Environmental isotopes and geochemistry in - lha
Transcription
Environmental isotopes and geochemistry in - lha
Académie d'Aix Marseille Université d'Avignon et des Pays de Vaucluse THESE Pour obtenir le grade de docteur de l'Université d'Avignon et des Pays de Vaucluse Ecole Doctorale : 380, Sciences et Agronomie Discipline: Sciences de la Terre (Earth sciences) Spécialité: Hydrogéologie (Hydrogeology) METHODES ISOTOPIQUES ET GEOCHIMIQUES POUR L'ETUDE DES EAUX SOUTERRAINES ET DE L'HYDROLOGIE DES LACS : CAS DU BASSIN DU NIL BLEU ET DU RIFT ETHIOPIEN Environmental isotopes and geochemistry in investigating groundwater and lake hydrology: cases from the Blue Nile basin & the Ethiopian Rift (Ethiopia) Présenté par (by) SEIFU KEBEDE Présentée et soutenue publiquement le 10 décembre 2004 JURY M. Edmunds J.L. Michelot Y.Travi T. Alemayehu K. Rozanski P. Aggarwal B. Blavoux Professor CR, HDR Professeur Asso. Professor Professor Doctor Professeur Université d’ Oxford, Center for Water Research Rapporteur Université de Paris Sud, LHGI Rapporteur Université d'Avignon, Labo. Hydrogéologie Directeur de Thèse Addis Ababa University, Geology Department Examinateur University of Krakow, Dept. Nuclear & Env. Physics Examinateur Head, Isotope Hydrology Section, IAEA Examinateur Université d' Avignon, Labo. Hydrogéologie Examinateur Résumé On utilise les isotopes de l’environnement (δ18O, δD, δ13C, 3H) et l’hydrogéochimie pour étudier le fonctionnement hydrologique des eaux souterraines et des lacs sur des secteurs sélectionnés en Ethiopie. Il s’agit de la dépression de l’Afar, du Rift Ethiopien et du bassin du Nil Bleu. On s’intéresse tout d’abord à la relation entre le climat et la composition des eaux météoriques. Les conclusions obtenues sont ensuite utilisées pour l’étude des eaux souterraines et des lacs. La variation saisonnière de δ18O et δD des eaux de pluie sur l’Ethiopie est principalement sous la dépendance du mouvement saisonnier de la ZITC, des origines des masses d’air, et des trajectoires associées, de l’humidité atmosphérique. Une fois que l’humidité issue des principales sources (Océans Indien et Atlantique ou évaporation continentale) atteint les reliefs éthiopiens, la composition isotopique de la pluie est modifiée par les effets locaux d’altitude, de température et de masse. Un exemple typique est donné par l’appauvrissement en δ18O de 0.1 ‰ par 100 mètres lorsque les masses humides se soulèvent le long du versant ouest des montagnes éthiopiennes. Toutefois, aucun de ces effets isotopiques ne paraît avoir une influence prédominante sur la variation spatiale ou temporelle de la composition isotopique des eaux météoriques. C’est pourquoi, la thèse recommande de considérer l’ensemble de ces effets qui peuvent s’opposer ou s’ajouter, plutôt que de mettre en valeur un seul effet, lorsqu’on interprète les signaux isotopiques (dans les eaux météoriques actuelles ou les archives isotopiques paléohydrologiques). L’identification de différents mécanismes de recharge pour les trois secteurs (Plateau Nord Ouest, Rift Principal et dépression de l’Afar) constitue un des principaux résultats. Le taux de fractionnement du à l’évaporation, avant la recharge, est le plus élevé dans l’Afar et le plus faible sur le Plateau Nord Ouest. Dans l’Afar la principale source de recharge provient des bras morts de cours d’eau partiellement évaporés ou d’écoulement de crues en provenance des escarpements qui bordent la dépression ou de la plaine de l’Awash. En couplant les méthodes géochimiques et isotopiques, ce travail précise également les mécanismes de recharge des eaux souterraines, leur temps de résidence et leur évolution géochimique dans le bassin supérieur du Nil Bleu. Bien que les basaltes du Cénozoïque soient le principal aquifère, plusieurs systèmes hydrogéologiques ont pu être identifiés et décrits sur la base des données hydrogéochimiques. Par ailleurs, dans deux secteurs (linéament volcanique de Yerer Tulu Welel -YTVL et graben du lac Tana-GLT) le dioxyde de carbone d’origine profonde joue un rôle important pour le contrôle de l’évolution chimique des eaux souterraines du type NaHCO3 avec un TDS élevé. Le bassin du Nil Bleu était autrefois considéré comme une région avec un système hydrogéologique simple constitué d’aquifères de roches cristallines L’application de la méthode du bilan isotopique à quelques lacs éthiopiens sélectionnés montre que la méthode est plus performante en comparant l’état hydrologique des lacs et en calculant les flux d’eau souterraine autour des lacs. On propose d’utiliser une droite d’évaporation hypothétique locale comme référence pour comparer les compositions isotopiques (actuelles ou anciennes) et obtenir ainsi des informations hydrologiques immédiates. Mots clés: isotopes de l’environnement, effets isotopiques, recharge en eau souterraine, bilan de lac, Nil Bleu, Rift Ethiopien, Ethiopie 2 Abstract This work uses environmental isotopes (δ18O, δD, δ13C, 3H) and geochemistry in groundwater and lake hydrological studies of selected sites from Ethiopia. The sites are the Afar Depression, the Main Ethiopian Rift and the Blue Nile Basin. The thesis first investigates the relationship between the seasonal and spatial variations in the isotopic composition of Ethiopian meteoric waters and the Ethiopian climate. It then makes use of this understanding in the groundwater and lake hydrological studies. The seasonal variation in δ18O and δD compositions of Ethiopian rainfall waters are mainly influenced by the seasonal drifting of the ITCZ and associated changes in sources of moisture or associated changes in moisture trajectory. Once the moisture mass from the major sources (Indian, Atlantic or continental) reaches the region, its δ18O and δD compositions is modified by the local altitude effect, the temperature effect and the amount effect. A clear example is the 0.1 ‰ per 100 meter depletion in δ18O as moisture mass moves upward over Ethiopian mountains facing west. The relation between spatial variation in mean air temperature and the spatial variation isotopic composition of meteoric waters has the form: δ18O = 0.21 Tair (°C) - 6.5. However, none of the isotope effect seems to dominate the other in influencing the spatial and temporal variation in isotopic composition of meteoric waters. Therefore, the thesis recommends that when one interprets the isotope signals (in modern meteoric waters or in paleo isotope record from archives) from the region one should consider the interplay of all effects that reinforce or cancel each other rather than singling out one isotope effect. One of the major results of the thesis is the identification of differences in ground recharge mechanisms of the three sectors (North Western Plateau, the MER, and Afar Depression) of the study region. The degree of evaporative fractionation prior to recharge is the highest in Afar Depression and the lowest in the NWP. In Afar the major source of groundwater recharge is from 'incompletely' evaporated losing streams or flush floods converging towards the Afar Depression from the bordering escarpments and from infiltration by Awash flood plain water. By coupling the isotopic and the geochemical methods this thesis also shows groundwater recharge mechanisms, its subsurface residence time and its geochemical evolution in the Upper Blue Nile Basin. Although the Cenozoic basalt is the principal aquifer in the upper Blue Nile Basin, multiple geochemically recognizable groundwater bodies/layers have been identified. This allows to describe different hydrogeological systems. Furthermore in two zones (the Yerer Tulu Welel Volcanic Lineament-YTVL and the Lake Tana Graben-LTG) carbon dioxide from deeper sources plays an important control on geochemical evolution of the high TDS NaHCO3 type groundwaters. The Blue Nile basin was previously considered as a region with simple hydrogeology underlain by crystalline aquifers. The isotope balance study of selected Ethiopian lakes shows that the isotopic lake balance method is more powerful in comparing the hydrological status of lakes and in computing groundwater flux around lakes. The thesis proposes a hypothetical local evaporation line for Ethiopia as a reference with which the isotopic composition (present or ancient from archives) of any lake could be compared to gain rapid hydrological information. The technique developed in this thesis has wider application in wet land and interconnected lake system studes including the analysis of degree of lake interconnectiveiy and wetland interconnectivity. The method is also used to quantify groundwater flux around lakes. Key words: environmental isotopes, isotope effects, groundwater recharge, lake water balance, Blue Nile Basin, Ethiopian Rift, Ethiopia 3 INTRODUCTION GENERALE i. Objectifs et démarche utilisés La présence de nombreux lacs, dépôts lacustres, et d’un flux de chaleur du à l’amincissement de la croûte dans le Rift d’Afrique de l’Est, a suscité de nombreuses investigations scientifiques depuis le début des années 1970. La majorité de ces études ont réalisé des mesures isotopiques et chimiques, avec comme objectif d’évaluer le potentiel des ressources géothermiques (Gonfiantini et al., 1973; UNDP, 1973; Scholes et Faber, 1976; Craig et al., 1977; IAEA projets en cours depuis 1994), ou pour l’analyse des changements environnementaux (Lamb et al., 2002), ou encore pour comprendre la dynamique des fluides crustaux dans la vallée du rift est africain (Darling, 1996). Quelques travaux indépendants ont utilisé les isotopes de l’eau pour étudier les interactions eaux de surface – eaux souterraines (Darling et al., 1996; Chernet, 1998; Mckienze et al., 2001; Ayenew, 1998; Kebede et al., 2002, Gizaw, 2002, les projets AIEA: http://www-naweb.iaea.org/napc/ih/tcs_list_region.asp ) et la climatologie actuelle (Rozanski et al., 1996). Ces travaux ont fourni une banque de données utile en particulier dans la vallée du rift éthiopien et la dépression de l’Afar. Les études antérieures ont permis de préciser les sources de recharge et l’hydrogéochimie dans les environs d’Addis Abeba (Gizaw, 2002), la dynamique des eaux souterraines autour des lacs de la vallée du rift éthiopien (Ayenew, 1998), les interactions eaux de surface – eaux souterraines dans le rift (Darling et al., 1996; Chernet, 1998; Chernet et al., 2001), les sources de pollution (Mckenzie et al., 2001; Reimann et al., 2003 ) et les ressources géothermales dans le rift éthiopien et la dépression de l’Afar (Craig et al., 1977; Darling, 1996). Ces études dispersées n’apportent cependant pas une vision générale des variations spatiales et temporelles des signaux isotopiques et de leurs contrôles climatologiques et hydrologiques. La principale question que l’on doit poser avant d’utiliser les isotopes de la molécule d’eau pour l’étude des eaux souterraines ou de l’hydrologie des lacs est, évidemment : quelle est la composition du signal d’entrée i.e. la composition de l’eau d’alimentation, que l’on va suivre dans le système. Ceci nécessite une connaissance précise de la variation spatiale des isotopes de l’eau et de leur relation avec les facteurs climatiques ou non climatiques. Cette connaissance est la base du traçage des eaux souterraines et de l’étude isotopique des lacs, et elle donne une relation isotope – climat actuel qui peut servir de référence pour interpréter les archives paléo-isotopiques. La plupart des données isotopiques obtenues jusqu’ici en Ethiopie concerne des surfaces limitées soit aux environs du rift éthiopien, soit dans la dépression de l’Afar et occasionnellement sur les escarpements bordant le rift. Ces deux régions dépendent d’un régime météorologique complexe où les deux moussons (Indienne, Atlantique), la topographie, beaucoup d’autres courants atmosphériques 4 (comme les Jets tropicaux d’Est, l’air froid et sec d‘Arabie) le fractionnement du à l’évaporation et l’orographie interagissent et jouent un rôle important dans la détermination de la composition isotopique du 'signal' d’entrée. Déterminer les relations isotope – météorologie est ainsi complexe et constitue un sujet important d’investigation des ressources en eau et de leur variabilité. Par ailleurs, les données isotopiques fiables étaient apparemment inexistantes sur le Plateau Nord Ouest Ethiopien. Beaucoup d’études antérieures indiquent un écoulement des eaux souterraines en direction du rift depuis les reliefs adjacents. Cependant, on sait peu de choses sur l’hydrogéologie, l’hydrogéochimie et la composition isotopique des eaux souterraines des plateaux adjacents. Ainsi, le mécanisme de transfert des eaux souterraines du plateau vers le rift n’est pas clair. L’examen des variations spatiales des isotopes de l’eau ou l’analyse des transfert d’eau souterraine depuis le plateau vers le rift ne sont pas les seuls objectifs qui nous ont conduit a travailler en partie sur le Bassin supérieur du Nil Bleu (Le Plateau Nord Ouest) dans cette thèse. Cette étude a démarré à la suite de la réalisation du « Master Plan » (BCEOM, 1998). Ce travail fournit une vision d’ensemble de l’état des eaux souterraines et les données physiques de base des aquifères régionaux. Il a laissé beaucoup de questions à résoudre sur les eaux souterraines du bassin du Nil bleu (définition des aquifères principaux, les origines de la recharge, l’écoulement souterrain, les relations eau de surface – eau souterraine, la qualité de l’eau souterraine. Ceci nous a conduit à utiliser les isotopes et l’hydrogéochimie pour examiner l’origine des eaux souterraines dans le Bassin supérieur du Nil bleu. En Ethiopie (limite nord de la ZITC et donc grande vulnérabilité vis-à-vis des conditions climatiques) ce n’est pas seulement l’évaluation des ressources en eau qui est utile, mais également sa variabilité. Plusieurs études existent sur la relation climat/variabilité des ressources en eau en Ethiopie. Les travaux sur les ressources en eau et la variabilité climatique de l’échelle millénaire à l’échelle saisonnière ont été soigneusement répertoriés par Nyssen et al. (2004). Au Quaternaire et à l’Holocène, la région a subi des variations dramatiques de la pluie, du niveau des lacs et du climat en général. La région a également subi une fluctuation majeure des variations inter annuelle et saisonnières de la pluie au cours de la dernière décennie. Savoir comment le climat a varié dans le passé peut être utilisé comme référence pour ses variations futures. Toutefois, comment interpréter le paléoclimat à partir des archives sédimentaires fait encore l’objet de nombreuses investigations globales. Dans les régions au climat complexe comme l’Ethiopie, la calibration entre la relation actuelle climat/hydrologie et la composition isotopique des lacs devrait fournir une information utile pour mieux interpréter les isotopes des archives isotopiques. Par ailleurs, la calibration hydrologieisotope-climat sur quelques lacs actuels sélectionnés est une approche pratique pour examiner d’autres lacs ou réservoirs peu connus. 5 Tous ces éléments nous ont finalement conduit à organiser le travail suivant quatre objectifs interdépendants, qui utilisent le même type de données. Les principaux objectifs sont: • Fournir ou améliorer le schéma général des variations temporelle et spatiale des compositions isotopiques et leur contrôle météorologique, dans les régions centre et nord de l’Ethiopie. • Examiner les mécanismes de la recharge dans les trois régions, à savoir : le Plateau Ethiopien Nord Ouest, Le Rift Ethiopien Principal et la Dépression de L’Afar. • Examiner la géométrie des aquifères, la circulation des eaux souterraines, leur recharge, et leur potentialité dans le bassin du Nil Bleu, en utilisant les techniques chimique et isotopique, et enfin, • préciser les relations entre la composition isotopique de l’eau des lacs actuels et les caractéristiques hydrologiques climatiques et hydrographiques de lacs éthiopiens sélectionnés (y compris les lacs du bassin du Nil Bleu). ii. Approche (méthodologie) Pour atteindre ces objectifs, cette thèse utilise les isotopes de l’eau et la géochimie des solutés. Pour étudier un système à l’aide des isotopes, on a besoin du signal d’entrée et du signal de sortie pour le caractériser. Ceci implique que l’application de la méthode dépend du type de système. Cette thèse tout en s’attaquant aux objectifs, testera aussi la pertinence de la méthode (en particulier l’application des isotopes à l’étude du bilan des lacs) sous le climat de l’Ethiopie et d’autres conditions spécifiques au site comme la salinité de l’eau des lacs et l’hydrographie. Isotope hydrogeology Composition of input signal Gw tracing Lake groundwater link proxy Modern isotope climate relations Modern lake isotope Input signal water balance pr ox y modern modelling / calibration Modern calibration / modeling Paleoclimate 6 Démarche logique, suivie et proposée par la thèse, pour l’étude des ressources en eau en Ethiopie. Les résultats de ce travail tendent à montrer l’importance des systèmes hydrologiques continentaux actuels pour calibrer et modéliser les paléoclimats en Ethiopie en suivant la logique indiquée sur ce diagramme. iii. Organisation de la thèse Le mémoire comporte cinq parties. Une courte introduction et un résumé des résultats précèdent les parties II et III ; elle est suivie par des articles scientifiques (soumis ou à l’impression). La première traite du premier objectif :- donner une meilleure vision de la composition isotopique (δ18O, δD, 3H) des eaux météoriques en Ethiopie, et des facteurs qui commandent leur variations spatiale et temporelle dans le cycle de l’eau. On s’intéresse en particulier à la variation saisonnière de δ18O et sa relation avec la ZITC, l’influence « feed back » de la surface du sol sur la composition isotopique des précipitations en Ethiopie, la variation spatiale des isotopes de la molécule d’eau et son contrôle. Cette première partie confirme que la composition du signal isotopique fourni par les précipitations en Ethiopie est influencée à la fois par des processus à grande échelle (eg. déplacement saisonnier de la ZITC et l’apport associé de son humidité) et les processus qui interviennent à la surface du sol (ré-évaporation à partir des bassins continentaux étendus ou activité locale de convection de vapeur, effet « d ‘ombre » sur la pluie, évaporation en cours de chute, effets orographique etc. ). Dans la seconde partie, on caractérise la composition isotopique des eaux souterraines du Rift Ethiopien, du Plateau Nord Ouest et de l’Afar, et on discute ensuite ces caractéristiques pour les trois régions. Les mécanismes de recharge dans ces trois secteurs importants d’Ethiopie (Plateau Nord Ouest, Rift Ethiopien Principal et dépression de l’Afar) sont comparés. On montre que la recharge des eaux souterraines est rapide, que les trajectoires de l’écoulement sont courtes, et que le fractionnement du à l’évaporation avant la recharge est peu important sur le Plateau Nord Ouest alors qu’il constitue un processus important en Afar. Les eaux souterraines du Rift Ethiopien Principal présentent des propriétés intermédiaires. La troisième partie se rapporte au troisième objectif. En utilisant les connaissances acquises dans la première partie, elle essaye de donner un schéma amélioré des ressources en eau du bassin du Nil bleu. On utilise essentiellement l’hydrochimie et l’hydrologie isotopique pour atteindre ces objectifs. Deux principaux bassins, structuralement déformés, et présentant une évolution chimique et une hydrodynamique homogènes, ont été identifiés. Ce sont celui de la zone de l’alignement volcanique de Yerer Tullu Welele et le graben du lac Tanna. Les isotopes de la molécule d’eau associés à quelques 3 H, le Carbone -13, et l’hydrochimie précisent les contrôles sur l’évolution chimique des eaux 7 souterraines dans le bassin. Dans ce chapitre on essaye de quantifier le potentiel en eaux souterraines à partir d’une approche physique simple. La quatrième partie se rapporte au quatrième objectif: - calibrer la relation entre la composition isotopique des lacs et le climat régional. On utilise la méthode du bilan isotopique (δ18O et δD) dans les études de bilan hydrologique des lacs. La Droite d’Evaporation Locale hypothétique sous les conditions climatiques de l’Ethiopie a tout d’abord été calculée et les compositions isotopiques de quelques lacs sélectionnés lui ont ensuite été comparés. La comparaison de la composition isotopique modélisée des lacs avec la composition isotopique mesurée fournit un moyen rapide de classification des lacs en : lacs à flux de sortie dominant, à évaporation dominante, à diminution de volume ou eau souterraine dominantes. La même approche peut aider à obtenir des informations sur les facteurs non climatiques (comme les effets hydrographiques, les effets de lacs en série, ou les effets de salinité) qui influencent le régime isotopique des lacs et ainsi des sédiments utilisés comme archives. La cinquième partie présente une synthèse du travail. Elle résume les principaux résultats, et les perspectives que l’on peut en tirer pour le futur. Elle permet d’esquisser les avantages et les limites de l’utilisation des méthodes isotopiques sous les conditions climatiques de l’Ethiopie et son contexte géologique. Figure i. Localisation de quelques sites importants 8 iv. Localisation de la zone d’étude, toponymie, régions Les sites étudiés(figure i) couvrent trois régions principales. Ce sont le Rift Ethiopien Principal, la Dépression de L’Afar et le Plateau Ethiopien Nord Est. La station GNIP/IAEA se situe au centre de l’Ethiopie à (2360 masl). Des données isotopiques sur les pluies, couvrant une courte période, ont été obtenues sur quatre stations dans la vallée du Rift (Sodo, Awassa, Kofele, et Agermariam). Le Nil Bleu prend sa source au lac Tana et draine le Plateau nord ouest. Tous les cours d’eau, à l’exception de l’Awash, s’écoulent depuis les reliefs centraux. Le lac de cratère Bishoftu qui contient des varves sédimentaires annuelles se trouve près d’Addis Abeba. Les champs géothermiques (exemple le centre géothermique d’Aluto Langano) et la zone des lacs constituent deux singularités du Rift Ethiopien Principal. v. Définitions, symboles, notations Isotopes de l’eau: L’eau est une molécule composée de deux atomes d’hydrogène et d’un atome d’oxygène. L’hydrogène possède deux isotopes stables 2H/D (espèce rare généralement appelée deutérium, 1H (espèce abondante) et un isotope radioactif 3H (tritium). L’oxygène possède deux isotopes stables 16 O (abondant) et 16 d’eau : H2 O (le plus abondant), 18 O (rare). Ces isotopes se combinent pour former quatre types D216O, H218O et D218O (plus rares. La composition de l’eau en ces isotopes varie dans les systèmes hydrologiques en fonction des conditions physiques et d’autres processus chimiques ou biologiques prévisibles. Rapports isotopiques: la composition de l’eau en ces isotopes s’exprime