfort represents an indispensable prerequisite for the quality of life within cities. It is essential for efficiency, well-being and health of citizens. Human thermal comfort within urban structures depends on the large-scale meteorological background conditions, which are modified within cities by their specific meteorological features in a way that the thermal level is enhanced. Based on simulations of the regional climate in SW Germany, the likelihood of longer lasting, more intense and more frequent heat waves has a pronounced reliability. This changed background conditions lead to more thermal stress situations for citizens, which can be quantified by the application of thermal indexes. Elevated thermal stress represents an important challenge for town planning to develop concepts, which enable human thermal comfort in urban structures despite of thermal background stress and limited possibilities for town planning due to widely existing urban structures in central European cities. The joint research project KLIMES is integrated in the way how to solve this problem. Focussing on mitigation of heat stress for citizens, aspects of (i) solar access in winter, (ii) air pollution control, (iii) noise control, (iv) unpleasant odour, (v) wind discomfort, and (vi) objectives of environmental protection should keep clearly in mind. Effects of hot and dry meteorological background conditions may be mitigated by suited planning concepts. If, however, the current meteorological background conditions are characterised by heat and a relatively high humidity, which leads to a specific thermal sensation called “sultry” in a colloquial language, the opportunities of town planning to maintain human thermal comfort are more limited than for hot and dry heat, where shading effects can be taken into account. They are not available in a similar extent during hot and humid weather, as the sky is frequently overcast during humid heat periods in Central Europe. Acknowledgement The author is indebted to the German Federal Ministry of Education and Research (BMBF) for funding the research projects KLIMES ALUF-1 and KLIMES ALUF-2 (FZK: 01LS05020). Many thanks to Jutta Holst for her human-biometeorological investigations at the “Rieselfeld” site and Fazia Ali-Toudert for her simulations of human thermal comfort by use of the ENVImet model. References ALI-TOUDERT, F., 2005: Dependence of outdoor thermal comfort on street design in hot and dry climate. - Ber. Meteor. Inst. Univ. Freiburg Nr. 15. ALI-TOUDERT, F., H. MAYER, 2006: Numerical study on the effects of aspect ratio and orientation of an urban street canyon on outdoor thermal comfort in hot and dry climate. - Building and Environment 41, 94-108. 116 ALI-TOUDERT, F., H. MAYER, 2007a: Thermal comfort in an east-west oriented street canyon in Freiburg (Germany) under hot summer conditions. - Theor. Appl. Climatol. 87, 223-237. ALI-TOUDERT, F., H. MAYER, 2007b: Effects of asymmetry, galleries, overhanging façades and vegetation on thermal comfort in urban street canyons. - Solar Energy. 81, 742-754. ARNFIELD, A.J., 2003: Two decades of urban climate research: A review of turbulence, exchanges of energy and water, and the urban heat island. - Int. J. Climatol. 23, 1-26. BRUSE, M., 2007: Simulating human thermal comfort and resulting usage patterns of urban open spaces with a multi-agent system. - Proc. 24rd Int. Conf. Passive and Low Energy Architecture (PLEA) 2007, 491-498. BRUSE, M., H. FLEER, 1998: Simulating surface-plant-air interactions inside urban environments with a three dimensional numerical model. - Environ. Modell. Softw. 13, 373-384. ELIASSON, I., 2000: The use of climate knowledge in urban planning. - Landscape Urban Plan. 48, 31-44. GRIMMOND, C.S.B., 2006: Progress in measuring and observing the urban atmosphere. - Theor. Appl. Climatol. 84, 3-22. HÖPPE, P., 1992: Ein neues Verfahren zur Bestimmung der mittleren Strahlungstemperatur im Freien. - Wetter und Leben 44, 147-151. HÖPPE, P., 1993: Heat balance modelling. - Experientia 49, 741-746. HÖPPE, P., 1999: The physiological equivalent temperature - a universal index for the biometeorological assessment of the thermal environment. - Int. J. Biometeorol. 43, 71-75. HÖPPE, P., 2002: Different aspects of assessing indoor and outdoor thermal comfort. - Energy and Buildings 34, 661-665. HSE, 2008: http://www.hse.gov.uk/temperature/thermal/explained.htm. JOHANSSON, E., R. EMMANUEL, 2006: The influence of urban design on outdoor thermal comfort in the hot, humid city of Colombo, Sri Lanka. - Int. J. Biometeorol. 51, 119-133. KATZSCHNER, L., M. BRUSE, CH. DREY, H. MAYER, 2007a: Untersuchung des thermischen komforts zur Abpuffereung von Hitze mittels eines städtebaulichen Entwurfs (BMBF Verbundprojekt KLIMES) - Ber. Meteor. Inst. Univ. Freiburg Nr. 16, 37-42. KATZSCHNER, L., H. MAYER, CH. DREY, M. BRUSE, 2007b: Strategies and concepts for thermal comfort discussions in urban planning to mitigate the impacts of climate extremes. - Proc. 24rd Int. Conf. Passive and Low Energy Architecture (PLEA) 2007, 103-108. KNEZ, I., S. THORSSON, 2006: Influences of culture and environmental attitude on thermal, emotional and perceptual evaluations of a public square. - Int. J. Biometeorol. 50, 258-268. LIN, T.-P., A. MATZARAKIS, 2008: Tourism climate and thermal comfort in Sun Moon Lake, Taiwan. - Int. J. Biometeorol. 52, 281-290. MAYER, H., 1988: Results from the research program ‘STADTKLIMA BAYERN’ for urban planning. - Energy and Buildings 11, 115-121. MAYER, H., 1993: Urban bioclimatology. - Experientia 49, 957-963. MAYER, H., 2006: Indizes zur human-biometeorologischen Bewertung der thermischen und lufthygienischen Komponente des Klimas. - Gefahrstoffe-Reinhaltung der Luft 66, 165-174. MAYER, H., P. HÖPPE, 1987: Thermal comfort of man in different urban environments. - Theor. Appl. Climatol. 38, 43-49. 117 MAYER, H., J. HOLST, P. DOSTAL, F. IMBERY, D. SCHINDLER, 2008: Human thermal comfort in summer within an urban street canyon in Central Europe. - Meteorol. Zeitschrift 17, in press. MEEHL, G.A., C. TEBALDI, 2004: More intense, more frequent, and longer lasting heat waves in the 21st century. - Science 305, 994-997. MILLS, G., 2007: Cities as agents of global change. - Int. J. Climatol. 27, 1849-1857. NIKOLOPOULOU, M., K. STEEMERS, 2003: Thermal comfort and psychological adaptation as a guide for designing urban spaces. - Energy and Buildings 35, 95-101. OKE, T.R., 2006: Towards better scientific communication in urban climate. - Theor. Appl. Climatol. 84, 179-190. OLIVEIRA, S., H. ANDRADE, 2007: An initial assessment of the bioclimatic comfort in an outdoor public space in Lisbon. - Int. J. Biometeorol. 52, 69-84. THORSSON, S., M. LINDQVIST, S. LINDQVIST, 2004: Thermal bioclimatic conditions and patterns of behaviour in an urban park in Göteborg, Sweden. - Int. J. Biometeorol. 48, 149-156. THORSSON, S., F. LINDBERG, I. ELIASSON, B. HOLMER, 2007: Different methods for estimating the mean radiant temperature in an outdoor urban setting. - Int. J. Climatol. 27, 1983-1993. Author's address: Prof. Dr. Helmut Mayer ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 118 119 Importance of urban meteorological stations the example of Freiburg, Germany Andreas Matzarakis and Helmut Mayer Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Abstract For more than two decades modern data collection systems have been used. The internet and the development of networks have significantly reduced the temporal delay between meteorological measurements and the visualization of measured variables. However, not only the free of charge, current information of the public on the meteorological conditions is of interest. The use of data from meteorological stations is also extremely valuable for research and education. The urban meteorological station Freiburg, which is run since May 1999 by the Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany represents an example for multiple uses. Meteorological information is provided by the station as current 10-min mean values and 10-min totals of precipitation, respectively. Bedeutung von meteorologischen Stadtstationen - das Beispiel Freiburg, Deutschland Zusammenfassung Moderne Datenerfassungssysteme, die schon seit mehr als zwei Jahrzehnten im Einsatz sind, sowie das Internet und die Entwicklung von Netzwerken haben es ermöglicht, die Visualisierung und Präsentation von meteorologischen Variablen fast zeitgleich mit ihren Messungen durchzuführen. Aber nicht nur die freie, aktuelle Information der Öffentlichkeit über die meteorologischen Bedingungen ist von Interesse und Bedeutung. Der Einsatz von meteorologischen Stationen ist auch für Forschung und Lehre sehr hilfreich. Die Meteorologische Stadtstation Freiburg, die das Meteorologische Institut der Albert-Ludwigs-Universität Freiburg seit Mai 1999 betreibt, ist ein derartiges Beispiel. Die dort gemessenen meteorologischen Daten werden als aktuelle 10-Minuten Mittelwerte bzw. 10-Minuten Summen beim Niederschlag im Internet bereitgestellt. 1. Introduction and aim The demand of the public for meteorological information is very high. Beside this, the demand for meteorological information is also of interest in different applications, e.g. in applied urban climatology (MATZARAKIS, 2001). Due to their objectives, synoptic and climate stations of national and other weather services are mostly located in areas, which usually do not represent the conditions characteristics of urban conditions. Despite of urban heat island and urban moisture excess, air temperature and air humidity are the meteorological variables showing a relatively reduced spatial variability at screen level. In contrast, the spatial pattern of radiation flux densities and wind conditions strongly responds to the three-dimensional urban structure. Therefore, measurements of radiation flux densities and wind speed are conducted within urban structures only for specific purposes (e.g. human-biometeorological assessments), while they are monitored above the urban roof level in the usual case. As a consequence, reference stations for the urban meteorological conditions are mostly set up on the top of build- 120 ings, which are higher than the mean roof level of a city, i.e. the meteorological measurements are conducted above the urban canopy layer (OKE, 2007). The setup and representative location of an urban meteorological station is one important question. The processing of the data is a technical one and the presentation of the current data might be most important. For the latter, the existence and use of the internet make it easy to visualize current meteorological data and information, which allows everyone to be informed about the current meteorological conditions. The technical part including the processing and visualization of the data can be solved by use of commercial software packages offered by meteorological instrument manufacturers or data logger companies. The other possibility is to use program languages and to develop software and html packages for data processing and internet visualization. Also, an urban meteorological station has to be planned and run in a long-term continuous way, which requires permanent service, update and calibration of meteorological instruments. Keeping all these requirements and problems in mind, the Meteorological Institute, Albert-Ludwigs-University of Freiburg (Germany), established an urban meteorological station Freiburg on the top of the chemistry building, which is one of the tallest buildings in Freiburg (Fig. 1). The meteorological measurements started in May 1999 for testing purposes and in September 1999 with full data processing and analyses. Fig. 1: Location of the urban meteorological station Freiburg, run by the Meteorological Institute, Albert-Ludwigs-University of Freiburg (source: Microsoft Virtual Earth) The idea behind the installation of an urban meteorological station was: • provision of current meteorological information of Freiburg in a clear way for the public and administration, 121 • provision of meteorological data for the education of students in meteorology and climatology, • provision of meteorological data for scientific studies in meteorology and climatology, • use as an urban meteorological anchor station. 2. Instrumentation and data processing Located in the northern part of the city centre of Freiburg (272 m a.s.l.), the urban meteorological station Freiburg was installed on the roof of the chemistry building (51 m a.g.l.) of the Albert-Ludwigs-University of Freiburg. The website of the station is available under http://www.mif.uni-freiburg.de up to now. The horizon is only limited in the east by the Black Forest. Currently, the following meteorological parameters are continuously measured: • incoming short-wave radiation in 2 m a.g.l. (pyranometer, type CM21 by Kipp & Zonen Company), • air temperature in 2 m a.g.l. (electrically ventilated Pt100 probe according to Frankenberger, self construction and production of the Meteorological Institute, Albert-Ludwigs-University of Freiburg), • air humidity in 2 m a.g.l. (principle: electrically ventilated psychrometer according to Fankenberger, self construction and production of the Meteorological Institute, Albert-Ludwigs-University of Freiburg), • horizontal wind speed and wind direction in 10 m a.g.l. using a combined measurement system (Lambrecht Company), • precipitation (tipping-bucket rain gauge, Vaisala Company), • air pressure at roof level (aneroid, Vaisala Company). The meteorological measurements on the roof of the tall building are conducted in the transition area between the urban canopy layer and the urban boundary layer. All sensors are scanned every 30 s. A data logger (Campbell 21X) aggregates 10-min mean and 10-min totals (precipitation), respectively. The transfer of the data from the station to a server of the Meteorological Institute takes place every 10 minutes via the network of the Albert-Ludwigs-University of Freiburg. The data are controlled and further meteorological parameters (e.g. vapour pressure or relative humidity) are calculated. The files are stored as daily files and are available also as raw files. After storage, the meteorological data are prepared for the online visualisation, which is performed by own written Delphi programs. The results are available on the website of the station. The temporal delay between data collection or aggregation and visualization takes approximately 2 minutes. The website of the urban meteorological station Freiburg presents the results in different forms: • current meteorological data as 10-min mean values and 10-min totals for precipitation, respectively, in form of a table (Fig. 2), 122 • meteorological data as 10-min mean values and 10-min totals for precipitation, respectively, during the past 48 hours (in CET) in form of diagrams (Figs. 3 and 4). Fig. 2: Introducing website of the urban meteorological station Freiburg, run by the Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany (http://www.mif.uni-freiburg.de) Fig. 3: Example for the visualization of the horizontal wind speed v measured during the past 48 hours at the website of the urban meteorological station Freiburg 123 Fig. 4: Example for the visualization of the wind direction dd measured during the past 48 hours at the website of the urban meteorological station Freiburg 3. Applications 3.1 Research The data of the urban meteorological station Freiburg can be used as basic information on extreme weather. For example, Figs. 5 to 8 contain the patterns for horizontal wind speed, gust speed, wind direction and precipitation during the severe storm “Emma”, which passed Freiburg on 1 March 2008. An almost constant wind direction (around 210 °, i.e. around SW), a peak 10-min mean wind speed of approximately 18 m/s and peak gust speed of approximately 24 m/s characterized “Emma” in Freiburg. Precipitation was recorded in the period with high wind speed (Fig. 8). The visualization of meteorological variables during hot summer weather on 15 July 2007 represents another example (Figs. 9 to 15) for the application of the meteorological data measured at the urban meteorological station Freiburg to describe extreme weather. The pattern of the incoming short-wave radiation reflects the cloudless conditions. Peak 10-min mean air temperature reached 35 °C. The patterns for vapour pressure and relative humidity show that the humidity conditions were not in a stress range for people. The course of the wind direction above the mean roof level of Freiburg exhibits two major directions: between 240 ° and 330 ° during the daylight hours and around 120 ° in the night from 9 pm to 6am CET. This wind direction pattern, which is characteristic of the regional circulation system “Höllentäler”, can be clearly seen in the wind rose. The katabatic wind “Höllentäler” does not occur in each night, as distinct differences of the surface temperature between the elevated Black Forest and the lower Freiburg area caused mainly by high pressure weather are the necessary requirement for its formation. As indicated by the course of the horizontal wind speed, air flow during the “Höllentäler” situation is enhanced, but in a discontinuous way. 124 Fig. 5: Horizontal wind speed v during the severe storm “Emma” on 1 March 2008 measured at the urban meteorological station Freiburg Fig. 6: Horizontal gust speed v during the severe storm “Emma” on 1 March 2008 measured at the urban meteorological station Freiburg Fig. 7: Wind direction dd during the severe storm “Emma” on 1 March 2008 measured at the urban meteorological station Freiburg 125 Fig. 8: Precipitation N during the severe storm “Emma” on 1 March 2008 measured at the urban meteorological station Freiburg Fig. 9: Incoming short-wave radiation G during hot summer weather on 15 July 2007 measured at the urban meteorological station Freiburg Fig. 10: Air temperature Ta during hot summer weather on 15 July 2007 measured at the urban meteorological station Freiburg 126 Fig. 11: Vapour pressure VP during hot summer weather on 15 July 2007 measured at the urban meteorological station Freiburg Fig. 12: Relative humidity RH during hot summer weather on 15 July 2007 measured at the urban meteorological station Freiburg Fig. 13: Horizontal wind speed v during hot summer weather on 15 July 2007 measured at the urban meteorological station Freiburg 127 Fig. 14: Wind direction dd during hot summer weather on 15 July 2007 measured at the urban meteorological station Freiburg Fig. 15: Wind rose during hot summer weather on 15 July 2007 measured at the urban meteorological station Freiburg and coloured according to different 6-h periods The urban meteorological station Freiburg has the character of an anchor station for the city of Freiburg. Therefore, the collected meteorological data can also be used for different applications in urban meteorological modelling (MATZARAKIS et al., 2007; ALITOUDERT and MAYER, 2007). In addition, the meteorological data from the urban meteorological station Freiburg are used for quality controls of meteorological data recorded at other adjacent stations (IMBERY, 2005). Institutions of the Albert-Ludwigs-University Freiburg also use the data from the urban meteorological station Freiburg, e.g. for the calibration of meteorological sensors or climate impacts analyses. 3.2 Education In terms of education, the students in Freiburg are in close contact with the actual meteorological situation and are asked to discuss and explain weather conditions or specific phenomena like strong change of wind direction or air temperature drops from the starting semester. The students have the possibility not only to see and learn the way how meteorological measurement systems operate but they can also use the collected data. 128 The data are processed and analysed by the students in order to learn how to work with data files. The students calculate monthly averages, analyse the different meteorological variables statistically, draw figures and use this data for other purposes where meteorological data are necessary. The meteorological data of the urban meteorological station Freiburg are also available for scientific studies of students, e.g. bachelor or master thesis. 3.3 Administration and other demands The meteorological data of the urban meteorological station Freiburg are also demanded for administrative purposes, e.g. by the technical office of the Albert-LudwigsUniversity of Freiburg in order to calculate the heating conditions for the university buildings based on the outdoor air temperature. Finally, for the urban climate analysis of the municipality of Freiburg (for the master plan 2020), the data of the urban meteorological station Freiburg station were used to characterize the meteorological background situation in Freiburg (RÖCKLE et al., 2003), e.g. to obtain results on human thermal comfort. 4. Conclusions The free availability of quality controlled meteorological data is of significant importance due to different reasons, e.g. for scientific research, the education of students or the information of the public in the internet. The urban meteorological station Freiburg, run by the Meteorological Institute, Albert-Ludwigs-University of Freiburg, since May 1999, meets these requirements in an excellent way. References ALI-TOUDERT, F., H. MAYER, 2007: Thermal comfort in an east-west oriented street canyon in Freiburg (Germany) under hot summer conditions. - Theor. Appl. Climatol. 87, 223-237. IMBERY, F., 2005: Langjährige Variabilität der aerodynamischen Oberflächenrauhigkeit und Energieflüsse eines Kiefernwaldes in der südlichen Oberrheinebene (Hartheim). - Ber. Meteorol. Inst. Univ. Freiburg Nr. 14. MATZARAKIS, A., 2001: Die thermische Komponente des Stadtklimas. - Ber. Meteorol. Inst. Univ. Freiburg Nr. 6. MATZARAKIS, A., F. RUTZ, H. MAYER, 2007: Modelling radiation fluxes in simple and complex environments - application of the RayMan model. - Int. J. Biometeorol. 51, 323-334. OKE, T., 2007: Siting and Exposure of Meteorological Instruments at Urban Sites. In: C. BOREGO and A.-N. NORMAN (Eds.), Air pollution modelling and its application XVII. 615631. RÖCKLE, R., C.-J. RICHTER, H.-C. HÖFL, W. STEINICKE, M. STREIFENEDER, A. MATZARAKIS, 2003: Klimaanalyse Stadt Freiburg. - Auftraggeber Stadtplanungsamt der Stadt Freiburg. Authors' address: Prof. Dr. Andreas Matzarakis ([email protected]) Prof. Dr. Helmut Mayer ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 129 Dependence of the thermal urban climate on morphological variables Andreas Matzarakis and Helmut Mayer Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Abstract Urban meso- and micro-scale heat island conditions can be explained by energetic causes and properties of the urban surface and atmosphere. Urban morphological parameters, i.e. sealing factor, sky view factor or aspect ratio of street canyons, can also be used to explain their influences on the well-known phenomenon of urban climate, the urban heat island (UHI). Based on near-surface air temperature (Ta) data and statistical parameters that were calculated from measured Ta values, single and multiple regression analyses were carried out to determine the influence of the sealing factor and sky view factor on Ta characteristics. Ta values measured in the temporary climate network in Munich (South Germany), which was established from 1981-1985 within the scope of the research project “Urban Climate in Bavaria”, represent the basis. The results of this re-analysis show that for the single regression the highest correlation can be reached with the sealing factor for the annual amount of total heat and mean annual Ta. In the multiple regressions the highest regression coefficients are found for the annual number of warm days, total heat and mean annual Ta. Einfluss von morphologischen Variablen auf das thermische Stadtklima Zusammenfassung Die Bildung der urbanen Wärmeinsel (UHI) kann über energetische Ursachen und physikalische Eigenschaften der urbanen Oberflächen und Atmosphäre erklärt werden. Die Morphologie der urbanen Oberflächen und Strukturen, z.B. Versiegelungsgrad, Himmelsichtsfaktor (Sky View Faktor) oder das Höhen/Breitenverhältnis von Straßen, bilden eine adäquate Grundlage zur Beschreibung und Erklärung des bekanntesten Phänomens in der Stadtklimatologie, der urbanen Wärmeinsel (UHI). Auf der Grundlage von Messwerten für die bodennahe Lufttemperatur Ta sowie aus Ta abgeleiteten statistischen Kenngrößen wurden einfache und multiple Regressionsanalysen mit den Versiegelungsgrad und dem Sky View Faktor durchgeführt. Die verwendeten Ta Daten stammen aus dem temporären Klimamessnetz, das im Rahmen des Forschungsvorhabens STADTKLIMA BAYERN von 1981 bis 1985 in München eingerichtet worden war. Die Ergebnisse dieser Reanalyse zeigen, dass bei der einfachen Regression die höchsten Korrelationskoeffizienten zwischen Versiegelungsgrad einerseits sowie jährlicher Wärmesumme und mittlerem jährlichen Ta erzielt wurden. Bei der multiplen Regression ergab sich der höchste Korrelationskoeffizient für die jährliche Anzahl an warmen Tagen und mittlerem jährlichen Ta. 1. Introduction Urban climate and its well-known phenomenon, the urban heat island (UHI), are of important interest due to different reasons. Therefore, considerable research has been conducted in recent years (www.urbanclimate.net or www.urban-climate.org). A main focus was given in the analysis on the quantification of the urban heat island and other related phenomena, especially the differences of the near-surface air temperature Ta between the city and its rural hinterland (OKE, 1973, 1987, JOHNSON et al., 1991, MATZARAKIS, 2001, HELBIG et al., 1999, KUTTLER, 2004a, b). Specific urban spaces, e.g. parks, have been the focus of interest regarding several parameters and factors, e.g. Ta 130 differences, moisture conditions or human-biometeorological indices (UPMANIS et al., 1998, MAYER and MATZARAKIS, 2006, MAYER et al., 2003, GULYAS et al., 2007, KUTTLER et al., 2007). Many parameters play a significant role in the formation of the urban climate, UHI and urban heat archipelago (UHA), if the intra-urban thermal conditions are analysed. UHI and UHA are mainly caused by energetic features and fluxes (OKE, 1981, TSO et al., 1990; MATZARAKIS, 2001), which strongly depend on the increase of sealed surfaces (in horizontal and vertical directions) and the storage of heat in the urban materials. Two important factors can be used to quantify the urban morphology for UHI and UHA analyses (e.g. SHASHUA-BAR and HOFFMANN, 2003): the sky view factor SVF and the relative area of sealed surfaces in urban spaces indicated by the sealing factor Fv. The impact of SVF is particularly important with respect to nocturnal cooling rates, where long-wave radiation is trapped by warm surfaces as opposed to being released to the cold sky hemisphere. Such a decrease in long-wave radiation loss is directly related to SVF and considered to be a major component of the UHI phenomenon (BARRING et al., 1985, OKE et al., 1991, MOIN and TSUTSUMI, 2004). Based on existing data sets for near-surface Ta, i.e. near the bottom of the urban canopy layer (UCL), the aim of the present study is to improve the existing knowledge of the impact of the selected urban morphological parameters SVF and Fv on the intra-urban thermal conditions. The statistical analyses were performed for the city of Munich (South Germany), where a temporary urban climate network consisting of 20 climate stations was established from 1981-1985 in the UCL within the scope of the big research project “Urban Climate in Bavaria” (STADTKLIMA BAYERN). Working hypothesis, objectives, applied investigation design and results of STADTKLIMA BAYERN are explained in detail in the literature (e.g. BAUMGARTNER et al., 1985; BRÜNDL et al., 1986; MAYER, 1986, 1987, 1988) 2. Methods SVF is a dimensionless parameterisation of the quantity of visible sky at a certain location. Represented as a value between zero and one, SVF will approach unity in perfectly flat and open terrain, whereas locations with obstructions such as buildings and trees will cause SVF to be proportionally less (BARRING et al., 1985, OKE et al., 1991, SVENSON, 2004). Fv describes the percentage of the built up area. In STADTKLIMA BAYERN, it was determined for a horizontal circle with a radius of 100 m around each of the 18 temporary climate stations in Munich (Fig. 1, Table 1). All climate stations were of a similar type (Fig. 2). All sensors used were of the same type (thermohygrographs) to measure and record Ta and the relative humidity RH continuously during the complete investigation period. The sensors were calibrated at regular intervals at each station. UHI and UHA, respectively, were not only detected by the spatial and temporal patterns of Ta but also through statistical values calculated from Ta (Tab. 1). The following statistical values were available (BRÜNDL et al., 1986): - mean annual air temperature (MIT) of Ta, 131 Fig. 1: Screen level urban climate network in Munich (South Germany) from 19811985 within the scope of the STADTKLIMA BAYERN project (according to BRÜNDL et al., 1986, modified) Fig. 2: Station “Sonnenstraße” as example for the stations in the temporary urban climate network from 1981-1985 in the UCL in Munich (South Germany) within the scope of the STADTKLIMA BAYERN project 132 Table 1: Numbers and name of the stations as well as their, specifications, sealing factors Fv and sky view factors SVF, stations in the temporary urban climate network in the UCL in Munich (South Germany) within the scope of the STADTKLIMA BAYERN project (BRÜNDL et al., 1986; MAYER, 1987) station no. station name Fv SVF 1 Universität 0.95 0.247 3 Einsteinstraße 0.75 0.382 4 Praterinsel 0.20 0.150 5 Kirchenstraße 0.80 0.410 6 Ständlerstraße 0.30 0.627 7 Füstenrieder Straße 0.60 0.614 8 Antonienstraße 0.75 0.336 9 Hinterbrühl 0.05 0.438 10 Sonnenstraße 0.90 0.326 11 Walliser Straße 0.20 0.385 12 Romanplatz 0.80 0.610 13 Franz-Nißl-Straße 0.45 0.649 14 Paulckestraße 0.60 0.507 15 Pasing Bahnhof 0.75 0.329 16 Wasserburger Landstraße 0.55 0.766 17 Bad-Kreuznacher-Straße 0.30 0.677 19 Knappertbuschstraße 0.70 0.410 - absolute 30 minute maximum (MAX) and minimum (MIN) of Ta per year, - mean annual number of tropical days (TT), i.e. days with a daily maximum air temperature Ta,max ≥ 30 °C, - mean annual number of summer days (ST), i.e. days with a daily maximum air temperature Ta,max ≥ 25 °C, - mean annual number of warm days (wT), i.e. days with a daily mean air temperature Ta,mit > 20 °C, - mean annual number of heating days (HT), i.e. days with a daily mean air temperature Ta,mit < 12 °C, - mean annual number of frost days (FT), i.e. days with a daily minimum air temperature Ta,min < 0 °C, - mean annual number of ice days (ET), i.e. days with a daily maximum air temperature Ta,max < 0 °C, - mean annual number of cold days (kT), i.e. days with a daily mean air temperature Ta,mit < -10 °C, 133 i=N - mean annual total of heat (WS), i.e. ∑T a ,mit ,i i =1 with N: number of days per year, and the limitation for daily mean Ta values: Ta ,mit ,i > 0 °C , i=N - mean annual total of cold (KS), i.e. ∑T a ,mit ,i with N: number of days per year, and i =1 the limitation for daily mean Ta values: Ta ,mit ,i < 0 °C . In the statistical regression analyses, both mean annual Ta values and mean annual thermal characteristics calculated from Ta are the depending variables (Y), while Fv and SVF, respectively, represent the independent variable (X). The regression analyses were carried out in form of a linear regression Y = a0 + a1 ∗ Fv (1) Y = b0 + b1 ∗ SVF (2) and in form of a multiple regression: Y = c0 + c1 ∗ Fv + c2 ∗ SVF 3. (3) Results The statistical calculations can be interpreted as a re-analysis of urban climate data measured and recorded approximately 20 years ago. The results for mean thermal characteristics (Table 2) reveal an intra-urban variability, which depends on station-specific morphological parameters. They are exemplarily indicated by simple fish-eye graphics (Fig. 3). For the urban climate stations in the inner city of Munich, which are characterised by high permanent buildings (with high density of residents) and low percentage of green areas (station 1), and urban spaces with a strong mixture of diverse buildings and a very low portion of green areas (station 10), the results show relatively high values of total heat and number of warm days, while the values for cold days as well as heating, frost and ice days are relatively low, i.e. a specific pattern of the UHA or urban hot spots was formed. Due to its location within an urban park, the statistical values for the station 4 point out to a cold spot. With increasing distance to the city centre, the stations with single or more buildings (e.g. stations 13 or 16) are characterised by lower statistical values than for the hot spots in the city centre. Compared to the conditions within parks in the outskirts, the stations (13 or 16) can also be seen as hot spots. The results of the linear regression analyses for Fv and SVF are contained in Table 3. The grey colour indicates a correlation coefficient r higher than 0.5. The results in Table 3 can be summarised as follows: - The strongest correlation exists between Fv and the mean annual number of warm days followed by (i) the correlation between Fv and the mean annual total heat and (ii) the correlation between Fv and mean annual air temperature Ta. 134 Table 2: Mean values of air temperature Ta (MIT), total of heat (WS) and cold (KS), absolute maximum (MAX) of Ta and absolute minimum (MIN) of Ta, mean annual numbers (n/a) of tropical days (TT), summer days (ST), warm days (wT), heating days (HT), frost days (FT), ice days (ET) and cold days (kT) at the stations in the temporary urban climate network in the UCL in Munich (South Germany) within the scope of the STADTKLIMA BAYERN project, averaging period: 1982-1984 (BRÜNDL et al., 1986; MAYER, 1987; MATZARAKIS, 2001) station MIT WS KS MAX MIN TT ST wT HT FT ET kT °C °C °C °C °C n/a n/a n/a n/a n/a n/a n/a 1 9.3 3572.9 -150.5 38.2 -13.4 8 38 36 219 81 29 1 3 8.3 3312.1 -192.6 36.8 -15.8 4 32 26 235 113 32 2 4 8.4 3260.0 -172.6 34.6 -14.2 2 20 17 217 97 30 1 5 8.9 3447.3 -161.5 37.5 -14.3 7 40 33 225 99 34 1 6 8.0 3152.0 -215.9 36.6 -14.8 5 32 21 237 91 29 1 7 8.7 3375.1 -192.6 37.8 -13.4 7 40 29 229 107 30 1 8 8.8 3420.4 -188.7 36.5 -15.8 4 29 28 227 95 34 2 9 7.4 2958.7 -224.2 36.2 -15.2 4 25 13 250 131 31 1 10 9.6 3643.2 -137.0 35.9 -13.4 7 42 37 218 80 24 0 11 7.9 3117.4 -219.7 37.9 -15.0 5 32 21 240 122 33 2 12 8.7 3422.6 -192.5 37.1 -15.1 5 36 27 227 109 32 2 13 7.7 3081.7 -251.9 36.2 -17.8 3 33 20 246 127 37 4 14 8.2 3224.4 -213.4 37.8 -15.0 7 38 24 236 114 32 2 15 8.6 3349.0 -187.1 37.1 -14.8 5 33 28 230 103 30 1 16 8.1 3193.2 -227.4 36.2 -19.5 6 36 20 238 122 31 2 17 8.6 3373.4 -203.6 37.2 -16.3 7 37 29 229 105 35 2 19 8.5 3339.7 -218.4 38.3 -17.8 7 40 28 230 107 34 3 - A trend exists that the statistical values of Ta have a higher correlation with Fv than with SVF. SVF mainly influences the radiation balance and the turbulent transport within the UCL. These effects seem to be lower than for other meteorological parameters in the formation of UHI and UHA, respectively. - There are no correlations with r > 0.5, neither with Fv nor with SVF, for the statistical values of absolute Ta-maxima and the mean annual numbers of tropical, ice and cold days. The results for the multiple regression analyses are compiled in Table 4. They show: - As expected, the correlation coefficients for the multiple regressions are higher than the correlation coefficients for linear regressions. 135 1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Fig. 3: Fish-eye graphics for the stations in the temporary urban climate network from 1981-1985 in the UCL in Munich (BAUMGARTNER et al., 1984) 17 19 136 Table 3: Regression coefficients a0, a1, b0 and b1 from the equations (1) and (2) as well as correlation coefficients r, basis: statistical thermal values for 17 stations in the temporary urban climate network in the UCL in Munich (South Germany) within the scope of the STADTKLIMA BAYERN project, averaging period: 1982-1984, Fv: sealing factor, SVF: sky view factor, Ta: air temperature (according to BRÜNDL et al., 1986; MAYER, 1987; MATZARAKIS, 2001) independent variable Fv a0 a1 r 7.5 1.6 0.780 mean annual air temperature Ta (°C) mean annual total of heat (°C) 3008.0 mean annual total of cold (°C) -233.2 absolute Ta-maxima (°C) 36.2 absolute Ta-minima (°C) -16.0 mean annual number / tropical days 3.7 mean annual number / summer days 26.6 mean annual number / warm days 14.4 mean annual number / heating days 242.1 mean annual number / frost days 122.3 mean annual number / ice days 33.0 mean annual number / cold days 1.9 529.6 63.7 1.2 1.0 3.1 13.6 20.0 -19.0 -28.7 -2.4 -0.4 independent variable SVF b0 b1 r 9.0 -1.2 -0.371 0.818 3467.5 -344.0 -0.332 0.581 -146.7 -108.9 -0.622 0.352 36.6 0.7 0.122 0.167 -12.9 -5.3 -0.539 0.488 4.7 1.7 0.171 0.640 29.1 11.3 0.331 0.835 29.4 -8.1 -0.211 -0.555 218.3 28.3 0.518 -0.528 87.5 40.0 0.460 -0.220 29.0 5.6 0.322 -0.109 0.6 2.2 0.401 Table 4: Regression coefficients c0, c1 and c2 from the multiple regression equation (3) and the multiple correlation coefficient r; basis: statistical thermal values for 17 stations in the temporary urban climate network in the UCL in Munich (South Germany) within the scope of the STADTKLIMA BAYERN project, aver-aging period: 1982-1984, Fv: sealing factor, SVF: sky view factor, Ta: air temperature (according to BRÜNDL et al., 1986; MAYER, 1987; MATZARAKIS, 2001) mean annual air temperature Ta (°C) mean annual total of heat (°C) mean annual total of cold (°C) absolute Ta-maxima (°C) absolute Ta-minima (°C) mean annual number of tropical days mean annual number of summer days mean annual number of warm days mean annual number of heating days mean annual number of frost days mean annual number of ice days mean annual number of cold days c0 7.9 3107.7 -184.0 35.6 -13.2 2.3 18.0 15.4 229.8 105.2 30.3 0.7 c1 1.5 507.0 52.5 1.4 0.4 3.4 15.6 19.7 -16.1 -24.8 -1.8 -0.1 c2 -0.7 -187.8 -92.7 1.1 -5.2 2.8 16.1 -2.0 23.3 32.4 5.1 2.2 r 0.811 0.837 0.780 0.402 0.543 0.558 0.790 0.837 0.696 0.642 0.360 0.403 137 - The correlation coefficient for the multiple regression related to the mean annual number of tropical days exceeds the 0.5 threshold. - As for the linear regressions, the correlation coefficients of the multiple regressions for the absolute Ta-maxima as well as the mean annual number of ice and cold days do not exceed the 0.5 threshold. - The mean annual number of warm days and the mean annual total of heat show the strongest multiple correlations with Fv and SVF, followed by the mean annual air temperature Ta and the mean annual number of summer. Based on Ta values from all stations in the temporary urban climate network in the UCL in Munich, Fig. 4 reveals an increase of mean annual Ta with increasing Fv and decreasing SVF. The correlation coefficients r for the linear regressions point out that the correlation between Ta and Fv is clearly stronger than for Ta and SVF. Munich, 1982 - 1984 10 Ta,mit = 9.0 - 1.2*SVF, r = 0.371 Ta,mit (°C) 9 8 Ta,mit = 7.5 + 1.6*Fv, r = 0.780 7 0.0 0.2 0.4 0.6 0.8 1.0 Fv, SVF Fig. 4: Relationships between the mean annual air temperature Ta as well as Fv and SVF based on data recorded from 1982-1984 at stations in the temporary urban climate network in the UCL in Munich (South Germany) within the scope of the STADTKLIMA BAYERN project 4. Conclusions Urban morphological parameters are dominating factors for the formation of the urban heat island UHI and the urban heat archipelago UHA, respectively. Typically, UHI is of a wide interest in numerous studies on urban climate. However, for several applications, e.g. human-biometeorological analyses, UHI or UHA are not the main cause for human heat stress in summer. Related to cities, which are not influenced by thermally induced regional and local circulation systems, the radiation heat is more important in this context. 138 Known urban morphology parameters are sealing factor, sky view factor, aspect ratio of street canyons, building density, material properties, and portion of green spaces. Each of these parameters has an influence on UHI and UHA, which was analysed in numerous investigations by use of different methodical designs (e.g. OKE, 1973, 1981, 1987). For the formation of the thermal urban climate, the configuration of morphology also plays an important role. In addition, it has to be taken into account whether the SVF is determined by a building or green open space in order to quantify the upper and lower hemisphere in urban structures. Another important characteristic is the orientation of buildings as well as the location of green open spaces and other vegetation in their position in the sky view hemisphere. Acknowledgement This article is dedicated to Prof. Dr. Dr. h.c. Albert Baumgartner, who passed away on 6 March 2008. He was the head of the Chair for Bioclimatology and Applied Meteorology at the Ludwig-Maximilians-University of Munich (South Germany) from 1972-1985. The big research project “Urban Climate in Bavaria” (STADTKLIMA BAYERN) was carried out under his scientific responsibility from 1980-1986. References BAUMGARTNER, A., H. MAYER, W. BRÜNDL, A. KOTZ, U. MODLINGER, E.-M. NOACK, 1984: STADTKLIMA BAYERN - Kurzmitteilung Nr. 8. – Ber. Lehrst. Bioklimat. Angew. Meteorol. Univ. München. BAUMGARTNER, A., H. MAYER, E.-M. NOACK, 1985: Abschlußbericht zum Teilprogramm „Thermalkartierung“ von STADTKLIMA BAYERN. - Bayer. Staatsministerium für Landesentwicklung und Umweltfragen, Reihe Materialien Nr. 39. BARRING L., J.O. MATTSSON, S. LINDQUIST, 1985: Canyon geometry, street temperatures and urban heat island in Malmö, Sweden. - J. Climatol. 5, 333-444. BRÜNDL, W., H. MAYER, A. BAUMGARTNER, 1986: Abschlußbericht zum Teilprogramm „Klimamessungen München“ von STADTKLIMA BAYERN. - Bayer. Staatsministerium für Landesentwicklung und Umweltfragen, Reihe Materialien Nr. 43. GULYÁS, Á., J. UNGER, A. MATZARAKIS, 2006: Assessment of the microclimatic and thermal comfort conditions in a complex urban environment: modelling and measurements. - Building and Environment 41, 1713-1722. HELBIG, A., J. BAUMÜLLER, M. KERSCHGENS, 1999 (eds): Stadtklima und Luftreinhaltung. - 2nd edition, Springer, Berlin. JOHNSON, G.T., T.R. OKE, T.J. LYONS, D.G. STEYN, I.D. WATSON, J.A. VOOGT, 1991: Simulation of surface urban heat islands under ‘ideal’ conditions at night. Part 1: Theory and tests against field data. - Boundary-Layer Meteorology 56, 275-294. KUTTLER, W., 2004a: Stadtklima, Teil 1: Grundzüge und Ursachen. - UWSF-Zeitschrift für Umweltchemie und Ökotoxikologie 16, 187-199. KUTTLER, W., 2004b: Stadtklima, Teil 2: Phänomene und Wirkungen. - UWSF-Zeitschrift Umweltchemie und Ökotoxikologie 16, 263-274. KUTTLER, W, S. WEBER, J. SCHONNEFELD, A. HESSELSCHWERDT, 2007: Urban/rural atmospheric water vapour pressure differences and urban moisture excess in Krefeld, Germany. International Journal of Climatology 27, 2005-2015. 139 MATZARAKIS, A. 2001: Die thermische Komponente des Stadtklimas. - Ber. Meteor. Inst. Univ. Freiburg Nr. 6. MAYER, H., 1986: Zielsetzung und Konzeption des Forschungsvorhabens STADTKLIMA BAYERN. - Mitt. Geogr. Gesell. München 71, 21-39. MAYER, H., 1987: Ergebnisse aus dem Forschungsvorhaben STADTKLIMA BAYERN. - Mitt. Geogr. Gesell. München 72, 119-160. MAYER, H., 1988: Results from the research program ‘STADTKLIMA BAYERN’ for urban planning. - Energy and Buildings 11, 115-121. MAYER, M., A. MATZARAKIS, 2006: Impact of street trees on the thermal comfort of people in summer - a case study in Freiburg (Germany). - MERCHAVIM 6, 285-300. MAYER, H., A. MATZARAKIS, M.G. IZIOMON, 2003: Spatio-temperal variability of moisture conditions within the Urban Canopy Layer. - Theor. Appl. Climatol. 76, 165–179. MOIN U.M., J. TSUTSUMI, 2004: Rapid estimation of sky view factor and its application to human environment. - Journal of the Human-Environment System 7, 83-87. OKE, T.R., 1973: City size and the urban heat island. - Atmospheric Environment 7, 769-779. OKE, T. R., 1981, Canyon geometry and the nocturnal heat island: comparisons of scale model and field observations. - Journal of Climatology 1, 237-254. OKE, T. R., 1987: Boundary Layer Climates. - Methuen, London. OKE, T.R., G.T. JOHNSON, D.G. STEYN, I.D. WATSON, 1991: Simulation of surface urban heat islands under ‘ideal’ conditions at night. Part 2: Diagnosis of causation. - Boundary-Layer Meteorology 56, 339-358. SHASHUA-BAR, L., M.E. HOFFMANN, 2003: Geometry and orientation aspects in passive cooling of canyon streets with trees. - Energy and Buildings 35, 61-68. SVENSON, M., 2004: Sky view factor analysis - implications for urban air temperature differences. - Meteorological Applications 11, 201-211. TSO, C.P., B.K. CHAN, M.A. HASHIM, 1990: An improvement to the energy balance model for urban thermal environment analysis. - Energy and Buildings 14, 143-152. UPMANIS, H., I. ELIASSON, S. LINDQUIST, 1998: The influence of green areas on nocturnal temperatures in a high latitude city (Göteborg, Sweden). - International Journal of Climatology 18, 681-700. Authors' address: Prof. Dr. Andreas Matzarakis ([email protected]) Prof. Dr. Helmut Mayer ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 140 141 Towards urban sustainability: trends and challenges of building environmental assessment methods Fazia Ali-Toudert Chair for Environmental Architecture, Faculty of Building, University of Dortmund* Abstract Many countries have developed their own building environmental assessment methods or customized existing ones. International standardization is also underway to ensure a common framework (ISO, EN). These methods present some similarities in scope, intent and structure; yet they may differ substantially in many core aspects including the environmental criteria considered, the quantification of performance, and the management of the whole assessment process. The present paper compares, both in form and technical content, a number of these systems (e.g. HQE®, BREEAM, LEED®, CASBEE, GBTool) with focus on their trends and perspectives and their capacity to move to the ultimate target of urban sustainability. This paper addresses the critical and current issue to know how to manage increasing complexity, i.e. induced by the extension from a single building to urban scale and by including the socio-economic dimension, together with ensuring more transparency, accuracy and reliability within simple assessment schemes. Zur urbanen Nachhaltigkeit: Trends und Perspektiven der umweltbezogenen Gebäudebewertungssysteme Zusammenfassung Viele Länder haben bereits ihre eigenen umweltbezogenen Gebäudebewertungssysteme entwickelt oder haben existierende Systeme an das eigene Land angepasst. Für einen einheitlichen Rahmen zur Bewertung der Nachhaltigkeit im Bauwesen ist auch eine internationale Standardisierung (ISO, EN) in Bearbeitung. Die Methoden weisen gewisse Ähnlichkeiten auf, z.B. im Anwendungsbereich sowie bei Zielen und Struktur. Sie zeigen jedoch auch wesentliche Unterschiede in Kerneigenschaften, wie z.B. die verwendeten umweltgerechten Bewertungskriterien, die Quantifizierung der Umweltgerechtigkeit und das Management des gesamten Bewertungsprozesses. Der vorliegende Artikel liefert einen Vergleich - bezogen auf Form and Inhalt - unterschiedlicher Systeme (HQE®, BREEAM, LEED®, CASBEE, GBTool), wobei der Fokus auf ihren Trends und Perspektiven sowie deren Erweiterungsfähigkeiten zum umfassenden Ziel der urbanen Nachhaltigkeit liegt. Dieser Beitrag untersucht die kritische Frage der Erfassung zunehmender Komplexität, nämlich die Erweiterung des Untersuchungsobjekts vom Einzelgebäude zu einer urbanen Struktur. Dabei werden ökonomische und soziologische Dimensionen sowie gleichzeitig die Sicherstellung größerer Transparenz, Genauigkeit und Zuverlässigkeit anhand vereinfachter Bewertungssysteme mitberücksichtigt. 1. Introduction Numerous studies have been dedicated to building environmental assessment methods, either by comparing several methods or by analysing thoroughly one specific method (e.g. GBC). ________________ *Current affiliation. This study was undertaken during a post-doctoral stay at CSTB, France. 142 The focus was put alternatively on their relevance, content, initial and evolving intentions and roles, differences, perspectives, etc. (see e.g. COLE, 2005). The author justly addressed a number of critical issues in the essence of such tools which are as many paths for re-thinking them. He states that current rating systems are facing the challenge to evolve in terms of simplicity, refining performance measures and indicators, improving verification methods, streamlining the certification process, the necessary support documentation together with their capability to manage more complexity in a simple and practical form. So, the present paper will not duplicate such studies but focuses on one of the most unaddressed challenging issues: The urban scale as new emphasis. 2. Building environmental assessment methods Prior to handle the urban issue, a brief review of the most mature and successful assessment methods was exemplarily undertaken, together with the French method HQE® tertiaire. These are the British, American, Japanese and international reference products, respectively: BREEAM, LEED®, CASBEE, GBC (GBTool). In the following analysis, the focus is put on the structuring of criteria because of their profound implications on the process and final evaluation of the building as an “ecological” product. A short statement of the convergences vs. divergences of these systems is drawn, with the ultimate task to commit a reflection whether it is relevant to extend these tools to urban scales and if so, how to achieve that goal. The review criteria were: i. applicability: scope & scale, type of project / building; ii. development approach: intention, update and management of the method; iii. system maturity: age, stability, representativeness, versatility; iv. technical content: performance topics, thoroughness, end user, aim of the tool, decision aid means; v. communicability: rating system's anatomy, performance criteria's anatomy, clarity; vi. measurability: quantification, benchmark, weighting, results representation; vii. usability: availability of information, assistance to user, cost of assessment; viii. verification & certification: the assessor, required documentation, phases of assessment, final report & certification. Only a few of them are discussed here. More details are available from the author on request. 2.1 Applicability: scope and type of projects and buildings The building is the main object of study of these methods. However, a noticeable trend for an extension to an urban scale is visible: - The GBC takes into account explicitly the urban issue in one specific topic of its building assessment scheme, i.e. “Site selection, project planning and development”. - HQE® and LEED® are developing new independent rating systems exclusively dedicated to the neighbourhood scale, i.e. HQE aménagement and LEED-ND. - CASBEE: i) by extending comfort and well-being issues to the open spaces surrounding the building, ii) in CASBEE-H (where H refers to Heat Island) which is an 143 adjusted version of CASBEE applied to large cities like Tokyo or Osaka, and iii) in "CASBEE for districts and regions" which is under development. - "BREEAM Developments", on the other hand, provides an assessment framework to guide the sustainable design of developments, to allow developers to demonstrate the sustainability features of their proposals to the local planning authority. At a national level, the building rating systems are differentiated depending on i) building type (residential, offices, schools, etc.) or on ii) the life phase of the building (planning, operation & maintenance, etc.): - CASBEE differentiates between each phase of building life in form of Tools 0 to Tool 3 (pre-design, new construction, existing building, renovation), however all building types are taken into account in one tool. - BREEAM handles all building life phases in one rating system; whereas each rating system is dedicated to one building type. - LEED® portfolio includes i) rating systems for specific building types and ii) for new and existing buildings. - HQE® tertiaire approach is dedicated to tertiary activity including offices and educational buildings. Further tools are under development. By contrast, the GBC (GBTool), which is exclusively academic and not commercial, has developed a generic system such, which explicitly recognizes regional specificities and offers a versatile possibility of use. 2.2 Development approach All investigated national rating systems (HQE®, BREEAM, LEED®, CASBEE) are commercial tools. They are more or less supported by their governments or private industry, sometimes within an academic frame. Their sensitivity to market imperatives explains the multiplicity of use-specific tools as mentioned above. By contrast, GBC is a primarily research project and by implication a voluntary tool. GBC suffers no limitations induced by marketing considerations. For instance, the strong commitment of the industry and federal agencies in the LEED project, explains partly its rapid growth and expansion in comparison to other tools. CASBEE in turn clearly displays the aim of its implementation in Asia. This calls attention to the necessity of a careful analysis according to market contexts. BREEAM as well as LEED® are particularly effective in the management of their products thanks to numerous technical committees and in the latter case to the consensual approach based on the vote of the large LEED-membership. 2.3 Technical content and management Fig. 1 shows a comparison of the structure, i.e. the main topics of each of the environmental assessment methods under consideration. The two columns on the right side show respectively i) a summary of the analysed methods and ii) the draft proposed by the international Standard ISO under development on the subject. 