généralement par le rapport de l’isotope lourds sur l’isotope léger ( 18O/16O, D/H). Dans la mesure où l’isotope lourd est très rare, le rapport est un nombre très petit. Ces fractions ne sont pas facilement utilisables pour des opérations mathématiques simples. La notation delta pour mille: Les rapports isotopiques de la molécule d’eau sont généralement comparés au rapport isotopique d’une eau standard de rapport isotopique connu. Le « Standard Mean Ocean Water » (Vienna-SMOW, VSMOW) est le standard le plus largement utilisé. Ainsi, les abondances en 18O et D s’expriment comme un rapport en notation delta pour mille (parts pour mille, ‰) différences relative au standard. Les nombres obtenus sont entiers et utilisables avec des opérations mathématiques simples. (18O / 16O ) Ech. ( D / H ) Ech. − 1 *1000 et δD ( pourmille ) = − 1 *1000 (18O / 16O ) SMOW ( D / H ) SMOW δ 18O( pourmille ) = 9 L’Excès en deutérium (d-excess) est équivalent à δD-8δ18O d’une eau météorique donnée. L’excès en deutérium moyen de l’ensemble des précipitations à l’échelle globale est de 10. L’Excès en deutérium initial dans les pluies s’écarte de 10 en relation avec les conditions d’évaporation à l’origine de la vapeur et de l’influence de la vapeur continentale. Ainsi, l’Excès en deutérium est souvent utilisé comme marqueur d'origine de la vapeur dans une région donnée. Enrichissement/Appauvrissement ou enrichi/appauvri: Cette terminologie est utilisée pour comparer les compositions en δ18O et δD de différents types d’eaux météoriques dans une région donnée. Les eaux qui contiennent de forts δ18O et δD par rapport aux autres eaux de la région sont souvent considérées comme 'enrichies', Les eaux avec de faibles δ18O et δD comme appauvries. Deux exemples simples : les volumes d’eau évaporée sont enrichis en isotopes lourds du fait de l’évaporation, ou les eaux souterraines sont souvent appauvries par rapport à la composition isotopique de la pluie locale du fait d’une recharge sélective. La DEMG (GMWL): Sur un diagramme δ18O-δD la vapeur qui se forme à partir de l’évaporation des océans, les eaux qui se forment par condensation de la vapeur océanique, ou les eaux souterraines sur les continents directement rechargées par les pluies sans modification majeure, ou les eaux des rivières isotopiquement non modifiées (à une échelle globale), se regroupent sur une droite de pente 8 et de décalage à l’origine de 10. Cette droite s’appelle la Droite des Eaux Météorique Globale ou droite de Craig. La DEML (LMWL): Les eaux météoriques ne se situent pas toujours sur la Droite Météorique. En fonction des condition évaporatoires sur la surface de l’Océan et des sources d’humidité, localement une déviation par rapport à cette droite peut exister. La droite que l’on obtient à partir de la composition des eaux météoriques dans une région donnée est appelée Droite des Eaux Météoriques Locales. Le diagramme des eaux de pluie d’été non évaporées sur Addis Abeba donne une droite : δD = 8δ18O +15. La totalité des pluies mensuelles donne la relation : δD = 7.2δ18O +12. La DEL (LEL): Les eaux météoriques sujettes à l’évaporation (lacs, rivières, mares, etc.) vont subir un fractionnement isotopique du fait de la perte d’eau sous forme de vapeur. Les eaux évaporées ont tendance à se situer sur une droite qui s’écarte de la Droite des eaux Météorique mondiale. La droite qu’elles ont tendance à former (sur un diagramme δ18O-δD) est appelée Droite d’Evaporation Locale. La pente de cette droite se situe entre 3.5 et 6 et dépend de l’humidité locale. L’effet d’altitude ou le pseudo effet d’altitude : Lorsqu’une masse d’air humide s’élève le long d’une barrière montagneuse, la température de l’air humide tend à se refroidir adiabatiquement. Beaucoup de facteurs peuvent provoquer l’appauvrissement en isotope lourd avec l’altitude. On peut citer la température. La condensation qui est causée par la chute des températures suivant l’augmentation de l’altitude conduit à un appauvrissement en isotope lourd. L’effet Rayleigh est une autre cause. Quand une masse d’air humide est contrainte à s’élever les isotopes les plus lourds 10 tendent à être retirés préférentiellement de la vapeur par les gouttes d’eau. Ceci produit un appauvrissement en isotope lourd avec l’altitude. L’effet d’altitude est souvent observé aussi sur le versant sous le vent d’une montagne. Le pseudo effet d’altitude est souvent confondu avec l ‘effet d’altitude. Dans les deux cas il y a appauvrissement en isotope lourd avec l’altitude. Le pseudo effet d’altitude est provoqué par un enrichissement par évaporation des gouttes de pluie au cours de leur chute sous le nuage. Il est souvent observé dans les vallées ou sur le versant sous le vent des chaînes de montagnes. Cet enrichissement par évaporation, différent de l’effet initial de Rayleigh dans le nuage, provoque aussi une diminution de l’excès en deutérium, marquant ainsi clairement cette situation. Tritium et Unité Tritium: Le tritium est un des isotopes de l’hydrogène. Il est radioactif avec une demi vie de 12.26 ans. Il est produit dans l’atmosphère par le bombardement cosmique de l’azote 14N + n => 3H+ 12C. La concentration en tritium des eaux s’exprime T/H. Ceci correspond à une fraction minuscule. Une méthode alternative consiste donc à utiliser l’Unité Tritium (UT). Le rapport T/H = 10-18 correspond à 1UT. Il est souvent utilisé pour dater les eaux souterraines jeunes. A l’heure actuelle en Ethiopie, dans les eaux de pluie exemptes de pollution nucléaire (cosmogéniques) les concentrations en tritium se situent entre 5 et 10 UT vi. Abbreviations utilisées dans la thèse NWPSEPMERGNIPBCLYTVLLTGJJASCLEL- The North Western Ethiopian Plateau The South Eastern Ethiopian Plateau The Main Ethiopian Rift Global Network for Isotopes in Precipitation of the IAEA Bishoftu Crater Lakes The Yerer Tulu Welel Volcanic Lineament The Lake Tana Graben June-July-August-September Calculated local evaporation line vi. Reférences Ayenew, T. 1998. The Hydrogeological system of the lake district basin, central Main Ethiopian Rift. Phd thesis, ITC publication Number 64, the Netherlands, 200p. Battistelli, A., Yiheyis, A., Calore, C., Ferragina, C., Abatneh, W., 2002. Reservoir engineering assessment of Dubti geothermal field, Northern Tendaho Rift, Ethiopia. Geothermics, 31: 381–406 BCEOM, 1999. Abay River Basin integrated master plan, main report, Ministry of Water Resources, Addis Ababa. Chernet, T., 1998. Etude des Mechanismes de mineralisation en fluorure et elements associes de la region des lacs du rift Ethiopien. Ph.D. Thesis, Avignon, France. Chernet , T., Travi, Y., Valles, V., 2001. Mechanism of degradation of the quality of natural water in the lakes region of the Ethiopian Rift Valley. Water Reser.35, 2819–2832. Craig, H., Lupton J.E., Horowiff, R.M., 1977. Isotope Geochemistry and Hydrology of geothermal waters in the Ethiopian rift valley. Scripps Institute of Oceanography, University of California report, 160p. Darling, G., Gizaw, B., Arusei, M., 1996. Lake-groundwater relationships and fluid-rock interaction in the East African Rift Valley: isotopic evidence. J. African Earth Sci. 22, 423-430. Darling, WG., 1996. The Geochemistry of fluid processes in the eastern branch of the east African rift system, Ph.D thesis, British Geological Survey, UK, 235p. Gizaw, B., 2002. Hydrochemical and Environmental Investigation of the Addis Ababa Region, Ethiopia. Ph. D dissertation, Faculty of Earth and Environmental Sciences Ludwig-Maximilians-University of Munich, 157p. 11 Gonfiantini, R., Borsi, S., Ferrara, G. and Panichi, C., 1973. Isotopic composition of waters from the Danakil Depression (Ethiopia). Earth and Planetary Science Letters 18: 13-21. IAEA TC projects ETH8005 ETH8006, ETH8 007 (1995 to present) . Ongoing and completed projects conducted by International Atomic Energy Agency, the Ethiopian Science and Technology Commission and the Ethiopian Geological Surveys, Various unpublished and expert visit reports, isotope data, etc; Ethiopian Geological Survey, Addis Ababa, Ethiopia. Kebede, S., Lamb,H., Telford,R., Leng, M. and Umer, M., 2002. Lake-Groundwater relationships, oxygen isotope balance and climate sensitivity of the Bishoftu Crater Lakes, Ethiopia. Advances in Global Change Research, 12: 261-275. Lamb, H., Kebede, S., Leng.M.J., Ricketts, D., Telford, R., Umer, M., 2002b. Origin and stable isotope composition of aragonite laminae in an Ethiopian crater lake. In: Odada, E., Olago, D. (Eds.), The East African Great Lakes Region: Limnology, Palaeoclimatology and Biodiversity, Advances in Global Research Series. Kluwer Academic Publishers, Dordrecht. McKenzie, J., Siegel, D., Patterson, W., McKenzie, J., 2001. A geochemical survey of spring water from the main Ethiopian Rift Valley, southern Ethiopia: implication for well head protection. Hydrogeol. J. 9, 265-272. Nyssen,J., Poesen, J., Moeyersons, J., Deckers, J.,Haile, M., Lang, A., 2004. Human impact on the environment in the Ethiopian and Eritrean highlands—a state of the art. Earth sciences reviews, 64: 273-320. Reimann, C., Bjorvatn,K., Frengstad, B., Melaku, Z., Tekle-Haimanot, R., Siewers, U., 2003. Drinking water quality in the Ethiopian section of the East African Rift Valley, part I: data and health aspects. The Sci. Tot. Env. 31, 65-80. Rozanski, K., Araguas-Araguas, L., Gonfiantini, R., 1996. Isotope patterns of precipitaion in the East African Region. In: Johnson, T.C., Odada, E. (Eds). The Climatology, Palaeoclimatoloy, Paleoecology of the East African Lakes, Gordon and Breach,Toronto, pp. 79-93. Schoell, M., Faber, E., 1976. Survey on the isotopic composition of waters from NE Africa. Geologisches Jahrbuch. 17, 197-213. UNDP, 1973. Geology, geochemistry and hydrology of hot springs of the East African Rift system within Ethiopia., UNDP report DD/SF/ON-11, N.Y. 12 PARTIE I Composition isotopique des eaux météoriques en Ethiopie (Plateau NW, Afar et Rift Ethiopien Principal) 13 Introduction Les concentrations en δ18O et δD des eaux météoriques en Ethiopie, ainsi que leur variations spatiotemporelles sont commandées par l’interaction d’une grande variété de facteurs. Parmi les plus importants on notera ; a) les facteurs d’échelle continentale ou globale tel le déplacement saisonnier de la ZITC (Zone Inter Tropicale de Convergence) et les déplacements des masses humides qui lui sont associées ; et b) les facteurs locaux ou régionaux qui influencent ou modifient les compositions isotopiques. Les facteurs locaux comprennent l’effet d’altitude, l’effet de l’évaporation locale, l’effet d’ombre (pseudo altitude) sur les versants « sous le vent, et l’action de la vapeur ré évaporée depuis le sol. Dans ce chapitre on va : 1. Discuter brièvement les mécanismes à l’origine de la pluie, qui ont une influence sur le marquage isotopique des eaux en Ethiopie, 2. discuter la relation entre la variation des teneurs isotopiques mensuelles des pluies et le déplacement saisonnier de la ZITC, 3. essayer de comprendre l’origine de la composition isotopique des eaux météoriques en Ethiopie en comparant avec des régions similaires en Afrique, 4. proposer des hypothèses relatives à l’importance de l’orographie sur la composition des eaux météoriques en Ethiopie, 5. décrire les variations spatiales des teneurs en isotopes de l’eau et leur relation avec les trajectoires des masses d’air et le climat local (température/précipitation/évaporation), 6. discuter les avantages et les contraintes liés à l’utilisation des isotopes de l’eau pour les études hydrologiques en Ethiopie, 7. présenter des données préliminaires sur la chimie des eaux de pluie pour essayer d’identifier les sources de vapeur. L’analyse de la relation entre les variables climatiques et la composition isotopique des eaux météoriques en Ethiopie servira de base à: a) la compréhension des mécanismes de recharge et pour le traçage du transfert des eaux souterraines depuis le Plateau Nord Ouest (NWP) vers le Rift Ethiopien Principal (MER) et la dépression de l’Afar (Partie II); b) l’évaluation des ressources en eau souterraine des aquifères hydrogéologiquement peu connus du bassin du Nil Bleu (Partie III); et, c) estimer le bilan isotopique de quelques lacs éthiopiens représentatifs (Partie IV). 14 1) Climate and rainfall derivation in Ethiopia There is a general agreement that the Ethiopian rainfall regime is under the influence of the Indian and the Atlantic Ocean monsoons (Griffits, 1972, Gemechu, 1977). The flow of the monsoon moisture to the region is controlled by the seasonal migration of the Inter Tropical Convergence Zone (ITCZ). In summer (June July August September- JJAS) the ITCZ is located in northern Ethiopia and the region is under the influence of the southwesterly and southerly monsoon flows (figure 1). The southwesterly and southerly flows bring moisture from three sources including, low level moisture which is pulled from the Congo vegetation basin, from the Atlantic Ocean and partly from the Equatorial Indian Ocean (Hemming, 1961; Suzuki, 1967; Gemechu, 1977; Camberlin, 1997; Nyssen et al., 2004). Open continental water bodies (such as tropical lakes and Lake Victoria) may also play an important role in feeding the low-level southwesterly flows (Kebede, 1964; Camberlin, 1997; Okeyo, 1992). The NWP and the Rift Valley including Afar get rainfall during this time. Between October and March the ITCZ is located south of Ethiopia. This results in northerly flow of dry and cold air from the Arabian continent. The coldest temperature is recorded in the highland region during this time. The same southward flow brings some moisture from Arabian Sea and Northern Indian Ocean to the eastern lowlands bordering the South Eastern Plateau (SEP) producing rainfall in October, November and December (Gemechu, 1977). In spring (March and April-MA) the ITCZ is moving northward crossing Ethiopia. This results in northeasterly and easterly moisture flows. These bring the spring rain to the region from the Northern Indian Ocean. Only the southe eastern plateau (SEP) and the southern and eastern sectors of the north western plateau (NWP) are influenced by this moisture. In the lowlands bordering Sudan and the central sector of the NWP the influence of the North Indian Ocean moisture is absent or is very weak (see histogram in figure 1). Seventy five percent of annual rainfall in the NWP and in the MER occurs during summer (JJAS) when the Inter-Tropical Convergence Zone (ITCZ) is located north of Ethiopia. The other 25% of rainfall occurs in spring (MA) when the ITCZ is still passing over Ethiopia northwards. Locally, the elevated terrain of the Ethiopian plateau influences the orographic enhancement of rainfall (Camberlin, 1997). Although the mountains in the Ethiopian plateau act generally as moisture enhancement they also lead to extremely complex pattern of rainfall, temperature and aridity (figure 2) over the region with pockets of humid climates alternating with arid ones within a few tens of kilometers (Nicholson, 1996). At the regional scale, the MER and the Afar depression which are 15 located in the leeward side of both the summer and the spring monsoons and they are characterized by arid to semiarid climate owing to the capture of moisture by the mountainous areas bordering them from East and West. Figure 1. Seasonal drifting of the ITCZ and its influence on rainfall regime of Ethiopia. Histograms show monthly rainfall distribution starting from January. The western most sector of Ethiopia gets only the JJAS rainfall the central sector is characterized by bimodal rainfall distribution getting the March April rainfalls and the JJAS rainfalls with a break in May and June. The eastern sector of Ethiopia gets its main rainfall from March to May and in October. The figure in the right shows the elevation map of Ethiopia and the mountains bordering the rift valley and Afar. The lower graph indicates the different air flow pattern over Africa in July and January. The east west arrow in July circulation indicate the direction of the AEJ. The north south dashed line (July) is the Congo convergence zone of the Indian and the Atlantic Ocean monsoons. Figures from Telford (1998) and Nicholson (1996). 16 Figure 2. Ethiopian mean annual rainfall (above) in mm and mean annual temperature map (below) in °C. The Afar rift and the MER get lower amount of rainfall owing to capture of moistures from the Indian Ocean and the Atlantic Ocean by the mountains bordering them. The temperature is also higher in Afar and the MER. There is no direct relation between rainfall amount and elevation in the NWP. In the SEP and the Rift rainfall increases with elevation. 17 The mountains also influence the local convective activities; they influence the local rainfall distribution and the timing of rainfall in the day. In the Northern plateau clouds are formed at the end of the morning because of evaporation and associated convection. In Eritrean highland (just north of Ethiopia) for example 80% of daily precipitation in summer occurs between 12 and 16 hours (Krauer, 1988). The same diurnal distribution of rainfall is common in the NWP particularly in September. This convective nature of rainfall explains why rainfall amounts are locally extremely variable in Ethiopia (Nyssen et al., 2004) particularly in the NWP. The configuration of the mountains also influences the spatial variations in mean annual temperature. The mean annual air temperature of the NWP is 16°C compared to 35°C in the Afar (figure 2). The coldest temperature region is located in the arid mountains in the NWP bordering Afar. The annual rainfall in NWP ranges between 1000mm and 2000mm while in the Afar it is less than 250 mm/year. Rainfall and temperature in the MER is intermediate between the NWP and the Afar. Despite its location between the Sahel belt and the Equatorial Africa there are some characteristics that make the climate regime of the northern Ethiopia distinct. Compared to similar latitude regions of Sahelian Africa the Ethiopian region gets prolonged and higher rainfall amounts due to orographic enhancement. Furthermore, because of its proximity to the Indian Ocean, the Ethiopian highland gets part of its moisture from the Indian Ocean unlike the western Sahel which gets its rain predominantly from the Atlantic Ocean. The low-level westerly flows also traverse a vast expanse of vegetated basin in the Congo before they reach the Ethiopian highland. This makes the Ethiopian highland to get part of its moisture from continental sources. Compared to the tropical eastern Africa, the NWP gets much of its rainfall during the Sahel summer (JJAS) and the little rains during March and April. The tropical eastern Africa gets much of its moisture from March to May and little rains from October to December (Camberlin and Okoola, 2003). The differences in rainfall distribution and amount among the three sectors of Africa are associated to the position of the ITCZ. The position of the ITCZ in turn influences the exact location of the source of moisture and the trajectories that moisture laden airs follow before they reach these regions. In addition to the low-level Atlantic Ocean monsoon and the Indian Ocean monsoon, there are many other airflow patterns at different altitude and from different directions over the Sahel and East Africa. These include the Tropical Easterly Jet and the African Easterly Jet in the upper level 1. The isotopic composition of rains associated with the Indian and Atlantic Ocean monsoons are relatively wellstudied (Rozanski et al., 1996; Taupin et al., 2000, HAPEX-Sahel project 1 Following the East West Line bordering the ITCZ an east west moisture flow exists in upper atmosphere. This east west zonal flow is called the AEJ. In sub tropics it is called the TEJ. It is to be noted that the TEJ is different from the Indian Ocean monsoon which flow towards the ITCZ at low levels. 18 http://directory.eoportal.org/pres_HAPEXSAHEL.html) while the Easterly Jets and their significance in influencing the isotope regime of Sahel rainfalls were also mentioned by some authors (eg. Joseph et al., 1992; Hailemichael et al., 2002). The following section of this thesis and Taupin et al. (2000) show the easterly jets are less important in influencing the isotopic composition of Ethiopian and Sahel rains respectively. The influence of the Jets on airflow pattern is beyond the scope of this thesis. 2. The isotope data As the MER and the Afar contain numerous lakes and high geothermal flux, they have been the subjects of paleo-hydrological, paleo-climatological and geothermal studies since the second half of 20th century. These studies have produced hydrogeochemical and environmental isotope data. Recently the International Atomic Energy Agency (IAEA) through its Technical Cooperation (TC) projects is conducting isotope hydrological studies in Ethiopian Rift and Adjacent plateaus. No previous stable isotope data has been apparently available from the NWP until the recently gathered and analyzed over 200 samples for δ18O, δD, δ13C, 3H and hydrochemistry for this thesis. The majority of the previously collected data is compiled and used with the new data for analyzing the relation between spatial isotope variations and climatic factors that control these variations. The isotopic composition of over 1000 groundwater wells, 60 rivers samples, 100 cold springs, 100 lakes, and 133 geothermal springs were compiled (CD included) and used in the analyses of spatial variations in terms of moisture sources and local climatic conditions. Rainfall isotope data for Addis Ababa station was downloaded from the IAEA/WMO/GNIP data base (http://isohis.iaea.org) and were analyzed to understand the relationship between temporal (seasonal or long-term) variations in δ18OδD and climate variables. Some of the previous works that were utilized to compile isotope data include IAEA-TC-projects (1996- an ongoing project) Gizaw (2002), Kebede et al. (2002a), Kebede et al. (2002b), McKenzie et al.( 2001) Beyene (2000) Ali (1999) Travi and Chernet (1998) Chernet (1998) Ayenew (1998) Rozanski et al. (1996) Darling (1996) Darling et al. (1996) Fontes et al. (1980) Craig et al.