144 Fig. 1: Comparative analysis of the structure of building environmental assessment methods, together with the related ISO project 145 Basically all these systems handle the major environmental issues of Energy, Water, Materials & Waste and Indoor Environmental Quality. However, differences are noticeable from one system to another in the consideration of the - physical context (site, land use, open spaces, transport, etc.), - quality of service (functionality, durability, long-term performance & maintenance, etc.), - human dimension in terms of social and economic aspects, - environmental loadings as main indicators of performance (greenhouse gases, pollution). For example, BREEAM focuses on the environmental loading indicators favoured by the consideration of topics such as "transport", "land use & ecology" rather than solely in terms of energy consumption. The HQE® approach is structured in 14 targets, which is a more fragmented scheme in comparison to other systems, yet still covering the main performance issues. GBC is a more flexible tool and the successive versions may vary largely as can be noticed between GBTool 2000 and GBTool 2005 as no trademark stability concern exists. CASBEE applies the recent ideas introduced by GBC 2000 of differentiating between the building as product and as services, by evaluating separately the environmental loadings on one hand and the quality of services on the other hand. CASBEE also pays more attention to the surroundings of the buildings and to socio-economic dimensions and hence initiates, at a national level, the extent of assessment boundaries to urban scale and sustainability matters as a whole. The Management of the project involves more stakeholders than the only design team and requires an explicit commitment. This issue is included differently depending on the system: - as a separate topic such in BREEAM, including all pre-design, construction, operation & maintenance; - included in the main topics of the rating systems such in LEED® (e.g. commissioning in the topic Energy); yet, all management aspects are not taken into account; - as a combination between a chapter “Environmental Management System: EMS” and single environmental targets (e.g. Targets 3 and 7). 2.4 Communicability Two main types of structure were identified: - A linear structure, as in BREEAM or LEED® where environmental performance is listed in form of individual checklists. Each of them consists of the aim or intent, awarded credits, compliance requirements and necessary documentation. This structure presents the advantage of clarity and ease of use. - An arborescent structure as in GBC project, where the performance criteria are organised in a series of topics and sub-topics. First intended for versatility, this structure presents the disadvantage to be less transparent. Both HQE® and CASBEE are inspired from this model. 146 2.5 Measurability All systems combine quantifiable and prescriptive criteria. All systems but HQE® tertiaire use a quantitative scale in form of cumulative points achieved for each performance criterion. In GBTool and CASBEE two partial totals are calculated which correspond to i) Quality and ii) Loadings, respectively. CASBEE reports a final score which is a ratio of both. This gives the possibility to a finer analysis of the building real impacts. A ranking or a building profile is then used to communicate final results. In GBTool the interpretation takes into account the regional and local specificities, since the benchmarks can be managed separately by national teams, with the assessment system remaining identical. Several levels of weightings are also possible in GBTool. This issue is critical, yet variable from one system to another and confirms the relative value of the final results provided by each method. In order to guarantee the compliance to performance criteria, the verification means must be explicitly defined. Here, the British and American tools provide more links to decision-aid sources. As well, the clear formulation of the required documentation makes the assessment easier and more reliable. This latter point also suffers some divergences from one system to another. All these aspects are major improvement areas of these tools. 2.6 Trends and perspectives The issue of sustainability assessment is strategic, either in Europe or at a wider international level. Most countries have developed their own tools or adjusted existing ones to their specific context. Yet, a common language is lacking, and several projects are underway which seeks to act as a common theoretical background for forthcoming methods. Fig. 2 shows the two main frameworks presently under development: ISO/TC59/SC 17 and CEN TC 350 projects, together with one example of country local standards, i.e. France. These projects are still limited to the building scale. 3. Sustainability assessment at an urban scale 3.1 A climate-conscious urban design method One main issue of increased complexity in assessment methods is the extension of their scope to urban context. The concern of sustainability of cities has focused the interest of several research fields for decades. Yet, the lack of a framework which coordinates all findings in readily understandable performance criteria prevents their effective implementation. Hence, the proposal of an “urban sustainability assessment method” inspired from existing building assessment methods is one way for bridging the gap between theory and practice at urban level, and between diverse disciplines on sustainability matters. The following material discusses some relevant points to build this new methodology. The starting point in building environmental design (1970's) was the concern for optimising the use of natural energies, the so-called bioclimatic architecture. Later in the assessment methods, the importance of energy is confirmed by its high weighting (for energy criteria: BREEAM 15 points, LEED® 10 points and HQE® high or very high level). 147 Fig. 2: Standardization of environmental issues applied to the building: ISO - EN – NF 148 Similarly, at urban level, the first attempts for structured design methods also dealt primarily with the climate and energy, see e.g. ALI-TOUDERT (2000) for a review of published methods. This is because of the critical issues of energy savings, human comfort, health and safety issues, all related to the availability of solar energy and wind access which are compromised by the urban density and to the formation of particular urban microclimates, etc. For instance, ALI-TOUDERT (2000) proposed a conceptualized methodology for integrating the climate in urban planning and urban design. 1. Urban Planning: i. the climate at regional & local levels to determine the basic design recommendations; ii. site selection to gain the optimal advantages from appropriate urban locations; iii. urban permeability (to wind) to keep connection with the natural environment and energies, and avoid the overheating of the city (Urban Heat Island UHI mitigation); iv. land use in terms of integration versus segregation of activities (residential, working areas, leisure, industry, etc.); v. landscaping which summarizes the positive effects of green at a large scale; vi. urban geometry as a link to the next design stage, as follows. 2. Urban design: i. openness to the Sky for solar and energetic control, i.e. to ensure solar access / protection; ii. urban porosity, which governs the ventilation rates in the urban spaces and hence indoors; iii. directionality, which discusses the optimal orientation of the street and buildings according to solar and wind needs; iv. urban reflectance, which governs the heat storage potential in the urban fabric: buildings & surfaces (UHI mitigation); v. building envelope, which acts as an interface between architectural and urban design strategies; vi. urban vegetation, which explains how the green may be the most useful for enhancing human comfort and energy savings. At design scale, the focus was put on geometrical indicators to ensure operative guidelines. These have been refined in a later research (ALI-TOUDERT, 2005). 3.2 Towards an urban sustainability assessment method The interest for sustainable cities as a generic keyword is manifold and combining all information sources for elaborating new appropriate rating systems is necessary: 149 1. The so-called environmental urban architecture, mainly supplied by architects and urban designers, and which progressively extends its physical limits (see e.g. THOMAS, 2003). 2. Urban climate research which provides tremendous information on the specific climate of urbanized sites and especially the urban heat island which is the main expression of climate change. These findings rely on a strong physical basis. 3. The current design practices where practitioners and other stakeholders try to extend “intuitively” building assessment methods to urban neighbourhoods by paying more attention to building’s surroundings. As previously mentioned, some attempts are made to bring on the market rating systems which scope is the neighbourhood or even the city as a whole, e.g. LEED-ND. Yet, these tools are in an experimental stage and need verification and feedback. U r b a n C li m a t e & P o l lu t i o n M itig a tio n o f th e U r b a n H e a t Is la n d U H I (G W P ) M a n a g e th e u rb a n m ic ro c lim a te s M a n a g e th e a v a ila b ility o f n a tu ra l e n e rg ie s (s o la r a c c e s s , u rb a n w in d s , s to rm w a te r, e tc .) R e d u c e th e h e a t s to ra g e in th e u rb a n fa b ric : B u ild in g s a n d u rb a n s u rfa c e s R e d u c e th e a n th ro p o g e n ic e n e rg y : T ra n s p o rt, in d u s tr ie , b u ild in g h e a tin g & a ir c o n d itio n n in g , e tc .) E n s u re a n d im p ro v e C o m fo rt & H e a lth in d o o rs a n d in o p e n s p a c e s : th e rm a l, w in d , a c o u s tic & v is u a l c o m fo rt T o p ic s : → S ite S e le c tio n → L o c a tio n e ffic ie n c y → L a n d U s e → U rb a n F o rm s & U rb a n S u r fa c e s R e d u c tio n o f P o llu tio n o f a ir , S o il a n d W a te r 1 . S a n ita ry Q u a lity o f A ir (C O , C O 2 ; C H 4 , N O x , O 3 n e a r s u rfa c e ; P M 2 .5 , 1 0 ) 2 . A c id ific a tio n o f S o ils a n d W a te r (S O 2 , N O x , N H 3 ) 3 . E u tro p h y o f S o ils (N O x , N H 3 , P O 4 -3 ) R e d u c e e m is s io n s o f h a rm fu l g a s e s a n d to x ic s u b s ta n c e s d u e to h e a tin g , c o m b u s tio n o f H y d ro c a rb u ra n ts , W a s te , e tc . R e s o u r c e c o n s u m p t io T o p ic s : → L a n d U s e → U rb a n In fra s tru c tu re → U r b a n F o rm s → S u s ta in a b le B u ild in g s R e d u c tio n o f L a n d C o n s u m p tio n P re s e rv a tio n o f B io d iv e rs ity a n d e c o lo g y o f s ite s S ite s e le c tio n a c c o rd in g to th e ir q u a lity , a n d p la n n in g s tr a te g ie s (d e n s ific a tio n / u rb a n e x p a n s io n , e tc .) T o p ic s : → S ite S e le c tio n → L o c a tio n e ffic ie n c y → L a n d U s e → E c o lo g y o f S ite R e d u c tio n o f N a tu r a l R e s o u r c e s C o n s u m p tio n 1 . E n e r g y (S o la r & e n e rg e tic c o n tro l, e n e rg y e ffic ie n c y , u s e o f re n e w a b le e n e rg y , e tc .) 2 . W a te r ( w a te r c o n s e rv a tio n , S to rm w a te r & w a s te w a te r m a n a g e m e n t a n d tre a tm e n t, e tc .) 3 . M a te ria ls & W a s te (e n e rg y c o n te n t, re c y c lin g & re u s e , w a s te tre a tm e n t & m a n a g e m e n t, e tc .) T o p ic s : → R e s o u rc e C o n s e rv a tio n S o c ia l In te g r a tio n d im e n s io n S o c io -E c o n o m c O b j e c t iv e s o f a s u s t a in a b le u r b a n d e v e lo p m e n t To give a picture of such an approach, Fig. 3 is a proposal of a basis structure for a sustainability assessment method at an urban scale. E q u ity , M ix e d p o p u la tio n s & c u ltu re s , m ix e d a c tiv itie s , S o lid a rity , S a fe ty , e x p re s s io n o f c u ltu re , e tc . L iv a b le s p a c e s , p ro x im ity o f a c tiv ité s , p ro m o tio n o f s o c ia l life , e tc . T o p ic s : → U r b a n F o rm s → Q u a lity o f U s e a n d S e rv ic e s → S o c ie ty & C u ltu re E c o n o m ic D im e n s io n G lo b a l c o s t, F in a n c ia l im p a c ts , E c o n o m ic d y n a m ic , A ffo rd a b ility o f h o u s in g & s e rv ic e s , e tc . P o li c y T o p ic s : → E c o n o m ic s G o u v e rn a n c e E n v iro n m e n ta l p o litc y , E c o -e d u c a tio n , A c tiv ity a n d W a s te M a n a g e m e n t, O p e ra tio n & M a in te n a n c e , Q u a lity o f s e rv ic e , e tc . T o p ic s : → M a n a g e m e n t → Q u a lity o f s e rv ic e Fig. 3: A framework for an urban sustainability assessment method 150 Basically, there will be continuity and no rupture on an environmental level, while moving from building to urban scale assessment, because the indicators of environmental loadings are identical: global warming potential GWP, resource consumption, ozone depletion ODP, Pollution of air, soil and water. Hence, the major topics applied to environmental building assessment are expected to be reused at urban level, such as the efficiency in energy or water resources use. Yet these topics need to be addressed differently according to the specificity of the current object of interest: the “urban fabric”. The content will be revised according to a number of major differences: 1. The object «city» consists of indoor spaces (buildings) and open spaces (streets, places and parks). Both are living spaces and support human activities, which require a high environmental quality, as well as they both effect more or less negatively the environment. 2. The urban climate and more precisely the urban heat island (UHI) is the main phenomenon characterizing the city from an environmental point of view. Consequently, a key issue for implementing a powerful assessment method at urban scale is to understand the mechanisms which lead to the formation of the UHI together with their dependence on planning and design choices. 3. The consumption of land as precious resource takes here a much more dominant place in comparison to building scale, since the site selection as well as the whole land use strategy relies on the availability of land (expansion, densification, infrastructure, etc.). These in turn will affect the need for other natural resources such energy, water or materials. 4. The social and economic dimensions assume an important role, since the city is by definition an organized framework for human activities and a concentration of capitals. This means that an extension of performance assessment to urban level extends automatically its limits beyond the environment to include all sustainability issues. A number of topics can be pre-defined as a working basis: - site selection, urban location efficiency, land use and ecology of sites: to avoid hazardous locations, improve site and climate quality, and preserve ecosystems; - urban forms & urban infrastructure and sustainable buildings: which include the optimisation of urban density, street network and pedestrian areas, building forms and arrangements, an integrated community development, based on diversity and a balanced mix of activities, proximity work/ habitat, etc.; - quality of life including the human wellbeing & health, social & economic integration; - conservation of resources: land, energy, water, materials and waste; - management and quality of service including commitment and eco-education, construction, operation and maintenance. 4. Conclusion Building assessment methods offer a good basis for elaborating new a scheme for urban scale assessment. Yet, attention must be called to necessary precautions, already observed at building level, from which the necessary clarity and accuracy in quantitative 151 assessment, scoring, context specificities, double counting, as well as a careful definition of the prescriptive criteria related to qualitative issues, etc. It is essential to offer systems that serve as a common framework for all stakeholders, and particularly to design teams which face the great challenge to manage conflicting design issues and manifold interests. Moreover, a distinction between performance assessment and market interests is also important. Indeed the latter might effects negatively the objective and rigorous setting of performance criteria and scoring. A multidisciplinary work is essential to build this method and a close collaboration between all environmental fields, urban climatologists and planners/designers is crucial. References HQE®: www.assohqe.org LEED®: www.usgbc.org/LEED ; BREEAM: www.breeam.org/ CASBEE: www.ibec.or.jp/CASBEE GBC: www.iisbe.org/iisbe. ALI-TOUDERT, F., 2000: Integration of the climatic dimension in Urbanism. - Master thesis. School of Architecture and urban design EPAU, Algiers. ALI-TOUDERT, F., 2005: Dependence of outdoor thermal comfort on street design. - PhD thesis. Rep. Meteor. Inst. Univ. Freiburg No. 15, http://www.freidok.unifreiburg.de/volltexte/2078. COLE, R., 2005: Building environmental assessment methods: redefining intentions and roles. Building Research and Information 35, 455-467. THOMAS, R., 2003: Sustainable urban design - an environmental approach. - Spon. London. Author's address: Dr. Fazia Ali-Toudert ([email protected]) Chair for Environmental Architecture (GB IV) Faculty of Building, University of Dortmund Baroper Str. 301, D-44227 Dortmund, Germany 152 153 The assessment of the thermal environment Gerd Jendritzky Meteorological Institute, Albert-Ludwigs University of Freiburg, Germany Abstract One of the fundamental issues in human biometeorology is the assessment and forecast of the thermal environment in a sound, effective and practical way. This is due to the need for human beings to adapt their heat budget to the thermal environment in order to optimise comfort, performance and health. Based on thermo physiology and heat exchange theory an overview is given on different assessment approaches up to the development of the “Universal Thermal Climate Index” within COST Action 730 (2004). Die Bewertung der thermischen Umgebungsbedingungen Zusammenfassung Die Bewertung und Vorhersage der thermischen Umweltbedingungen des Menschen in einer physiologisch korrekten, wirkungsvollen und praktischen Weise stellt eines der wichtigsten Fragestellungen in der Human-Biometeorologie dar. Dies ergibt sich aus der Notwendigkeit, den Wärmehaushalt des Menschen den thermischen Umweltbedingungen anzupassen, um Gesundheit, Wohlbefinden und Leistungsfähigkeit zu gewährleisten. Ausgehend von der Wärmbilanz des Menschen wird ein Überblick über verschiedene Bewertungsansätze gegeben bis hin zur Entwicklung des „Universellen thermischen Klimaindex UTCI“ im Rahmen der COST Action 730 (2004). 1. Introduction The close relationship of humans to the thermal component of the atmospheric environment is evident and belongs to everybody’s daily experience. Thus, issues related to thermal comfort, discomfort, and health impacts are the reason that the assessment and forecast of the thermal environment in a sound, effective and practical way is one of the fundamental subjects in human biometeorology. Thereby the term “thermal environment” comprises both the consideration of the atmospheric heat exchange conditions (stress) and the physiological response (strain). Balancing the human heat budget, i.e. adaptation of the organism to variable environmental (atmospheric) heat exchange conditions is controlled by a (for healthy people) very efficient autonomous thermoregulatory system (see 2.1) that is additionally supported by behavioural adaptation (as e.g. eating and drinking, activity and resting, clothing, exposure, housing, migration) controlled by discomfort sensations. These capabilities enable the (healthy) human being to live and to work in almost all climates on earth albeit not ever under comfortable conditions. The thermal comfort conditions appear to be rather similar over the globe (FANGER, 1970), but show big inter-individual differences. DE DEAR and BRAGER (2002) found that people who are able to adapt themselves by clothing to the prevailing weather outdoors will prefer different indoor temperatures dependent to the thermal outdoor experience of the past days. This resulted in an adaptive thermal comfort model which enlarges significantly the potential 154 to avoid or at least reduce air-conditioning needs thus conserving energy. On the other hand the differences in thresholds for thermal stress, when e.g. mortality will increase signify-cantly, indicate that societies are acclimatized to the local climate (“acclimatization” as special term for adaptation to climate) to a certain degree. However, the typical seasonal behaviour in time series of health data shows clearly that – on a population level – acclimatization doesn’t take place completely! In this context the often published statement that in a future warmer world the reduction in winter mortality will more than compensate the increase in summer mortality should be disputed due to completely different cause-effect relationships. When looking on traditional buildings obviously humanity always used local experience in climate related building design as a measure to adapt to the thermal environment. By the possibility of air-conditioning this knowledge seems to be less under the focus of architects. On the other hand probably due to the climate change discussion and to the health impacts of extreme events such as the Chicago heat wave 1995 and the extreme summer 2003 in Europe, there is an increasing awareness of the significance of the thermal environment for health and well-being including for the need of adaptation. The main application areas are listed in the context of the UTCI (Universal Thermal Climate Index) development (see 3). In spite of promising evolution in the last decade numerous weaknesses in science and application in the field of the thermal environment there are still some weaknesses to be dealt with, such as: - The predominant majority of the operational used assessment procedures do not fulfil the requirements discussed in 2.1 and 2.2, i.e. the physiological and physical basics. From this it follows that simple indices can only be of limited value. The develop-ment of UTCI as an international standard intends to improve the situation. - When looking for dose-response relationships between thermal conditions and the health outcome, a huge fuzziness must be accepted because data usually from 1st order weather stations (e.g. from a rural airport) are applied as rough proxy to describe the exposure of people living indoors in unknown buildings in unknown floors under the influence of an unknown heterogeneous urban heat island. There is a strong research need to get a better idea on the actual exposure. - Because the skill of the numerical weather forecast models differs between the meteorological variables, there are some problems in the mean range predictability of complex procedures in particular with water vapour and clouds (mean radiant temperature Tmrt). - In spite of the almost unmanageable amount of publications in urban climatology, there is a big gap in realization in urban planning and building design. - Only few approaches are available dealing with the quantitative consideration of acclimatization (e.g. the adaptive thermal comfort model or HeRATE (Health related adaptation to the thermal environment)). More research is needed to gain more insight on the population level in order to improve both research and application. - The multidisciplinary interfaces between the acting people (scientists, stakeholders, planners etc.) from the disciplines involved in the issues “thermal environment and adaptation” is a challenge. 155 2. Thermoregulation, human heat budget, and thermal assessment procedures 2.1 Thermoregulation For the human being it is crucial to keep the body core temperature at a constant level (37°C) in order to ensure functioning of the inner organs and of the brain thus optimising his/her comfort, performance and health. In contrast, the temperature of the shell, i.e. skin and extremities, can vary strongly depending on the environmental condition which is one of the mechanisms to adapt to it keeping heat production and heat loss, at least over a longer period (“steady-state”), in equilibrium. Heat is produced by metabolism as a result of activity, sometimes increased by shivering or slightly reduced by mechanical work where applicable, e.g. when climbing. The surplus heat must be released to the environment. The body can exchange heat by convection (sensible heat flux), conduction (contact with solids), evaporation (latent heat flux), radiation (long- and short-wave), and respiration (latent and sensible). From the mathematical point of view, the human organism can be separated into two interacting systems of thermoregulation: (1) the controlling active system which includes the thermoregulatory responses of shivering thermo genesis, sweat moisture excretion, and peripheral blood flow (cutaneous vasomotion) of unacclimatized subjects, and (2) the controlled passive system dealing with the physical human body and the heat transfer phenomena occurring in it and at its surface (Fig. 1). That accounts for local heat losses from body parts by free and forced convection, long-wave radiation exchange with surrounding surfaces, solar irradiation, and evaporation of moisture from the skin and heat and mass transfer through non-uniform clothing. Under comfort conditions the active system shows the lowest activity level indicating no strain. Increasing discomfort is associated with increasing strain and according impacts on the cardiovascular and respiratory system. The tolerance to thermal extremes depends on personal characteristics (HAVENITH, 2001, 2005): age, fitness, gender, acclimatization, morphology, fat. Age and fitness are the most important predictors whereas both are closely correlated. High age and/or low fitness level means low cardiovascular reserve which causes low thermal tolerance. The strain for the organism due to thermal stress can be quantified e.g. by an Physiological Strain Index PSI that for the heat load area is based on heart rate and Tcore (Hyperthermia) (MORAN et al., 1998) and for the cold stress area on Tskin and Tcore (Hypothermia) (MORAN et al., 1999). 