( 1977) Schoell and Faber (1976) Gonfiantini et al. (1973) UNDP (1973) etc. 3. Results and Discussion 3.1. Seasonal variation in isotopic composition of Ethiopian rainfall waters 19 Isotopic composition (δ18O, δD and 3H) of rainfall has been measured somewhat regularly at an IAEA/WMO station at Addis Ababa (2300masl, 16°C mean annual temperature, 1260mm/yr longterm mean annual precipitation) since 1965. Short-term rainfall δ18O and δD compositions were occasionally measured at few other Ethiopian rift valley stations. The short-term δ18O and δD compositions of rainfalls from the MER (see location map i) show a similar pattern of seasonal variation and comparable values of δ18O and δD composition to those of the Addis Ababa rainfalls. Box 2 (Appendix 1) gives the tritium content of the Addis Ababa rainfalls and estimates the missing data. There is a good relation between the seasonal variation in isotopic composition of Ethiopian rainfalls and the seasonality in Ethiopian climate. The δ18O and d-excess of the Addis Ababa rainfalls and the two-year rainfall isotope data from MER stations are characterized by a notable seasonal variation (though not as pronounced seasonal variation as in temperate and high latitude regions) (figure 3). The summer rainfalls waters are relatively depleted in δ18O and they have higher d-excess than the spring rainfalls reflecting typical characteristics of the Sahel rains. The weighted average δ18O and δD composition of the summer rainfall waters of the Addis Ababa IAEA station is -2.5‰ in δ18O, -5‰ in δD, and 15 in d-excess. The spring rainfalls have a weighted mean composition of +1‰ in δ18O, +20‰ in δD and 10 in d excess. In a δD-δ18O plot (not shown) the monthly rains of Addis Ababa is defined by the relation: δD = 7.2δ18O + 12. A similar plot on non-evaporated summer rains at Addis Ababa has the relation: δD = 8δ18O + 152. The Maximum-Minimum-Average plot (figure 4) shows that, the most depleted δ18O were recorded in the dry season when the northerly dry and cold winds reach the Ethiopian highland. 2 Since the summer moisture is the main water available for runoff, recharge and lake inflows this work uses the relation δD =8δ18O +15 as the Local Meteroic Water Line (AAMWL). This is more meaningfull than the LMWL constructed from the annual rains since the small rains do not produce major runoff, recharge or lake inflows. 20 20 0 50 15 150 10 200 5 250 Rainfall in mm δ18 Ο and d-excess 100 300 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec -5 350 400 Rainfall in mm Awasa Sodo Agermariam Kofele Addis Ababa d-excess Figure 3. Monthly variation in mean δ18O and d-excess of the Addis Ababa and the MER rainfalls. d excess plot is for Addis Ababa station. Sources of data: IAEA GNIP data base of Addis Ababa station (1965-1998) for Addis Ababa; IAEA-TC Projecs (1998 to 1999) for the four stations in the MER.. Rainfall histogram is drawn from Addis Ababa rainfall (1900-2000). The depletion in the summer rainfall relative to the spring rainfall is related to the difference in source of moisture and to local meteorological processes. The summer rainfall (75% of rainfall in Addis Ababa) is derived from the admixture of the Atlantic Ocean and the South/Equatorial Indian Ocean air masses (figure 1). The small variability in δ18O of the summer rains (figure 4) and section 3.3 suggest either nearly constant ratio of contribution of the two sources over the last 40 years (which is unlikely since the Ethiopian rainfall amount has varied at least by ±20% during this time (Conway, 2000) while the interannual variation in isotopes nearly remain constant) or that one of the two monsoons is the predominant source for Addis Ababa summer rains. However section 3.2 will show that the isotopic composition of the summer rain are in agreement with the meteorological evidence which states the Congo basin and Atlantic are important sources of moisture in summer. 8 18 δ 0 6 4 2 0 -2 -4 -6 -8 -10 Jan Feb Mar Apr May June Jul Aug Sept Oct Nov Dec 18 Figure 4. Monthly Maximum-Minimum-Mean plot of the δ18O ‰ of the Addis Ababa rainfalls (1965-2002). Highly variable δ is observed in June and October which marks the northward and south ward passage of the ITCZ over central part of Ethiopia. O content 21 The spring rainfalls are the most enriched compared to the summer rains. During this time, the oceanic moisture reaches the area from Northern Indian Ocean. The enrichment of the rainfalls during spring time may be related to three factors, a) as the Ethiopian highland is geographically closer to the North Indian Ocean and the moisture that reaches the area represents the initial stage of condensation which did not undergo major rainout fractionation effect (Joseph et al, 1992); b) the high temperature, the low atmospheric humidity and the low amount of rainfall during this time favors evaporation of rainwater leading to enriched rainfalls and low d-exces; c) the high sea surface temperature over northe indian ocean favors the formation of enriched vapor coming to Ethiopia. The dry season (November to February) is characterized by relatively enriched and low d excess compositions. The enriched δ18O and the low d excess compositions of this period reflect evaporation while raining owing to low rainfall during this time. Unlike the other Tropical East Africa 3 which shows small isotope variation during dry seasons (Rozanski, et al., 1996: Rietti-Shati, et al., 2000), the Addis Ababa rainfalls show small variation during July and August (see length of the line in figure 4). The highest variability in δ18O is observed in October and June. The most plausible explanation for this variability is related to the northward and southward passage of the ITCZ over central Ethiopia. This creates variable local atmospheric condition. Strong convection and high altitude condensation with in the air column associated with the front of the ITCZ while it is moving north results in the most depleted δ18O. In June where the ITCZ is not yet established, small local convection produces enriched rains. Likewise, in October if the ITCZ is not yet moved southward in the proceeding month strong convection within the air column and associated cold air mass from the Arabian continent would favor formation of depleted rains. When it is already moved southwards evaporation of rain while falling or local convective clouds results in enriched rains. Seasonal variation in deuterium excess is also influenced by the source of moisture and the local conditions. High deuterium excess is recorded in the summer rainfall and in September. The mean weighted d-excess increases continuously from about 10 at the onset of the North Indian Ocean monsoon in March to 16 at the end of the summer monsoon. The continuous increase in d-excess through the summer may indicate the continuous increment in the recycled component of the local moisture. The low d excess in the dry season rains reflect evaporation of rainwater while raining and subsequent evaporative enrichment. The high d-excess in the summer rains also reflect that the rains during this time are less evaporated while raining owing to high rainfall amount and the saturated atmosphere. 22 In September increase in δ18O is not accompanied by decrease in d excess. Two superimposed processes may result in this characteristic. While the enrichment reflects more involvement of local moisture as the ITCZ is moving southwards, the high d excess suggests the influence of both recycling and type of rain which is often solid form (afternoon hail storm). In solid precipitation isotopic disequilibrium may cause high d-excess (Gonfiantini et al., 2001)4. The same climate-isotope relation in Addis Ababa rainfalls explains the pattern of the seasonal variation in rainfall isotopic composition of the short-term MER stations. Figure 3 shows that among all the months September register nearly identical δ18O composition in all the stations perhaps reflecting that the rains were formed under similar rainfall formation mechanisms and from similar sources. Since the summer monsoon is retreating at this time and local afternoon convective storms are replacing it, the rains isotopic composition reflects local convection and moisture source from local recycled moisture. The summer rainfall δ18O variation in Ababa mirrors the variation in Sahel rains compositions (enriched at the beginning and depleted at the end of the summer). The later as reported by Taupin et al. (2002). The West African rains in Cameroon have also similar pattern of variation in summer (Njitchoua et al., 1999). However a slight difference exists between Addis Ababa rains and the West African rains after the end of the main rainy season. In Addis Ababa the southward migration of the ITCZ pulls the dry and cold Arabian air which favors formation of depleted rains in October. The most depleted compositions are also recorded in October. In west Sahel, the ITCZ after the rainy season pulls dry but warm air locally called 'Harmatan' from the Sahara. This produces enriched and low d excess rains. Other notable feature of seasonal variation in δ18O is that most depleted compositions are observed in October. This corresponds not to the amount of rain but to the physical condition associated with the southward migration of the ITCZ and the penetration cold air from Arabian continent. Furthermore the small rainy season (March- April) have high δ18O than the dry season rains (between October and February). This implies on seasonal basis rainfall amount is not the only factor that influences the isotopic composition of the rains. 3 4 the East African Rainfall stations include: Ndola, Dar es Salam, Kampala, Harare, Antananarivo, Entebbe 'Ice formation, if occurring, is supposed to take place by freezing the water droplets without affecting the isotopiccomposition.However, the isotopic fractionation in the subsequent vapour condensation on the ice surface, deviates from the equilibrium value because the light molecules H216O may be privileged for their higher diffusivity in air. This effect tends to offset the thermodynamic equilibrium by which the isotopically heavy molecules are preferentially fixed in condensed phases, and may determine a significant increase of the deuterium excess, because of the relatively small difference in diffusivity coefficients between HD16O and H218O.' Gonfiantini et al., 2001. 23 3.2. Origin of δ18O-δD of Ethiopian meteoric waters and its comparison with Sahel and East Africa As previously noted (Sonntag et al., 1979: Joseph et al., 1992; Rozanski et al., 1996; Darling and Gizaw, 2002) and as newly observed (part III and IV) the Ethiopian rainfall waters (and pristine meteoric waters in general) are somewhat 'unique' in their isotopic compositions compared to the Sahel and the East African rainfalls (meteoric waters) compositions. The major observations are: 1) Despite the low mean annual temperature and the high altitude location of Ethiopia, the weighted mean annual isotopic composition of Addis Ababa rainfalls does not show depletion compared to other East African rainfalls (Rozanski et al., 1996). 2) Groundwaters (or rains ) in Ethiopia are enriched compared with the western Sahel (Joseph et al., 1992). 3) There is an imbalance between the mean isotopic composition of rainfalls and the isotopic compositions of groundwaters in Ethiopia while in other East African region the two show comparable compositions1 (Darling and Gizaw, 2002; Gizaw, 2002). 4) The Ethiopian plateau groundwaters do not show altitude commensurate depletion despite the high altitude location of the region compared to the Sahel (part III this work). However the general pattern of seasonal variation in δ18O of summer rains is similar in both regions. 5) The Addis Ababa summer rainfalls particularly the rainfalls of the month of September (just after the retreat of the summer monsoon) contain the highest d-excess and relatively enriched δ18O. 6) The general pattern of JJAS variation in δ18O and δD (figure 3) resembles that of the Sahel rainfalls than the East African rains. The seasonal isotopic variation of Sahel west Africa was documented by various authors including Taupin, et al., 2002; Taupin, 2000. Unlike the other East African stations which show small isotope variability during the dry seasons (Rozanski et al., 1996), the Addis Ababa station shows small isotope variability during main rainy season in summer (figure 4). 7) Ethiopian Lakes are more enriched than other East African Lakes of comparable Evaporation/Inflow ratios (Part III). At least three major hypotheses (original hypotheses by: Sonntag et al., 1979 or Rozanski et al., 1996; Joseph et al., 1992 and Darling and Gizaw, 2002) exist to explain the enriched isotopic characteristics of the Addis Ababa rainfall compared to the Eastern and Sahelian Africa. These are: 1) Influence of the Congo vegetation or continental open water bodies: Earlier (Sonntag et al., 1979) suggested that moisture advection from transpiration by the Congo vegetation basin could influence the isotopic composition of rainfalls in north and northeast of the basin. Rozanski et al. (1996) relates the enriched δ18O of the December to May rainfalls of Addis Ababa to the mixing of transpired moisture from the Congo vegetated basin to the December 24 to May rain bearing moisture. Since transpiration is a non-fractionating process it returns enriched moisture to the atmosphere making the Ethiopian rains enriched. While meteorological observations supports the mixing of vapor from the Congo vegetation and other continental water bodies (see section 1) to the moisture laden westerly air coming from Atlantic to Ethiopia in JJAS, how it mix with the maritime moisture is not clear. It is not clear also how the vapor from Congo vegetation basin influences the December to May Ethiopian rains. Between December and May the ITCZ is still in southern Ethiopia and Central Ethiopia is getting its moisture from the Indian Ocean. 2) Influence of the North Indian Ocean via the TEJ/AEJ in the summer rainfalls: Based on isotopic composition of groundwaters along the Sahelian Africa starting from Djibouti to Senegal, Joseph et al., 1992 hypothesized that the African Easterly Jet (AEJ) and the Tropical Easterly Jet (TEJ) may transport an important amount of moisture from Indian Ocean or from the Arabian see westward across Sahel Africa in summer. They observed that the groundwaters in Western Shaelian African are more depleted in δ18O than the groundwaters and rainfall waters of Eastern African regions (Djibouti and Ethiopian Highlands). According to this hypothesis both the March-April (from Indian Ocean monsoon) and the JJAS rainfalls (from the Zonal flows) over Ethiopian highland represents the first condensations stage of the Indian Ocean or Arabian Sea moisture. This makes the Ethiopian meteoric waters enriched compared to meteoric waters of the West Sahelian Africa. The latter receives rains from the Zonal flows at the end of their condensation stage. The question that follow this hypotheses are a) How far is the sampled groundwaters representative of the continental scale meteorological processes? Local evaporation effect prior to recharge seems for example an important hydrological process in Djibouti and the Afar Depression (the areas from where Joseph et al., have taken groundwater samples) making the Ethiopian groundwaters enriched; and, b)does the generally known meteorologically based monsoon flow patterns and the ITCZ drifts support this idea? It is widely agreed that (at least in Ethiopia) the summer monsoon comes from the Atlantic or the Congo basin or from the Southern Indian Ocean than from the Indian Ocean alone. Furthermore recent closer monitoring of event based rainfall isotope monitoring (Taupin et al., 2000) shows the Indian Ocean moisture is unimportant in the Sahel rains. The next section of this thesis will show a clear West to East flow of the summer monsoon over the Ethiopian Plateau facing west as demonstrated by continuous depletion of heavy 25 isotopes eastward on the NWP. These exclude the importance of the Easterly Jets, which should normally deplete westward, as important moisture source in Ethiopia. 3) The Addis Ababa anomaly?: Although the state of sampling condition is not directly implied as a cause of the isotopic characteristics of the Addis Ababa rainfall, a recent work by Darling and Gizaw (2002) shows that there is an imbalance in the weighted rainfall isotopic composition and the isotopic composition of 'unmodified' groundwaters in Ethiopia. According to these authors, in all stations in Eastern Africa, except Addis Ababa, the comparison of the weighted mean annual rainfall isotopic composition with that of unmodified groundwaters shows a comparable range 5. However, if a comparison were made between the weighted mean summer rainfall (the rainfall which is available for recharge and runoff in the NWP) isotopic composition (δ18O = -2.5‰ and δD = -5 ‰) and groundwater isotopic compositions of the Ethiopian plateau the problem of 'imbalance' should not have existed. As will be demonstrated in Part III, the Ethiopian Lakes are enriched than other East African Lakes of similar evaporation to inflow ratios reflecting a general enrichment of the Ethiopian meteoric waters feeding them. This shows again the relative enrichment of Ethiopian meteoric waters. Based on the new and the previous observations and the new questions, the following lines of approach and hypotheses can be used/made about the origin of the isotopic composition of the meteoric waters of the Ethiopian highland. Approaches • Observations have still to be improved. Comparisons between the three sectors were often made on short isotope records or on few water samples. 5 100 80 60 Ave rage Summe r Rainfal l at Addi s Ababa δδδD 40 20 0 * * Ave rage March-Apri l Rainfal l at Addis Ababa -20 -40 -10 -8 -6 -4 -2 0 2 8 10 Lakes and rivers draining them 4 6 GMWL -60 18 δδ O B A. Low TDS cold groundwaters high TDS Na-HCO3 waters from YTVL and LTG Figure A (Darling and Gizaw, 2002 and Gizaw, 2002) shows imbalance between groundwater and annual average δ18O and δD of rains figure B (part II this work) shows the groundwaters from NWP plots around the average summer rains. The imbalance between rains and groundwater composition occurs therefore only if the average isotopic compositions of annual rain are plotted against isotopic composition of groundwaters. 26 • The rainfall isotopic composition of the Ethiopian highland and that of eastern Africa should be compared on seasonal basis than on the mean annual rains as the two regions are not always influenced by similar moisture trajectories. The offset in the main rainy seasons between the Northern Ethiopian highland and other east African region indicate the two regions are influenced by different moisture trajectories. Therefore the difference in the mean annual rainfall δ18O compositions of rains of the two sectors should be compared not in terms of condensation history but in terms of the kind of land surface (mountains, open water bodies, vegetation, desert etc) on the moisture pathways and in terms of evaporation conditions at source. If the effects of condensation along moisture trajectory were to be compared between Ethiopian and the East African stations one has to do the comparison on the March-April rains which are in phase in the two regions. • Likewise, if comparison were to be made between the Sahel and the Ethiopian meteoric waters to understand the history of moisture trajectory, it should be made on the isotopic composition of the summer rains or on the isotopic composition of pristine groundwaters recharged out of the summer rains rather than on the mean annual rains isotopic compositions. Furthermore the differences in the nature of rainfall formation (monsoonal, convective, orographic, and frontal) along the different part of the Sahel should be considered. Hypotheses • The Ethiopian March April rains are enriched than the eastern African equivalent because the former rains represent the initial stage of condensation for the moist easterly flows from the Indian Ocean. One would for example see the March-April rainfalls in Kampala are more depleted than the March-April Addis Ababa rains (Rozanski, et al., 1996) showing northeastsouthwest depletion of δ18O along moisture trajectory. Furthermore the Ethiopian region gets its rain from northern part of the Indian Ocean while the East African stations get their rain from southern and equatorial Indian Ocean. Sea surface temperature difference on the Indian Ocean could result in differences in isotopic composition of the easterly moistures. At Addis Ababa the March-April months are preceded by long dry period compared to the east African stations, which have wet seasons in October and December. Evaporative effect and isotopic exchange with the dry atmosphere enriches the Addis Ababa March-April rains compared to East African stations. • The fact that the Ethiopian meteoric waters do not show altitude-commensurate depletion compared to the Sahel meteoric waters can be related to a variety of factors. These include a) part of the summer rainfall in Ethiopia is derived from the Indian Ocean monsoon which did not undergo previous condensation while the influence of Indian Ocean is minimal in western 27 Africa; b) as already shown by meteorological evidences and as hypothesized by Sonntag et al., 1979; part of the summer monsoon in Ethiopia is fed by continental open water bodies and the vapor from Congo vegetation basin. The preceeding discussions favor the second factor. • Since Ethiopia is uplifted compared to Sahel Africa, orographic effect can play an important role in lifting up the enriched ground level moisture derived from local evapo-transpiration or from the low-level westerly flows from Congo. This enriches the maritime monsoon air mass (Atlantic/Indian) and therefore the summer rains of Ethiopia. The elevated terrain in Ethiopia and the heating of the plateau can trigger and maintain local convective activity. The apparent lack of altitude commensurate depletion of the meteoric waters of northern and central Ethiopia compared to the Sahel could be therefore partly the result of the maintenance of the convection of ground level enriched/recycled moisture due to orography. Enrichment of local convective rains resulting from mixing of ground level vapor is not uncommon in western Sahel (Taupin et al., 2002). However, in the Sahel, this type of rains happens temporarily when the ITCZ fails to move north promoting local convection and storms (Taupin et al., 2002). In this sense it can be hypothesized that in tropical warm mountains, topography can play two opposing roles. While the decrease in temperature with altitude (and reduction of evaporation effect) leads to depletion of the heavy water isotopes, the lifting up of enriched ground level moisture from evapo-transpiration by mountains enriches the air mass in heavy isotopes. The isotope altitude effect or the continentality effect is therefore the balance between the two. The lack of strong depletion of isotopes despite high altitude (or with altitude) is not only restricted to the Ethiopian mountains. Similar pattern of isotope distribution in tropical mountains is observed in Tibetan Plateau (Zhang et al., 2004; Sugimoto personal communication, IAEA groundwater conference 2003), and the Kenyan mountains (Rietti-Shati et al., 2000). The latter attributes the lack of strong 'altitude effect' and the enrichment of meteoric waters at high altitude in Kenya to the influence of recycling of local moisture. Zhang et al. (2004) attributes the lack of isotope continentality effect north of Himalayas to the contribution of locally produced vapor in the mountainous region. The role of orographic convection or orographic clouds was not a widely pronounced issue in the literature. Recently Liotta, et al. (2004) demonstrated how orography plays an important role in changing the original isotopic mark (particularly d excess) of precipitation in Sicily. 28 One of the evidences for existence of the influence of the local (or regional) continental recycled moisture in the Ethiopian meteoric waters is the d-excess composition of the Addis Ababa rainfalls and that of the groundwaters of the NWP. High deuterium excess is often attributed to land surface-atmosphere interaction via moisture contribution from evapotranspiration (Rietti-Shati et al., 2000; Gat et al., 1994). Figure 5 shows how this process produces high d-excess (and enriched δ18O). The Ethiopian rainfalls (at Addis Ababa) are characterized by high deuterium excess (>10) ranging from 10 in pre-summer monsoon and increases up to 16 during the retreat of monsoon. The continuous increase in d-excess starting from the onset of rainfall in March up to the end of monsoon in September may be partly related to the continuous increase in the volume of locally recycled moisture in the rainfalls. Figure 5 (modified from Gat et al., 1994). The isotopic composition of evaporated surface water (δw), the original water body, soil moisture or leaf-water prior to evaporation (δp), and the evaporated water vapor (δE) all plot along the same line called the local evaporation line. The meteoric water that forms from the condensation (δp) of the 'monsoon only' moisture (δ a) is separated from δa by the enrichment factor (ε*). When the local evaporate (δE) mixes with the monsoon vapor (δa) due to orographic lifting of the evaporate. A new enriched vapor (δa') that plot in triangular zone bounded by the evaporation line, the MWL and above the MWL is formed. The rains that condense from this new vapor plot along a new line parallel to the MWL but with a higher d-excess and enriched δ18O rainfalls. The composition of the new vapor δa' depends on the degree of mixing between the monsoon moisture and the 'local convective moisture'. Higher amount of local moisture produces isotopically enriched rainfall with high d excess. Generally it can be said that while differences in sources of moisture make the Ethiopian meteoric waters to have different isotopic composition as compared to eastern Africa; Orographically triggered convection of local or low level moisture from Congo basin and the Atlantic enriches the Ethiopian summer moisture mass. 3.3. Independent geochemical evidence on source of moisture Often as a complementary source of information in understanding sources of moisture, rainfall chemical composition was used in Western Africa (eg. Savenije , 1995). The assumption is that salinity decrease that is observed in inward continental regions (and during the rainy season) compared to the coastal regions (the start of rainy season) is cause by continuous recycling of moisture along moisture trajectory. This assumption however sounds simplistic because the salinity decrease in rainfall may be caused by wash out and continuous cleaning of the atmosphere starting from the onset of rain bearing system through its development and end. It appears that elemental ratios such as Na/Cl, 29 or ratios of other elements tell more about changes in moisture sources than salinity used alone. Appendix I-2 discusses the chemical composition of weekly monitored rainfall of the summer 2003 rainfalls of Addis Ababa. The rainfall chemistry (box 1 in Appendix I) measured at Addis Ababa shows a continuous decrease in the salinity and all major ions starting from the onset of the summer rainfall. It starts to increase again around the end of the monsoon. As to what factor (a continuous clean up of the atmosphere or a continuous recycling of moisture) this is related is not clear. However the Na/Cl of the September rainfalls is different from the other summer month's ratio reflecting differences in sources of moisture and the nature of rainfall formation. There is a likelihood of mixing of near surface moisture into raising air mass during strong September afternoon convection. The isotopic composition of September rainfall is also relatively enriched and has the highest d-excess reflecting isotope disequilibrium during raindrop formation. Comparison between the Ethiopian summer rainfalls and the summer rains from Sahelian West Africa (Goni et al., 2001) shows that the former are dilute in all ions but keeps similar Na/Cl ratio. This suggests similarity in sources of moisture for the two regions but different trajectories over which the moistures pass before they reach the regions. The furthest distance of Ethiopia from the Atlantic could explain the relatively dilute salinity of the Addis Ababa rains compared to the West Sahel rains. 3.4. Long-term variation in isotopic composition of rainfalls This part attempts to show the existence of long-term trends in δ18O and d-excess of the Addis Ababa rainfall (Figure 6 and Figure 7). To see the long term, trend analyses is made on seasonal basis. The year is divided into four parts: the dry season (September to February), the spring rainfalls (March and April), the summer rainfall (June to September) and the month of May (the transition between the spring and the summer rains). Some notable features from the figures are: • The regular pattern of variation in δ18O and d-excess of the Summer rainfall (with a slight decreasing trend in the d-excess) • Quasi- regular to regular pattern of δ18O and d-excess of spring rainfall with no major increasing or decreasing trend except some points that fall out of the major trend. • Highly irregular pattern of δ18O and d excess of the dry season rains • Enriched δ18O and low d-excess summer rains centered over the mid 1980s, a period of lowest rainfall in Ethiopia (the enrichment is related to local evaporation effect due to low rainfall) • Generally a decreasing trend in d excess of summer rainfalls 30 8 6 4 18 δ O 2 0 -2 -4 -6 j-93 j-95 j-97 j-99 j-01 janv-95 janv-97 janv-99 janv-01 M ay j-91 j-89 j-87 j-85 j-83 Spring janv-93 Automn j-81 j-79 j-77 j-75 j-73 j-71 j-69 j-67 j-65 -8 Summer 20 15 10 5 0 janv-91 janv-89 janv-87 janv-85 janv-83 janv-81 janv-79 janv-77 janv-75 janv-73 janv-71 janv-69 -10 janv-67 -5 janv-65 D excess in spring and summer rainfalls Figure 6. Long-term variation in amount weighted δ18O of Addis Ababa precipitation. Summer Rainfall Spring Rainfall Trend in d-excess of summer rainfall Trend in d-excess of spring rainfall Figure 7. Long-term variation in deuterium excess of the Addis Ababa rainfall. These observations lead to the following general remarks: • The trend in δ18O of the summer rainfalls are dominantly linked to global or continental scale processes with minor effect of local scale processes. The absence of strong inter-annual variability with in the summer or the spring rains may also imply the persistence of source of moisture over the last five decades. Since the rainfall has been changing during the last decades the persistence reflects that the summer monsoon has predominantly one source rather than two or more sources mixing at sub-equal proportion. Had there been two or more sources of summer moisture in Addis rainfalls the isotope trend would have been likely irregular. 31 • The irregular pattern of the dry season rainfalls isotopic compositions are influenced by local climatic factors during rainfall (such as the time of the day where rainfall takes place, the rainfall intensity, the temperature etc). • The slight decrease in deuterium excess irrespective of nearly constant δ18O over the last sixty years may indicate the decrease in the percentage of the contribution of recycled continental moisture to the maritime air mass because of changes in land surface characteristics such as massive deforestation over Ethiopian highland (Ethiopian forest cover diminished from 47% of the land cover to 3% in the last 50 years). The isotope data do not allow conclusive remarks because over the recent years analytical techniques have improved and the variations in isotopic composition may also be related to improvement in precisions in measurements. 3.5. Spatial variation in isotopic composition of meteoric waters The altitude, the temperature and the amount effects On the windward face of a mountain, the δ18O and δD composition of rainfalls decrease with increasing altitude. This phenomenon is termed as the 'altitude effect'. The altitudinal variation of isotopic composition provides a suitable basis to trace source of groundwater recharge. A moisture mass that ascends a barrier results is fractionation of isotopes leading to a depleted composition at higher grounds. Because of complexities in topography and circulation of rainfall bearing moistures, and complexity in local convective activities getting a single isotope altitude gradient for Ethiopia sounds imprecise. In the MER and the Afar Depression, for example rain-producing moisture descends into the lowlands from the adjacent highlands. Although in some localities of the Rift and the Afar depletion of isotopic composition with altitude is observed (eg. McKenzie et al., 2001; Gizaw, 2003), the cause of such depletion is not the traditional/conventional altitude effect. The depletion of δ18O with altitude in a leeward side of mountains, if present is often the outcome of the pseudo altitude effect which is equivalent to Foehn effect in meteorology. The pseudo altitude effects are often the result of high temperature in the leeward side of mountains and the likelihood of evaporation that causes isotope fractionation of rainfalls6. In regions where rainfall isotopic record is not available across an altitude gradient, low TDS or low chloride or low temperature and modern groundwaters can be used to calibrate the altitude isotope relations. 6 The implication of this in isotope groundwater tracing is worth noting. Occasionally the moisture coming from the windward direction a mountain passes into the lee-ward side leading to occasional heavy rainfalls and flooding. This kind of meteorological processes were observed in Afar and the Ethiopian rift valley several time (Gemechu, 1977). This kind of strong monsoon flow may lead to formation of depleted rainfall in a region otherwise should receive enriched moisture. The occasional flood forming monsoon events which are able to pass over the mountain barriers may result in isotopically depleted groundwater in Afar and Djibouti. This kind groundwater recharge sources have been previously noted by Fontes et al. (1980) in Djibouti. 32 Figure 8 uses groundwaters as proxies of rainfall isotopic composition. The figure shows on the NWP there is a -0.1 ‰ depletion of δ18O and in the MER the Afar in general there is a δ18O depletion of 0.14 ‰ per 100m. The -0.1‰ depletion in δ18O in windward face of the NWP is consistent with many tropical mountains including Cameroon (Njitchoua et al.,1999) mount Kenya Rietti-Shati, et al. 2000), Costa Rica (Lachniet and Patterson, 2000). A recently proposed relation of -0.5‰/100 in southern part MER (McKenzie et al., 2001) using groundwater isotopic compositions as proxies of rainfall compositions seems to be inconsistent with many tropical isotope altitude relations. The apparent disagreement of the later reflects the effects of leeward evaporative enrichment before recharge in amplifying the altitudinal variation in isotopes if groundwaters were used as a basis of calibrating the altitude effect. Figure 8 attributes altitude effect for the NWP and a pseudo altitude effect for the Ethiopian rift and bordering escarpment. The d excess map shows that there is an increase from the western sector of the NWP to its eastern mountainous sector overlooking Afar Depression then it decreases significantly as one move towards the Afar depression. The increase in d-excess with altitude is a common feature around windward face of mountains. The increase in d excess in with altitude in the western face of the NWP could be related to continuous increase in orographic clouds in the region or to local moisture recycling or to reduction in evaporative fractionation of rains at high altitude. Comparison of the isotope map (figure 8) with the rainfall and temperature maps of Ethiopia (figure2) shows that there is no direct relation between rainfall amount and isotopic depletion in the NWP. In the NWP, the cold and less humid eastern mountainous region overlooking the Afar Depression is characterized by the most depleted waters while the high rainfall region in western lowlands of the NWP are characterized by enriched meteoric waters. The gradual depletion of the monsoon moisture as it moves eastward over the NWP owing to condensation dominates the amount effect in the region. The depleted composition in the eastern highlands of the NWP is therefore the result of rainout effect and low temperature. This implies that the 'amount effect' in tropical stations (Rozanski et al., 1993) is suitable to compare inter-annual variation or seasonal variations or rainfall events in δ18O composition at one particular station than in comparing spatial variations in δ18O. On monthly time scale also there is no well-defined relation between amount of rainfall and the isotope depletion in Addis Ababa (Gizaw, 2002) and in the west Sahel (Taupin et al., 2002). Comparison of the isotope map and the temperature map shows that a good relation exists between local temperature and isotopic composition of meteoric waters. In the NWP for example the most 33 depleted meteoric waters are located in the coldest region and the most enriched water in the western warm lowlands. The relation between air temperature and the δ18O composition has the form: δ18O = 0.21 T (°C) - 6.5 34 Wollo highland 4000 ALTITUDE IN M -5 Metekel lowland Blue Ni le Gorge 3000 P =2 0 0 0 m m / yr T = 18 °C 2000 P = 8 0 0 m m / yr T= 14 °C -4 Afar depression -3 -2 18 Gojam plateau δ O 5000 -1 1000 0 P=800 mm 0 P = 18 0 0 m m / yr T= 2 4 °C P = 2 5 0 m m / yr T= 3 5 °C 1 Figure 8. Spatial δ18O (top) and d-excess (middle) map of cold groundwaters and simplified sketch of altitudinal variation in δ18O and dexcess on representative cold groundwaters (lower) along an East-West transect from western Ethiopia (Metekel Lowland) to Djibouti. In the lower figure, the circles indicate oxygen isotope content and the numbers near the circles show the d-excess contents. The filled in arrows indicate the direction of the summer monsoon moisture movement. P and T are mean annual rainfall and temperature respectively. The moisture continuously depletes in heavy isotopes(δ18O) as it moves over the western face of the NWP because of the altitude effect. In the leeward face of the mountain facing the Afar depression the moisture mass descends and gets warmer. This leads to irregular pattern of δ18O variation with altitude in the leeward face. In some places evaporation while rainfall in the leeward side results in a pseudo altitude effect of -0.14 ‰ per 100m. An additional complication in escarpment facing Afar is the Indian Ocean monsoon which comes from East. There is a 0.25‰ depletion of δ18O for a 1°C decrease in air temperature in the NWP (figure in the upper right corner). 35 This relation is different from the Dansgaard's (Dansgaard, 1964) local air temperature- oxygen isotope relation: δ18O = 0.7 (°C)-13.6 but similar to many other tropical mountian's temperature isotope relation- eg. 0.3‰ depletion per 1°C decrease in temperature in Pacific slope of Costa Rica (Lachniet and Patterson, 2002). Geographical variation in δ18O and d-excess and their relation to climate variables Using groundwaters as a basis of comparison, figures 8 and 9 compares the δ18O and d-excess compositions of three sectors namely the NWP, the MER, and the Afar Depression. The spatial variation of isotopes mirrors the climate differences of the three regions. Generally groundwaters of the NWP plot above GMWL and they have the highest d-excess. The Afar groundwaters are characterized low d-excess. Geothermal waters of Afar are relatively depleted in δ18O but they are characterized by low d-excess. The main Ethiopian rift groundwaters have an intermediate composition. This geographical variation in isotopic composition is related to the variation in air temperature, to the rainfall derivation mechanisms and to the degree of importance of evaporative enrichment prior to recharge in the three sectors. The most depleted and high d-excess groundwaters are located in the highlands of the NWP. The rainfall regime in this region is influenced by the summer monsoon penetrating the region from southwest. The groundwaters in eastern mountainous sector (the Wollo Highland) of the NWP are characterized by isotopically the most depleted (figure 8 and 9) representing recharge by isotopically depleted rains. In the eastern mountainous region of the NWP the rainfall moisture is at its final stage of condensation as the summer monsoon and orographically enhanced moisture moves eastward and northeastward over the mountains facing west. This results in isotopically depleted rains in the eastern sector of the NWP. The isotopic enrichment in the Afar Depression groundwaters and their low deuterium excess is most likely related to evaporation prior to recharge because of high temperature in the leeward side. The altitudinal depletion in δ18O in the western escarpment bordering Afar should be the result of this 'pseudo altitude effect'. Thermal groundwaters of the Afar which are depleted in δ18O (δ18O <-3.5‰) but characterized by d-excess less than 10 seems to be older and recharged from paleo evaporated waters. The majority of thermal waters of the Afar and Djibouti depression have similar δ18O content to some plateau waters (-3 to -5) but they have the lowest d excess. Positive oxygen shift alone does not explain the isotopic content of the Afar thermal waters because projection towards the Meteroic water gives original water with -4.5 to -6‰ δ18O, a composition which is not observed in modern Ethiopian cold groundwaters. 36 Generally the NWP is characterized by high d-excess (the mountainous areas in the eastern part of the NWP). The MER contains more water points with high d excess than the Afar depression. This will be later (part II) interpreted as a frequent involvement of plateau type water in the MER but not frequently in Afar. An interesting zone in this regard is the branching point between the MER and the Afar. In the zone three major fault systems including the western boundary faults of the Afar, the NE oriented faults of the MER and the East-West old fault of the YTVL intersects. The zone is characterized by emergence of many geothermal springs. The structures favor groundwater transfer from the plateau to the rift. The high d-excess around the branching point reflects the involvement of plateau type water than local recharge from local precipitation. 60 50 MER Lakes 40 thermal waters in NWP 30 Lake Groundwater mixing line δD 20 10 0 Geothermal oxygen shift in MER thermal waters -10 -20 Geothermal waters in Afar and Dijibouti -30 -40 -7 -5 -3 -1 1 3 5 7 9 18 δ O NWP MER AFAR Figure 9. δ18O, δD plot of groundwaters of the NWP, the MER and the Afar Depression. The Afar groundwaters are characterized by low d excess, the NWP groundwaters have the highest d excess. The MER waters plot in between. The highly enriched waters of the three sectors are the result of lake water mixing or recharge from modern evaporatively fractionated infiltration. The majority of evaporatively fractionated waters occur around the major lakes in the MER. 3.6. Conveniences of δ18O and δD in Ethiopian water resources and paleo-climate studies Isotope groundwater tracing uses the advantage of presence of temporal (seasonal, long-term) or spatial variation (geographic, altitudinal) in isotopic composition of recharging waters as its basis. In the Ethiopian case the following properties of the isotope-metrology-geography relations are found to be useful in water resources studies. These are: 37 a) The clear difference in the δ18O of the summer rainfall and the spring rainfalls helps which of the rainfall is available for groundwater recharge, for runoff, and for lake inflows b) The strong evaporation effect prior to recharge in the MER and the Afar Depression owing to high evaporation helps to distinguish sources of groundwater recharge and to trace the presence of unmodified plateau waters in the Rift and Afar aquifers c) The altitude-δ18O relation on the NWP facing west may be used to trace the recharge areas the mechanism of regional groundwater flow in NWP d) The fact that the global scale process such as the seasonal drifting of the ITCZ and associated change in moisture source is not completely masked by local or continental moisture feed back processes may be to the advantage for paleo-climate studies that seek global scale changes tied with oceans e) The fact that each moisture trajectory has distinct isotope signal should provide useful information in meteorological studies in Ethiopia. At least from the isotope variation (figure 4) one can see two or three clear zones. The NWP which is characterized by a west-east gradient of depletion of isotopes, the quasi irregular δ18O but low d-excess zone of Afar, and/or the MER which has an intermediate characteristics. These differences in isotope pattern mirrors meteorology. Therefore isotope can be integrated in to meteorological monitoring networks of Ethiopia to understand the complex moisture regimes in the country f) The major runoff producing rains in the NWP are those in JJAS. This allows many of the Ethiopia lakes to receive one regime of isotopic composition of inflows. Furthermore the seasonal variation in isotopic composition of Ethiopian rainfalls is not as pronounced as in high latitude or temperate regions. This allows lake water isotopic composition at different months to follow a single evaporation line (LEL). This allows rapid computation of isotope water budgets of lakes in Ethiopia. In high latitude regions the strong seasonal variation in isotopic composition of lake inflows make isotope water budget computation difficult because lakes in high latitude regions often follow two LEL on in winter and the other in summer (Gibson et al., 1996) Despite all these advantages, some care should be taken (as will be demonstrated in part III) in using isotopes of water in groundwater tracing. These include: a) The isotopic signal of groundwaters in the Ethiopian region contains in itself the signature of the complexity in