2.2 The heat budget The heat exchange between the human body and the thermal environment (Fig. 2) can be described in the form of the energy balance equation which is nothing but the application of the first theorem of thermodynamics: M − W − [Q H (Ta , v ) + Q * (Tmrt , v )] − [Q L (e, v ) + QSW (e, v )] − Q Re (Ta , e ) ± S = 0 M: W: S: QH: Q*: QL: metabolic rate (activity) mechanical power (kind of activity) storage (change in heat content of the body) skin: turbulent flux of sensible heat radiation budget turbulent flux of latent heat (diffusion water vapour) (1) 156 QSW: turbulent flux of latent heat (sweat evaporation) respiration: QRe: respiratory heat flux (sensible and latent) Human Physiology Model Brain Controllers a) Control System Behaviour Environment Heat Exchange Tski Skin Temperature Threshold + Sweating - Skin Blood Flow Heat Exchange T cor + - Shivering Core Temperature Threshold b) Passive System Fig. 1: Schematic presentation of a physiological model of human thermoregulation – a: control system, b: passive system (FIALA et al., 2001; HAVENITH, 2001) 157 Avenues of Heat Exchange sweat evaporation Sun or other radiation respiration direct radiation infra-red radiation clothing M convection External work infra-red radiation reflected radiation conduction Fig. 2: The human heat budget (HAVENITH, 2001) The meteorological input variables include air temperature Ta, water vapour pressure e, wind velocity v, mean radiant temperature Tmrt including short- and long-wave radiation fluxes, in addition to metabolic rate and clothing insulation. In eq. (1) the appropriate meteorological variables are attached to the relevant fluxes. However, the internal (physiological) variables (Fig. 1), such as the temperature of the core and the skin, sweat rate, and skin wettedness interacting with the environmental heat exchange conditions are here not explicitly mentioned. 2.3 Thermal assessment procedures Besides applying just air temperature, more than 100 simple thermal indices - most of them two-parameter indices - have been developed in the past about 150 years in order to describe the complex conditions of heat exchange between the human body and its thermal environment. For warm conditions such indices consist usually in combinations of Ta and different measures for humidity. For cold conditions the combination consists in Ta and v, so-called wind chill index WCI (TIKUISIS and OSCEVSKI, 2001, 2003; SHITZER, 2006). Simple indices are simple to calculate and to forecast. In addition, they are most easily understood by the general public and other stakeholders (such as health service providers). However, due to their simple formulation compared to eq. (1) (by neglecting relevant fluxes or variables, respectively), these indices can never fulfil the essential requirement that for each index value there must always be a unique thermophysiological effect, regardless of the combination of the meteorological input values. Thus, their use is limited, results are often not comparable and further influences as variable thresholds etc. have to be defined artificially. These comments do not refer to WCI, which informs about the risk of short-term freezing of bare skin. Comprehensive reviews on such approaches can be found e.g. in FANGER (1970), LANDSBERG (1972), DRISCOLL (1992), and PARSONS (2003). 158 Fig. 3: Scheme for the thermophysiological assessment of the thermal environment; PMV: Predicted Mean Vote, PT*:Perceived Temperature, PET: Physiologically Equivalent Temperature, OUT_SET*: Outdoor Standard Effective Temperature, AT: Apparent Temperature, WCT: Wind Chill Temperature, Tsk: mean skin temperature, SR: sweat rate, Esk: evaporative heat loss, Wsk: wetness of the skin, Icl: insulation of clothing, clo: clothing value, Ta: air temperature, Tmrt: mean radiant temperature, v: wind velocity, e: water vapour pressure Thus, consequently dealing with the thermo-physiologically significant assessment of the thermal environment requires the application of a complete heat budget model that takes all mechanisms of heat exchange into account as described in eq. (1). Such models (Fig. 3) possess the essential attributes to be utilised operationally in most biometeorological applications in all climates, regions, seasons, and scales. This is certainly true for MEMI (HÖPPE, 1984, 1999) and the Outdoor Apparent Temperature (STEADMAN, 1984, 1994). However, it would not be the case for the simple Indoor AT, which is the basis of the US Heat Index, often used in outdoor applications neglecting the addition "Indoor". Other good indices include the Standard Predictive Index of Human Response approach (GAGGE et al., 1986), and Out_SET* (PICKUP and DE DEAR, 2000; DE DEAR and PICKUP, 2000), which is based on GAGGE's work. BLAZEJCZYK (1994) presented the man-environment heat exchange model MENEX, while the extensive work by HORIKOSHI et al. (1995, 1997) resulted in a Thermal Environmental Index. FANGER's (1970) PMV-equation can also be considered among the advanced heat budget models, if GAGGE`s et al. (1986) improvement in the description of latent heat fluxes by the introduction of PMV* is applied. This approach is generally the basis for the operational thermal assessment procedure Klima-Michel-model (JENDRITZKY et al., 1979, 1990) of the German national weather service DWD (Deutscher Wetterdienst) with the output parameter "perceived temperature PT" (STAIGER et al. 1997) that considers a certain degree of adaptation by various clothing. This procedure is running operationally taking quantitatively the acclimatisation approach HeRATE (KOPPE and 159 JENDRITZKY, 2005) into account. HeRATE is a conceptual model of short-term acclimatisation that modifies absolute PT thresholds by superimposition of the (relative) experience of the population in terms of PT of the previous weeks (Fig. 4). Among others, this procedure has the advantage that the index can be used without modification in different climate regions and during different times of the year without the need to artificially define seasons and to calibrate it to a particular city. Nevertheless, so far DWD is the only national weather service to run a complete heat budget model (Klima-Michelmodel) on a routine basis to a larger extent for its applications in human biometeorology. Fig. 4: Acclimatisation related PT thresholds for the example of SW-Germany in 1984 based on the HeRATE approach (KOPPE and JENDRITZKY 2005) 3. The near future: The Universal Thermal Climate Index UTCI 3.1 Questions Although each of the published heat budget models is in principle appropriate for use in any kind of assessment of the thermal environment, none of the models is accepted as a fundamental standard, neither by modellers nor by users. On the other hand, it is difficult to accept that after more than 30 years experience with heat budget modelling and easy access both to IT and meteorological data, people still use oversimplified and thus unreliable indices or even just air temperature. Some years ago, the International Society on Biometeorology (ISB) recognised the issue presented above and established a Commission ”On the development of a Universal Thermal Climate Index UTCI” (JENDRITZKY et al., 2002) (www.utci.org). Since 2005 these efforts could be reinforced by the COST Action 730 (Cooperation in Science and Technical Development) of the European Science Foundation ESF that provides the basis that at least the European scientists plus experts from abroad can join together on 160 a regular basis in order to achieve significant progress in deriving such an index (COST UTCI 2004). The aim is an international standard based on scientific progress in human response related thermo-physiological modelling of the last four decades (FIALA et al., 2001, 2003) including the acclimatisation issue. This work is done under the umbrella of the WMO-Commission on Climatology (CCl) and will finally be made available in a WMO “Guideline on the Thermal Environment” probably by 2009 so that everybody dealing with biometeorological assessments, in particular NMHSs (National Meteorological and Hydrological Services), but also universities, public health agencies, epidemiologists, environmental agencies, city authorities, planners etc. can then easily apply the state-of-the-art procedure for their specific purposes. The guideline will provide numerous examples for applications and solutions for handling meteorological input data. The Universal Thermal Climate Index UTCI (working title) must meet the following requirements: 1. 2. 3. 4. thermo-physiologically significant in the whole range of heat exchange applicable for total body considerations and for local skin cooling (frost bites) valid in all climates, seasons, and scales useful for key applications in human biometeorology The fields of applications considered as particularly significant for users are e.g.: 1. 2. 3. 4. public weather service PWS (Forecast of thermal conditions) public health system PHS (HHWS, epidemiology) precautionary planning (Urban planning, building design) climate impact research in the health sector (Thermal assessments) 3.2 Approaches Mathematical modeling of the human thermal system goes back 70 years. In the past four decades more detailed, multi-node models of human thermoregulation have been developed, e.g. STOLWIJK (1971), KONZ et al. (1977), WISSLER (1985), FIALA et al. (1999, 2001), HUIZENGA et al. (2001) and TANABE et al. (2002). These models simulate phenomena of the human heat transfer inside the body and at its surface taking into account the anatomical, thermal and physiological properties of the human body (see Fig 1). Environmental heat losses from body parts are modeled considering the inhomogeneous distribution of temperature and thermoregulatory responses over the body surface. Besides overall thermo-physiological variables, multi-segmental models are thus capable of predicting 'local' characteristics such as skin temperatures of individual body parts. Validation studies have shown that recent multi-node models reproduce the human dynamic thermal behavior over a wide range of thermal circumstances (FIALA et al., 2001, 2003; HAVENITH, 2001; Huizenga et al., 2001). Many of these models have been valuable research tools contributing to a deeper understanding of the principles of human thermoregulation (FIALA et al., 2001). However, there is still a need for better understanding of adaptive responses and their physiological implications. The passive system of the IESD-Fiala model (FIALA et al., 1999, 2001) is a multisegmental, multi-layered representation of the human body with spatial subdivisions. Each tissue node was assigned appropriate thermo-physical and thermophysiological 161 properties. The overall data replicates an average person with respect to body weight, body fat content, and Dubois-area. The physiological data aggregates to a basal whole body heat output and basal cardiac output, which are appropriate for a reclining adult in a thermo-neutral environment of 30°C. In these conditions, where no thermoregulation occurs, the model predicts besides others a basal skin wetness of 6%; a mean skin temperature of 34.4°C; and body core temperatures of 37.0°C in the head core (hypothalamus) and 36.9°C in the abdomen core (rectum) (FIALA et al. 1999). Verification and validation work using independent experiments from air exposures to cold stress, cold, moderate, warm and hot stress conditions, and a wide range of exercise intensities revealed good agreement with measured data for regulatory responses, mean and local skin temperatures, and internal temperatures for the whole spectrum of boundary conditions considered. The experts of the COST Action 730 WG on Thermo-physiological Modeling meanwhile agreed to base the UTCI model on the Fiala approach, which will be substantially advanced by including not yet used data from other groups. The UTCI model must meet all the above listed requirements in application. From practical considerations the advanced Fiala multi (actually some hundred) - segmental model cannot be applied explicitly on a routine basis. Thus, the future UTCI computations will make use of lookup tables derived from simulations with the help of the Fiala model that cover all imaginable combinations of air temperature, wind, humidity, and mean radiant temperature plus clothing. In an operational procedure the non-meteorological variables metabolic rate MET and thermal resistance of clothing are of great importance. The UTCI Commission has not yet finally defined a representative activity to be that of a person walking because the sparse literature provides inconsistent results. Probably a walking speed of lower than 4 km/h, which would mean a metabolic rate of 2.3 MET (135 W/m²), will be selected. Clothing isolation Icl will be considered as an intrinsic clo-value in the range of Icl = 0.4 – 1.7 clo (1 clo = 0.155 Km2/W) dependent on air temperature. This should cover the kinds of clothing worn by people who are adapted to their local climate. The need to address specific characteristics of clothing, such as significant ventilation between body surface and inner surface of clothing, is still subject of discussion. 4 Conclusions The basic state of knowledge in the field of the thermal environment and human health allows for the delivery of a number of advisory services in order to enhance the capability of societies and individuals to properly adapt to climate and climate change. As regards risk factors, biometeorology has to inform and advise the public and decision makers in politics and administration with the aim of recognizing and averting health risks at an early stage, in the framework of preventive planning, for example by making recommendations for ambient standards, by evaluation of site decisions, and by consultation on adaptive behaviour. This will facilitate adaptation both of the general public and of individuals thus improving health and well-being of populations. Science can contribute to identify the most appropriate approaches, measures, technologies, and policies to improve the adaptive capacity to climate and climate change in particular in the fields HHWSs and precautionary planning in urban areas. The significance of these issues also in the context of the climate change problem is obvious. Evi- 162 dently, the services essential for the good health, safety and well-being of national communities require the application knowledge, practices, research and technology based on the state-of-the-art in order to design and deliver appropriate biometeorological information and advisories to the public which may support people in proper adaptation. Acknowledgement COST Action 730 on UTCI is funded for four years within the EU-ESF COST Programme. Many thanks also to all scientists from different fields out of 19 European countries incl. Israel plus Australia, Canada, New Zealand and WMO, who contribute to the development of UTCI. Representative for these are the WG chairs Prof. Dr. George Havenith (UK), Prof. Dr. Philipp Weihs (Austria), Dr. Ekatarina Batchvarova (Bulgaria), and as vice-chair of the ISB commission on UTCI Prof. Dr. Richard de Dear (Australia). References BLAZEJCZYK, K., 1994: New climatological- and physiological model of the human heat balance outdoor (MENEX) and its applications in bioclimatological studies in different scales. Zeszyty IgiPZ PAN 28, 27-58. COST UTCI (2004): Towards a universal thermal climate index UTCI for assessing the thermal environment of the human being. - MoU of COST Action 730. www.utci.org pp. 17 DRISCOLL, D.M., 1992: Thermal comfort indexes. Current uses and abuses. - Nat. Weather Digest 17, 33-38. DE DEAR, R., J. PICKUP, 2000: An outdoor thermal environment index (OUT_SET*) - Part II Applications. - In: De Dear R, Kalma J, Oke T, Auliciems A (eds.), Biometeorology and Urban Climatology at the Turn of the Millenium. Selected Papers from the Conference ICBICUC'99 (Sydney, 8-12 Nov. 1999). WMO, Geneva, WCASP-50, 258-290. DE DEAR, R.J., G.S. BRAGER, 2002: Thermal comfort in naturally ventilated buildings: revisions to ASHRAE Standard 55. - Energy and Buildings 34, 549-561. FANGER, P.O., 1970: Thermal Comfort. - Danish Technical Press, Copenhagen. FIALA, D., K.J. LOMAS, M. STOHRER, 1999: A computer model of human thermoregulation for a wide range of environmental conditions: The passive system. - Journal of Applied Physiology 87, 1957-1972. FIALA, D., K.J. LOMAS, M. STOHRER, 2001: Computer prediction of human thermoregulatory and temperature responses to a wide range of environmental conditions. - Int. J. Biometeorol. 45, 143-159. FIALA, D., K.J. LOMAS, M. STOHRER, 2003: First principles modeling of thermal sensation responses in steady-state and transient conditions. - ASHRAE Transactions: Research Vol. 109, Part I, 179-186. GAGGE, A.P., A.P. FOBELETS, P.E. BERGLUND, 1986: A standard predictive index of human response to the thermal environment. - ASHRAE Trans. 92, 709-731. HAVENITH, G., 2001: An individual model of human thermoregulation for the simulation of heat stress response. - Journal of Applied Physiology 90, 1943-1954. HAVENITH, G., 2005: Temperature regulation, heat balance and climatic stress. - In: W. Kirch, B. Menne, R. Bertollini (eds.) Extreme Weather Events and Public Health Responses. Springer, Heidelberg, 69-80. 163 HÖPPE , P., 1984: Die Energiebilanz des Menschen. - Wiss. Mitt. Meteorol. Inst. Uni München Nr. 49. HÖPPE, P., 1999: The physiological equivalent temperature - a universal index for the biometeorological assessment of the thermal environment. - Int. J. Biometeorol. 43, 71-75. HORIKOSHI, T., T. TSUCHIKAWA, Y. KURAZUMI, N. MATSUBARA, 1995: Mathematical expression of combined and separate effect of air temperature, humidity, air velocity and thermal radiation on thermal comfort. - Archives of Complex Environmental Studies 7, 9-12. HORIKOSHI, T., M. EINISHI, T. TSUCHIKAWA, H. IMAI, 1997: Geographical distribution and annual fluctuation of thermal environmental indices in Japan. Development of a new thermal environmental index for outdoors and its application. - Journal of the Human-Environment System 1, 87-92. HUIZENGA, C., H. ZHANG, E. ARENS, 2001: A model of human physiology and comfort for assesssing complex thermal environments. - Building and Environment 36, 691-699. JENDRITZKY, G., W. SÖNNING, H.J. SWANTES, 1979: Ein objektives Bewertungsverfahren zur Beschreibung des thermischen Milieus in der Stadt- und Landschaftsplanung ("Klima-MichelModell"). - Beiträge d. Akad. f. Raumforschung und Landesplanung Nr. 28, Hannover. JENDRITZKY, G., 1990: Bioklimatische Bewertungsgrundlage der Räume am Beispiel von mesoskaligen Bioklimakarten. In: Jendritzky G, Schirmer H, Menz G, Schmidt-Kessen W (Hrsg.): Methode zur raumbezogenen Bewertung der thermischen Komponente im Bioklima des Menschen (Fortgeschriebenes Klima-Michel-Modell). - Beitr. Akad Raumforschung Landesplanung, Hannover, Nr. 114, 7–69. JENDRITZKY, G., A. MAAROUF, D. FIALA, H. STAIGER, 2002: An update on the development of a universal thermal climate index. - Proc. 15th AMS Conf. Biomet. Aerobiol. and 16th ICB02, 129-133. KONZ, S., C. HWANG, B. DHIMAN, J. DUNCAN, A. MASUD, 1977: An experimental validation of mathematical simulation of human thermoregulation. - Comput Biol Med 7, 71-82. KOPPE, C., G. JENDRITZKY, 2005: Inclusion of short-term adaptation to thermal stresses in a heat load warning procedure. - Meteorol. Z. 14, 271-278. LANDSBERG, H.E., 1972: The assessment of human bioclimate - a limited review of physical parameters. - World Meteorological Organization, Technical Note No. 123, WMO-No. 331, Geneva. MORAN, D.S., A. SHITZER, K.B. PANDOLF, 1998: A physiological strain index to evaluate heat stress. - Am. J. Physiol. Regul. Integr. Comp. Physiol. 275, R129-R134. MORAN, D.S., J.W. CASTELLANI, C. O'BRIEN, A.J. YOUNG, K.B. PANDOLF, 1999: Evaluating physiological strain during cold exposure using a new cold strain index. - Am. J. Physiol. Regul. Integr. Comp. Physiol. 277, R556-R564. PARSONS, K.C., 2003: Human thermal environments: the effects of hot, moderate, and cold environments on human health, comfort and performance. - Taylor & Francis, London, New York PICKUP, J., R. DE DEAR, 2000: An outdoor thermal comfort index (OUT_SET*) - Part I - The model and its assumptions. In: de Dear R, Kalma J, Oke T, Auliciems A (eds.): Biometeorology and Urban Climatology at the Turn of the Millenium. Selected Papers from the Conference ICB-ICUC'99 (Sydney, 8-12 Nov. 1999). WMO, Geneva, WCASP-50, 279-283. SHITZER, A., 2006: Wind-chill-equivalent temperatures: regarding the impact due to the variability of the environmental convective heat transfer coefficient. - Int. J. Biometeorol. 50, 224-232. 164 STAIGER, H., K. BUCHER, G. JENDRITZKY, 1997: Gefühlte Temperatur. Die physiologisch gerechte Bewertung von Wärmebelastung und Kältestress beim Aufenthalt im Freien in der Maßzahl Grad Celsius. - Annal. Meteorol. 33, 100-107. STEADMAN, R.G., 1984: A universal scale of apparent temperature. - J. Climate Appl. Meteor. 23, 1674-1687. STEADMAN, R.G., 1984: Norms of apparent temperature in Australia. - Aust. Met. Mag. 43, 116. STOLWIJK, J.A.J., 1971: A mathematical model of physiological temperature regulation in man. - NASA contractor report, NASA CR-1855, Washington DC TANABE, S.I., K. KOBAYASHI, J. NAKANO, Y. OZEKI, M. KONISHI, 2002: Evaluation of thermal comfort using combined multi-node thermoregulation (65MN) and radiation models and computational fluid dynamics (CFD). - Energy and Buildings 34, 637-646. TIKUISIS, P., R.J. OSCZEVSKI, 2002: Dynamic model of facial cooling. - J. Appl. Meteor. 41, 1241-1246. TIKUISIS, P., R.J. OSCZEVSKI, 2003: Facial cooling during cold air exposure. - BAMS July 2003, 927-934. WISSLER, E.H., 1985: Mathematical simulation of human thermal behavior using whole body models. - In: Shitzer A, Eberhart RC (eds.) Heat transfer in medicine and biology - analysis and applications, Plenum Press, New York and London, 325-373. Author's address: Prof. Dr. Gerd Jendritzky ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 165 Temporal patterns of nitrogen oxides measured at selected air quality monitoring stations in south-west Germany Helmut Mayer1), Jutta Holst1) and Dieter Ahrens2) 1) 2) Meteorological Institute, Albert-Ludwigs-University of Freiburg State Agency for Environment, Measurements and Nature Conservation Baden-Württemberg Abstract Long-term series of NO and NO2 concentrations measured at four selected air quality monitoring stations in the Federal State of Baden-Württemberg (SW Germany) are statistically analysed to obtain mean diurnal, weekly and annual courses of NO, NO2, NO/NO2 and NOx. The analyses are related to the rural station “Welzheimer Wald” and the three urban stations “Frei-burg”, “Mannheim-Süd” and “Stuttgart-Bad Cannstatt”. The discussion of the mean temporal patterns of nitrogen oxides includes chemical properties of NO and NO2 as well as the station-specific strength of anthropogenic emissions. Studies exemplarily conducted for “Freiburg” and “Mannheim-Süd” show decreasing trends for mean annual NO2 concentrations in the period 1985 to 2006. The significance of NO2 in impact-related air quality indices is explained for the indices DAQx and LAQx. Zeitliche Muster von Stickoxiden an ausgewählten Luftmessstellen in Südwestdeutschland Zusammenfassung Langfristige Zeitreihen von NO und NO2 an vier ausgewählten Luftmessstellen in BadenWürttemberg (Südwestdeutschland) werden statistisch untersucht, um mittlere Tages-, Wochenund Jahresgänge von NO, NO2, NO/NO2 und NOx zu erhalten. Die Immissionsanalysen beziehen sich auf die ländliche Station „Welzheimer Wald“ sowie die drei städtischen Stationen „Freiburg“, „Mannheim-Süd“ und „Stuttgart-Bad Cannstatt“. Die Diskussion der mittleren zeitlichen Muster von Stickoxiden schließt chemische Eigenschaften von NO und NO2 sowie die stationsspezifische Stärke anthropogener Emissionen ein. Exemplarische Analysen für die Stationen „Freiburg“ und „Mannheim-Süd“ weisen auf abnehmende Trends für NO2 Jahresmittelwerte im Zeitraum 1985 bis 2006 hin. Die Rolle von NO2 in wirkungsbezogenen Luftqualitätsindizes wird anhand der Indizes DAQx und LAQx aufgezeigt. 1. Introduction According to the EU directive 1999/30/EC from 22nd April 1999, which implicated the amendment of the 22nd Federal Immission Control Directive in Germany, limit values for NO2 become effective from 1st January 2010. For the protection of human health, the annual limit value is 40 µg/m3 per calendar year. The critical level for the protection of vegetation is set to 30 µg/m³ NOx (= NO + NO2) per calendar year. In order to estimate how realistic these limit values can be met, an overview on NO2 patterns is of interest. In this study, this is discussed for selected air quality monitoring stations in the Federal State of Baden-Württemberg (SW Germany). 166 2. Data and sites The analyses were conducted for the three selected urban air quality monitoring stations “Freiburg”, “Mannheim-Süd” and “Stuttgart-Bad Cannstatt” as well as for the rural air quality monitoring station “Welzheimer Wald” (500 m a.s.l.), which is located within a forest approximately 35 km north-east of Stuttgart and far from anthropogenic emission sources (Fig. 1). Mannheim- Welzheimer Stuttgart-Bad Cann- Freiburg Fig. 1: Location of the air quality monitoring stations “Freiburg”, “Mannheim-Süd”, “Stuttgart-Bad Cannstatt” and “Welzheimer Wald” in the Federal State of Baden-Württemberg (LUBW, 2008) Characteristics of the three cities are listed in Table 1. Freiburg and Mannheim are located in the southern and middle upper Rhine plain, while major parts of Stuttgart are located in a basin open to eastern direction, but surrounded by rolling hills. Freiburg is situated in a bay of the Black Forest and is strongly influenced by thermal wind systems. Among the three cities, the anthropogenic emission situation is lowest in Freiburg, where emissions caused by vehicular traffic are dominating. In Mannheim, emission from crossing interstate highways and additional strong emission from industrial sources increase the vehicular traffic related emissions. In Stuttgart, the emission conditions are mainly governed by local vehicular traffic (LUBW, 2006). 