rainfall derivation mechanisms and moisture circulation patterns over Ethiopian terrain. So care should be taken in interpreting isotopic signals of groundwaters and in tracing sources of recharge. Complementary geological and hydrogeological information should be used 38 b) In the Ethiopian rift valley and the Afar, evaporative enrichment prior to recharge is an important process. At the same time these regions contain a number of isotopically highly enriched lakes. Often enriched isotopic compositions in aquifers adjacent to lakes in the MER and the Afar are attributed to lake water mixing into them (eg; Darling et al., 1996; Kebede et al., 2002). Care should be taken or other independent methods (such as hydrogeological or geochemical knowledge) be used in such qualitative lake groundwater interaction tracing. This is because groundwater to which lakes are not mixed may also show isotopic enrichment owing to pre recharge evaporation. Similarly in Afar for example geothermally 'oxygen shifted' groundwaters and groundwaters from 'evaporatively fractionated paleo-recharge' may not be distinguishable based on isotopes of water alone c) Under the present day condition, the spatial relation between isotopes and meteorology shows that the isotopic composition reflects most the local temperature than the local amount of precipitation. Furthermore, local process ('environmental factors') such as evaporation and rain shadow effect plays an important role in modifying the original isotopic composition of moisture. If isotopes were to be used as proxies for paleo-climate in such regions one has to consider the importance 'paleo- local' factors. Part IV elaborates this further if lake sediments were used as paleo proxies 4. General remarks Part I relates the spatial and temporal variation in common isotopes of Ethiopian meteoric waters to the global and local meteorological processes. Seasonal variation in the δ18O and δD is related to the seasonal drifting of the ITCZ and the associated changes in sources of moisture. The summer rainfalls (June-Septermber) are depleted compared to the spring (March-April) rainfalsl. The difference in the isotope signals of the two seasons helps to identify which rainfall is available for groundwater recharge. The most depleted composition is recorded in the cold dry season. Local processes are also important (if not the most important in the Afar Depression) in modifying the spatial variations in δ18O and δD signals of Ethiopian meteoric waters. Altitudinal variation in δ18O is -0.1‰/100 m on the North Western Plateau . The windward altitude effects in the North Western Plateau and the leeward evaporation in the Afar and the Main Ethiopian Rift principally influence the spatial variation in isotopic composition of the Ethiopian meteoric waters. Under the present day condition there is no direct relationship between the spatial variation in rainfall amount and the spatial variation in isotopes in the North Western Plateau. The arid cold mountains bordering Afar contain the most depleted groundwaters compared to the high rainfall high temperature regions of the Western and Southwestern Ethiopian lowlands. This reflects the amount effect can be dominated by other isotope effects such as Rayliegh disitlation. This inturn implies the in paleo 39 climate reconstruction care should be taken in directly associating the amount effect as the major isotope effect. Yet there is a clear relation between spatial temperature map and spatial δ18O over the NWP and the Afar depression. A 1°C decrease in air temperature depletes the 18O content by about 0.21‰ on western face of the North Western Plateau. The isotope meteorology of Ethiopia shares its characteristics with Sahel and the East African regions. This makes the central sector of Ethiopia to be isotopically sensitive to changes in paleo moisture source switching with the paleo-position of the ITCZ. Under the present day condition, however, a complete seasonal switching from the Atlantic/Congo monsoon to a complete Indian Ocean monsoon or changes in proportion of moisture sources in each monsoon would change the δ18O from -2.5‰ to 0 ‰. In high latitude and temperate climate such switching in sources of moisture would bring much larger isotope variations. In Ethiopia neither of the isotope effects (altitude effect, temperature effect, amount effect, switching of the source, and continentality effect) seems to be chage the isotopic composition by large (greater than 2 per mil) amount. This implies, in Ethiopian paleo-climate studies therefore major changes in isotopic compositions such as the -6.5‰ decrease in δ18O in Pliocene meteoric waters (Hailemichel et al., 2002; Levin et al., 2004) should be interpreted not only in terms of changes in the sources of moisture and rainfall intensity but also in terms of other factors such as temperature lowering (reduced evaporation effect) and humidity increase that reinforce the depletion as well us in terms of major displacement in the position of the ITCZ. Modern modeling on the Bishoftu Crater Lakes (Kebede et al., 2002) shows under high rainfall condition the amount effect, lowering of temperature, the increase in humidity, and depletion in isotopic composition of ambient vapor reinforce each other and produce isotopically the most depleted lake water and therefore their sediment derivatives. As compared to high latitude regions, the seasonal or inter annual variation in isotopic composition of rainfalls at Addis Ababa is small and the variation that exist is the result of interplay of factors. Under such conditions, suitable paleo proxies are those materials which amplify the input signal by local predictable processes (eg. lakes amplify the input signal due to evaporation and register it in their sediments). Event and seasonal based comparison of the isotopic composition of the Sahel and East Africa rains would give much better information about the origin of the isotopic composition of these regions. Rainfall geochemical monitoring is an additional low-cost approach to understanding the Ethiopian rainfall derivation mechanisms. A closer look at orographic clouds and Foehn winds may provide more information about the origin of isotopic compositions of Ethiopian waters. 40 References Ali, A., 1999. River-groundwater interactions in the Sekelo-Akaki basin. Unpublished M.Sc. thesis, Addis Ababa University. Ayenew, T. 1998. The Hydrogeological system of the lake district basin, central main Ethiopian Rift. Phd thesis, ITC publication Number 64. Beyene, K., 2000. Chemical and isotopic studies of geothermal prospects, in the southern Afar region, Ethiopia. Proceedings of the World Geothermal Congress. Page: 977-983, Tohoku, Japan, May 28-June 10, 2000 Camberlin, P., 1997. Rainfall anomalies in the source region of the Nile and their connection with the Indian summer monsoon. J. Climate, 10, 1380-1392. Camberlin, P., and Okoola, R.E., 2003. The onset and cessation of the ‘‘long rains’’ in eastern Africa and their inter-annual variability. Theor. Appl. Climatol. 75, 43–54. Chernet, T., 1998. Etude des Mechanismes de mineralisation en fluorure et elements associes de la region des lacs du rift Ethiopien.Université d’Avignon, France, Ph.D. thesis. Craig, H., Lupton, J.E., and Horowiff, R.M., 1977. Isotope Geochemistry and Hydrology of geothermal waters in the Ethiopian rift valley. Scripps Inst of Oceanogr. Univ.of California report, 160pp. Dansgaard, W. (1964). Stable isotopes in precipitation, Tellus, 16:436-468. Darling, W.G., 1996. The Geochemistry of fluid processes in the eastern branch of the east African rift system. Ph.D thesis, British Geological Survey. Pp51-235. Darling, W.G., Gizaw, B., and Arusei, M.K., 1996. Lake groundwater relationships and fluid rock interaction in the east African rift valley, isotopic evidence. Journal of African Earth Sciences, 22, 423-431. Darling, WG, and Gizaw, B.,2002. Rainfall–groundwater isotopic relationships in eastern Africa: the Addis Ababa anomaly. Study of environmental change using isotopic techniques. C&S Papers Series, IAEA, pp 489– 490. Fontes, J.Ch., Florkowiski, T., Saliege J. F., and Zuppi, G.M., 1980. Environmental Isotope study of groundwater systems in the Republic of Djibouti. In Arid Zone Hydrology, IAEA; 237-262. Gat, J.R., and Rietti-Shati, M.,1999. The Meteorological vs. the hydrological altitude effect on the isotopic composition of meteoric waters. In: Isotope techniques in water resources management IAEA-CSP2/C 1998 proceeding of a symposium held in Vienna, 10-14 May 1999 (CD-ROM). Gat, J. R., Bowser, C. J., and Kendall, C., 1994. The contribution of evaporation from the Great Lakes to the continental atmosphere: Estimate based on stable isotope data: Geophysical Research Letters, v. 21, p. 557–560. Gemechu, D., 1977. Aspects of Climate and Water Budget in Ethiopia. Addis Ababa Univ. Press. 71 pp. Gibson, J.J., Edwards, T.W.D., and Prowse, T.D., 1996b. Development and validation of an isotopic method for estimating lake evaporation. Hydrological Processes, 10: 1369–1382. Gizaw, B., 2002. Hydrochemical and Environmental Investigation of the Addis Ababa Region, Ethiopia. Ph. D dissertation, Faculty of Earth and Environmental Sciences Ludwig-Maximilians-University of Munich, 157p. Gonfiantini, R., Borsi, S., Ferrara, G., and Panichi, C., 1973. Isotopic composition of waters from the Danakil depression (Ethiopia). Earth Planet. Sci. Let. 18, 13-21. Gonfiantini, R., Roche, M-A.,Olivry, J-C., Fontes, J-Ch., Zuppi, G.M., 2001. The altitude effect on the isotopic composition of tropical rains Chemical Geology 181:147–167. Goni, I.B. Fellman, E., and Edmunds W.M., 2001. Rainfall geochemistry in the Sahel region of northern Nigeria Atmospheric Environment 35: 4331-4339. Griffiths, J.F., 1972. Climates of Africa. In: Griffiths, J.F.(ed), World Survey of Climatology. Elsevier, Amsterdam. Hailemichael, M., Aronson, J., Samuel, S., Tevesz, M., Carter, J., 2002. D18O in mollusk shells from Pliocene Lake Hadar and modern Ethiopian Lakes: implications for hisory of the Ethiopian monsoon. Palaeogeogrphy, Palaeoclimatoloty, Paleoecology, 186: 81-99. Hemming, C.F., 1961. The ecology of the coastal area of Northern Eritrea. Journal of Ecology, 49: 55–78. IAEA TC project (ETH005 to ETH 008 project since 1995). An ongoing project conducted by International Atomic Energy Agency, the Ethiopian Science and Technology Commission and the Ethiopian Geological Surveys. London, 1982, 364 pp. IAEA, Isotope Hydrology Section GNIP data base, http://isohis.iaea.org, access date 2003. Joseph, A., Frangi, P., and Aranyossy, J.F., 1992. Isotopic composition of Meteoric water and groundwater in the Sahelo-Sudanese Zone. Journal of Geophysical research 97 : 7543-7551. Kebede, T., 1964. Rainfall in Ethiopia. Ethiopian Geography Journal, 2: 28-36. Kebede, S., Ayenew, T., and Umer, M., 2002a. Application of Isotope and water balance approaches for the study of the hydrogeological regime of the Bishoftu crater lakes, Ethiopia. SINET: An Ethiopian Journal of Science 24(2), 151-166. 41 Kebede, S., Lamb,H., Telford,R., Leng, M., and Umer, M., 2002b. Lake-Groundwater relationships, oxygen isotope balance and climate sensitivity of the Bishoftu Crater Lakes, Ethiopia. Advances in Global Change Research, 12: 261-275. Krauer, J., 1988. Rainfall, erosivity and isoerodent map of Ethiopia. Soil Conservation Research Project, Research Report 15, University of Berne, Switzerland. 132 pp. Lachniet, M., and Patterson, W., 2000. Stable isotope values of Costa Rican surface waters. Journal of Hydrology, 260: 135-150. Levin, N., Quade, J., Simpson, S., Semaw, S., and Rogers, M., 2004. Isotopic evidence for Plio-Pleistocene environmental change at Gona, Ethiopia, Earth and Planetary Science Letters 6970: 1-18. Liotta, M., Favara, R., Valenza, M., 2004. The role of orographic clouds in changing the original isotopic mark of precipitation: the case of north-western Sicily. International workshop on the Application of Isotope Techniques in Hydrological and Environmental studies, September 6-8, UNESCO, Paris, France. Mamo, S., and Yokota, S., 1998. Estimation of Koka reservoir leakage by hydrogeological and isotope techniques. In: Proceedings of Eighth International Congress International Association for Engineering Geology and the Environment. 21-25 Sept. 1998., Vancouver Canada. McKenzie,M.J., Siegel, D., Patterson, W., McKenzie, D. J., 2001. A geochemical survey of spring water from the main Ethiopian rift valley, southern Ethiopia: implication for well head protection. Hydrogeology Journal, 9: 265-272. Nicholson, S.E., 1996. A review of climate dynamics and climate variability in eastern Africa. In: The Liminology, Climatology and paleoclimatology of the East African Lakes (T.C. Johnson and E. Odada, eds), 2556, Gordon and Breach, Amsterdam, The Netherlands. Njitchoua R., Sigha-Nkamdjou, L., Devera, L. Marlin, C., Sighomnou, D., Nia, P. , 1999. Variations of the stable isotopic compositions of rainfall events from the Cameroon rain forest, Central Africa. Journal of Hydrology, 223: 17–26 Nyssen,J., Poesen, J., Moeyersons, J., Deckers, J.,Haile, M., and Lang, A., 2004. Human impact on the environment in the Ethiopian and Eritrean highlands—a state of the art. Earth sciences reviews, 64: 273-320. Okeyo, A.E., 1992. A two dimensional model of the mesoscale climate over Kenya. Rev. African meteorol. Soc., 1: 20-44. Rietti-Shati, M., Yam, R., Karlen, W., and Shemesh, A., 2000. Stable isotope composition of tropical highaltitude fresh-waters on Mt. Kenya, Equatorial East Africa. Chemical Geology 166: 341–350. Rozanski, K., Araguas-Araguas L., and Gonfiantini R., 1996. Isotope Patterns of Precipitation in the East African Region, In: The Liminology, Climatoloy,and Paleoclimatology of the East African Lakes, Johnson, T. C. and E. Odada (eds.), Gordon and Breach, Toronto, pp79-93. Rozanski, K, Araguás-Araguás, L., and Gonfiantini, R., 1993, Isotopic patterns in modern global precipitation, in Climate change in Continental Isotopic Records, Geophysical Monograph 78, American Geophysical Union, 136. Savenije, H.H.G., 1995. Does moisture feedback affect rainfall significantly? Physics and Chemistry of the Earth,20:507-513 Schoell, M., and Faber, E., 1976. Survey of Isotopic composition of Waters from NE Africa. Geol. Jb., 17: 197213. Sonntag, E., Klitzsch, E., Lohnert, E.P., El-Shazly, E.M., Munnich, K.O.., Junghans, Ch., Thorweihe, U., Weistroffer,K., and Swailem, F.M., 1979. Palaeoclimatic information from deuterium and oxygen-18 in carbon14 dated north Saharian groundwaters, in: Isotope Hydrology 1978, Vol.II, International Atomic Energy Agency, Vienna. Suzuki, H., 1967. Some aspects of Ethiopian climates. Ethiopian Geographical Journal 5: 19– 22. Telford, R.J., 1998. The palaeoenvironmental record of Holocene environmental change in the Ethiopian Rift Valley. Unpublished Ph.D. Thesis, University of Wales. Tessema, Z., 2003. Hydrogeology of Awasa Area. Proceedings of the International Symposium on Isotope Hydrology and Integrated Water Resources Management, 19 to 23 May 2003, Vienna, Austria. Travi, Y., and Chernet, T., 1998. Fluoride contamination in the lakes region of the Ethiopian Rift. In: Applications of isotope techniques to investigate Groundwater Pollution. IAEA-TECDOC-1046, Vienna, Austria. Taupin, J.D., Coudrain-Ribsein, A., Gallaire, R., Zuppi, G.M., and Filly, A., 2000. Rainfall characteristics (δ18O, δD , ∆H and ∆Hr) in Western Africa, regional scale and influence of irrigated areas. J. Geophysical Research, 105: 11911-11924. Taupin, J-D., Gaultier, G., Favreau, G., Leduc, C., Marlin, C., 2002. Variabilité isotopiques des précipitations sahéliennes à échelles du temps à Niamey (Niger) entre 1992 et 1999: implication climatique. C.R.Geoscience, 334: 43-50. UNDP, 1973. Geology, geochemistry and hydrology of hot springs of the East African Rift system within Ethiopia., UNDP report DD/SF/ON-11, N.Y. 42 White, J. and Gedzelman, S., 1994. The isotopic composition of atmospheric water vapor and the concurrent meteorological conditions. Journal of Gephysical Research, 89: 4937-4939. Zhang, X., Liu, L., and Tian, L., 2004. Variation of δ18O in Precipitation along vapor transport paths. Advances in Atmospheric Sciences, 21: 562-572. Zuber, A., 1986. Mathematical models for the interpretation of environmental radioisotopes in groundwater systems. In: P.Fritz and J.Ch. Fontes (eds). Handbook of environmental isotope geochemistry, volume 2, terrestrial environment, p 1-59. 43 PARTIE II Mécanismes de la recharge des eaux souterraines dans le Rift Ethiopien Principal, l’Afar et le Plateau Nord Ouest 44 Introduction Cette partie a pour objet de caractériser la composition isotopique des eaux souterraines du Plateau Nord Ouest, du Rift Ethiopien Principal et de la dépression de l’Afar; elle montre le mécanisme de la recharge et essaye de montrer les mécanismes du transfert des eaux souterraines d’une région à l’autre. Les principaux résultats sont présentés dans le manuscrit qui suit cette partie introductive. La figure 1 présente les compositions isotopiques des eaux souterraines de trois secteurs. On peut en tirer les principales observations suivantes: • La majorité des eaux souterraines thermales dans l’Afar et à Djibouti se caractérisent par un excès en deutérium (d) faible (<10). Un examen plus attentif montre que généralement (pas toujours cependant) il décroît lorsqu’on se déplace vers l’escarpement qui borde la plaine de l’Afar à l’Ouest. • Les eaux souterraines du Plateau Nord Ouest possèdent l’excès en deutérium le plus fort et les valeurs en δ18O relativement les plus appauvries. • Les eaux souterraines du Rift Ethiopien Principal montrent des valeurs de d intermédiaires. • Les eaux souterraines thermales du Plateau Nord Ouest possèdent un fort excès en deutérium et sont les plus appauvries en isotopes lourds. • Les concentrations originelles en δ18O des eaux géothermales de la dépression de l’Afar vont de -4.5 à -6 ‰ pour δ18O. L’estimation est faite à partir de l’intersection de la droite des teneurs en δ18O des eaux thermales et de la droite Météorique Mondiale. • Les eaux souterraines géothermales de l’Afar fortement appauvries et celles du Plateau Nord Ouest montre généralement des valeurs similaires de δD mais des δ18O différents. 60 50 GMWL 40 MER Lakes thermal waters in NWP 30 Lake Groundwater mixing line δD 20 10 0 Geothermal oxygen shift in MER thermal waters -10 -20 Geothermal waters in Afar and Dijibouti -30 -40 -7 -5 -3 -1 1 3 5 7 9 18 δ O NWP MER AFAR Figure 1. Diagramme δ18O, δD pour les eaux souterraines du Plateau Nord Ouest, du Rift Ethiopien Principal et de la dépression de l’Afar vs la DMM. Les eaux souterraines de l’Afar se caractérisent par un excès en deutérium faible, celles du Plateau Nord Ouest l’excès en deutérium le plus fort. Les eaux du Rift Ethiopien Principal se situent entre les deux. Les eaux fortement enrichies des trois secteurs sont issues de mélange avec des eaux lacustres ou de l’infiltration actuelle d’eaux préalablement évaporées. Dans le Rift Ethiopien Principal, la majorité des eaux enrichies par évaporation se situe autour des principaux lacs. 45 • La majorité des échantillons d’eau (en particulier les eaux souterraines du Rift Ethiopien Principal qui ne sont pas influencées par un mélange avec de l’eau des lacs et la déviation géothermique) se situent dans l’intervalle -1 à -2‰ pour δ18O et -10 à 0‰ pour δD. Si la détermination de l’origine de chaque échantillon est nécessaire, il est recommandé d’effectuer les interprétations à partir des conditions spécifiques du site. On peut cependant, sur la base des observations proposer les généralisations suivantes : • Sur le Plateau Nord Ouest, l’évaporation avant la recharge est faible et cette dernière intervient surtout à partir des pluies d’été (voir dans la première partie les compositions isotopiques des pluies). • Dans la dépression de l’Afar, le signal isotopique originel en oxygène est modifié soit par le fractionnement du à l’évaporation préalable à la recharge, soit pour les eaux thermales à la déviation géothermique. L’effet de l’évaporation directe sur les eaux de l’aquifère est aussi possible dans les puits peu profonds de l’Afar. D’une manière générale les eaux souterraines de l’Afar se placent sous la Droite Météorique d’Addis Abeba et quelques unes d’entre elles sous la droite Météorique Mondiale tout en gardant une pente 8. La Figure 2 indique comment une évaporation partielle avant la recharge et un petit effet géothermique peuvent produire ce type d’eaux souterraines. • Le mécanisme de recharge qui induit pour les eaux souterraines de l’Afar une pente 8 mais les place sous la droite Météorique peut être utilisé comme témoin actuel d’un mécanisme paléohydrologique. Les eaux souterraines fossiles dans les régions arides se placent sous la Droite Météorique Mondiale ou sous la Droite météorique locale actuelle tout en conservant une pente 8. • Comparées à celles de l’Afar, la majorité des eaux souterraines du Rift Ethiopien Principal sont moins affectées par le fractionnement du à l’évaporation. Ceci implique que l’évaporation est relativement moins importante en raison d’une température plus basse que dans l’Afar. De plus, l’apport fréquent dans le Rift des eaux de type « Plateau » maintient un excès en deutérium plus élevé dans les eaux. • Le calcul de la composition isotopique originelle des eaux thermales les plus appauvries de l’Afar et de Djibouti donne des valeurs de δ18O entre -4.5 et -6‰ (intersection avec la Droite Météorique Locale, non indiqué pour éviter une surcharge sur la figure 1). Cette composition est rare pour les eaux souterraines froides du Plateau Nord Ouest. Ceci pourrait impliquer que dans certains compartiments de la dépression de l’Afar, certaines eaux géothermales seraient plus anciennes et rechargées sous des conditions climatiques plus froide qu’actuellement. • La similarité de la composition en δD des eaux thermales de l’Afar et du Plateau Nord Ouest (ou la similarité entre le δ18O des eaux de la recharge des eaux thermales de l’Afar et du Plateau Nord Ouest) peut indiquer que ces eaux ont été rechargées sous des conditions climatiques semblables. Si on recalcule la température de recharge de ces eaux en utilisant l’effet isotopique de température actuel (part I, section 3.5) on obtient environ 9°C . C’est environ 3°c plus froid que le minimum moyen de température de la région la plus froide d’Ethiopie. Mais, il faut rester prudent quant à l’utilisation de cette valeur en attendant d’avoir des datations sur les eaux fortement appauvries. De plus, l’estimation de la température de recharge prend en considération seulement l’effet de température. L’effet de masse et l’origine de la vapeur peuvent s’ajouter pour produire des valeurs faibles de δ18O. Une seule datation sur une source thermale qui peut représenter le système appauvri (source Abé près de la frontière Djibouti-Ethiopie) a donné un âge C-14 de 1200 ans (Fontes et al., 1980). • Le large éventail de valeurs des concentrations en δ18O (-5 à +1‰) dans la dépression de l’Afar montre que la recharge est issue d’origines variées, de différentes époques et différentes altitudes, et que les 46 eaux souterraines sont compartimentées. La recharge depuis les eaux de crues et les « Wadis » drainant l’escarpement à l’Ouest, qui ont subies un fractionnement du à l’évaporation, semblent être la principale source de recharge des eaux souterraines de l’Afar. Ceci pourrait aussi suggérer que le Plateau Nord Ouest bordant l’Afar n’est pas l’unique zone de recharge. L’influence des eaux de type « plateau » (excès en deutérium fort) dans la dépression de l’Afar décroît lorsqu’on s’éloigne de l’escarpement vers la plaine de l’Afar. Les aquifères alluviaux près du pied de l’escarpement semblent plus influencés par les eaux de type « plateau » que les aquifères et les eaux géothermales de la plaine de l’Afar. • • Les eaux souterraines thermales du Plateau Nord Ouest se sont rechargées sous un climat plus froid que l’actuel. Les recharges en haute altitude ne peuvent pas à elles seules expliquer les valeurs observées. En prenant l’effet d’altitude actuel sur le Plateau Nord Ouest(-0.1‰/100m ) il faudrait une altitude de 5000 m, qui n’existe pas dans la région, pour obtenir l’appauvrissement de ces eaux thermales. Le mécanisme de transfert des eaux souterraines du Plateau Nord Ouest vers la dépression de l’Afar ne correspond pas à une configuration simple d’une zone de recharge vers un exutoire, mais il semble que beaucoup de chenaux d’écoulement se déversant en différents points et de nombreuses sources de recharge (pluie locale, écoulements de crue, eaux ayant subi l’effet géothermique) existent dans les systèmes hydrogéologiques de la plaine de l’Afar. MWL Eaux météoriques actuelles disponibles pour la recharge. Eaux souterraines ayant subies fractionnement du à l’évaporation Eaux météoriques plus anciennes disponibles pour la recharge Figure 2. Diagramme schématique (‘modèle d’évaporation’) montrant l’origine des eaux souterraines froides et thermales de l’AfarDjibouti. Les eaux de ruissellement actuelles qui apparaissent près de l’escarpement faisant face à l’Afar subissent une évaporation incomplète (ou concentrées par évaporation) lors de leur transfert vers la plaine et leur infiltration vers les eaux souterraines. Ces eaux qui possèdent une composition isotopique variable ( du fait de leurs origines spatio-temporelles différentes) subissent un fractionnement et se déplacent sous la Droite Météorique locale suivant les flèches parallèles. Une modification semblable de la composition isotopique originelle intervient sur les eaux rechargées sous des conditions climatiques plus froides (cercles les plus bas) au cours de la recharge. Au final, cela conduit à des eaux souterraines qui se situent sous la Droite Météorique Locale ou la Droite Mondiale suivant une pente 8. La longueur des flèches dépend du degré d’évaporation avant la recharge. Avec un tel mécanisme, calculer l’altitude de recharge avec seulement l’oxygène-18 peut ne pas être toujours dénué de sens. La déviation géothermale de l’oxygène dans des portions d’aquifère peut produire un effet isotopique similaire dans l’Afar et Djibouti- la seule différence étant que les flèches pourraient suivre une direction horizontale. Les eaux souterraines non modifiées qui suivent l’écoulement régional depuis le plateau se surimposent à ces processus (ce qui semble peu important dans l’Afar). Remarque générale Une distinction existe entre le mécanisme de recharge dans l’Afar, le Rift Ethiopien Principal et l’Afar. L’importance du fractionnement du à l’évaporation préalable à la recharge augmente en allant du plateau Nord Ouest vers le Rift Principal et l’Afar. L’influence des eaux type « Plateau » est plus fréquente dans le Rift Ethiopien principal que dans l’Afar. La présence d’eaux souterraines anciennes dans les systèmes d’eaux souterraines thermales de l’Afar ne peut être exclue. Le mécanisme de recharge observé dans l’Afar implique que l’oxygène seul ne peut pas toujours être utilisé de manière fiable pour déterminer les altitudes de recharge. References Fontes J-Ch., Florkowski T, Saliege JF, Zuppi GM. 1980. Environmental Isotope study of groundwater systems in the Republic of Djibouti. In Arid Zone Hydrology IAEA; 237-262. 47 Submitted to Hydrological processes Journal Tracing Sources of Recharge to Groundwaters in the Ethiopian Rift and Bordering Plateau: Isotopic Evidence. Seifu Kebede1,* , Yves Travi1, Tamiru Alemayehu2, Tenalem Ayenew2, Pradeep Aggarwal3 1 Laboratory of Hydrogeology, University of Avignon, 33 Rue Louis Pasteur, 84000, Avignon, France Department of Geology and Geophysics, Addis Ababa University, POBox: 1176, Addis Ababa, Ethiopia 3 International Atomic Energy Agency, Isotope Hydrology Section, Wagramer Strasse 5, POBox 100, Austria. 2 Abstract Characterization of isotopic composition Ethiopian meteoric waters were made and spatial isotope variation in meteoric waters of northern half of Ethiopia were analyzed and then used in tracing origin of groundwater in the Ethiopian rift and the bordering North Western Ethiopian Plateau. The δ18O composition of the Ethiopian rainfall is one of the most enriched waters compared to rainfalls of other IAEA-GNIP stations. Seasonal variation in isotopes of meteoric waters of the region is mainly controlled by differences in sources of moisture owing to the seasonal drifting of the ITCZ and local factors such as amount and temperature effects. A weak geographic variation in δ18O observed in the region. However a strong spatial variation in d-excess exists. Environmental isotopes are found to be a good tool in obtaining a reasonable first hand information on groundwater dynamics and surface water groundwater relations. The fact that the North Western Plateau groundwater aquifers are less important in recharging the Afar depression groundwater aquifers is demonstrated. KEY WORDS: Ethiopia, environmental isotopes, isotope geography variation, groundwater recharge, surface water groundwater relation, geothermal waters * Correspondence to: S. Kebede, email: [email protected] 48 1. INTRODUCTION A lack of adequate rainfall and limited availability of perennial fresh surface water bodies in the Ethiopian Rift, cause the local population to rely mainly upon groundwater as the major source of water supply. Numerous lakes occur in the region, but cannot be used for direct water supply due to their high salinity. In order to develop the groundwater resources for sustainable development in the Rift, improved understanding of groundwater recharge, relationships between surface and groundwater, and groundwater flow is necessary. Environmental isotopes have proven to be an important tool for such investigations. The study region is divided into three geological and climatic domains: The Main Ethiopian Rift (MER) - the narrow graben to the south that contain the Rift lakes, the Afar depression - the triangular low laying plain in the northeast portion of Ethiopia, and the northwest plateau (NWP) (figure 1). The MER and the Afar depression are part of the Great Ethiopian Rift. Extensive lakes occupied the MER and the Afar during the Quaternary period and the lakes later shrink to their present day size. The MER and the Afar are characterized by several internal drainage basins. High heat flow and geothermal activities characterize the rift. Figure 1. Location map of Ethiopia and the study region. The region that contains the majority of the lakes is the MER; the triangular depression that is drained by the Awash River is the Afar rift. The lakes from North to south in the rift are: Abbé (end of Awash River), Koka, Ziway, Langano, Abijata, Shalla, Awassa, Abaya, Chamo, and Turkana. 49 Rainfall in Ethiopia is controlled primarily by monsoon winds either from the Atlantic or from the Indian oceans depending on the season (Griffiths, 1972; Lacaux et al., 1992; figure 2). Seventy five percent of annual rainfall on the NWP and in the MER occurs during the summer (June-September) when the Inter-Tropical Convergence Zone (ITCZ) is located north of Ethiopia (Nicholson, 1996). During this period, the Indian and the Atlantic Ocean monsoons converge in the ITCZ. The Atlantic monsoon is also believed to rely on moisture advection from the tropical vegetation basin of Congo through the southwesterly low-level moisture flow (Kebede, 1964; Shinoda, 1986). The other 25% of rainfall occurs in Spring (March-April) when the ITCZ passes southwards. During this period, the north Indian Ocean is believed to be the main source of moisture. During the rest of the year (OctoberFebruary), the climate is characterized by prevailing dry and relatively cold weather. The mean annual air temperature of the NWP is 16°C compared to 35°C in the Afar. The annual rainfall in NWP ranges between 1000mm and 2000mm. In Afar rainfall amount is less than 300 mm/year. Rainfall amount in the MER is intermediate between the NWP and the Afar. Figure 2. Seasonal drifting of the ITCZ over Africa and its control on the rainfall regime of Ethiopia, modified from Lacaux et al.,1992. This paper presents δ18O, δD and some 3H compositions of rains, lakes and rivers, groundwaters and geothermal waters from previous studies and from our own recent work. The data from previous works which have focused on local and specific problems were used in the contexts understanding the controls of spatial variation in isotopic composition of meteoric waters. This knowledge is later used in tracing regional groundwater flow. The main objectives of this paper are therefore ; a) to characterize the isotopic composition of meteoric waters and their spatial and temporal variability so as to define conditions of groundwater tracing in the study sector and; b) to gain an insight into 50 groundwater recharge and dynamics in the rift region and the NWP . The second objective mainly focuses on groundwater links between the Afar, the MER and the bordering NWP. This will provide useful information for groundwater resources management and for paleo-climate studies. 2. ISOTOPIC CHARACTERISTICS OF THE ETHIOPIAN METEORIC WATERS 2.1. The isotope data As the MER and the Afar contain numerous lakes and high geothermal flux, they have been the subjects of paleo-hydrological, paleo-climatological and geothermal studies since the second half of 20th century. These studies have produced hydrogeochemical and environmental isotope data. Recently the International Atomic Energy Agency (IAEA) through its Technical Cooperation (TC) projects is conducting isotope hydrological studies. No previous stable isotope data has been apparently available from the NWP until we recently gathered and analyzed over 150 samples for δ18O and δD. Some of the previous works that were used in compiling the isotope data include IAEA-TC-ETH006 project (2002) Kebede et al. (2002a) Kebede et al. (2002b) McKenzie et al.( 2001) Beyene (2000) Ali (1999) Travi and Chernet (1998) Chernet (1998) Ayenew (1998) Rozanski et al. (1996) Darling (1996) Darling et al. (1996) Fontes et al. (1980) Craig et al.( 1977) Schoell and Faber (1976) Gonfiantini et al. (1973) UNDP (1973) etc. The isotopic composition of a total of 276 groundwater wells, 38 rivers, 67 cold springs, 74 lakes, 133 geothermal springs and a 35 years long intermittently measured IAEAGNIP rainfall isotope data from Addis Ababa station were compiled and used in this work. 2.2. Isotopic characteristics of Ethiopian rainfall waters A long-term (1965 to present) δ18O, δD and Tritium record in rainfall is available at Addis Ababa- an IAEA/WMO station within the global network of isotopes in precipitation (GNIP) - at an altitude of 2360masl in the NWP. A two years isotope record of rainfall is available at 9 stations in the MER. These include a monthly rainfall isotope samples from four stations in lake Awasa catchment between 1999 and 2000 (Tessema, 2003; IAEA-TC project, 2002) and a weekly samples from five stations in the Ziway-Shall lake basin between 1994 and 1995 (Chernet, 1998). No major conclusion can be made about the isotopic composition of rainfall in the Afar as any long-term data is missing. One of the important observations in the Addis Ababa rainfall in particular and the Ethiopian meteoric waters in general is their δ18O isotopic enrichment. The δ18O and δD compositions of Addis Ababa rainfalls are enriched compared to those of Eastern and Sahelian regions of Africa. Due to the high altitude, the furthest distance from moisture sources (oceans) and the relatively low average air temperature, Addis Ababa should normally be expected to have depleted isotopic values compared to nearby regions of Africa. Several hypotheses have been forwarded to explain this phenomenon. The different factors which enrich the Ethiopian meteoric waters include the contribution of already enriched moisture from transpired moisture of the vegetated Congo basin (Rozanski et al., 1996), the 51 contribution of Indian Ocean moisture which represents the first stage of rainfall condensation in the region (Joseph et al., 1992) and local moisture recycling. A satisfactory explanation for the climatic factors responsible for the enriched isotopic composition of Addis Ababa rains is yet to be found. Figure 3 shows a seasonal variation in δ18O, d-excess and rainfall amounts of the Addis Ababa IAEAGNIP station and rainfall stations in the MER. The δ18O and d-excess of these rainfalls is characterized by a notable seasonal variation. The summer rainfalls are relatively depleted (with a weighted average δ18O of -2.5‰) and are characterized by higher d-excess (d-excess>14); the spring rainfalls are enriched (with a weighted average δ18O = +1.5‰) and have lowest d-excess (d-excess<12). The difference in the δ18O and δD signals of the rainfall of the two seasons provides a suitable opportunity to trace and study sources of groundwater recharge. The d-excess in the Addis Ababa rainfall increases continuously from its lowest value in March to its maximum in September. Although the length of record is short in the MER, the general pattern in δ18O seasonal variation and the range of δ18O values in the MER rainfalls is similar to that of the Addis Ababa station (figure 3). In all stations, the months of July, August and October are characterized by the most depleted δ18O composition. 0 20 50 15 150 10 200 5 250 Rainfall in mm δ18 Ο and d-excess 100 300 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 400 -5 Rainfall in mm 350 Awasa Sodo Agermariam Kofele Addis Ababa d-excess Figure 3. Monthly variation in mean δ18O and d-excess of the Addis Ababa and the MER rainfalls. Awassa, Sodo, Agermariam and Kofele are the four stations in the Awassa basin. The seasonal variation in δ18O composition and d-excess of the Addis Ababa rainwaters seems to be controlled by the seasonal drifting of the ITCZ and associated difference in moisture sources. The region gets about 25% of rainfall between March and May. During this time the ITCZ is located south of Ethiopia which leads to derivation of moisture to the region from the North Indian Ocean. The 52 enrichment of δ18O during this time may be related to the fact that these rainfalls represent the first condensation stage of oceanic moisture (Joseph et al., 1992). In addition, as the atmosphere is dry and humidity is low during this period re-evaporation of raindrops may lead to enrichment in δ18O. During summer (J-J-A-S) the ITCZ is located north of Ethiopia. During this time moisture is brought into the region from the mixture of the two monsoons. The region receives 75% of rainfall during this season. The relatively depleted δ18O in summer rainfalls compared to the spring rainfalls is due to ‘the amount effect’ and due to the low sea surface temperature characteristics of the Atlantic Ocean. Depleted δ18O in October may be related to temperature effect. This is because the lowest temperature prevails in October when northeastern dry and cold wind penetrates into the region occasionally bringing moisture from the Red Sea. 2.3. Spatial variation in isotopic compositions of meteoric waters: evidence from groundwaters Geographic variation: Using isotopic composition of groundwaters as a basis of comparison, figures 4 and 5 compare the isotopic composition of meteoric waters from the MER, the Afar Depression and the NWP. Generally groundwaters of the North Western Ethiopian plateau are characterized by high deuterium excess. Most of the groundwaters in the NWP plot above GMWL. The Afar groundwaters are characterized by relatively enriched δ18O and low d-excess. The MER groundwaters have an intermediate nature. This geographical variation in isotopic compositions of meteoric waters is related to the variation in air temperature, rainfall condition and the degree of importance of evaporative enrichment prior to recharge in the three sectors. The most depleted groudwaters are located in the mountainous regions of the NWP. The difference in δ18O compositions and d-excess of meteoric waters was used as a basis for tracing regional groundwater link among the three sectors. 53 Figure 4. δ18O (above) and d-excess (below)compositions of Ethiopian meteoric waters and their geographic variations. Regardless of the drastic climatic difference in the region, there is only a weak latitudinal/longitudinal change in δ18O composition of groundwaters (figures 4). However it can be generalized that the NWP meteoric waters are characterized by higher deuterium excess (13 to 19). The majority of the Afar groundwaters plot below the GMWL maintaining a slope of 8. Their deuterium excess varies between 0 and 10. The MER groundwaters plot both above and below the GMWL showing intermediate property compared to the NWP and the Afar groundwaters. While a very wide range of δ18O is observed in NWP groundwaters (δ18O varies between -6 and 0‰), the range of δ18O composition in those groundwaters of the MER that are unaffected by evaporated water mixing is somewhat homogeneous (δ18O varies between -4.3 and -2‰). The majority of groundwaters of the NWP plot towards the depleted end of the δ18O composition of the Addis Ababa summer rainfalls. In the Afar some exceptionally depleted waters (δ18O<-4‰) and few groundwaters which plot above the GMWL are observed. 54 40 30 20 δD 10 0 -10 NWP thermal groundwaters -20 -30 Afar thermal groundwaters -40 -7 -6 -5 -4 -3 -2 -1 0 δ 18 O NWP MER AFAR Figure 5. Isotope plot of cold and thermal groundwaters from the NWP, the MER and the Afar depression. Figure 5 shows the existence of a clear distinction in isotopic composition of the geothermal waters of the NWP and the Afar depression. The NWP geothermal waters exclusively plot above the GMWL with high d-excess (>15) and highly depleted δ18O (<-4.5‰). Compared to the cold groundwaters of the same region, the Afar geothermal waters are characterized by a wide range of δ18O composition. The majority of them plot below the GMWL. Compared to the shallow groundwaters and cold springs of the same region, the Afar thermal waters are characterized by depleted isotopic composition. The MER geothermal waters plot both above and below the GMWL showing an intermediate characteristic compared to the NWP and the Afar geothermal systems. Isotope altitude relations: On the windward face of a mountain, the δ18O and δD compositions of rainfalls decrease with increasing altitude. This phenomenon is termed as the 'altitude effect'. Isotopealtitude relation is one of the strong approaches in groundwater tracing particularly in identifying recharge altitude. Generally, the depletion of δ18O in precipitation with elevation ranges from about 0.1 to 0.5‰/100 meters and is strongly correlated with ambient surface air temperature (Gat and Rietti-Shati, 1999). Despite the marked variation in temperature, rainfall amount, humidity and hydrology with elevation, the 'altitude effect' is not a pronounced phenomenon in the rift valley and bordering escarpment. A weak and sometimes reversed isotope-altitude relation has been obtained from the rainfalls collected over the two rainy season along an altitudinal transect in the central section of MER in the ZiwayShalla lake basin. 55 Groundwaters from the western sector of the NWP shows δ18O enrichment compared to the δ18O composition of groundwaters of the eastern mountainous sector of the NWP. Generally a 0.1‰/100 meters depletion in δ18O is observed in the NWP (figure 4). In the Afar depression and bordering western escarpment also a 0.14‰/100 meters depletion in δ18O of groundwaters is observed (figure 4). Two different processes cause the altitudinal variations in δ18O of meteoric waters of the two sectors. In the NWP the altitudinal variation is caused by a continuous depletion of δ18O as the Atlantic moisture mass is forced up towards the eastern mountainous region of the NWP by orographic lifting. Under the present day condition the Atlantic and Indian Oceans air masses that are drawn to the ITCZ descend adiabatically 2 or more km of elevation into the Afar depression and the MER where they become hot and dry (Griffiths, 1972). Thus, the 0.14 ‰/100 meters depletion in δ18O observed in the MER and the Afar depression region is caused by evaporation of meteoric waters in the arid leeward side of the mountains bordering the rift. This phenomenon is often called 'the pseudo-altitude effect'. 2.4. Isotopic composition of evaporated waters and their spatial variations Surface waters from the region are enriched in their δ18O and δD compositions (figure 6). These waters plot below the GMWL defining a line called the Local Evaporation Line (LEL). This isotopic enrichment provides an opportunity to trace lake-groundwater relations. The slope of LEL varies depending on the humidity and isotopic composition of ambient vapor. No major distinction can be made between LELs of the NWP and the MER lakes. Evaporated waters from the MER and NWP plot on a LEL defined by δD=5.7 δ18O+6 while the Afar lakes plot on a line defined by a LEL of δD= 3.5δ18O+7.5. The difference of the Afar lakes isotopic composition from the NWP and the MER lakes is most likely related to differences in the environmental parameters, or difference in their water budgets (mainly water residence time in lakes) or difference in their salinity. All of the rivers on the NWP (except those that drain the lakes) plots close to the meteoric water line with slope 8 ruling out the effect of strong en route evaporation. 56 100 80 +6 18 O ,7 δ 5 δD= + 7,5 ,5δ18 O δD = 3 60 δD 40 20 0 -6 -4 -2 -20 0 2 4 6 8 10 12 -40 18 δ O NWP MER Afar GMWL LEL, MER & NWP LEL, Afar Figure 6. Isotope plot of evaporated waters (rivers and lakes from the three sectors). 