167 Table 1: Characteristics of the cities taken into account (STALA, 2004, LUBW, 2006)) city elevation residents arban area vehicles emission by traffic emission by other sources Freiburg 240 m a.s.l. 213998 15306 ha 105747 1177 t/a 1116 t/a Mannheim 95 m a.s.l. 307499 14496 ha 179468 2804 t/a 7585 t/a Stuttgart 235 m a.s.l. 590657 20736 ha 351239 2646 t/a 2716 t/a Data for the air pollutants nitrogen monoxide (NO) and nitrogen dioxide (NO2) were provided as half-hourly mean values by the State Agency for Environment, Measurements and Nature Conservation Baden-Württemberg. After a quality control they were analysed according to different objectives. Due to the differing start of the continuous recordings of air pollutant concentrations (“Freiburg” and “Mannheim-Süd”: 1980, “Stuttgart-Bad Cannstatt”: 1981, “Welzheimer Wald”: 1984), the investigation period differed between the selected air quality monitoring stations. However, it covers such a long time for all stations, that it makes sense to determine representative temporal patterns of NO and NO2 as well as NO/NO2 and NOx. 3. Results 3.1 Mean diurnal courses On principle, NO is formed and emitted by three different methods (MÖLLER, 2003): a) intermediate through denitrification (from nitrate) and abiotic nitrite decomposition, b) combustion of substances containing N, c) reaction N2 and O2 at high air temperature, d) processes in chemical industry transforming substances. The mean residence time of NO in the atmosphere is within the range of hours, but below 1 d (MÖLLER, 2003). As method a) is the only one, by which NO is emitted at the rural sites, NO at the station “Welzheimer Wald” does not show a distinct mean diurnal course (Fig. 2). In addition, the mean NO concentrations were relatively low. On average over 1984-2000, mean NO concentration amounted to 2 µg/m3 (Table 2). In contrast to this rural station, the mean NO concentrations were clearly higher at the three urban stations. Due to influences of enhanced NO emissions, the mean NO increase was more pronounced for “Mannheim-Süd” and “Stuttgart-Bad-Cannstatt” than for “Freiburg”. The mean diurnal course of NO at the urban stations shows a different behaviour between working days (Monday to Friday) and the weekend (Saturday and Sunday). On working days, the mean diurnal NO course has the form of a double-wave. Pronounced peak values occur during the morning rush hour and distinctly lower peak values during the evening rush hour (see also HELBIG et al., 1999; MAYER, 1999), i.e. NO was mainly directly emitted into the atmosphere by the vehicular traffic (method c). The major reason for the lower NO peak values observed during the evening rush hour is the chemical reaction of NO with ozone (O3), which is known as “O3 tritration”. This reaction cannot occur in a comparative manner during the morning rush hour, as the O3 168 concentration is relatively low in the second half of the night. The behaviour of NO in the ambient air at the weekend is caused by the changed pattern of the vehicular traffic. It leads to lower half-hourly mean NO concentrations than on working days. 120 Mo Tu We Th Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 Welzheimer Wald, 1984-2000 100 Fr Sa So 3 NO (µg/m ) 80 60 40 20 0 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 CET (hrs) Fig. 2: Mean diurnal courses of nitrogen monoxide NO at all days of the week for four selected air quality monitoring stations in the Federal State of BadenWürttemberg (according to MAYER and KOPPE, 2001) Table 2: Mean absolute and relative values for NO and NO2 as well as the ratio NO/NO2 and NOx (= NO + NO2) Freiburg 1980-2000 Mannheim-Süd 1980-2000 Stuttgart-Bad Cannstatt 1981-2000 Welzheimer Wald 1984-2000 NO 17 µg/m3 (41%) 39 µg/m3 (95%) 41 µg/m3 (100%) 2 µg/m3 (5%) NO2 30 µg/m3 (63%) 45 µg/m3 (94%) 48 µg/m3 (100%) 15 µg/m3 (31%) 0.6 (75%) 0.8 (100%) 0.8 (100%) 0.1 (13%) 3 18 µg/m3 (20%) NO/NO2 NOx 3 47 µg/m (53%) 3 84 µg/m (94%) 89 µg/m (100%) The mean residence time of NO2 in the atmosphere amounts to 5 to 7 days (BAUMBACH, 1990; MÖLLER, 2003). At the beginning of the investigation period, approximately 10% of the NO2 concentration in the ambient air was directly emitted, mainly due to the vehicular traffic (BAUMBACH, 1990), while the remaining 90% were formed by chemical reactions, e.g. O3 tritration, and resulted from transport processes in the atmospheric boundary layer. At present, the portion of directly emitted NO2 is estimated to about 25%, which is caused by modifications of motors (RABL and SCHOLZ, 2005) and increasing number of catalysers (SCHOLZ et al., 2007). 169 80 Mo Tu We Th Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 70 Welzheimer Wald, 1984-2000 3 NO2 (µg/m ) 60 Fr Sa So 50 40 30 20 10 0 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 CET (hrs) Fig. 3: Mean diurnal courses of nitrogen dioxide NO2 at all days of the week for four selected air quality monitoring stations in the Federal State of BadenWürttemberg (according to MAYER and KOPPE, 2001) Compared to NO, the specific features of NO2 led to a higher mean and mean extreme NO2 concentrations at the rural station “Welzheimer Wald” (Table 2). In addition, a mean diurnal NO2 course is indicated with slight differences between working days and the weekend (Fig. 3). On working days, mean peak values of NO2 occur in the evening and in an attenuated form in the late morning. The major mean NO2 peaks in the evening mainly result from the O3 tritration and regional transport from far located sources. As for NO, the mean NO2 level at the three urban stations increases from “Freiburg” (30 µg/m3) over “Mannheim-Süd” (45 µg/m3) to “Stuttgart-Bad Cannstatt” (48 µg/m3) with only a slight difference between the latter sites (Table 2), which are characterised by stronger anthropogenic emissions. At the working days, the diurnal courses of NO2 at the three urban stations show a double wave characterised by greater mean daily amplitudes and more pronounced mean peak values than for the rural station. At “Freiburg”, the mean NO2 peak value in the late morning was slightly higher than in the evening, while the mean evening NO2 peak value was clearly higher in “Mannheim-Süd”. No comparative local effects could be observed in “Stuttgart-Bad Cannstatt”. Similar to NO, the mean diurnal NO2 pattern at the urban stations reflects the changed conditions in vehicular traffic at weekend, which leads to lower NO2 concentrations. Noticeable is the distinct mean NO2 maximum at the beginning of the night. The form of the mean diurnal courses of the ratio NO/NO2 (Fig. 4) at the urban stations is somewhat similar to those of NO, i.e. it is different between working days and weekend. In addition, the mean peak values in the late morning of the working days are more 170 pronounced than those in the late evening. At the rural station “Welzheimer Wald”, the mean diurnal course of NO/NO2 shows only a slight maximum around noon. 2.0 Mo Th We Th 1.6 Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 Welzheimer Wald, 1984-2000 NO/NO2 Fr Sa So 1.2 0.8 0.4 0.0 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 CET (hrs) Fig. 4: Mean diurnal courses of the ratio NO/NO2 at all days of the week for four selected air quality monitoring stations in the Federal State of BadenWürttemberg (according to MAYER and KOPPE, 2001) 200 Mo Tu We Th 160 Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 Welzheimer Wald, 1984-2000 3 NOx (µg/m ) Fr Sa So 120 80 40 0 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 12:00 0:00 CET (hrs) Fig. 5: Mean diurnal courses of NOx (= NO + NO2) at all days of the week for four selected air quality monitoring stations in the Federal State of BadenWürttemberg (according to MAYER and KOPPE, 2001) 171 Mean NO/NO2 values averaged over the investigation periods are below 1.0 (Table 2). The ratios increase with anthropogenic nitrogen emissions, i.e. they are lowest (0.1) at “Welzheimer Wald” and highest (0.8) at “Mannheim-Süd” and “Stuttgart-Bad Cannstatt”. Mean NO/NO2 peak values are above 1.0 for the three urban stations, while mean NO/NO2 peak value at “Welzheimer Wald” is 0.2 due to very low NO concentration level. The form of the mean pattern of diurnal NOx courses (Fig. 5) is similar to those of NO, as the portion of NO2 to NOx mainly leads to an increase of the NOx values compared to the NO values. The highest mean NOx concentration averaged over the entire stationspecific investigation periods was 89 µg/m3 at “Stuttgart-Bad Cannstatt” followed by 84 µg/m3 at “Mannheim-Süd”, 47 µg/m3 at “Freiburg” and 18 µg/m3 at “Welzheimer Wald”, i.e. the relative portions of the mean NOx concentrations related to “StuttgartBad Cannstatt” were 20% at “Welzheimer Wald”, 53% at “Freiburg” and 94% at “Mannheim-Süd” (Table 2). 3.2 Mean annual courses The mean annual courses of NO (Fig. 6), NO2 (Fig. 7), NO/NO2 (Fig. 8) and NOx (Fig. 9) are based on monthly mean values. Their form, which is characterised by lower values in summer and higher values in winter, is similar for each component. The main reasons are (i) the better exchange conditions in the atmospheric boundary layer in summer due to a higher incoming short-wave radiation (BAUMBACH, 1990; HELBIG et al., 1999; ROST and Mayer, 2004; HOLST et al., 2006), whereas weather with stable thermal stratification frequently occurs in winter, and (ii) consequences of O3 titration, which is most pronounced in summer (MÖLLER, 2003). 80 Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 Welzheimer Wald, 1984-2000 70 3 NO (µg/m ) 60 50 40 30 20 10 0 I II III IV V VI VII VIII IX X XI XII month Fig. 6: Mean monthly values of nitrogen monoxide NO at four selected air quality monitoring stations in the Federal State of Baden-Württemberg (according to MAYER and KOPPE, 2001) 172 80 Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 Welzheimer Wald, 1984-2000 70 3 NO2 (µg/m ) 60 50 40 30 20 10 0 I II III IV V VI VII VIII IX X XI XII month Fig. 7: Mean monthly values of nitrogen dioxide NO2 at four selected air quality monitoring stations in the Federal State of Baden-Württemberg (according to MAYER and KOPPE, 2001) 2.0 Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 Welzheimer Wald, 1984-2000 NO/NO2 1.6 1.2 0.8 0.4 0.0 I II III IV V VI VII VIII IX X XI XII month Fig. 8: Mean monthly values of the ratio NO/NO2 at four selected air quality monitoring stations in the Federal State of Baden-Württemberg (according to MAYER and KOPPE, 2001) 173 140 Freiburg, 1980-2000 Mannheim-Süd, 1980-2000 Stuttgart-Bad Cannstatt, 1981-2000 Welzheimer Wald, 1984-2000 120 3 NOx (µg/m ) 100 80 60 40 20 0 I II III IV V VI VII VIII IX X XI XII month Fig. 9: Mean monthly values of NOx (= NO + NO2) at four selected air quality monitoring stations in the Federal State of Baden-Württemberg (according to MAYER and KOPPE, 2001) Related to the anthropogenic emission conditions, the mean annual amplitudes are stronger for the three urban stations than for the rural station. With respect to the substances, the mean annual amplitudes are more evident for NO, NO/NO2 and NOx than for NO2, as NO2 is a substance, which is mainly formed by chemical processes and predominantly not emitted directly. 3.3 Temporal evolution of NO2 Related to the annual limit value for NO2, which becomes effective from 1st January 2010, the temporal evolution of NO2 was analysed. For a better understanding of the results, Fig. 10 shows the temporal evolution of the NOx emission in Germany and the Federal State of Baden-Württemberg. Since 1986, the NOx emission in Germany decreased from 3810 kt/a in 1986 to 1434 kt/a in 2005. In the period with NOx emission data for both investigation areas, NOx emission in Baden-Württemberg amounts to about 10% of NOx emission in Germany. The temporal evolution of the mean annual NO2 concentration in the ambient air for the two selected stations “Freiburg” and “Mannheim-Süd” is characterised by a statistically significant (99%) decreasing linear trend (Fig. 11). The slopes of the regression lines point out to a stronger decrease for “Mannheim-Süd” than for “Freiburg”. Indicated by the coefficients of determination for the linear regressions, mean annual NO2 values exhibit a distinct interannual variability. For example, in “Freiburg” annual mean NO2 concentration was relatively low (19.6 µg/m3) in 1994, but higher (23.0 µg/m3) in 2006. 174 4000 3500 NOx emission in Germany 2500 NOx emission in Baden-Württemberg 3 NOx (10 t/a) 3000 2000 1500 1000 500 0 1975 1986 1989 1992 1995 year 1998 2001 2004 Fig. 10: Temporal evolution of the NOx emission in Germany (UBA, 2007) and in the Federal State of Baden-Württemberg (LUBW, 2006) 70 Freiburg: NO2 = -0.7595*year + 35.0 R2 = 0.6886 60 Mannheim-Süd: NO2 = -1.0587*year + 54.1 R2 = 0.7369 3 NO2 (µg/m ) 50 40 30 20 10 0 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 year Fig. 11: Temporal evolution of mean annual NO2 concentration in the ambient air at two selected air quality monitoring station in the Federal State of BadenWürttemberg from 1985 to 2006 175 In the investigation period, highest mean NO2 concentration at “Mannheim-Süd” was measured in 1989 (60.5 µg/m3), while the lowest was recorded in 2006 (32.0 µg/m3). Annual weather conditions have a pronounced influence on the NO2 concentration. This can be seen in 1997, where long-lasting stable conditions in the atmospheric boundary layer (inversions) in winter caused higher mean annual NO2 concentrations at both stations, but particularly at “Mannheim-Süd”. The exemplary results in Fig. 11 lead to the assumption, that the annual limit NO2 value for the protection of human health will be met for urban background conditions in SW Germany in the future, if the evolution of NOx emissions furthermore shows a decreasing trend. 3.4 Significance of NO2 in the assessment of the integral air quality The evaluation of single air pollutants according to limit values is the normal case to control the air quality. In addition, different demands exist to assess the integral air quality, as people are not exposed to only one pollutant in the ambient air, but to a conglomerate of air pollutants. For this purpose, air quality indices were developed in the last 10 years. The daily air quality index DAQx and the long-term air-quality index LAQx belong to the impact-related air quality indices. The reasons and the methods to determine both air quality indices are explained in detail in the literature (MAYER et al., 2004; ROST and MAYER, 2004; MAYER, 2006, MAYER et al., 2008). DAQx considers the air pollutants CO, NO2, O3, PM10 and SO2. The air pollutant with the worst single assessment determines DAQx. In SW Germany, NO2, O3 and PM10 are the most significant air pollutants forming DAQx. Freiburg, 1995-2003 5 4 DAQx DAQx 3 2 DAQx_NO2 1 DAQx_PM10 DAQx_O3 0 0 30 60 90 120 150 180 210 240 270 300 330 360 DOY Fig. 12: Mean annual courses of the daily air quality index DAQx as well as the substance-specific DAQx values for NO2, O3 and PM10 at the air quality monitoring station “Freiburg” in the period 1995 to 2003 (according to ROST and MAYER, 2004) 176 In order to get an information on the influence of the different contributing species on the integral assessment of the local air quality, mean annual courses of the daily air quality index DAQx as well as the substance-specific DAQx values for NO2, O3 and PM10 are presented in a comparative way for the air quality monitoring stations “Freiburg” (Fig. 12) and “Mannheim-Süd” (Fig. 13). While DAQx is strongly dominated by O3 in summer at both stations, particularly at “Freiburg” caused by thermal wind systems, leading air pollutants for DAQx in winter are PM10 and NO2. The stronger influence of direct, anthropogenic emissions at “Mannheim-Süd” causes higher DAQx values for PM10 and NO2. A systematic difference between the DAQx values for PM10 and NO2 cannot be observed for “Freiburg”, whereas the tendency of higher DAQx values for PM10 than for NO2 becomes apparent at “Mannheim-Süd” in winter. Mannheim-Süd, 1995-2003 5 4 DAQx_PM10 DAQx DAQx 3 DAQx_NO2 2 1 DAQx_O3 0 0 30 60 90 120 150 180 210 240 270 300 330 360 DOY Fig. 13: Mean annual courses of the daily air quality index DAQx as well as the substance-specific DAQx values for NO2, O3 and PM10 at the air quality monitoring station “Mannheim-Süd” in the period 1995 to 2003 (according to ROST and MAYER, 2004) In an analysis of LAQx covering the period 1995 to 2003, ROST and MAYER (2004) have taken into account a collective of 13 air pollution monitoring stations in BadenWürttemberg consisting of 11 urban and 2 rural stations. NO2 turned out to be the LAQx forming substance most frequently (above 50%), followed by PM10 (above 20%). The procedure to determine LAQx includes modifying substances. In the study by ROST and MAYER (2004), they have to be considered in almost all cases. Benzene was the dominating modifying substance (nearly 40%), followed by PM10 (nearly 30%) and NO2 (nearly 20%). 177 4. Conclusions Due to the limit values for NO2, which become effective from 1st January 2010 in the EU countries, a study was conducted on patterns and behaviour of mean NO2 values at selected air quality monitoring stations in SW Germany (Baden-Württemberg). As the investigation aimed for the annual NO2 limit value, mean annual NO2 concentrations were mainly analysed. The results show the importance of NO2 for the air pollution control, which increases with anthropogenic emissions. References BAUMBACH, G., 1990: Luftreinhaltung - Berlin, Springer-Verlag. HELBIG, A., J. BAUMÜLLER, M.J. KERSCHGENS, 1999: Stadtklima und Luftreinhaltung. - Berlin, Springer-Verlag. HOLST, TH., J. ROST, H. MAYER, 2006: Muster regional differenzierter atmosphärischer Austauschbedingungen bei Überschreitungen des Tagesgrenzwerts für PM10 von 50 µg/m3. – Forsch. Ber. Meteor. Inst. Univ. Freiburg. LUBW, 2006: Luftschadstoff-Emissionskataster Baden-Württemberg 2004. - Karlsruhe, Landesanstalt für Umwelt Messungen und Naturschutz Baden-Württemberg. LUBW, 2008: Messstellen Luftmessnetz Baden-Württemberg. - http://mnz.lubw.badenwuerttemberg.de/messwerte/langzeit/mp_akt/ms_akt_index.htm. MAYER, H., 1999: Air pollution in cities. - Atmospheric Environment 33, 4029-4037. MAYER, H., 2006. Indizes zur human-biometeorologischen Bewertung der thermischen und lufthygienischen Komponente des Klimas. - Gefahrstoffe-Reinhaltung der Luft 66, 165-174. MAYER, H., CH. KOPPE, 2001: Langzeitentwicklung von Luftverunreinigungen an ausgewählten Luftmessstationen in Baden-Württemberg - Teil II: Trendanalysen für NO, NO2 und NO/NO2 im Zeitraum 1980 bis 2000. - Ber. Meteor. Inst. Univ. Freiburg Nr. 10007162/33. MAYER, H., J. HOLST, D. SCHINDLER, D. AHRENS, 2008: Evolution of the air pollution in SW Germany evaluated by the long-term air quality index LAQx. – Atmospheric Environment 42, in press. MAYER, H., L. MAKRA, F. KALBERLAH, D. AHRENS, U. REUTER, 2004: Air stress and air quality indices. - Meteorol. Zeitschrift 13, 395-403. MÖLLER, D., 2003. Luft. - Berlin, Walter de Gruyter. RABL, P., W. SCHOLZ, 2005: Wechselbeziehungen zwischen Stickstoffoxid- und Ozon-Immissionen, Datenanalysen aus Baden-Württemberg und Bayern 1990-2003. - Immissionsschutz 1/2005, 21-25. ROST, J., H. MAYER, 2004: Berechnungen zur räumlichen und zeitlichen Variabilität des planungsrelevanten FoBiG Luftqualitätsindexes an ausgewählten Luftmessstationen in BadenWürttemberg - Teil III: Langzeitindex. – Forsch. Ber. Meteor. Inst. Univ. Freiburg. SCHOLZ, W., C. KESSLER, D. AHRENS, A. NIEDERAU, 2007: Anstieg des NO2/NOx-Verhältnisses an Luftmessstationen in Baden-Württemberg 1995-2005. Immissionsschutz 2/2007, 68-72. STALA, 2004: Bodenflächen, Einwohner, Beschäftigte, Wohnungen, Kfz. - Stuttgart, Statistisches Landesamt. UBA, 2007: Nationale Trendtabellen für die deutsche Berichterstattung atmosphärischer Emissionen 1990-2005, Version 1.05 - Dessau, Umweltbundesamt. 178 Authors' address: Prof. Dr. Helmut Mayer ([email protected]) Dr. Jutta Holst ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany Dr. Dieter Ahrens ([email protected]) State Agency for Environment, Measurements and Nature Conservation Baden-Württemberg Postfach 10 01 63, D-76231 Karlsruhe, Germany 179 Climatic tourism potential in the North Sea and Black Forest region – a comparison between REMO and DWD data Christina Endler and Andreas Matzarakis Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Abstract Within the project CAST (“climate trends and sustainable tourism development in coastal and mountain regions”), the climatic tourism potential will be analysed for present and future climate conditions for two different climatic sensitive regions in Germany: the North Sea and the Black Forest. Furthermore, analysing tourism trends, developing new products for sustainable tourism as well as establishing a communication and knowledge platform including e.g. adaptation to climate change are focussed in the project CAST and realized in collaboration with stakeholders. For the time being the analysis of the climatic tourism potential is based on the regional climate simulations (A1B) carried out by the regional climate model REMO from the Max-Planck-Institute for Meteorology in Hamburg with a spatial resolution of 10 km. Challenges and difficulties in modelling the extensive physics of the atmosphere and its interactions with the climate system are known. The question, how well can REMO simulate the present climate, will initially be answered by the evaluation with climate data of the German Weather Service (DWD) for Norderney, Titisee, and Freiburg for the reference periods 1961-1990 and 1971-2000, respectively. Besides precipitation and air temperature REMO offers some difficulties in modelling wind, cloud cover, and relative humidity. Orographical regions are characterized by an underestimation of precipitation and fog conditions and an overestimation of stormy days, whereas lowlands are characterized by an overestimation of precipitation, stormy days, and fog conditions. The analysis of DWD data allows a careful interpretation of the modelled data and an evaluation of model uncertainties. Klimatisches Tourismuspotenzial für die Regionen Nordsee und Schwarzwald – ein Vergleich zwischen REMO und DWD Daten Zusammenfassung Im Rahmen des Projektes KUNTIKUM („Klimatrends und nachhaltige Tourismusentwicklung in Küsten- und Mittelgebirgsregionen“) wird das klimatische Tourismuspotenzial für das gegenwärtige und zukünftige Klima für zwei verschiedene klimatisch sensible Regionen Deutschlands untersucht: Nordsee und Schwarzwald. Weiterhin werden in Zusammenarbeit mit Tourismusakteuren Tourismustrends analysiert und Strategien sowie neue Produkte für den nachhaltigen Tourismus entwickelt. Ein zusätzlicher Schwerpunkt des Projektes KUNTIKUM ist die Erstellung einer Wissens- und Kommunikationsplattform, die u. a. Anpassungsstrategien an den Klimawandel beinhalten. Für die Analyse des klimatischen Tourismuspotenzials werden vorerst die regionalen Klimasimulationen (A1B) des Regionalmodells REMO vom Max-Planck-Institut für Meteorologie in Hamburg, mit einer räumlichen Auflösung von 10 km, verwendet. In der Modellierung ist bekannt, dass Modelle die komplexe Physik der Atmosphäre und die Wechselwirkungen im Klimasystem nur teilweise realistisch abbilden können. Wie gut REMO das heutige Klima simulieren kann, wird zunächst mit Hilfe der Klimastationsdaten des Deutschen Wetterdienstes (DWD) für Norderney, Titisee und Freiburg für die Referenzperioden 19611990 bzw. 1971-2000 evaluiert. REMO offenbart neben Niederschlag und Lufttemperatur einige Schwierigkeiten in der Modellierung von Wind, Bewölkung und relativer Luftfeuchte. Dabei werden im orografisch strukturierten Gelände Niederschlag und Nebel unterschätzt und stürmische Tage überschätzt. In der Tiefebene hingegen werden Niederschlag, stürmische und neblige 180 Tage überschätzt. Die Analyse der DWD Daten ermöglicht eine sorgfältige Interpretation der Modellergebnisse und eine Abschätzung der Modellunsicherheiten. 1. Introduction To simulate past, present, and future climate conditions, dynamical models are necessary. Numerical partial differential equations reproduce – as well as possible - the interactions of several subsystems of the climate system. Nowadays, the grid mesh is about 250 to 500 km and 9 to 20 layers in the vertical. Physical-chemical micro- and meso scale processes like cloud formation, turbulence effects or land breeze effects are not primary modelled as processes but parameterized. The IPCC report 2001 stated the need for a substantial reduction of uncertainties in regional climate modelling and an increase of the spatial resolution of the results. Since that time the computer power increased substantially and the model physics and dynamics have been further developed in order to simulate the climate change at spatial resolutions substantially below 50 km. The new generation of non-hydrostatic regional climate models even has the potential to simulate the regional climate at space scales down to 1 km. However, the dynamics of the earth system at higher spatial scales has to be taken into account adequately. Initially, we used the regional climate model REMO in a high resolution of 10 km for Germany. These model calculations have been conducted on behalf of the Federal Environment Agency (Umweltbundesamt, UBA) (ZEBISCH et al., 2005). Hence, UBA would like to support positively the research topics concerning climate change and adaptation. Thus, the model runs are available for interested applicants. In order to derive the model results for application purposes, to avoid error propagation, and to interpret model uncertainties an evaluation of these results is necessary. One application field is the tourism industry. Tourism is one the determining factor for economy. Tourism and climate are closely linked. Climate and climate changes influence the tourism industry, respectively. Consequently, demand and supply depend on given climate conditions. Policy makers have to act accordingly and adapt to modified economical and ecological conditions under climate change. Therefore, scientific information has to be produced in a useful and easy understandable way (DE FREITAS, 2003; MATZARAKIS et al., 2004). Within our project we calculate the climatic tourism and recreation potential for the two investigation areas North Sea and Black Forest. Background of the computed tourism and recreation potential for the presence are climate data from the German Weather Service (DWD). These data shall be compared to data carried out by REMO (JACOB, 2001; JACOB et al., 2007) and CLM (Climate Local Model) (BÖHM et al., 2006). Within our project REMO data as well as CLM data will be used for the analysis of tourism potential for 1961-1990 and 1971-2000, respectively, and for future climate trends (2021-2050). Hence, the SRES scenarios A1B, B1, and A2 will be considered. REMO and CLM data build the base for thermal, physical, and aesthetic computations being used for the validation of both thermal comfort and tourism and recreation potential. Moreover, frequency classes and frequencies of extreme weather events are generated based on monthly and 10-day-intervals. The derived results, in terms of climate tourism 181 information schemata (CTIS), and maps shall be allocated for stakeholders (MAT2007). ZARAKIS, 2. Data and Methods For this study the A1B scenario carried out by REMO (JACOB, 2001; JACOB et al., 2007) is initially used. The model region covers Germany and the Alps. The data has a spatial resolution of 10 km and a temporal resolution of hours. The data are available from 1950 to 2100. Thereby, the periods 1961-1990 as well as 1971-2000 of the A1B scenario are used as reference periods for future climate changes. Furthermore, climate data for exclusive stations are available from mid 20th century up to now. We have chosen the time span from 1961-2000 for comparison. The following parameters of REMO and DWD data, respectively, are the basis for the computation of physiologically equivalent temperature (PET) being background for thermal comfort and discomfort (HÖPPE, 1999; MATZARAKIS et al., 1999): • • • • • • date, longitude, latitude and elevation, air temperature, vapour pressure, wind velocity, cloud cover (DWD) and global radiation (REMO). PET is computed by the model RayMan (MATZARAKIS et al., 2007). Additionally, precipitation and snow cover are also included in the analysis. The values refer to 14 and 14:30 CET for REMO and DWD, respectively, except precipitation being the total annual precipitation amount. The snow cover carried out by DWD refers to 6 UTC. For the analysis of the climatic tourism potential particular thresholds based on meteorological parameters are defined (Table 1). Table 1: Parameters relevant for tourism and recreation concerning their thresholds and authors Parameter Thresholds Author thermal comfort heat stress cold stress sunshine fog "sultriness" dry day rainy day 18 °C < PET < 29 °C PET > 35 °C PET < 0 °C cloud cover < 5/8 relative humidity > 93 % vapour pressure > 18 hPa precipitation < 1 mm precipitation > 5 mm stormy day wind velocity > 8 m/s ski potential snow cover > 10 cm MATZARAKIS (2007) MATZARAKIS and MAYER (1996) MATZARAKIS (2007) GÓMEZ MARTÍN (2004) MATZARAKIS (2007) SCHARLAU (1943) MATZARAKIS (2007) MATZARAKIS (2007) BESANCENOT (1990), GÓMEZ MARTÍN (2004) BENISTON (1997), KULINAT and STEINECKE (1984), BREILING and CHARAMZA (1999), ROTH et al. (2005) 182 Evaluating REMO data with selected DWD data additional parameters are included (shown in Table 2 to 7). Keeping in mind that REMO has a spatial resolution of 10 km the averaging influences the values. The exclusive DWD stations are located in the corresponding grid box. Thereby it is not focused on the definite consistence of the values rather on the validation. 3. Results and Discussion The comparison of REMO and DWD data is realized for exclusive regions and stations for both the North Sea (Norderney) and Black Forest (Titisee and Freiburg). The elevations of the DWD stations compared to REMO vary concerning the grid mesh of 10 km. Table 2: Comparison of different selected parameters concerning REMO and DWD data for the periods 1961-1990 and 1971-2000 for Norderney; corresponding notations are: PETa: annual PET, Tmrt: mean radiant temperature, RR: precipitation, RH: relative humidity, VP: vapour pressure, V: wind velocity Norderney 1961-1990 1971-2000 Parameter REMO DWD REMO DWD PETa [°C] PETmax [°C] PETmin [°C] PET < 0 [d] PET = 18-29 [d] PET > 30 [d] PET > 35 [d] mean air temperature [°C] Tmax [°C] Tmin [°C] mean radiant temperature [°C] Tmrt_max [°C] Tmrt_min [°C] annual precipitation amount [mm] RR < 1 mm [d] RR > 1 mm [d] RR > 5 mm [d] averaged relative humidity [%] RH > 93 % [d] averaged cloud cover [octas] cloud cover < 5 octas [d] averaged vapour pressure [hPa] VP > 18 hPa [d] averaged wind velocity [m/s] Vmax [m/s] V > 8 m/s [d] 5.0 33.3 -14.3 115 13 0 0 9.9 23.0 -5.7 18.1 48.9 -15.7 1013.8 196 169 66 82.9 88 5.3 144 10.5 8 7.2 23.4 142 5.1 32.5 -19.0 119 19 0 0 9.0 25.8 -11.4 25.4 54.6 -21.0 769.5 235 130 52 83.7 48 5.3 132 10.2 10 6.9 26.5 115 5.1 34.4 -14.3 113 14 0 0 9.9 23 -5.7 18.2 49.3 -15.7 1014.7 197 169 65 83.0 90 5.3 145 10.6 10 7.1 23.4 137 5.6 32.1 -19 109 23 0 0 9.3 26.4 -11.4 25.8 53.6 -21.0 742.9 238 127 50 83.2 47 5.2 136 10.3 11 6.5 26.5 94 183 Norderney has an actual and modelled height of 11 m asl and 1 m asl, respectively. Titisee offers an actual height of 846 m asl and a modelled height of 984 m asl. Freiburg has an actual and modelled height of 268 and 228 m asl, respectively. Especially in highly orographical areas like the Black Forest significant elevation differences can occur. The exclusive comparative parameters for REMO and DWD data for the two periods 1961-1990 and 1971-2000 for Norderney are given in Tables 2 and 3. Table 3: Comparison of monthly and seasonal precipitation concerning REMO and DWD data for the periods 1961-1990 and 1971-2000 for Norderney Norderney precipitation [mm] monthly January February March April May June July August September October November December seasonal winter (DJF) spring (MAM) summer (JJA) autumn (SON) 1961-1990 REMO DWD 1971-2000 [%] REMO DWD [%] 82.0 70.4 49.3 44.3 55.4 62.0 84.8 107.9 115.9 111.7 131.3 98.6 58.9 40.7 52.7 40.9 49.0 62.4 75.9 73.3 72.1 80.1 88.0 75.4 0.7 0.6 1.1 0.9 0.9 1.0 0.9 0.7 0.6 0.7 0.7 0.8 81.2 76.8 51.5 41.5 54.1 66.0 89.7 101.0 126.8 103.7 122.5 99.8 58.9 41.4 53.5 36.5 47.0 60.5 66.1 65.2 80.1 79.2 82.7 71.7 0.7 0.5 1.0 0.9 0.9 0.9 0.7 0.6 0.6 0.8 0.7 0.7 83.7 49.7 84.9 119.6 58.3 47.5 70.5 80.1 0.7 1.0 0.8 0.7 105.8 49.0 85.6 117.7 65.3 45.7 63.9 80.7 0.6 0.9 0.7 0.7 Keeping in mind that the data modelled by REMO are too warm and wet, the mean air temperature differs from DWD data by 0.9 K (1961-1990) and 0.6 K (1971-2000), respectively. The annual PET does not show distinctive deviations in both periods. Despite a warmer air temperature and underestimation of cloud cover, the thermal comfort and mean radiant temperature are underestimated. However, the averaged cloud cover has the same magnitude. REMO obviously overestimates wind velocities and fog conditions. The latter is based on the relative humidity. The distribution of precipitation (Table 3) shows an overestimation of winter and autumn precipitation, whereas the spring and summer precipitation are consistent with DWD data. The exclusive comparative parameters for REMO and DWD data for the two periods 1961-1990 and 1971-2000 for Titisee are given in Table 4 and 5. 184 Table 4: Comparison of different selected parameters concerning REMO and DWD data for the periods 1961-1990 and 1971-2000 for Titisee; corresponding notations are: PETa: annual PET, Tmrt: mean radiant temperature, RR: precipitation, RH: relative humidity, VP: vapour pressure, V: wind velocity, SN: snow cover Titisee 1961-1990 1971-2000 Parameter REMO DWD REMO DWD PETa [°C] PETmax [°C] PETmin [°C] PETd < 0 [d] PETd 18-29 [d] PETd > 30 [d] PETd > 35 [d] mean air temperature [°C] Tmax [°C] Tmin [°C] mean radiant temperature [°C] Tmrt_max [°C] Tmrt_min [°C] annual precipitation amount [mm] RR < 1 mm [d] RR > 1 mm [d] RR > 5 mm [d] averaged relative humidity [%] RH > 93 % [d] averaged cloud cover [octas] cloud cover < 5 octas [d] averaged vapour pressure [hPa] VP > 18 hPa [d] averaged wind velocity [m/s] Vmax [m/s] V > 8 m/s [d] SN > 10 cm [d] 8.9 47.3 -23.7 108 63 20 7 11.0 32.8 -16.5 25.5 62.6 -23.1 1046.1 205 160 70 64.4 13 4.1 199 8.9 9 4.2 14.0 17 18 9.1 42.9 -26.7 100 74 18 7 5.8 24.0 -25.2 23.3 54.4 -31.6 1328.6 219 147 83 83.7 94 5.5 134 8.3 1 1.3 22.6 4 70 9.1 47.3 -23.7 107 65 22 8 11.2 33.5 -16.5 25.8 62.6 -23.1 1032.4 205 160 68 64.4 13 4.0 201 9.0 11 4.2 14 17 21 9.7 44.1 -26.2 93 74 22 10 6.1 24 -25.2 24.0 55.6 -28.5 1384.2 217 149 85 82.6 86 5.4 137 8.3 2 1.3 22.6 4 59 Threshold values based on PET like cold stress, heat stress, and thermal comfort are reproduced well by REMO. High discrepancies occur in mean air temperature (ΔT=5.2 K and ΔT=5.1 K for 1961-1990 and 1971-2000, respectively) in fog, sunshine, wind, and snow conditions. The relative humidity is underestimated about 20% by REMO and thus the number of days with fog is sevenfold underestimated. The number of days less than 5 octas is overestimated due to a reduction of the averaged cloud cover by 0.6 octas (1961-1990) and 1.4 octas (1971-2000), respectively. The averaged wind velocity is tripled by REMO. Moreover, the snow cover is marginally simulated. This discrepancy may be due to an underestimation of precipitation and an overestimation of air temperature by REMO. The distribution of precipitation shown in Table 5 indicates allseasonally an underestimation. 185 Table 5: Comparison of monthly and seasonal precipitation concerning REMO and DWD data for the periods 1961-1990 and 1971-2000 for Titisee Titisee precipitation [mm] monthly January February March April May June July August September October November December seasonal winter (DJF) spring (MAM) summer (JJA) autumn (SON) 1961-1990 REMO DWD 1971-2000 [%] REMO DWD [%] 92.3 85.7 62.0 75.3 99.2 98.9 91.0 86.2 78.3 81.1 96.9 99.1 129.4 106.1 108.7 102.3 112.7 112.8 102.3 112.2 79.3 93.0 131.6 138.2 1.4 1.2 1.8 1.4 1.1 1.1 1.1 1.3 1.0 1.1 1.4 1.4 96.9 83.4 61.3 71.1 95.3 95.5 87.2 88.5 81.3 67.3 98.2 106.5 146.0 104.9 110.0 99.1 109.9 123.0 105.3 95.2 91.9 111.4 138.8 148.7 1.5 1.3 1.8 1.4 1.2 1.3 1.2 1.1 1.1 1.7 1.4 1.4 92.4 78.8 92.0 85.4 124.6 107.9 109.1 101.3 1.3 1.4 1.2 1.2 109.1 75.9 90.4 82.3 132.9 106.3 107.8 114.0 1.2 1.4 1.2 1.4 The exclusive comparative parameters for REMO and DWD data for the two periods 1961-1990 and 1971-2000 for Freiburg are given in Table 6 and 7. The parameters measured by DWD and modelled by REMO are inconsistent for the mean air temperature, relative humidity, days with sunshine, fog and sultriness. While thermal comfortable conditions and cold stress agree well, heat stress is overestimated by REMO. The annual PET varies about 2 K and indicates the greatest difference of the studied stations. The modelled mean air and radiant temperature of Freiburg differs about 4 K for both time spans as well. While averaged wind velocity and relative humidity show almost the magnitude, the number of stormy and foggy days differs. Cloud cover is underestimated by 1.4 octas in both time spans as well, thus heat stress, sunshine and sultriness can be overestimated. Although the precipitation is overestimated the numbers of dry and rainy days are consistent. The monthly distribution of precipitation shown in Table 7 denotes predominately an overestimation of winter and slightly of summer and autumn precipitation. 186 Table 6: Comparison of different selected parameters concerning REMO and DWD data for the periods 1961-1990 and 1971-2000 for Freiburg; corresponding notations are: PETa: annual PET, Tmrt: mean radiant temperature, RR: precipitation, RH: relative humidity, VP: vapour pressure, V: wind velocity, SN: snow cover Freiburg Parameter PETa [°C] PETmax [°C] PETmin [°C] PET < 0 [d] PET = 18-29 [d] PET > 30 [d] PET > 35 [d] mean air temperature [°C] Tmax [°C] Tmin [°C] mean radiation temperature [°C] Tmrt_max [°C] Tmrt_min [°C] annual precipitation amount [mm] RR < 1 mm [d] RR > 1 mm [d] RR > 5 mm [d] averaged relative humidity [%] RH > 93 % [d] averaged cloud cover [octas] cloud cover < 5 octas [d] averaged vapour pressure [hPa] VP > 18 hPa [d] averaged wind velocity [m/s] Vmax [m/s] V > 8 m/s [d] SN > 10 cm [d] 4. 1961-1990 REMO 13.1 52.2 -21.3 59 84 38 19 14.6 38.1 -13.8 26.3 65.7 -18.5 1150.0 232 133 68 62.3 7 4.1 196 10.9 33 3.3 10.7 3 10 1971-2000 DWD 11.2 41.9 -20.2 61 90 13 2 10.7 30.8 -15.0 30.5 55.7 -25.8 954.8 235 130 62 72.4 24 5.5 121 9.9 11 2.9 12.3 12 3 REMO 13.3 52.2 -21.3 59 84 42 21 14.8 38.1 -13.8 26.5 65.7 -18.5 1130.5 234 131 66 62.3 7 4.1 196 11.0 36 3.3 11.7 4 13 DWD 11.3 39.2 -20.2 56 90 12 1 11.2 30.8 -15.0 31.0 55.7 -18.5 929.5 238 127 61 71.2 18 5.5 117 10.0 12 3.1 12.3 12 3 Conclusions The difficulties in modelling air temperature and precipitation by REMO are known. Surprisingly, REMO simulates the climate for Norderney (North Sea) well. Regional discrepancies may be due to averaging (10 km x 10 km area) as well. The North Sea is characterized by lowland and by little differences in elevation. Thus, the radiation budget is rather consistent and uniformly distributed and not affected by orography. Consequently, derived parameters do not vary enormously. 187 Table 7: Comparison of monthly and seasonal precipitation concerning REMO and DWD data for the periods 1961-1990 and 1971-2000 for Freiburg Freiburg precipitation [mm] monthly January February March April May June July August September October November December seasonal winter (DJF) spring (MAM) summer (JJA) autumn (SON) 1961-1990 REMO DWD 1971-2000 [%] REMO DWD [%] 94.1 84.3 50.3 72.9 123.9 145.6 116.7 96.0 80.3 75.4 94.7 115.7 60.1 54.5 64.4 80.3 105.7 116.1 95.6 102.6 71.9 65.6 72.3 65.8 0.6 0.6 1.3 1.1 0.9 0.8 0.8 1.1 0.9 0.9 0.8 0.6 100.2 80.5 54.0 61.1 121.1 138.7 113.3 105.4 78.0 58.7 97.2 122.3 53.1 54.0 58.3 73.7 102.8 109.0 99.2 84.0 75.9 75.3 70.8 73.6 0.5 0.7 1.1 1.2 0.8 0.8 0.9 0.8 1.0 1.3 0.7 0.6 98.0 82.4 119.4 83.5 60.1 83.5 104.8 69.9 0.6 1.0 0.9 0.8 99.9 78.7 119.1 78.0 64.8 78.2 97.4 74.0 0.6 1.0 0.8 0.9 An underestimation of thermal comfort and overestimation of precipitation and stormy days are modelled by REMO. Concerning orographical areas like the Black Forest, more difficulties in reproducing the climate conditions occur. In higher levels precipitation and snow depth are underestimated, whereas wind velocity is overestimated. In lower regions the precipitation and threshold values based on PET overestimated. A general model problem is the relative humidity. The relative humidity and derived fog conditions are notably underestimated in orographical areas and overestimated in lowlands. The parameters considered in this study are derived from the raw data by REMO. Therefore, these derived parameters show uncertainties as well. The analysis of DWD data allows thus a careful interpretation of the modelled data and an evaluation of model uncertainties. The presented study represents the first step in comparison of climate condition for tourism and recreation approaches. Additionally, analysis and comparison of the climatic tourism potential are planned using CLM data as well. Acknowledgement This research study is supported by the Federal Ministry of Education and Research (Bundesministerium für Bildung und Forschung, BMBF) within the framework of the “klimazwei” initiative. Thanks to the German Weather Service (DWD) for the preparation of climate data of Germany. 188 References BENISTON, M., 1997: Variations of snow depth and duration in the Swiss Alps over the last 50 years: links to changes in large-scale climatic forcing. - Climatic Change 36, 281-300. BESANCENOT, J. P. ,1990: Climat et tourisme. Masson édit. - Collection géographie, Paris. BÖHM, U., M. KÜCKEN, W. AHRENS, A. BLOCK, D. HAUFFE, K. KEULER, B. ROCKEL, A. WILL, 2006: CLM - The climate version of LM: Brief Description and long term application. COSMO Newsletter Nr.6, 225-235. BREILING, M., J. CHARAMZA, 1999: The impact of global warming on winter tourism and skiing: a regionalised model for Austrian snow conditions. - Regional Environmental Change 1, 4-14. DE FREITAS, C. R., 2003: Tourism climatology: evaluating environmental information for decision making and business planning in the recreation and tourism sector. - Int. J. Biometerol. 48, 45-54. GÓMEZ MARTÍN, M. B., 2004: An evaluation of the tourist potential of the climate in Catalonia (Spain): A regional study. - Geogr. Ann. 86 A, 249-264. IPCC, 2001: Climate Change 2001: The Scientific Basis. Contribution of the Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change (HOUGHTON, J.T. et al., eds.). - Cambridge and New York 2001. HÖPPE, P., 1999: The physiological equivalent temperature - a universal index for the biometeorological assessment of the thermal environment. - Int. J. Biometeorol. 43, 71-75. JACOB, D., 2001: A note on the simulation of the annual and inter-annual variability of the water budget over the Baltic Sea drainage basin. - Meteorol. Atmos. Phys. 77, 61–73. JACOB, D., L. BÄRRING, O. B. CHRISTENSEN, J. H. CHRISTENSEN, S. HAGEMANN, M. HIRSCHI, E. KJELLSTRÖM, G. LENDERINK, B. ROCKEL, C. SCHÄR, S. I. SENEVIRATNE, S. SOMOT, A. VAN ULDEN, B. VAN DEN HURK, 2007: An inter-comparison of regional climate models for Europe: Design of the experiments and model performance. - Climatic Change 81: 31-52. KULINAT, K., A. STEINECKE, 1984: Geographie des Freizeit- und Fremdenverkehrs. - Erträge der Forschung Band 212, Darmstadt. MATZARAKIS, A., 2007: Entwicklung einer Bewertungsmethodik zur Integration von Wetterund Klimabedingungen im Tourismus. – Ber. Met. Inst.Univ. Freiburg 16, 73-79. MATZARAKIS, A., H. MAYER, 1996: Another kind of environmental stress: Thermal stress. – WHO Newsletter No. 18, 7-10. MATZARAKIS, A., C. R. DE FREITAS, D. SCOTT, 2004 (eds.): Advances in Tourism Climatology. - Ber. Meteor. Inst. Univ. Freiburg Nr. 12. MATZARAKIS, A., H. MAYER, M. G. IZIOMON, 1999: Applications of a universal thermal index: physiological equivalent temperature. - Int. J. Biometeorl. 43, 76-84. MATZARAKIS, A., F. RUTZ, H. MAYER, 2007: Modelling Radiation fluxes in simple and complex environments - Application of the RayMan model. - Int. J. Biometeorol. 51, 323-334. ROTH, R, N. PRINZ, A. KRÄMER, C. SCHNEIDER, J. SCHÖNBEIN, 2005: Nachhaltige Entwicklung des Schneesports und des Wintersporttourismus in Baden-Württemberg. Ein Leitfaden für Politik, Sport, Kommunen und touristische Leistungsträger. - Wirtschaftsministerium BadenWürttemberg. SCHARLAU, K., 1943: Die Schwüle als meßbare Größe. - Bioklimat. Beibl. 10, 19. 189 ZEBISCH, M., T. GROTHMANN, D. SCHRÖTER, C. HASSE, U. FRITSCH, W. CRAMER, 2005: Climate change in Germany - Vulnerability and adaptation of climate sensitive sectors. - Climate Change 10/05. Umweltbundesamt, Dessau. 2005 Authors' address: Dipl.-Met. Christina Endler (christina.endler@meteo,uni-freiburg.de) Prof. Dr. Andreas Matzarakis (andreas. [email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany 190 191 Regional determination of historical heavy precipitation for the reconstruction of extreme flood events Paul Dostal1), Florian Imbery1), Katrin Bürger2) and Jochen Seidel3) 1) 2) Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany Department of Physical Geography, Albert-Ludwigs-University of Freiburg, Germany 3) Department of Hydrology and Geohydrology, University of Stuttgart, Germany Abstract The reconstruction of historical extreme hydrometeorological events and their triggering precipitation patterns contributes to a validation of extreme value statistics. This can mitigate several uncertainties in the flood risk analysis, e.g. in calculating possible discharges with extreme value statistics, which are based on short reference data series (BÜRGER et al., 2006 a, b). It is shown that the presented reconstruction of the extreme flood of 1824 in the Neckar catchment can be used in a recent flood risk management and to validate the results of trend and extreme value analysis of hydrometeorological time series. Bestimmung historischer regionaler Starkniederschläge für die Rekonstruktion extremer Hochwasserereignisse Zusammenfassung Mit der detaillierten Rekonstruktion von historischen Hochwasserereignissen und deren auslösenden großräumigen Wetterlagen steht ein zusätzliches Hilfsmittel für die Analyse von Hochwasserwahrscheinlichkeiten zur Verfügung. Damit kann ein wichtiger Beitrag geleistet werden, bisherige Unsicherheiten in der statistischen Bewertung von relativ kurzen Datenreihen für die Bestimmung des Hochwasserrisikos zu verringern (BÜRGER et al., 2006 a, b). Es wird gezeigt, dass die hier vorgestellte Rekonstruktion des Neckarhochwassers 1824 sowohl in einem aktuellen Hochwassermanagement als auch für die Validierung aktueller Entwicklungen in der Trendanalyse und der Extremwertstatistik hydrometeorologischer Zeitreihen eingesetzt werden kann. 1. Introduction Extreme flood events in the context of the climate change discussion have received increasing attention in Germany in recent years (IPCC, 2007). In particular, the Elbe flood in 2002 indicates changes in the frequency and characteristics of floods which may reflect climatic transitions. This led to an initiative by the German Federal Ministry of Education and Research (BMBF) for a ”Risk Management of Extreme Flood Events“ to avoid or limit the adverse impact of floods in Germany. The research group Xfloods from the Meteorological Institute and the Department of Physical Geography, both AlbertLudwigs-University of Freiburg (Germany), is part of this initiative with a special focus on extreme floods in the Federal State of Baden-Württemberg, SW Germany. The research project integrates the information of historical data to identify and quantify extreme flood events of the past 500 years as a basis for flood risk management as well as for the calculation of maximum regional precipitation. Different data were extracted to gain area precipitation of the Neckar Catchment for the 1824 flood from historical re- 192 cords as well as local annuals and chronicles from 1500-1900 and supplemented by instrumental observations available since the middle of the 18th century. The project can contribute to a safer handling of extreme floods in the future. This applies in particular to the analysis and modelling of millenarian flood events, which have been little considered previously. In such a way, the knowledge of the past can be integrated in the flood protection for tomorrow. An overview of the Neckar catchment is shown in Fig 1. Fig. 1: Neckar catchment with its main tributaries Nagold, Enz, Kocher and Jagst 2. Methodological concept and data overview 2.1 Data For the reconstruction of the historical strong precipitation event of 1824, a model with various regression tools and geostatistical methods was developed. Fig. 2 contains an overview of the used data and the different steps of processing the precipitation patterns and discharges of the Neckar. 193 Fig. 2: Flowchart of data processing for the reconstruction of the 1824 extreme flood event on the Neckar catchment 2.2 Meteorological data from October 1824 For the calculation of the historical area precipitation (AP1824), sets of historical meteorological data, like precipitation (N), air Temperature (T), sea level pressure (SLP), wind speed (u) and wind direction (dd) were used. With a European wide dataset of historical meteorological data (BARRIENDOS and RODRIGO, 2006), synoptic weather charts were developed for October 1824. Using synoptic weather charts it was possible to find similar weather situations in the reference dataset with the modern analogue method (BARDOSSY and FILIZ, 2003). Weather charts help to understand the superior weather situation for this hydrometeorological extreme event and improve the results of the Kriging calculation for the spatial rainfall (ISAAKS and SRIVASTAVA, 1989; HINTERDING, 2001). The data for the event of 1824, which were used for the synoptic weather reconstruction and the area precipitation calculation, are contained in Table 1. To calculate the spatial precipitation, reference data from the German Weather Service (Deutscher Wetterdienst DWD) were used. The reference data are meteorological data from 1900-2006 provided from the DWD. As a reference basis the whole meteorological data amount of South Germany (Rheinland-Pfalz, Hessen, Saarland, Bavaria and Baden-Württemberg) were used with a total of more than 1000 meteorological stations. The meteorological data were available in several resolutions, e.g. 10-min means and totals (precipitation), respectively, as well as values from climate stations observed three times a day (7 am, 2 pm and 9 pm LT). 194 Table 1: Meteorological stations since the 18th century as data basis for the 1824 flood analysis Location Measuring Time Country Freiburg 1780 - 2006 Germany Karlsruhe 1799 - 2006 Germany Stuttgart 1800 - 2006 Germany Freudenstadt 1824 - 2006 Germany Rottweil 1818 - 1827 Germany Ellwangen 1818 - 1825 Germany Giengen 1820 - 2006 Germany Genkingen 1820 - 1826 Germany Augsburg 1812 - 2006 Germany Basel 1755 - 2006 Switzerland Bern 1824 - 2006 Switzerland Straßburg 1800 - 2006 France Paris 1824 - 2006 France Armagh 1795 - 1880 Ireland Edinburgh 1785 - 2006 Scotland Barcelona 1780 - 2006 Spain Cadiz 1786 - 2006 Spain Madrid 1786 - 2006 Spain Milano 1763 - 2006 Italy Padova 1766 - 2006 Italy Palermo 1790 - 2006 Italy Prague 1781 - 2006 Czech Rep. Reykjavik 1816 - 2006 Iceland Stockholm 1756 - 2006 Sweden Uppsala 1722 - 2006 Sweden 2.3 Synoptic situation On the basis of historical measurements of air pressure (see Table 1), the synoptic situation over Central Europe and parts of the northern Atlantic was reconstructed from 26 October to 1 November 1824. These weather patterns were compared with recent weather data from the DWD, based on the monthly weather bulletins of the DWD and the precipitation statistics for Southwest Germany (BARTELS, 1997), with the aim to find a contemporary similar weather situation. The best conformity was determined for the period from 26 to 28 October 1998. The weather situation of October 1998 also caused strong precipitation and floods in the Neckar catchment. 195 2.4 Regression model 1998 – 1824 Assuming that the regional distributions of precipitation in October 1824 and 1998 have a high similarity, the regional precipitation pattern in the Neckar Catchment for the flood event of 1824 was developed as a function of the 1998 precipitation and the orographic conditions in the research area (BARDOSSY and FILIZ, 2003; JACOBEIT et al., 2003). A simple linear regression between historical precipitation measurements and precipitation data from 1998 was not suitable to give a sufficient result. Hence, a linearlogarithmic regression model in the form ⎞ ⎛ 4 ⎟ ⋅ N1998 N1824 = ⎜⎜1.5 + 4 ⎟ ln(DEM) ⎠ ⎝ (1) with N as precipitation amount for each event and DEM as the altitude (m a.s.l.) of the measuring location, derived from a digital elevation model, was used. As shown by the coefficient of determination (R2), this function provided a good correlation between measured and modelled precipitation values (R2 = 0.88). Assuming a high distribution of the historical measuring points in the Neckar Catchment, the total rainfall amount of 220 points in the Neckar catchment was calculated on the basis of this equation. 