3. APPLICATION OF ISOTOPES IN GROUNDWATER STUDIES 3.1. Sources of recharge and groundwater dynamics The existence of seasonal variation in δ18O of rainfalls, the differences in the d-excess of meteoric waters of the three sectors, and the enrichment of lakes in δ18O and δD provide a good opportunity to understand the sources of groundwater recharge, to trace lake-groundwater interaction, and to trace regional groundwater flows. The wide range of δ18O composition in the NWP groundwaters (figure 5) and their low salinity is the result of lack of appreciable mixing between the different pockets of aquifers. The groundwater in this sector is characterized by shallow and rapid circulation. The lack of isotopic enrichment in groundwaters of the NWP rules out the importance of evaporative concentration prior to recharge. This may be due to fast recharge through the fractures of the basaltic plateau. The existence of appreciable amount of tritium in these waters testifies that the waters are recharged during recent times. This shallow circulation and the young age of the groundwater system on the NWP may be the cause of the frequent drying of groundwater wells following prolonged drought. Groundwaters from the NWP plot towards the depleted end of the Addis Ababa summer rainfall isotopic composition. This testifies that the spring rainfall is unimportant in recharging the groundwaters of the NWP. The depletion of groundwaters of the NWP in their δ18O compared to the weighted average δ18O of the summer rainfalls shows that selective recharge by heavy rainfalls of July and August is the principal source of groundwater replenishment. For the Afar groundwaters, three sources of recharge can be distinguished. Those groundwaters which plot above the GWML are most likely recharged by high intensity local rainfalls. Those groundwaters 57 which plot below the GMWL are most likely recharged by pre-evaporated meteoric waters (mainly evaporated flood waters). Those groundwaters which show exceptional δ18O depletion and which plot below the GMWL are most likely paleowaters. The clear distinction between the d-excess value of groundwaters of the Afar and that of the meteoric groundwaters of the NWP (figure 5) testifies that there is a poor subsurface hydraulic link between the two sectors. The effect of plateau type groundwaters on the aquifers of the Afar depression decreases as one move from the escarpment down to the Afar plain. In the MER however all sort of recharge is possible. These include hydraulic connection of aquifers with the lakes, recharge from rift rainfall and the recharge on the high lands bordering the rift. The relative importance of these different sources of recharge depends on the local geological and topographic conditions. 3.2. Lake groundwater relation One of the clear applications of stable isotopes of water in the region is in unraveling lakegroundwater relations. Hydraulic link between the lakes and adjacent groundwaters and the mixing ratio between them can be discerned from δ18O-δD plot. The presence of a clear hydraulic link between lakes and adjacent groundwaters has been observed around Awasa lake (figure 7a), the Debrezeyit Crater Lakes (figure 7b; Kebede et al., 2002a), the Ziway-Shalla lakes (figure 7c; Ayenew, 1998; Chernet, 1998; Craig, 1977) and the lake Afrera (figure 7d). These testify the importance that the MER and the NWP lakes play in recharging the groundwaters adjacent to them. a c 80 80 60 60 D 40 40 D 20 20 0 -6 -4 -2 -20 0 2 4 6 8 10 -6 -4 0 -20 0 -2 -40 2 4 6 8 10 -40 18O 18O d b 100 100 80 80 60 60 D 40 D 40 20 20 0 -6 -4 -2 -20 0 2 4 6 8 10 -6 -4 -2 0 -20 0 2 4 6 8 10 -40 -40 180 18O 58 Figure 7. Isotope plot to show lake groundwater relations, a-lake Awasa; b-Bishoftu Crater Lakes; c-the Ziway Shalla lakes; and d-the Afrera salt lake. Groundwaters(open circles) from the up-gradient areas plot on the meteoric water line while the groundwaters from the down-gradient zone plot on the lake water- groundwater mixing line testifying hydraulic link between the lakes(filled circles) and adjacent groundwater aquifers. 3.3. The problem of paleo-waters Obtaining the groundwater residence time is important from the point of view of water resources management. In the Afar and the MER little information is available regarding the age of groundwaters. The presence of paleowaters in the region is barely known. This is on one hand due to the lack of measurements of 14C activity and on the other hand due to complication of the 14C dating due to degassing of unknown quantity of dead carbon from the mantle. Various workers have conducted tritium measurement on the thermal and cold groundwaters. Tritium measurement indicates that surface waters and lakes contain appreciable amount of modern waters (figure 8). The majority of cold groundwaters from the three sectors of the region contain modern meteoric waters. With exception of thermal springs in the MER (Beyene, 2000), the majority of thermal springs and thermal wells contain no appreciable tritium (figure 8). Lakes contain tritium concentration similar to the modern rainfall testifying the importance of rainfall in their water budget and the importance of evaporative enrichment. 30 frequencyof 3H 25 20 a 15 10 >30 22<< =24 14<< =16 < =0 0 6<< =8 5 TU b 3 2 >30 22<< =24 14<< =16 0 6<< =8 1 < =0 frequencyof3H 4 TU 20 15 c 10 >30 22<< =24 14<< =16 0 6<< =8 5 < =0 frequencyof3H 25 TU 59 Figure 8. Frequency of tritium concentration in waters of the three sectors; a-cold groundwaters, b-lakes, and c-thermal groundwaters. Those geothermal and cold groundwaters of the Afar depression which are highly depleted (δ18O<-4‰) are most likely recharged under a wetter and colder climate. In NWP geothermal and high TDS groundwaters which show the most depleted δ18O are most likely recharged under a colder climate conditions. 4. CONCLUSIONS The Ethiopian meteoric waters show unique δ18O composition. Despite the highest altitude and low mean annual temperature of the region, the meteoric waters are enriched in δ18O compared to meteoric waters of other regions of South east and Sahelian Africa. The Ethiopian meteoric waters are characterized by a weak spatial variation in δ18O. However a remarkable variability in d-excess is observed. In the MER and the Afar depression, isotope-altitude relation is complex and weak. The relation varies seasonally and geographically. In determining sources of recharge to groundwaters both the δ18O and d-excess characteristics should be considered. In general, the NWP groundwaters are characterized by shallow circulation, short subsurface residence, short subsurface flow path, and discontinuous pockets of aquifers with little evidence of the existence of regional groundwater flow. Of the two rainfall seasons the summer rainfalls are the major source of groundwater recharge. Meteoric waters with similar composition to NWP are observed in the aquifers of southern and central sector of the MER. There is no strong evidence for the existence of a major subsurface groundwater link between the deep thermal groundwater of Afar and the present meteoric recharge of the NWP. However infiltration from evaporated flood waters and wadis draining the escarpment seems the major sources of recharge in Afar. This has a wide ranging implication for geothermal resources development of the Afar depression. Many previous studies speculated that the NWP is the principal source of recharge to the Afar thermal waters. Isotopic evidence does not strongly support this hypothesis. Lakes in the central sector of the MER play a major role in recharging adjacent aquifers. Further research on the relation between atmospheric processes and isotopic signal of the Ethiopian meteoric waters is required to fully benefit from isotope approach in groundwater tracing and in paleohydrology. Dating of the deeper groundwater/geothermal waters should also be an area of further 60 investigation. The discussions and conclusions that were made in this regional investigation should provide useful inlet for future detailed groundwater studies in the region. References Ali A. 1999. River-groundwater interactions in the Sekelo-Akaki basin, Unpublished M.Sc. thesis, Addis Ababa University; 120. Ayenew T. 1998. The Hydrogeological system of the lake district basin, central main Ethiopian Rift. Phd thesis, ITC publication Number 64, the Netherlands; 200. Beyene K. 2000. Chemical and isotopic studies of geothermal prospects, in the southern Afar region, Ethiopia. In Proceedings of the World Geothermal Congress Tohoku: Japan; 977-983 Chernet T. 1998. Etude des Mechanismes de mineralisation en fluorure et elements associes de la region des lacs du rift Ethiopien. Ph.D. thesis, Université d’Avignon, France; 200. Craig H, Lupton JE, Horowiff RM. 1977. Isotope Geochemistry and Hydrology of geothermal waters in the Ethiopian rift valley. Scripps Institute of Oceanography University of California report; 160 Darling WG. 1996. The Geochemistry of fluid processes in the eastern branch of the east African rift system, Ph.D thesis, British Geological Survey, UK; 235. Darling WG, Gizaw B, Arusei MK. 1996. Lake groundwater relationships and fluid rock interaction in the east African rift valley, isotopic evidence. Journal of African Earth Sciences 22, 423-431. Fontes JCh., Florkowiski T, Saliege JF, Zuppi GM. 1980. Environmental Isotope study of groundwater systems in the Republic of Djibouti. In Arid Zone Hydrology IAEA; 237-262. Gat JR, Rietti-Shati M. 1999. The Meteorological vs. the hydrological altitude effect on the isotopic composition of meteoric waters. In Isotope techniques in water resources management, IAEA-CSP2/C IAEA: Vienna (CDROM). Gonfiantini R, Borsi S, Ferrara G, Panichi C. 1973. Isotopic composition of waters from the Danakil depression (Ethiopia). Earth and Planetary Science Letters 18: 13-21. Griffiths JF. 1972. Climates of Africa. In World Survey of Climatology, Griffiths JF(ed). Elsevier, Amsterdam. IAEA TC project ETH006,2002. An ongoing project conducted by International Atomic Energy Agency, Unpublished expert reports, the Ethiopian Science and Technology Commission and the Ethiopian Geological Surveys: Addis Ababa. Joseph A, Frangi, P, Aranyossy, JF. 1992. Isotopic composition of Meteoric water and groundwater in the Sahelo-Sudanese Zone. Journal of Geophysical research 97: 7543-7551. Kebede S, Ayenew T, Umer M. 2002a. Application of Isotope and water balance approaches for the study of the hydrogeological regime of the Bishoftu crater lakes, Ethiopia. SINET: An Ethiopian Journal of Science 24(2), 151-166. Kebede S, Lamb H, Telford R, Leng M, Umer M. 2002b. Lake-Groundwater relationships, oxygen isotope balance and climate sensitivity of the Bishoftu Crater Lakes, Ethiopia. Advances in Global Change Research 12: 261-275. Kebede T. 1964. Rainfall in Ethiopia. Ethiopian Geographical Journal 2: 28-36. Lacaux JP, Delmas R, Kouadio G, Cros B, Andreae MO. 1992. Precipitation chemistry in the Mayombé forest of equatorial Africa: Journal of Geophysical Research 97: 6195-6206. McKenzie MJ, Siegel D, Patterson W, McKenzie DJ. 2001. A geochemical survey of spring water from the main Ethiopian rift valley, southern Ethiopia: implication for well head protection. Hydrogeology Journal 9: 265-272. Nicholson S. 1996. A review of Climate Dynamics and Climate Variability in Eastern Africa. In The Liminology, Climatoloy,and Paleoclimatology of the East african Lakes, Johnson TC, Odada E (eds). Gordon and Breach: Toronto; 25-56. Rozanski K, Araguas-Araguas L, Gonfiantini R.1996. Isotope Patterns of Precipitaion in the East African Region, In The Liminology, Climatoloy,and Paleoclimatology of the East african Lakes, Johnson TC, Odada E (eds). Gordon and Breach: Toronto; 79-93. Schoell M, Faber E. 1976. Survey of Isotopic composition of Waters from NE Africa. Geol. Jb. 17: 197-213. 61 Shinoda M. 1986. Rainfall distribution and monsoon circulation in tropical Africa n the 1979 summer: Their comparison between East and West Africa. Journal of the. Meteorological Society of Japan 64: 540-561. Travi Y, Chernet T. 1998. Fluoride contamination in the lakes region of the Ethiopian Rift. In Applications of isotope techniques to investigate Groundwater Pollution, IAEA-TECDOC-1046, Vienna, Austria. UNDP. 1973. Geology, geochemistry and hydrology of hot springs of the East African Rift system within Ethioiopia., UNDP report DD/SF/ON-11, N.Y; 300. Tessema Z. 2003. Hydrogeology of Awasa Lake Catchment. Isotopic and Hydrochemical approach. In Proceedings of the Interanational Symposium on Isotope Hydrology and Integrated Water Resources Management, IAEA-CN-104/P-73, Vienna, Austria. 62 PARTIE III Isotopes de l’Environnement et Hydrogéochimie pour l’étude des eaux souterraines du bassin du Nil Bleu 63 1. Introduction Dans cette partie on résume brièvement l’hydrologie d’ensemble du bassin du Nil Bleu, avec une description rapide de la géologie, de la topographie, du climat et de l’occupation des sols du bassin. Dans le manuscrit situé après ce texte, les résultats des investigations géochimiques et isotopiques du bassin supérieur du Nil Bleu (bassin d’Abay) sont ensuite présentés. L’étude isotopique des eaux souterraines utilise le schéma général de l’origine des signaux isotopiques dans les eaux météoriques d’Ethiopie, développé dans la partie I. Une courte note complémentaire sur l’hydrogéologie du bassin du lac Tana est fournie dans l’encadré 3 (Appendice 1). Une des questions principales pour établir une base convenable au futur développement des ressources en eau, concerne la compréhension de la dynamique des eaux souterraines. Ceci implique l’identification des aquifères, des lignes d’écoulement des eaux souterraines, des conditions de recharge, de la qualité des eaux souterraines etc. Les objectifs de cette partie de la thèse sont de définir les conditions d’écoulement, d’identifier les conditions de la recharge, et d’examiner l’évolution géochimique des eaux souterraines dans le bassin du Nil Bleu. 2. Caractéristiques du site Géologie Les roches du socle cristallin, les roches volcaniques, les roches sédimentaires consolidées ou non consolidées, constituent la base géologique du bassin. Le socle Précambrien représente les roches les plus anciennes. Cet ensemble contient une grande variété de roches sédimentaires, volcaniques et intrusives, métamorphisées à des degrés divers. Elles affleurent dans les basses plaines de la zone périphérique. Le Paléozoïque se caractérise par l’absence de dépôts et d’événement géologique majeur. Formes du relief et drainage Les formes du relief du bassin sont le reflet de sa géologie et des processus superficiels. La topographie actuelle du bassin du Nil Bleu (figure 1) est le résultat du soulèvement cénozoïque et de l’empilement de laves basaltiques épaisses, érodées et incisées ensuite par le Nil Bleu. Le soulèvement centré sur la vallée du Rift et le Plateau Nord Ouest forme le côté ouest du dôme cénozoïque. D’une manière générale, l’altitude du plateau décroît vers les basses terres de la frontière soudanaise. On observe également d’autres formes de relief remarquables, mises en place par les éruptions volcaniques. Les complexes volcaniques “Choke” au centre du bassin, “Guna” et “Debretabor” au Nord, représentent trois ensembles volcaniques visibles. Le graben du lac Tana et la zone volcanique linéamentaire de Yerer Tulu Welel constituent deux autres éléments remarquables. La forte incision des rivières et l’érosion de l’empilement de laves cénozoïques ont conduit à la formation des gorges du Nil Bleu et permettent aux roches sédimentaires mésozoïques et au socle d’affleurer. Des lambeaux isolés de plateaux, d’extension limitée, sont omniprésents dans la région. L’extension latérale de ces unités lithologiques est limitée par ces incisions profondes et les failles régionales majeures qui 64 déplacent les unités géologiques. L’érosion et le transport solide depuis le plateau ont comblé quelques grabens tectoniques majeurs fournissant un fort potentiel en eaux souterraines au Soudan (eg. le rift du Nil Bleu). Localement, l’accumulation de sédiments au pied des montagnes forme également des cônes alluviaux qui peuvent contenir de fortes réserves en eau souterraine. Quelques formations alluviales importantes se rencontrent dans le graben de Diddessa et le graben du Lac Tana. Figure 1. Topographic map of the Blue Nile basin in Ethiopia Climat et précipitation La variation saisonnière du régime atmosphérique et la topographie complexe du Plateau Nord Ouest conduisent à distinguer clairement au moins quatre zones climatiques (figure 2) dans le bassin du Nil Bleu. Ce sont: un climat tropical de mousson avec un court hiver sec dans le Sud Ouest; un climat pluvieux chaud à tempéré avec un hiver sec dans le centre; un climat tempéré à précipitation de type bi-modal mais faible à l’Est et au Nord Est ; un climat chaud semi-aride dans les zones basses de l’Est. 65 Figure 2. Rainfall regions of the Blue Nile basin in Ethiopia. Connaissance hydrogéologique antérieure Généralement, les formations sédimentaires Mésozoïques sont considérées comme de bons aquifères (Chernet, 1990). Ces sédiments affleurent seulement localement soit dans le bassin du Nil Bleu soit dans les bassins structuraux du Linéament de Yerer Tulu Welel et du Graben du Lac Tana. Du fait du travail de dissection et de fragmentation de l’érosion, la couverture basaltique est considérée comme un système aquifère perché, faiblement capacitif et d’extension limitée (BCEOM, 1999). La décomposition de l’hydrogramme montre que le coefficient d’infiltration varie de 3 à 20% (Figure 3). Le coefficient d’infiltration le plus élevé et la plus forte contribution des eaux souterraines vers les eaux de surface s’observent dans la région centrale des montagnes de Gojam autour du plateau volcanique de Choke et dans le Graben du lac Tana. Les fortes précipitations sur le bouclier volcanique et la grande extension latérale des aquifères dans cette partie du plateau favorisent le stockage et la circulation. On peut citer trois sources à fort débit : les sources de Bure Baguna, les sources à fort TDS d’Andesa, et les sources à faible TDS de Jiga, qui émergent au pied des boucliers volcaniques. Les régions situées en rive gauche, en particulier les parties est et sud du Nil Bleu présentent un faible coefficient d’infiltration et une faible capacité de stockage. Ceci est vraisemblablement du à la forte fragmentation des aquifères et aux faibles précipitations dans cette partie du bassin. Le coefficient d’infiltration relativement faible dans la partie sud du bassin est lié à la faible perméabilité des roches de socle qui occupent cette région. A partir des données sur les coefficients d’infiltration (BCEOM, 1998, figure 3) et la distribution des hauteurs de précipitation (figure 2) le volume total annuel des eaux souterraines qui circulent dans le 66 Nil bleu est grossièrement estimé à 1.4 billions de mètres cubes. La pertinence de la figure 3 dépend de la qualité des données sur le débit des rivières. Figure 3 . Map showing the infiltration coefficient of the Blue Nile Basin, river discharge data from BCEOM, 1999. 3. Résultats et Remarques générales Les connaissances géologiques couplées avec les nouvelles données ont été utilisées pour schématiser l’écoulement souterrain dans deux secteurs hydrogéologiquement importants du bassin supérieur du Nil bleu. Les eaux souterraines du bassin du Nil Bleu sont souvent de bonne qualité avec un TDS faible. La pollution liée à l’activité humaine est limitée aux puits non protégés. Malgré des formations mézozoïques similaires à celle rencontrées dans le Sahel, le soulèvement et la couverture basaltique du cénozoïque rendent les roches mézozoïques d’Ethiopie très fragmentées et leurs eaux souterraines inutilisables. En quelques endroits (linéament YTW et Graben du Lac Tana) la tectonique met en contact ces sédiments et la couverture volcanique rendant le Mézozoïque directement accessible aux circulations souterraines. Les flux de CO2 profond au niveau du linéament de Yerer Tulu Welel et du graben du Lac Tana induisent l’émergence d’eaux thermales de type Na-HCO3. Ces eaux sont le résultat d’une interaction entre une source chaude, les couches sédimentaires et la couverture volcanique, situation qui caractérise ces deux zones. Un examen plus attentif de la géologie du linéament YTW et les informations géochimiques sur l’origine de la composition des eaux souterraines montrent que les sédiments mézozoïques peuvent jouer un rôle important comme source de gaz CO2 gas non seulement au niveau du Linéament, mais aussi dans le Rift Ethiopien. Références BCEOM, 1998. Abay River Basin integrated master plan, main report, Ministry of Water Resources, Addis Ababa. Chernet, T., 1990. Hydrogeology of Ethiopia and water resources development, Ministry of Mines and Energy report, 157p. 67 In press in Journal of Applied Geochemistry Groundwater recharge, circulation and geochemical evolution in the source region of the Blue Nile River, Ethiopia Seifu Kebede1,2,*, Yves Travi1, Tamiru Alemayehu2, Tenalem Ayenew2 1 Laborarory of Hydrogeology, University of Avignon, 33 Rue Louis Pasteur, 84000, Avignon, France 2 Department of Geology and Geophysics, Addis Ababa University, POBox: 1176, Addis Ababa, Ethiopia Abstract Geochemical and environmental isotope data were used to gain the first regional picture of groundwater recharge, circulation and its hydrochemical evolution in the upper Blue Nile River basin of Ethiopia. Q-mode Statistical Cluster Analysis (HCA) was used to classify groundwaters into objective groups and to conduct inverse geochemical modeling among the groups. Two major structurally deformed regions with distinct groundwater circulation and evolution history were identified. These are the Lake Tana Graben (LTG) and the Yerer Tullu Wellel Volcanic Lineament Zone (YTVL). Silicate hydrolysis accompanied by carbon dioxide influx from deeper sources plays a major role in groundwater chemical evolution of the high TDS Na-HCO3 type thermal groundwaters of these two regions. In the basaltic plateau outside these two zones, groundwater recharge takes place rapidly through fractured basalts, groundwater flow paths are short and they are characterized by low TDS and are Ca-Mg-HCO3 type waters. Despite the high altitude (mean altitude~2500 masl) and the relatively low mean annual air temperature (18 °C) of the region compared to Sahelian Africa, there is no commensurate depletion in δ18O compositions of groundwaters of the Ethiopian Plateau. Generally the highland areas north and east of the basin are characterized by relatively depleted δ18O groundwaters. Altitudinal depletion of δ18O is 0.1‰/100 m. The meteoric waters of the Blue Nile River basin have higher d-excess compared to the meteoric waters of the Ethiopian Rift and that of its White Nile sister basin which emerges from the equatorial lakes region. The geochemically evolved groundwaters of the YTVL and LTG are relatively isotopically depleted when compared to the present day meteoric waters reflecting recharge under colder climate and at high altitude. Key words: groundwater geochemistry, stable isotopes, Q-mean Hierarchical Cluster Analysis, carbon dioxide, Blue Nile, Ethiopia. * Corresponding author: Tel.: 0033-4-90144492; Fax: 0033-4-90144489; email: [email protected] 68 1. Background Surface water of the Blue Nile River basin of Ethiopia is not widely used for water supply because of its marked seasonality and lack of proper technology to retain it. In the basin, groundwater is the most important source of water and is the dominant source for domestic supply, especially in the dry areas where surface waters are scarce (UN, 1989). Groundwater well drilling programs have been initiated over the last decades, but groundwater provision is often unsuccessful because of poor groundwater productivity of wells, difficult drilling conditions, drying of wells and springs after prolonged drought, or sometimes due to poor quality. This is hampered by lack of understanding of groundwater systems. Information on groundwater recharge, storage, circulation, and chemical evolution is barely known. Groundwater development is being conducted without a good understanding of its role in the hydrology of the basin. Contrary to the Blue Nile River basin, many important hydrogeochemical researches have been conducted in the Ethiopian Rift System. The presence of many lakes, lacustrine deposits, heat flow owing to Rifting and accompanied thinning of the crust in the East African Rift System (EARS) have attracted major geoscientific investigations since the second half of the 20th century. Many of the geochemical investigations (Craig et al., 1977; Darling, 1996; Darling et al., 1996; Gizaw, 1996; Chernet et al., 2001; Reimann et al., 2003) showed the role of water-rock interaction in influencing the water quality, salinity and fluoride composition of groundwaters and thermal systems of the EARS. In many instances the water-rock interaction is induced by volatile gases from the mantle and by the high heat flow beneath the EARS. Groundwater circulation pattern, groundwater recharge source identification and the interaction between lakes and groundwaters have also been subject of many important studies in the EARS (Schoell and Faber, 1976; Craig et al., 1977; Darling et al., 1996; Ayenew, 1998; McKenzie et al., 2001). Many of these studies show riftward groundwater flow from adjacent highlands. However little was known about the hydrogeology, hydrogeochemistry and isotopic compositions of the groundwaters of the adjacent plateaus to substantiate the hypotheses of the plateau-rift groundwater connections Therefore, understanding the hydrogeochemistry, and hydrogeology of the Blue Nile River basin has at least two fold importance. The first is directly linked to the understanding of the role of the aquifers of the relatively humid NWP (in which the Blue Nile is a part) in recharging the groundwaters and thermal waters of the arid regions of the Ethiopian Rift. The second is related to groundwater resources assessment in the Blue Nile River basin where clean water provision is still not attained. 