2.5 Spatial interpolation with geostatistic methods The spatial variation of short-term precipitation is highly variable in space. For the spatial interpolation of the precipitation event, a kriging ruler was used. Ordinary kriging calculates a weighted sum of known observed precipitation amounts to estimate an unknown value using the equation: n z 0 = ∑ ωi z i (2) i =1 The observated values are selected from the sample data set within a userspecified neighborhood. Covariances are derived for observations within the neighborhood using the model variogram. The kriging process calculates a weight for each sample datum in the neighborhood based on its statistical distance, as opposed to Euclidean distance, to each and every datum in the neighborhood as well as to the estimate location. Then it employs a weighted linear combination of the data in the neighborhood to estimate the new value. The ordinary kriging system combines a set of matrices as a system of n+1 linear equations. They must be solved that the condition of unbiasedness holds for each neighbourhood of n values. It is enforced by requiring the weights derived for each sample in the neighbourhood to sum to one: n ∑ω i =1 i =1 (3) As summarised by ISAAKS and SRIVASTAVA (1989), for each estimated location weights are derived for every sample in the neighbourhood in the following way: ω = C −1 ⋅ D (4) where C is the covariance between sample data and D is the covariance between sample data and estimated location. The covariances are stored in matrices that, when multiplied together, produce a system of equations. In actuality, another equation is added to 196 the set using a technique known as the Lagrange parameter, because the weights are constrained to sum to one. The estimated value and its error variance becomes a function of n+1 variables, n weights, and the Lagrange multiplier - a dummy variable that forces the sum of weights to equal one. The solution of n+1 equations produces a set of weights that minimises the modeled error variance under the constraint that the weights must sum to one. C and D are determined for every estimation location. The C matrix considers the effects of clustering and redundancy in data samples. If two or more points are close together in the inverse distance method, the number of points in the cluster adversely affects the weighting. If the covariances between these clustered samples is large, the effect of the clustering in D is lessened by the inverse of the covariances in C. Kriging, therefore, regards distance, clustering or redundancy, and spatial dependency of samples when estimating each new location. Ordinary kriging tries to create a mean residual or error that is equal to zero and minimises the variance of the errors. In fact, the mean error and the true variance of the errors for the estimates are always unknown. Though it is possible to work with an average error and minimise the variance of the modeled error. The error variance is derived using the variances calculated during the weighted linear combination. The minimised error variance for an estimated location is related by: Var 2 = C 2 − ω ⋅ D (5) The mean-squared error is a prediction error. It is probabilistic in nature based on the relationship between the variogram model and its parameters modeled from samples of the stationary random process. Prediction variance, or kriging variance, is generated simultaneously for each kriged location. Both the kriged and variance surfaces can be visualised side by side. The kriging variance is likely to be lower when the sampling density is higher, although highly variable covariances in clustered areas do occur. Such a surface provides invaluable information about the fit of the model to the sample data. It also provides information on the effects of the distribution of the sampling scheme relative to the applied model (see Fig. 3). Fig. 3: Flowchart of spatial modelling of heavy precipitation at the Neckar catchment in October 1824 197 2.6 Runoff simulation LARSIM (Large Area Run-off Simulation Model) is a river basin model (BREMICKER et al., 2005), which was developed for the systematic modelling of runoff generation and flood-routing. LARSIM can be applied both as a water balance model for continuous simulation and as an event-based flood forecast model. It uses proven and comparatively simple model components, which are satisfying for practical purposes. For example, LARSIM is used for the flood prediction at the Neckar catchment. For validation, the modelled spatial precipitation pattern for the flood event of October 1824 was used for the input in LARSIM for comparing modelled to historical run-off data (DOSTAL et al., 2007). 3. Results To analyse the flood triggering weather situation of October 1824 and, therefore, the AP1824, a similar weather situation had to be found in the reference data. Examining the reference data, the best correspondence was the weather situation of 27 and 28 October 1998. This was a weather situation, which brought nearly in the total Neckar catchment a great amount of precipitation with devastating consequences at the Neckar and the Upper Rhine tributaries due to heavy runoffs. In comparison to the historical flood event of 1824, the flood of 1998 was not so severe because of different weather and environmental conditions (e.g. minor precipitation totals) in the run up of this extreme event. 3.1 The extreme weather situation of October 1824 The weather situation of October 1824 can be described as follows: At the beginning of October 1824, the weather was predominantly cloudy. Around 20 October 1824, higher air temperature indicated fair weather. Starting on 26 October, a dramatic change in the weather was signalised when thunderstorms arose. The weather process causing the flood is described in the historical sources as follows: ”The evening of 26th October thunderstorms, accompanied by rain, broke out in several regions of Southwest Germany, but no unusual rise of the water levels in the rivers could be noticed. On the 27th October, there was some rainfall in the upper Neckar valley. In the evening of 28th October unusually strong rain weather began, which lasted 36 hours with few interruptions up to the early morning hours of 30th October and caused this exceptional inundation. The rain quantity was very high in the Black Forest. The regions of the river Enz and the lower Neckar were especially affected by this inundation. In many places, the rainy weather with exception of smaller interruptions continued till 2th November.“ (SCHÜBLER, 1825) The reconstructed weather charts start at 24 October 1824 and show the development of the weather situation until the end of this hydrometeorological extreme event (see Fig. 4). 198 Fig. 4: Synoptic situation in Europe from 26 to 29 October 1824 3.2 Calculation of the precipitation in October 1824 In the Neckar catchment and its surrounding area are 220 rainfall stations located. The precipitation of October 1824 was modelled referring to all this rainfall stations. The modelling was done with a linear regression. The coefficient of determination (R2 = 0.88) points out to a very good comparability to the historical measured data. In combination with the reference data of the flood event of 1998 and the data of the maximum precipitation (BARTELS, 1997) for the Neckar catchment, the precipitation totals are shown in Table 2. Table 2: Precipitation totals for the year 1824 and modelled precipitation data (in mm) Location Freudenstadt Wangen/Neckar Hohenheim Stuttgart Genkingen Tübingen Giengen Model 197 121 128 111 92 103 92 Measurements 194 149 127 124 92 89 89 199 3.3 Modelling the area precipitation of the 1824 event Aim of the presented project was to calculate the area precipitation for a 1 km2 grid to fit the modelled data in the water balance model LARSIM, which is also based on a 1 km2 grid. As seen in Fig. 5, the highest precipitation with values up to 230 mm in 36 hours occurred in the northern parts of the Black Forest. Secondary maxima were located in the northern and eastern parts of the Neckar catchment. This spatial precipitateon distribution can be explained by the atmospheric circulation pattern and local orographic features. The location of the high and low pressure cells in the large-scale synoptic reconstruction implicates south westerly warm and humid air mass flow to Central Europe as a causal mechanism. South-westerly large-scale air flows are typically modified by the mountain ranges of the Vosges mountains and the Black Forest. Particularly the northern parts of the Black Forest frequently receive high rainfall amounts during such weather conditions due to the orographic features on the western site of the Upper Rhine Valley (see Fig. 5). Fig. 5: Modelled precipitation pattern in southwest Germany for the period 28 to 29 October 1824 3.4 Modelled Neckar discharges for the flood event of October 1824 Table 3 and Fig. 6 present the modelled discharges of the river Neckar and its tributaries derived from the calculated precipitation data, which were integrated in the waterbalance model LARSIM. The discharges show values which are significantly higher than Qextreme discharges for the Neckar (LFU, 2000, 2005). In all probability, the flood 200 event of October 1824 was one of the most extreme flood events during the last 300 to 500 years. Table 3: Discharges for several locations at the Neckar in October 1824 modelled with LARSIM Neckar HQ1824 (m3/s) 137 HQ100 (m3/s) 260 HQextreme (m3/s) 390 Neckar 223 348 522 Horb Neckar 416 549 800 Plochingen Neckar 1345 1145 1600 Nagold Location River Rottweil Oberndorf Nagold 186 181 273 Pforzheim Enz 559 504 757 Vaihingen Enz 544 513 766 Besigheim Enz 594 586 821 Lauffen Neckar 2309 1877 2550 Gaildorf Kocher 411 350 490 Stein Kocher 1001 709 993 Elpershofen Jagst 472 387 582 Untergriesheim Jagst 984 525 771 Rockenau Neckar 4115 2665 3600 Ziegelhausen Neckar 4272 2806 3700 Heidelberg Neckar N/A 2818 3750 River mouth Neckar 4187 2833 3750 Fig. 6: Discharge for the Neckar during the flood in October 1824 from historical administrative documents (KÖNIGLICHES MINISTERIUM DES INNEREN, 1896) and simulated discharges 201 4. Conclusions The application of data from historical sources, like in the actual flood risk management, increases the protection of potential inundation areas. The presented paper reveals the possibilities and the methods to reconstruct hydrometeorological extreme events which lay far in the past. These results show the potential of historical flood analysis for flood risk management. Better understanding of extreme flood events and better protection of endangered areas, goods and humans are the tangible outcomes. The importance of analysing historical floods is the possibility to demonstrate the consequences of such extreme events in a selected river catchment, such as the Neckar. This knowledge can be incorporated into flood risk management. By combining the historical precipitation and flood data with contemporary river channel morphology, current hazards and impacts can be predicted. This will lead to a better understanding of flood processes, as well as their characteristics. Improvements in the analysis of historical extreme events should be extended to other river catchments in order to mitigate the consequences of catastrophes such as the Elbe flood in 2002. Acknowledgement The authors thank the RIMAX project, funded by the German Federal Ministry of Education and Research, for financial support of their work (under grant no. 0330685). References BARDOSSY A., F. FILIZ, 2003: Identification of flood producing atmospheric circulation patterns. - In: V.R. THORNDYCRAFT, M. BENITO, M. BARRIENDOS, M.C. LLASAT (eds.), Paleofloods, Historical Data & Climatic Variability. Proc. PHEFRA Int. Workshop, Barcelona, Spain, 307-312. BARRIENDOS, M., F. RODRIGO, 2006: Study of historical flood events of Spanish rivers using documentary data. - Hydrological Sciences Journal 51, 765-783. BARTELS H., 1997: Starkniederschlagshöhen für Deutschland: KOSTRA. - Deutscher Wetterdienst, Offenbach. BREMICKER M., A. LUCE, I. HAAG, A. SIEBER, 2005: Das Wasserhaushaltsmodell LARSIM Modellgrundlagen (Hochwasser Vorhersagezentrale Baden-Württemberg); [available from www.hvz.baden-wuerttemberg.de/pdf/LARSIM DE 2005-06-24]. BÜRGER K., J. SEIDEL, F. IMBERY, P. DOSTAL, 2006a: RIMAX-Projekt Xfloods: Analyse historischer Hochwassereignisse für ein integratives Konzept zum vorbeugenden Hochwasserschutz. - Umweltwissenschaften und Schadstoff-Forschung (UWSF) 18, 27-29. BÜRGER, K., P. DOSTAL, J. SEIDEL, F. IMBERY, H. MAYER, R. GLASER, 2006b: Hydrometeorological reconstruction of the 1824 flood event in the Neckar River basin (southwest Germany). - Hydrological Sciences Journal 51, 864-877. DOSTAL P., K. BÜRGER, J. SEIDEL, F. IMBERY, 2007: Lernen aus der Vergangenheit. Historische Hochwasseranalyse. Ein Beitrag zum heutigen Hochwasserschutz. - Berichte zur Deutschen Landeskunde Bd. 81, 233-245. HINTERDING A., 2001: InterNied - a geostatistical interpolation procedure four hourly measured precipitation data. - In: P. KRAHE, D. HERPERTZ (eds.), International Commission for the Hydrology of the Rhine Basin (CHR), Report I-20, 41-47. 202 ISAAKS E.H., R.M. SRIVASTAVA, 1989: An introduction to applied geostatistics. - New York, Oxford University Press. IPCC Report, 2007: Climate Change 2007: Synthesis Report. www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4_syr_introduction.pdf] [available from JACOBEIT J., R. GLASER, J. LUTERBACHER, M. NONNEBACHER, H. WANNER, 2003: Links between flood events in central Europe since AD 1500 and the large-scale atmospheric circulation. - In: V.R. THORNDYCRAFT, M. BENITO, M. BARRIENDOS, M.C. LLASAT (eds.), Paleofloods, Historical Data & Climatic Variability. Proc. PHEFRA Int. Workshop, Barcelona, Spain, 269-274. KÖNIGLICHES MINISTERIUM DES INNEREN, 1896: Verwaltungsbericht der Königlichen Ministerialabteilung für den Straßen- und Wasserbau für die Rechnungsjahre vom 1. Februar 1893/1894 und 1894/1895, II Abt. Wasserbau. LANDESANSTALT FÜR UMWELTSCHUTZ BADEN-WÜRTTEMBERG (LFU), 2000: Das Hochwasser vom Oktober/November 1998 in Baden-Württemberg. - Oberirdische Gewässer/Gewässerökologie, Karlsruhe. LANDESANSTALT FÜR UMWELTSCHUTZ BADEN-WÜRTTEMBERG (LFU), 2005: Abflusskennwerte in Baden-Württemberg. - Oberirdische Gewässer/Gewässerökologie, Karlsruhe. SCHÜBLER, G., 1825: Über die ungewöhnliche Überschwemmung zu Ende Octobers des vorherigen Jahres und die dabei in verschiedenen Gegenden Württembergs gefallene Regenmenge. – Annalen der Physik und Chemie 3, zweites Stück. Authors' address: Dr. Paul Dostal (paul.dostal@meteo,uni-freiburg.de) Dr. Florian Imbery ([email protected]) Meteorological Institute, Albert-Ludwigs-University of Freiburg Werthmannstr. 10, D-79085 Freiburg, Germany Dr. Katrin Bürger ([email protected]) Department of Physical Geography, Albert-Ludwigs-University of Freiburg Werthmannstr. 4, D-79085 Freiburg, Germany Dr. Jochen Seidel ([email protected]) Department of Hydrology and Geohydrology, University of Stuttgart Pfaffenwaldring 61, D-70569 Stuttgart, Germany 203 An analysis of cloud observations from Vernadsky, Antarctica Amélie Kirchgäßner British Antarctic Survey, Natural Environmental Research Council, United Kingdom Abstract This paper presents results of a comprehensive analysis of cloud observations made at the Antarctic base Faraday/Vernadsky between 1960 and 2005. The annual total cloud cover has increased significantly during this period with the strongest and most significant positive trend found in winter, and positive tendencies observable in all seasons. The increase in total cloud cover is neither reflected in the low cloud amount nor in the number of records for low, medium or high clouds. It is therefore thought that the increase in total cloud cover is caused by an increase in the amount of medium and/or high clouds. Instead records for low cloud amount show a redistribution from cases of extreme cloud cover (0, 1, 7 and 8 okta), which account for up to 90% of annual records, to cases of moderate cloud cover. This indicates a shift from low-level stratiform towards convective clouds. Eine Analyse von Wolkenbeobachtungen an der Station Vernadsky, Antarktis Zusammenfassung In diesem Artikel werden die Ergebnisse einer umfassenden Analyse von Wolkenbeobachtungen an der antarktischen Forschungsstation Faraday/Vernadsky zwischen 1960 und 2005 präsentiert. Im Jahresmittel hat die Gesamtbedeckung über den Beobachtungszeitraum signifikant zugenommen, wobei der stärkste und signifikanteste Trend im Winter zu beobachten ist, nicht signifikante positive Tendenzen herrschen jedoch das gesamte Jahr hindurch. Die Zunahme des Gesamtbedeckungsgrades drückt sich weder im Bedeckungsgrad niedriger Wolken noch in der Häufigkeit der Beobachtungen von niedrigen, mittelhohen oder hohen Wolken aus. Daher kann angenommen werden, dass die Zunahme des Gesamtbedeckungsgrades durch eine Zunahme des Bedeckungsgrades durch mittelhohe und/oder hohe Wolken verursacht wird. Vielmehr zeigen die Beobachtungen niedriger Bewölkung eine Umverteilung von Fällen extremer Wolkenbedeckung (total bedeckt: 7 oder 8 Achtel, wolkenlos: 0 oder 1 Achtel), die bis zu 90% der jährlichen Beobachtungen ausmachen, hin zu Fällen moderater Wolkenbedeckung. Dies weist auf eine Verschiebung von niedriger stratiformer Bewölkung zu mehr konvektiver Bewölkung hin. 1. Motivation Clouds play an important role in the radiation budget and hence the energy balance of the Earth’s atmosphere. Additionally to their direct impact they also have an influence through a variety of feedback mechanisms. Clouds are very efficient absorbers of infrared radiation. They therefore have a strong natural greenhouse effect and can contribute 204 to the warming of the atmosphere. At the same time clouds efficiently reflect incoming solar radiation and thereby cool down the earth’s atmosphere. Changing practically any aspect of clouds, be it cloud type, location, liquid water content, base height, phase, habit, spectrum or life time, influences whether the cooling or the warming aspects of their properties dominate. It is a vital topic of current research to improve our understanding of how clouds will change as a consequence of global warming, and how this in turn, via feedback mechanisms, will influence future climate conditions. Over the past 50 years, the west coast of the Antarctic Peninsula has been one of the most rapidly warming parts of the planet, with the largest warming occurring in the winter season (TURNER et al., 2005; VAUGHAN et al., 2001; KING et al., 2004). These different studies find warming rates which are one order of magnitude bigger than the mean rate of global warming as reported by the Intergovernmental Panel on Climate Change (SOLOMON et al., 2007)). Hence it is of special importance to investigate the role clouds play particularly in this region. 2. Data 2.1 Synoptic observations The first permanent research base on the Argentine Islands (65°15’S, 64°16’W) was built in 1947 on the site that had previously occupied a research base during the British Graham Land Expedition. On the 6th of February 1996 the station - until then known as “Faraday”- was officially handed over to the Ukraine and is since then known under its present name “Vernadsky”. Fig. 1 shows the location of the base in Antarctica, and in more detail its position on Galindez Island. The period under investigation covers the years from 1960 to 2005. Cloud observations in general were begun at Faraday in 1947, but only since 1960 has the whole range of cloud parameters - as recorded presently - been observed regularly. Data used for this study consists of observations of various cloud parameters, such as total cloud cover and low cloud amount. These parameters are usually observed and recorded at threehourly intervals, though only six-hourly observations were used as the observations at 00, 06, 12, and 18 UTC were recorded most consistently. The data set covers 16802 days, so - looking at 6 hourly data - this means records potentially comprise 67208 recordings per parameter. Of these 67208 potential observations 191 are missing, the data set is therefore 99.7% complete. The reduced number of records in the second half of the 1980s is most likely connected to the installation of an Automatic Weather Station (AWS). A high number of missed observations in 1992 occurred during the replacement of this first AWS with a newer model in March of that year. By way of trial several calculations have been carried out using only observations from 12UTC, i.e. an observation time providing daylight conditions all year round. These tests verified that neither the general difficulty of cloud observations during darkness nor the switch over to night time observations by non-meteorologists after the introduction of the AWS change the statistical characteristics of the data set (Fig. 3). 205 Fig. 1: Map of Antarctica showing details of location of base Vernadsky Observations of cloud parameters are recorded according to the World Meteorological Organization’s Manual on Codes (WMO, 1995). All trends presented in this study are calculated using a standard least-squares method. The methodology used to calculate the significance levels is based upon SANTER et al. (2000). 3. Total cloud cover Over the period under investigation the average total cloud cover at Vernadsky was 6.5 oktas with a standard deviation of 0.2. If monthly mean values of cloud cover are averaged over the complete period, a mean annual course becomes apparent. Its main feature is a minimum in June with 6.0 oktas total cloud cover, which is imbedded in otherwise quite homogeneous cloud cover conditions of up to 6.8 okta (February and November), but there’s no distinct secondary minimum. The interannual variability is very small with values ranging from 6.1 (1985) to 6.9 oktas (1972). Fig. 2 shows the annual and seasonal mean total cloud cover for the period under investigation together with linear trends. The total cloud cover has increased during the years from 1960 to 2005, and this positive trend is present in all seasons. The trend is strong- 206 est during winter months (0.014 oktas/yr) and about half as strong annually. The annual trend is significant at the 5%-level; the trend for the winter months is significant at the 1%-level. Fig. 2: Annual and seasonal mean total cloud cover (in oktas) at Vernadsky for period 1960 to 2005 207 The increase in total cloud cover described above is found to be a result of a decrease of observations recording 0 or 1 okta. The annual number of these observations has decreased significantly, a trend that is also visible (and significant) in summer, autumn and is especially distinct in winter (Table 1). While generally this category doesn’t contain many observations during winter months on average 12%, but up to 21%, of observations record either 0 or 1 okta total cloud cover. Other categories of total cloud cover don’t show any significant trends. Table 1: Annual and seasonal trends in observations recording 0 or 1 okta total cloud cover at Vernadsky during the period 1960 to 2005 (*** significant at 1%, ** significant at 5%, * significant at 10%) 0 + 1 oktas Annual mean -0.12*** Spring (SON) -0.06 Summer (DJF) -0.08* Autumn (MAM) -0.10* Winter (JJA) 4. Trend (obs/yr) -0.23*** Low cloud amount Low clouds are the only cloud group for which the WMO guidelines require the observation of the cloud amount. The long-term annual average of low cloud amount at Vernadsky is 4.3 (± 0.7) oktas and thus more than 2 oktas smaller than the total cloud cover. Annual mean low cloud amount shows a much larger interannual variability than the total cloud cover with values ranging from 3.3 oktas in 1998 to 5.9 oktas in 1972. The interseasonal variability in low cloud amount is also slightly larger than that seen in total cloud cover with amplitudes in individual years ranging from 1.3 oktas in 1998 to 4.3 oktas in 1997. The long-term average over the whole period reveals an annual cycle, which is not generally visible in individual years. It is formed of a minimum of 3.8 okta in July and a maximum in April (4.7 oktas). No significant trends were observed during the period of interest for the annual or seasonal mean low cloud amount. On a monthly basis, however, a significant negative trend was observed for the month of December. The average low cloud amount in December has decreased by 11.5% over the entire period, which corresponds to almost one okta. This trend is significant at the 10% level. Fig. 3 shows the distribution of low cloud amount observations to the different categories from 0 to 8 oktas based on 6 hourly annual and seasonal records as well as for annual records of 12UTC-observations. It can be seen that including observations carried out during darkness does not change the characteristics of the data set. 208 35 30 25 Annual 20 Spring Summer Aut umn Wint er 15 12UTC 10 5 0 0 1 2 3 4 5 6 7 8 O kt as Fig. 3: Percentage of observations recording 0 to 8 okta of low cloud amount Studying the development of occurrences of the different possible observational categories the picture is less distinct than that found in total cloud cover. More detailed analysis shows that the number of extreme cases recorded (0, 1, 7 or 8 oktas) has decreased significantly throughout the year, while observations of moderate low cloud amount have increased. The significant decrease in extreme categories (0, 1, 7 or 8 oktas) is only present in the sum of these categories. The numbers of observations in the individual categories all show negative tendencies, but none of these individual trends is significant. Table 2: Annual and seasonal trends in observations of extreme (left) and moderate cases (right) of low cloud amount at Vernadsky during the period 1960 to 2005 (*** significant at 1%, ** significant at 5%, * significant at 10%) Trend (obs/yr) Σ (0,1,7,8 oktas) Categories (oktas) Trend (obs/yr) Annual mean -0.30* 2, 3, 4 0.09***, 0.04*, 0.04** Spring (SON) -0.28* 2, 4, 5 0.11***, 0.03*, 0.05* Summer (DJF) -0.33*** 2, 3, 4, 6 0.09***, 0.06***, 0.07***, 0.08* Autumn (MAM) -0.27** 2, 3, 6 0.06***, 0.04**, 0.06* Winter (JJA) -0.29* 2, 3, 4 0.10***, 0.06*, 0.04* 209 The decrease of observations of “extreme” categories is balanced by significant increases in the frequency of moderate cases of cloud observations. A significant increase in the annual number of observations recording 2, 3 and 4 oktas of low cloud amount was found. This shift can be observed during all seasons. Results of the trend analysis for extreme and moderate cases of low cloud amount are compiled in Table 2. 5. Conclusions Observational data from Vernadsky show a significant trend towards a higher total cloud cover. Annually the total cloud cover has increased by 0.008 okta, which corresponds to an increase by almost 5% over the complete period from 1960 to 2005. This trend is especially present in winter when the total cloud cover has increased by more than 8 percent over 46 years, a trend that is significant at the 1% level. A tendency towards increased total cloud cover is found throughout the year, during spring, summer and autumn the trend is not significant though. This development manifests itself mainly in a significant decrease of observations with 0 and 1 okta total cloud cover throughout the year. These trends are significant at the 1% level for annual observation numbers (-0.12 obs/year) and number of recordings in winter (-0.23 obs/year). The increase in total cloud cover is not reflected in observations of low cloud amount. These show no significant trend either annually or in any of the season. There’s evidence though that the average amount of low clouds in summer is decreasing, a trend which is significant in December. No indication was found that events of low clouds occur more frequently. Hence it can be assumed that the increase of total cloud amount is not due to an increase of low cloud amount but to an increase in medium or high clouds, especially in winter when the total cloud cover shows its largest increase – while low cloud amount doesn’t change, and in December (summer) when total cloud cover increases while low cloud amount actually shows a decrease. Regarding medium and high clouds there’s no evidence suggesting that the frequency of their occurrence has increased. The observed shift in low cloud amount to more cases of moderate cloud amount can be interpreted as an indication that a decrease in sea ice to the west of the Antarctic Peninsula (ZWALLY et al., 2002) leads to more cases of convective clouds. References KING, J.C., J. TURNER, G.J. MARSHALL, W.M. CONNOLLEY, T.A. LACHLAN-COPE, 2004: Antarctic Peninsula Climate Variability And Its Causes As Revealed By Analysis Of Instrumental Records. - In: E. Domack, A. Burnett, P. Convey, M. Kirby and R. Bindschadler (Eds.): Antarctic Peninsula Climate Variability: A historical and Paleoenvironmental Perspective. American Geophysical Union, 17-30. SANTER B.D., T.M.L. WIGLEY, J.S. BOYLE, D.J. GAFFEN, J.J. HNILO, D. NYCHKA, D.E. PARKER, K.E. TAYLOR, 2000: Statistical significance of trends and trend differences in layer-average temperature time series. - J. Geophys. Res. 105, 7337-7356. SOLOMON, S., D. QIN, M. MANNING, Z. CHEN, M. MARQUIS, K.M.TIGNOR, H.L. MILLER (EDS), 2007: Climate Change 2007. The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. - Cambridge University Press, Cambridge, United Kingdom, and New York, NY, USA, 996 pp. 210 TURNER, J., S.R. COLWELL, G.J. MARSHALL, T.A. LACHLAN-COPE, A.M. CARLETON, P.D. JONES, V. LAGUN, P.A. REID, S. IAGOVKINA, 2005: Antarctic climate change during the last 50 years. - Int. J. Clim. 25, 279-294. VAUGHAN, D.G., G.J. MARSHALL, W. M. CONNOLLEY, J.C. KING, R.M. MULVANEY, 2001: Devil in the detail. - Science 293, 1777-1779. WORLD METEOROLOGICAL ORGANIZATION, 1995: Manual on Codes, Volume 1. - WMO, Geneva, 492 pp. ZWALLY, H.J., J.C. COMISO, C.L. PARKINSON, D.J. CAVALIERI, P. GLOERSEN, 2002: Variability of Antarctic sea ice 1979-1998. - J. Geophys. Res. 107, 1029-1047. Author's address: Dr. Amélie Kirchgäßner ([email protected]) British Antarctic Survey High Cross, Madingley Road CB3 0ET Cambridge, United Kingdom 211 Usefulness of seasonal temperature forecasts for Germany Christina Koppe Deutscher Wetterdienst, Germany Abstract In this paper first verification results for the new seasonal forecast for Germany are presented. Even if the operational seasonal forecast for Germany includes the parameters 2m temperature and total precipitation this paper focuses on temperature only. The forecast bases on the seasonal forecasting system 3 of the European Centre for Medium-Range Weather Forecasts (ECMWF). The hindcast dataset which extends from 1981-2005 was verified against the ERA40 dataset for the years from 1981-2001 and against the operational analysis from 2002 onwards. The skill of the seasonal forecast was estimated by using the hit rate. Hit rates have been calculated for a two category deterministic forecast (ensemble mean is above normal and below normal) and a three category probabilistic forecast (highest probability for cold, normal or warm event). In general, the skill of seasonal forecast is considered to be very low in temperate climates. In this paper it was, however, demonstrated that depending on the question that is asked and under certain preconditions, there are cases in which the hit rates of the seasonal forecasts are significantly above the expected value. Anwendbarkeit von saisonalen Temperaturvorhersagen in Deutschland Zusammenfassung In diesem Artikel werden erste Verifikationsergebnisse des neuen saisonalen Vorhersagesystems des Europäischen Zentrums für Mittelfristige Wettervorhersagen (ECMWF) für Deutschland vorgestellt. Die Jahreszeitenvorhersage für Deutschland, welche vom DWD herausgegeben wird, bezieht sich auf die zwei Parameter Lufttemperatur und Niederschlag. Hier werden jedoch ausschließlich die Verifikationsergebnisse für die Lufttemperatur vorgestellt. Für die Verifikation wurde der sogenannte „Hindcast“ Datensatz, welcher den Zeitraum von 1981-2001 umfasst, verwendet. Verifiziert wurde dieser gegen die ERA40 Reanalysen für die Jahre 1981-2001 und gegen die operationelle Analyse für die Jahre ab 2002. Die Vorhersagegüte wurde mit Hilfe der „Trefferquote“ abgeschätzt. Die Trefferquote wurde sowohl für eine „Zwei-Klassen-Vorhersage“ (Ensemblemittelwert ist größer null oder kleiner null) als auch für eine probabilistische „Drei-Klassen-Vorhersage“ (Wahrscheinlichkeit dafür, dass ein Monat im Mittel zu warm, normal oder zu kalt ausfallen wird) berechnet. In der Regel ist die Vorhersagegüte von Jahreszeitenvorhersagen in gemäßigten Klimaten sehr gering. In dieser Arbeit konnte jedoch gezeigt werden, dass unter bestimmten Bedingungen und in Hinblick auf bestimmte Fragestellungen, die Trefferquote von Jahreszeitenvorhersagen signifikant über dem Erwartungswert liegt. 1. Introduction Seasonal forecasts are an attempt to provide useful information about the expected climate in the coming months. The seasonal forecast is not a weather forecast! Whilst weather is reflected in atmospheric conditions that change continuously, climate can be 212 considered as a statistical summary of weather events occurring in a given season or period (ECMWF, 2007). Due to the chaotic nature of the atmosphere, deterministic weather forecasting is restricted to a time window of about two weeks. However, some of the components that influence the weather show slow variations on long time scales (seasons and years). Therefore, also long-term predictions (climate forecasts) are to some degree possible. The most important of these components is the El Nino Southern Oscillation (ENSO) cycle. Other causes of climate variability are unusually warm or cold sea surface temperatures, snow cover or soil wetness (ECMWF, 2007). Seasonal forecasts provide a range of possible climate changes that are likely to occur in the coming months. It is important to bear in mind that, because of the chaotic nature of the atmospheric circulation, it is not possible to predict the daily weather variations at a specific location months in advance. It is not even possible to predict exactly the average weather, such as the average temperature for a given month (ECMWF, 2007). Nevertheless seasonal forecasts are able to provide information about the likely conditions averaged over a specific period. And this kind of information – even if the uncertainty is high – can help decision makers from different sectors to optimise their results. Especially decision makers from the energy sector already use seasonal forecasts in order to calculate the energy demand of the coming months. But also other climate sensitive sectors are able to deal with the uncertainty inherent in such forecasts and benefit from monthly average information. There are two basic approaches for the prediction of the climate during the coming months. One way to derive long-term predictions are statistical analyses of past weather and climate. The other way is to use the numerical weather prediction method. The seasonal forecasting model used in this study is based on the second approach. The aim of this paper is to show if seasonal forecasts for Germany have some skill and under which preconditions the skill of this forecasts is enhanced. 2. Materials and Methods The seasonal forecast for Germany is issued by the German Meteorological Service (DWD) and is based on the seasonal forecasting system of the European Centre for Medium-Range Weather Forecasts (ECMWF, www.ecmwf.int). Even if the forecast issued by DWD also comprises estimates about the expected precipitation this paper focuses only on temperature. The seasonal forecasting model of the ECMWF is a coupled ocean-atmospheric model and belongs to the group of numerical weather prediction models. It runs once a month and has usually a lead time of 7 months. Three times a year the model is run up to 13 months in advance and once a year, in November, the model run is extended to 14 months in order to cover the next calendar year. The horizontal resolution of the seasonal model of ECMWF is approximately 125 km. The basic idea of coupled ocean-atmospheric models is to estimate the future state of the atmosphere based on the state of the oceans and the atmosphere at the beginning of the forecast (initial conditions) and a set of mathematical equations which describe the physics and the dynamics of the atmosphere. The initial conditions for a forecast are calculated based on weather and ocean observations. Due to small inaccuracies in the observations there are uncertainties in the initial conditions. These uncertainties in the 213 initial conditions do not affect short-range weather forecasts. They increase, however, the uncertainty of medium- to long-range forecasts. Therefore, ensemble forecasting systems are often used for these kinds of forecasts to overcome such problems. Ensemble-based forecasting systems like the seasonal forecasting system of ECMWF model possible uncertainties in the initial conditions by adding small perturbations. These perturbations are in the same order of magnitude as the uncertainties. Thereby an ensemble of slightly different initial conditions is created. Each member of this ensemble is then used as starting point for a forecast. An ensemble-based forecasting system generates not only one forecast for a specific starting date and lead time, but a series of forecasts. The ensemble of the seasonal forecasting system of the ECMWF consists of 41 forecasts with slightly different initial conditions and slightly different model parameters in order to represent the uncertainties in the model formulation. For the model runs that extend to 13 or 14 months and for the hindcast dataset the ensemble-size is reduced to 11. Systematic model errors also become noticeable for long-range forecasts. Therefore, the forecasted values are in general related to a so-called model climate. That means that deviations from the model climate are calculated and used as input for further analyses. The model-climate of the ECMWF seasonal forecasting system has been created based on model runs with “historical” initial conditions and comprises the years 1981-2005 (hindcast dataset). The hindcast dataset is identical to the real-time forecasts, except for the ensemble size. Based on this hindcast data the model climate is created for each month of the year and each lead time. The hindcast dataset is also used for verification. The period that is used for the model climate is not identical with the period that is used as climatological normal (WMO, 1989). The current climatological normal comprises the years 1961-1990. With respect to temperature there are significant differences in the climatology that is based on the climatological normal and the one that is based on the period 1981-2005. Therefore, it is necessary to correct the anomaly of the model climate with respect to the average difference in the monthly mean between the reference climate and the model climate. The monthly averages of this reference period were defined as “normal” conditions. As seasonal forecasts are not able to predict the weather or the climate of the following month exactly, the most important information that a user can derive from seasonal forecasts is, if a month is predicted to be “colder than normal” or “warmer than normal”. This highlights the importance of the definition of “normal” conditions. For operational purposes a second reference climate period has been chosen for Germany. This second reference period comprises the years 1991-2005 and was chosen because climate is currently changing and because most of the years after 1991 belonged to the warmest ever observed (WMO, 2007). The grid point information that is derived from the ECMWF seasonal model with a horizontal resolution of 125 km is merged to regional averages for four regions in Germany (Fig. 1). In addition, the average temperature anomalies for the whole country are calculated. Due to the limited number of grid points per region, grid points with altitudes over 500 m have been excluded. Apart from averaging seasonal forecasts over a greater region it is important not to use daily values but also temporal averages. Often 214 averages over three months (seasons) are used. Many users of seasonal forecasts, however, require monthly information. Therefore, we use monthly means in this study. N E W N: North W: West E: East S Fig. 1: Forecast regions in Germany The seasonal forecast that is currently issued by DWD consists of deterministic information and of probabilistic information. The hindcast or re-forecast dataset was verified against reanalysis data (ERA40) for the years 1981-2001 and against the operational analysis of ECMWF for the years from 2002 on. As the uncertainty of seasonal forecasts is high, the verification for this product was done with respect to the question: “Is the forecasting system able to predict the direction of the anomaly correctly?” As measure for the quality of the forecast the hit rate was chosen. hit rate = number of correct forecasts number of correct forecasts + number of false forecasts If the forecast and the analysis both show a warm (cold) anomaly the forecasts was classified as “correct”. If the anomaly of forecast and analysis were different the forecast was classified as “false” (Table 1). Table 1: Contingency table for the calculation of the hit rate analysis forecast warm cold warm correct false cold false correct Apart from this deterministic information we estimated the probability that a month will be in the lower (“cold”), middle (“normal”) or upper tercile (“warm”) of the temperature distribution of both reference periods. The spread of the middle tercile (= normal conditions) for both reference periods for Germany is displayed in Fig. 2. For most of the months the upper and lower boundaries of the middle tercile are higher for the refer- 215 ence period 1991-2005 than for the period 1961-1990. This shows impressively the effects of a changing climate. 20 20 temperature (°C) normal 1961-1990 normal 1991-2005 15 15 10 10 5 5 0 0 Jan Feb Mrz Apr Mai Jun Jul Aug Sep Okt Nov Dez month -5 -5 Fig. 2: Spread of the middle tercile (category “normal”) for the period 1961-1990 (bright grey) and the period 1991-2005 (dark gray) for Germany. Values that are smaller than the lower boundary of the middle tercile belong to the lower tercile (category “cold”) and values that are greater than the upper boundary of the middle tercile belong to the upper tercile (category “warm”) To verify the tercile forecasts the hit rate as described above has been calculated with the only difference that a third category “normal” has been added. Days were grouped in the category with the highest probability. The expected value for an arbitrary forecast is for the tercile forecast a hit rate of 33%. For the forecast that only accounts for two categories the expected value is 50%. Hit rates that are greater than these values indicate that the forecast has some skill. 3. Results Hit rates have been calculated for the four regions and for the whole country. For the deterministic part of the product (deviations from normal) the data were classified in two classes: above normal and below normal. Hit rates for the tercile forecasts were calculated for the three categories: “cold”, “normal” and “warm”. In general hit rates are higher for the deviations from the 1961-1990 climate than for the 1991-2005 anomalies. The hit rates for the two category forecast have been calculated for different cases (Table 2). The first case was that the anomaly of the hindcast dataset was compared with the anomaly of the re-analysis dataset (“act. month”) without further restrictions. For 216 the prediction of the anomaly with respect to the 1961-1990 means, the hit rates were around 60% for all lead times. Hit rates that were calculated for the deviations from the 1991-2005 averages, however, are only slightly above 50% and show a lower skill. Forecasts with a lead time of 5 months show no skill with a hit rate below the expected value. Table 2: Hit rates for the two category forecasts (above normal / below normal) for Germany between 1981 and 2005 with respect to the reference periods 19611990 (above) and 1991-2005 (below). Hit rates greater than 50% indicate that the forecast has some skill. L0, L1, …L6: forecast lead time = 0 months, 1 month, …6 months; act. month: model run of the actual month; dev. sig.: only significant anomalies have been considered; sig. + prev.: if model runs of previous month and actual month group the target month in the same category and there was a significant anomaly; dev. >0,5 (>1,0): if predicted anomaly was greater than +/- 0,5°C (+/- 1,0°C); dev. >0,5 (>1,0)+prev.: if predicted anomaly was greater than +/- 0,5°C (+/- 1,0°C) for two model runs in a row 61-90 act. month dev. sig. sig + prev. dev. >0,5 dev. >1,0 dev. >0,5+prev. dev. >1,0+prev. L0 69% 60% 70% 79% 83% 88% 89% L1 63% 53% 62% 66% 76% 73% 73% L2 64% 56% 65% 67% 66% 71% 86% L3 61% 54% 61% 65% 78% 73% 88% L4 61% 52% 61% 69% 66% 77% 70% L5 61% 53% 70% 66% 65% 75% 59% L6 59% 52% 91-05 act. month dev. sig. sig.+ prev dev. >0,5 dev. >1,0 dev. >0,5+prev. dev. >1,0+prev. (*) few cases L0 68% 56% 69% 78% 87% 84% 100% L1 51% 46% 52% 55% 67% 57% 60% L2 53% 45% 55% 58% 61% 65% 63% L3 58% 51% 53% 58% 66% 71% 67% L4 52% 44% 43% 56% 52% 65% 100% L5 48% 41% 47% 53% 64% 66% 75% L6 51% 43% 68% 62% (*) 58% 60% (*) If only significant deviations (“dev. sig.”) from the reference climate are considered the hit rates for both reference periods decrease further. Apart from the forecast with a lead time of 0 months the forecast for the anomaly with respect to the 1991-2005 climate shows no skill at all. In this context it should be considered, that if only significant deviations from normal are considered the expected value is with around 48% slightly lower than 50%. But even if this point is accounted for, there is no skill in the forecasts. If, however, the hit rate is calculated only for those cases, when two model runs in a row predict the same anomaly (“sig. + prev”), the hit rates increase. And also the forecast for the reference period 1991-2005 shows some skill for the first three months. A further improvement in the skill can be obtained, if only deviations that are higher 217 (lower) than 0,5°C or 1,0°C (-0,5°C or -1,0°C) are taken into account. Under these conditions hit rates can reach more than 60% also for the 1991-2005 reference period. The hit rates of the hindcast dataset for the tercile forecast are in most of the cases higher than 33% and thus indicate that the forecast has some skill (Table 3). In general the hit rates for the terciles that have been calculated based on the climate of the period 1961-1990 are higher than for the terciles of the 1991-2005 climate. The reason for this higher skill with respect to the 1961-1990 period is probably that the skill partly results from climate change. As for the two-category forecast the hit rate increases if two subsequent model runs group one month in the same category. In general, a higher probability for a category further increases the hit rate and with it the forecast certainty. The predictability for an event further depends on the category (not shown). For the reference period 1961-1990 the hit rates are highest for the prediction of a “warmevent”. Hit rates for warm events are between 77% and 83% for the two-category forecast and between 60% and 71% for the three-category forecast, respectively. Is the 1991-2005 period used as reference period, then the hit rates for cold events were higher than the hit rates for warm events. They range from 59% to 73% for the twocategory forecast and from 55% to 66% for the three-category forecast. The hit rate for the other “events” (e.g. cold anomalies for the 1961-1990 reference period and warm anomalies for the 1991-2005 period) is generally below the expected value. Table 3: Hit rates for the tercile forecasts for Germany between 1981 and 2005 with respect to the reference periods 1961-1990 (above) and 1991-2005 (below). Hit rates greater than 33% indicate that the forecast has skill. L0, L1, …L6: forecast lead time = 0 months, 1 month, …6 months; act. month: model run of the actual month; prev. month: if model runs of previous month and actual month group the target month in the same category; Max > 0,4 …Max > 0,7: if maximum probability for one category is greater than 0,4, …0,7 61-90 act. month prev. month max > 0,4 max > 0,5 max > 0,6 max > 0,7 L0 55% 58% 57% 62% 71% 75% L1 42% 48% 45% 49% 56% 55% L2 42% 52% 45% 52% 52% 50% L3 44% 50% 46% 55% 65% 89% L4 42% 44% 44% 50% 50% 50% L5 39% 45% 41% 44% 57% 74% L6 42% 91-05 act. month prev. month max > 0,4 max > 0,5 max > 0,6 max > 0,7 L0 50% 57% 50% 60% 63% 65% L1 41% 46% 44% 43% 46% 52% L2 39% 41% 41% 46% 44% 50% L3 36% 42% 39% 39% 33% 33% L4 40% 43% 41% 46% 44% 46% L5 39% 42% 41% 40% 42% 47% L6 40% 44% 50% 56% 63% 40% 45% 43% 41% 218 4. Conclusions and discussion It has been shown that the seasonal forecast of ECMWF for Germany is under certain preconditions able to predict the direction of the temperature anomaly with a greater confidence that would be expected by chance. Hit rates for the reference period 19611990 are higher than for the reference period 1991-2005. The reason could be that the former period was generally colder than 1991-2005. Thus, most of the predicted anomalies for the 1961-1990 period were warm anomalies. This could mean that a certain amount of the skill of this forecast is due to climate change. This paper highlights the importance of the chosen reference period. Depending on which period is chosen there are differences in the skill of predicting a certain event. The skill to predict “warm events” is rather good for the reference period 1961-1990. But there is hardly no skill for the prediction of “cold” or “normal events”, when this reference period is used. For the reference period 1991-2005 it is the other way around. There is skill to predict “cold events” but no skill for “warm” and “normal events”. In general, the skill of seasonal forecasts is considered to be very low in temperate climates. In this paper it was, however, demonstrated that depending on the question that is asked and under certain preconditions, there are cases in which seasonal forecasts have skill also in Germany. It is, therefore, important that a user of seasonal forecasts knows about the skill and knows about the preconditions that enhance the skill of the forecast. This means that the user of such a forecast needs to be educated and advised. Due to the high uncertainty the seasonal forecast only has a very limited usefulness for the general public. But even if the skill of the seasonal is quite low (only a few percent higher than chance) it may be sufficiently high for some user groups. For an energy provider a forecast that has a hit rate, that is only a few percent higher than chance, can result in a significant benefit. In addition, most users do not make their decisions based on the seasonal forecast alone but on other information that also has a high uncertainty. So they are able to deal with the uncertainty that is inherent in seasonal forecasts. References ECMWF, 2007a: Seasonal forecast user guide (System 3). http://www.ecmwf.int/products/ forecasts/seasonal/documentation/index.html (Version 02.04.2007). WMO, 2007: World Climate News, No. 31, June 2007. WMO, 1989: Calculation of Monthly and Annual 30-year Standard Normals. WMO-TD/No. 341. Geneva, 11 pp. Author's address: Dr. Christina Koppe ([email protected]) Deutscher Wetterdienst Stefan-Meier-Straße 4-6, D-79104 Freiburg, Germany 219 220 Einladung zur festlichen Vortragsveranstaltung anlässlich des Jubiläums 50 Jahre Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg 221 Die festliche Vortragsveranstaltung anlässlich des Jubiläums 50 Jahre Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg beginnt am Dienstag, den 1. April 2008, um 10:00 Uhr im Festsaal im „Haus zur Lieben Hand“ der Albert-Ludwigs-Universität Freiburg (Löwenstraße 16) Musikalische Umrahmung durch das Quartett Tertiaire 222 Programmablauf Musikalischer Auftakt Grußworte Prof. Dr. Helmut Mayer, Meteorologisches Institut der Albert-Ludwigs-Universität Freiburg Prof. Dr. Hans-Jochen Schiewer, Prorektor der Albert-Ludwigs-Universität Freiburg Prof. Dr. Dr. h.c. Gero Becker, Prodekan der Fakultät für Forst- und Umweltwissenschaften der Albert-Ludwigs-Universität Freiburg Festvorträge zur Stadtklimatologie Prof. Dr. Wilhelm Kuttler, Institut für Geographie, Abteilung Angewandte Klimatologie und Landschaftsökologie, Universität Duisburg-Essen, Campus Essen: „Entwicklungslinien in der Angewandten Stadtklimatologie“ Dipl.-Ing. Babette Köhler, Stadt Freiburg, Referat für Stadtentwicklung und Bauen, Integrierte Stadtentwicklung: „Zukünftige Anforderungen der Stadtplanung an die Angewandte Stadtklimatologie“ Musikalisches Zwischenspiel Festvorträge zur Forstlichen Meteorologie Prof. DI Dr. Herbert Hager, Institut für Waldökologie, Universität für Bodenkultur Wien: „Entwicklungslinien in der Forstlichen Meteorologie“ PD Dr. Ulrich Kohnle, Forstliche Versuchs- und Forschungsanstalt Baden-Württemberg, Abteilung Waldwachstum, Freiburg: „Zukünftige Anforderungen aus der Forstwissenschaft und Forstwirtschaft an die Forstliche Meteorologie“ Musikalischer Ausklang Stehempfang 223 Photos from the festive presentation on 1st April 2008 in Freiburg (I) 224 Photos from the festive presentation on 1st April 2008 in Freiburg (II) 225 Photos from the festive presentation on 1st April 2008 in Freiburg (III) 226 Photos from the festive presentation on 1st April 2008 in Freiburg (IV) 227 Berichte des Meteorologischen Institutes der Albert-Ludwigs-Universität Freiburg Nr. 1: Nr.2: Nr. 3: Nr. 4: Nr. 5: Nr. 6: Nr. 7: Nr. 8: Nr. 9: Nr. 10: Nr. 11: Nr. 12: Nr. 13: Nr. 14: Nr. 15: Nr. 16: Nr. 17: Fritsch, J.: Energiebilanz und Verdunstung eines bewaldeten Hanges. Juni 1998 Gwehenberger, J.: Schadenpotential über den Ausbreitungspfad Atmosphäre bei Unfällen mit Tankfahrzeugen zum Transport von Benzin, Diesel, Heizöl oder Flüssiggas. August 1998 Thiel, S.: Einfluß von Bewölkung auf die UV-Strahlung an der Erdoberfläche und ihre ökologische Bedeutung. August 1999 Iziomon, M.G.: Characteristic variability, vertical profile and modelling of surface radiation budget in the southern Upper Rhine valley region. Juli 2000 Mayer, H. (Hrsg.): Festschrift „Prof. Dr. Albrecht Kessler zum 70. Geburtstag“. Oktober 2000 Matzarakis, A.: Die thermische Komponente des Stadtklimas. Juli 2001*) Kirchgäßner, A.: Phänoklimatologie von Buchenwäldern im Südwesten der Schwäbischen Alb. Dezember 2001*) Haggagy, M.E.-N.A.: A sodar-based investigation of the atmospheric boundary layer. September 2003*) Rost, J.: Vergleichende Analyse der Energiebilanz zweier Untersuchungsflächen der Landnutzungen “Grasland“ und „Wald“ in der südlichen Oberrheinebene. Januar 2004*) Peck, A.K.: Hydrometeorologische und mikroklimatische Kennzeichen von Buchenwäldern. Juni 2004*) Schindler, D.: Characteristics of the atmospheric boundary layer over a Scots pine forest. Juni 2004*) Matzarakis, A., de Freitas, C.R., Scott, D. (eds.): Advances in Tourism Climatology. November 2004*) Dostal, P.: Klimarekonstruktion der Regio TriRhena mit Hilfe von direkten und indirekten Daten vor der Instrumentenbeobachtung. Dezember 2004*) Imbery, F.: Langjährige Variabilität der aerodynamischen Oberflächenrauhigkeit und Energieflüsse eines Kiefernwaldes in der südlichen Oberrheinebene (Hartheim). Januar 2005*) Ali Toudert, F.: Dependence of outdoor thermal comfort on street design in hot and dry climate. November 2005*) Matzarakis, A., Mayer, H. (Hrsg.): Extended Abstracts zur 6. Fachtagung BIOMET des Fachausschusses Biometeorologie der Deutschen Meteorologischen Gesellschaft e.V.. März 2007*) Mayer, H. (ed.): Celebrating the 50 years of the Meteorological Institute, Albert-Ludwigs-University of Freiburg, Germany. May 2008*) *) Bericht online verfügbar unter: http://www.meteo.uni-freiburg.de/forschung/publikationen/berichte/index_html 228