69 This work uses geochemical and isotope hydrological approaches to provide an initial schematic geohydrological model on groundwater recharge, circulation, chemical evolution and its subsurface residence time in the poorly known hydrogeologic system of the upper Blue Nile River basin of Ethiopia. The present isotope data are the first set of data ever obtained in the Northwestern Ethiopian Plateau. The specific objectives of this work are: 1) to characterize the isotopic (δ18O, δD, δ13C, 3H) compositions of the groundwaters of the Blue Nile River basin; 2) to determine sources and mechanisms of recharge of groundwaters in the Blue Nile River basin; 3) to determine the dominant geochemical processes that influence groundwater chemical composition; and, 4) to schematize groundwater flow patterns and nature of aquifers in selected important hydrogeological regions of the basin. 2. Study site description, geology and hydrogeology The Blue Nile River basin is located in the Northwestern Ethiopian Plateau. The Main Nile River gets 70 % of its flows from the Blue Nile emerging from the Ethiopian Plateau and the remaining from the White Nile emerging from the Equatorial Lakes. About 44% of Ethiopian population lives in the Blue Nile basin (BCEOM, 1990). The basin has an elevation ranging from 500m in the western lowland to over 4000 m in the east and northeast. Spatial variation in rainfall amount is controlled by topography. Annual rainfall varies between 1000 mm in the lowland to 2000 mm in the highland. The Atlantic Ocean is the main source of rainfall in summer (June, July, August and September). The eastern mountainous region of the basin receives the summer rain and rainfall originating from Indian Ocean in April and March. 2.1. Geology The geology of the Blue Nile basin has been studied by various authors (Yemane et al.,1985; Assefa, 1991; Abate et al., 1996; Abebe et al.,1998; Chorowiz et al.,1998; Pik et al., 1998; Kebede et al., 1999; Asrat et al., 2001; Feseha, 2002). Crystalline basement rocks, volcanic rocks, and sediments make up the geology of the basin (figure 1). The oldest rocks in the region form the Precambrian basement. They are exposed in the low-lying plain in the western part of the basin. The various rocks forming the basement are broadly classified in to two petrographically and structurally distinct units (Kebede et al., 1999). These are the high grade gneisses and the 70 volcano-sedimentary green schist assemblages with associated ultramafic rocks. The Paleozoic is characterized by erosion and lack of any major rock formation. # Chilga sub graben L a ke T an T $ a # Debre Tabor Sub Graben #$ T T$ # T $ # # # # LTG # # T $ # # # # A M Ú ÊÊ Ú B O # # Ú Ê T #$ YTVL # # ÊÊ Ú Ú ÚF Ê # A ## U T T$ $ # TÊÚ $ Ú Ê TÊÚ $ Ú Ê TÊÚ $ ABL L T ÚAbaba AddisÊ Ú Ê Ú Ê YTVL ÚÊ Ê Ú Ú Ê Ú Ê DL Ú Ê 0 100KM Legend Basement Mesozoic Sediment Infiltration coefficients: 1-4% 4-6% 6-8% Trap Basalts Rhyolites & trachytes Quaternary Basalts 8-14% Artesian or thermal Na-HCO3 springs 14-20% Quaternary volcanoes Aluvium Figure 1. Simplified geological and hydrogeological map of the Blue Nile basin (Modified from Abebe et al., 1998, Chorowiz et al., 1998, and BCEOM, 1999). The western part of the LTG (Chilga sub graben) is highly faulted. Rock blocks dipping inward towards lake Tana are common around the lake. The zone bounded by the big bracket is the YTVL The prominent features of the YTVL are the Ambo fault which makes the northern boundary of the YTVL and the two NW-SE transfer zones DL (Didesa Line) and the ABL (Ambo-Butajira Line). Thermal spring sites and Quaternary volcanoes are almost exclusively located in the LTG and YTVL. The Quaternary basalts south of lake Tana have the highest infiltration coefficient and metamorphic basement have the lowest infiltration coefficient. Mesozoic sedimentary rocks are exposed in the Blue Nile gorge and the gorges of its major tributaries. The succession is about 1200 m thick. It includes from the bottom to the top, five units: lower sandstone (or Adigrat sandstone), the lower muddy sandstone (Gohation formation), Antalo limestone, the upper muddy sandstone (or mugger sandstone and gypsum) and the upper sandstone (Debrelibanos sandstone). 71 The Cenozoic is characterized by extensive faulting accompanied by widespread volcanic activity and uplift. The outpouring of vast quantities of basaltic lava accompanied by the eruption of large amounts of ash resulted in basaltic plateau often called trap series basalts. Several shield volcanoes, also consisting of alkali basalts and fragmental material, cover the center and the upper part of the Blue Nile basin. Over two third of the upper Blue Nile is covered by Cenozoic basalts and ashs. The mineralogical compositions of the basalts are spatially variable but all the basalt types contain dominantly olivines and clinopyroxenes with minor but variable amount of plagioclase, K-feldspars, and brown glass (Pik et al., 1998). The K-feldspars and brown glass is abundant in basalts of the southern sector of the plateau and plagioclase dominates in the basalts of the western and northern sectors. Quaternary lacustrine and fluvio-colluvial sediments and superficial deposits occur intermittently covering the basement and filling river channels. There are two prominent tectonically deformed regions on the plateau. These are the Lake Tana Graben (LTG) and the Yerer-Tullu Wellel Volcanic Lineament zone (YTVL). The majority of geothermal springs, Quaternary volcanoes, and quaternary basalt flows that exist in the Blue Nile basin are located in these two zones. These tectonic structures play an important role in controlling groundwater flow paths and groundwater chemical evolution. Although the groundwater flow conditions, the recharge conditions, and groundwater origin were unknown, the YTVL and the LTG were previously identified as a potential site of low enthalpy geothermal energy (Abebe, 2000). This observation was based on the presence of favorable geological structures, rainfall amount and heat flow. The LTG is a circular depression characterized by faulted blocks dipping towards lake Tana from all direction (figure 1). The faulted blocks in the western part of the lake have an average width of 1-4km and strike NNESSW. Gently inward-directed dips of the Tertiary basalt toward the center of the Tana basin are present to the west, north and east of the lake. In eastern (Deberetabor subgraben) and the north western (Chilga subgraben) subcatchments of the LTG, the late Miocene ligniteferous lacustrine deposits exist embedded in the trap series basalts (Chorowiz et al., 1998; Feseha, 2002). The Miocene lacustrine deposits contain mainly reworked volcaniclastics, thin layers of lignites, claystones, and siltstones. In places these sediments may have a thickness of 130m. Basaltic volcanism continued in the region until 10000yrs and basaltic lava cover the Miocene sediments. 72 The YTVL is an east-west trending zone that partly crosses the Blue Nile basin. It has a length of 800km and a diameter of 80km. The YTVL is a kind of half graben bounded by the Ambo fault from the north (Abebe et al., 1998). The Ambo fault has a throw of about 500 m. The major lineaments in the YTVL zone are the Didesa Lineament (DL) and the Ambo-Butajira Lineament (ABL). These lineaments are deep faults that cut across the YTVL. Along the YTVL, three main rock successions crop out: the Precambrian basement, the Mesozoic sedimentary rocks, and the Cenozoic volcanics. The volcanics are predominant whereas the basement and the sedimentary rocks are locally exposed. The sedimentary rocks (sandstones and limestones) thin out towards the southern part of YTVL. The Quaternary volcanics which cover the YTVL are mainly rhyolites and trachytes with abundant alkali-feldspars, alkali amphiboles and quartz. Faulting in the YTVL (the Ambo fault and associated lineaments) for instance juxtaposes the Mesozoic sediments and the volcanic cover favoring the formation of high discharge, low temperature thermal springs in the region. 2.2. Hydrology and hydrogeology The Blue Nile drainage is the result of river incision of the Cenozoic basaltic uplifted land. The Blue Nile River captures much of its runoff from the highlands in the southern and central part of the basin. The Blue Nile River is characterized by very high discharge during the wet season and very low discharge during the dry season. This reflects that the river discharge is dominated by inputs from rainfall and surface runoff than the groundwater component. With the exception of the eastern sector of the basin just east of the water divide of the Blue Nile basin where thick intermountain alluvial sediments bear high groundwater yield, the majority of groundwaters in the Blue Nile basin are abstracted from the fractured basaltic or metamorphic rocks. The well depth in the basaltic plateau ranges from 30 meters to 120 meters. The majority of cold springs emerge from the basaltic plateau. Because of dissection and fragmentation by river erosion, the basaltic cover is considered as perched groundwater systems with low storage and small aerial extent (BCEOM, 1999). Hydrograph separation shows that infiltration coefficient ranges from 3 to 20% of the total rainfall in the basaltic plateau (BCEOM, 1999). The highest infiltration coefficient and the highest groundwater contribution to surface water occur in the central Gojam highland region surrounding the Choke shield volcano and in the Lake Tana Graben. High rainfall on the shield volcano and the large lateral extent of the aquifers in that part of the plateau favors good groundwater storage in that region. Three high discharge springs: the Bure Baguna springs, the Andesa high TDS springs, and the Jiga low TDS springs emerge at the foot of the shield volcano. Regions on the left bank, particularly the eastern and 73 the south-eastern part of the Blue Nile River have generally low infiltration coefficient and low groundwater storage. This is most likely because of the strong dissection and fragmentation of the aquifers and the low rainfall amount in this part of the basin. Generally the Mesozoic sedimentary formations are thought to be good aquifers (BCEOM, 1999). The Mesozoic sediments, however, are only locally exposed. The relatively low infiltration coefficient in the southwestern part of the basin is related to the low permeability of the basement rock underlying that region. Few pumping tests data (BCEOM, 1999) in the region shows that the transmisivity is highly variable ranging from 1 to 700m2/day. The Quaternary basalts surrounding Lake Tana are characterized by high transmisivity (100-200m2/day) compared to the basalts of the trap series. Quaternary alluvial sediments have the highest transmisivity (in places more than 700m2/day). The metamorphic rocks in the western lowland have the lowest transmisivity (as low as 1m2/day). 3. Methodology and materials The methods used to achieve the objectives includes; a) direct analysis of the raw isotope hydrological and geochemical data; b) statistical classification of the data set accompanied by associating the statistical classes of the waters with hydrogeological variables; and, c) geochemical modeling. Furthermore δ13C, PCO2, pH and carbonate species compositions of the groundwaters were used to gain additional insight on groundwater geochemical evolution. 3.1. The chemical and environmental isotope data The majority of the water samples were collected from the upstream part of the Blue Nile River basin (figure 2). A total of 140 water samples were collected from groundwater wells, springs, lakes and rivers between November 2001 and August 2002. The samples were analyzed for their major ion concentration as well as for their isotope contents (δ18O, δD). Selected representative groundwater samples were analyzed for δ13C and tritium (3H) at International Atomic Energy Agency (IAEA) and the University of Avignon respectively. Chemical analyses were carried out at Laboratory of Hydrogeology, University of Avignon (France) while isotope compositions were measured at the IAEA Laboratory, Vienna. The δ18O and δD compositions were reported in per mil (‰) notation calibrated against the V-SMOW. Tritium concentration is reported in tritium unit (TU). The δ13C is reported in ‰ notation calibrated against PDB. Cation species were analyzed using 74 Atomic Absorption Spectrometer. Anion species were analyzed using a Dionex Ion Chromatograph equipped with automatic sampler. Silica (SiO2) was analyzed using colorimetric methods. Bicarbonate (HCO3-), carbonate (CO32-), pH, and temperature were measured in situ. Missing CO32- are estimated from pH and the activity of HCO3- using the equation K2=[CO32-][H+]/[HCO3-]. The partial pressure of PCO2 is estimated using the equation KCO2=[H2CO3]/PCO2. Saturation indices were calculated from chemical activities and ionic strength. Thirty two groundwater chemical data points from previous study (BCEOM, 1999) were included in the data set. The result of the analyses is presented in annex 1. N W E S ( Ñ L. T Ñ( $ Ta nÑzaGyÑue( $ $(ÑÑW T n ( ( Ñ no ca r ÑÑ Ñ Ñ Ñ ÑÑ Ñ ( Ñ Ñ Ñ ( (( Ñ Ñ ( ( ( Ñ Ñ( ( Ñ ( sa Dides T $ (( Ñ Ñ T $ T $ Ñ ÑÑ (Ñ ( ( ( ( (( ( ( Legend + Samples for chemistry Samples for isotopes Major Na-HCO3 springs T(ÑWon chi $ ( Ñ Ñ Ñ T $ Woliso ( ( (( Ñ (Ñ ( TT ( $$ T s Ababa Addi$ bo TA(m ( $ Ñ i op Et h ia ar M i ft R n g ( ( (Ñ (( ( ( ( Ñ Ñ ( Ñ Ñ Ñ Ñ( (( ( Ñ Ñ (( Ñ Ñ ( ( ( l Vo Bu re Ñ er ld h ie Nile Rive T $ ( Ñ( T $ S ke lu e C ho Ñ B ( a ( ( Ñ ( Andesa Ñ Ñ (Ñ ( (Ñ Ñ B ( ( ÑÑ lue Ñ ( ÑÑÑ Ñ Nile ri ( v Ñ Ñ Ñ ( Ñ ( ( ( Ñ Ñ ( Ñ Ñ ( (( ÑÑ (Ñ ( Ñ Ñ Ñ Ñ ( ( Ñ Ñ ( (( a Tan Ñ ( was h B m a s in and th e Ethiopian Rift S yste Th e A in Ñ Ñ Ñ Figure 2. Location map of water sampling points and other important sites. The region south of the Ethiopian rift margin is the Main Ethiopian Rift. Names of some localities are also shown. 3.2. Q-mode statistical cluster analysis (HCA) Statistical classification of geochemical data by Q-mode hierarchical cluster analysis (HCA) has proven to provide a suitable basis for objective classification of water composition into hydrochemical facies and for geochemical modeling (Alberto et al., 2001; Barbieri et al., 2001; Meng and Maynard, 2001; Swanson et al., 2001; Güler et al., 2002; Güler and Thyne, 2004). HCA is a semi-statistical technique intended to classify observations (e.g. water chemistry) so that the members of the resulting groups or subgroups are similar to each 75 other and distinct from the other groups. The characteristics of the groups or sub groups are not pre determined but can be obtained after the classification. The results obtained in HCA and the robustness of the HCA are justified according to their values in interpreting the data and in indicating patterns. It is therefore not the number of members of a group that determines the robustness of HCA. It is possible that many single member groups that do not belong to any of the multi member groups are placed in separate groups by themselves. This classification is useful specially to understand geological controls on water chemistry under a condition where useful geochemical data are available but clear hydrogeologic models have not yet been developed (Swanson et al., 2001). The advantage of HCA is that many variables such as physical, chemical or isotopic composition can be used to classify waters. In order that the variables have equal weight the raw chemical data should first be logtransformed and standardized. This restricts the influence of or the biases caused by the variables that have the greatest or the smallest variances or magnitudes on the clustering results. A detailed description of the advantages and uses of the HCA in hydrogeochemistry and the mathematical formulation behind HCA is thoroughly discussed in Swanson et al. (2001) and in Güler et al .(2002). The ability of the HCA to classify groundwater chemistry into coherent groups that may be distinguished in terms of aquifer type, subsurface residence time and degree of human impact on water chemistry provides a good opportunity to conduct hydrogeochemical modeling and understand groundwater geochemical evolution among the different groups or subgroups. In this study, we used the HCA to classify waters into objective groups and to conduct geochemical modeling among the different facies. A Microsoft EXCEL add-in module XLSTAT4.3 was used to conduct the HCA. 3.3. Inverse geochemical modeling Inverse geochemical modeling has widely been conducted in groundwater chemical evolution studies (Plummer et al., 1983; Kenoyer and Bowser, 1992; Varsànyi and Kovàcs, 1997; Hidalgo and Cruz-Sanjuliàn, 2001; Wang et al., 2001). It is a useful approach to determine the type and amount in mole of minerals that dissolve or precipitate along a groundwater flow path. In the cases where information is available on the hydrogeology of the basin, the flow paths can be selected based on the hydrogeological knowledge. The initial and the final member can be chosen by taking into account the location of the point, the hydraulic heads of the aquifer system and observed trends in chemical evolution of the water (Kenoyer and Bowser, 1992; Varsànyi and Kovàcs, 1997; Hidalgo and Cruz-Sanjuliàn, 2001). In areas where information on groundwater flow direction is lacking, the 76 initial and final waters can be selected from the HCA groups. This is based on the logical assumption that waters which fall in a statistical group may have similar residence time, similar recharge history, and identical flow paths or reservoir (Swanson et al., 2001; Güler and Thyne, 2004). PHREEQC computer code (Parkhurst and Appelo, 1999) was used to simulate the geochemical evolution among the average composition of statistical clusters. 4. Results and discussion 4.1. Chemistry and isotopic compositions of the waters and their spatial variation Complementary geochemical and isotope hydrological data show that in general there are two types of groundwater systems in the upper Blue Nile basin. These are the low salinity, Ca-Mg-HCO3 type, isotopically relatively enriched cold (13-25°C) groundwaters from the basaltic plateau and the high TDS, Na-HCO3 type, isotopically relatively depleted low temperature (25-40 °C) thermal groundwater systems from the deeply faulted grabens. 77 +M Ca SO 4+ Cl+ NO 3 100 g 0 CO 3 0 CO 3+ H 0 10 0 0 0 0 10 Ca 100 4 SO Mg +K Na 100 10 10 0 0 0 0 0 Cl+NO3 100 Figure 3. Piper plot of the chemical data. The majority of the groundwaters from the basaltic plateau are characterized by low TDS (generally less than 500mg/L). Calcium (Ca2+) and magnesium (Mg2+) dominate the cation species. They are characterized by CaMg-HCO3 type water in the Piper plot (figure 3). In the general groundwater chemical evolution model (Plummer et al., 1990; Adams et al., 2001; Edmunds and Smedley, 2000), these types of waters are often regarded as recharge area waters which are at their early stage of geochemical evolution. Rapidly circulating groundwaters which have not undergone a pronounced water-rock interaction may have also similar characteristics. 78 The majority of the low temperature thermal groundwater springs from the YTVL and the LTG have high TDS (generally greater than 1000mg/L). Sodium (Na +) and potassium (K+) dominate their cation species and bicarbonate (HCO3-) is the dominant anion. These groundwaters fall in the Na-HCO3 type groundwaters in the Piper plot. This is because with further hydrolysis of silicate minerals by the Ca-Mg-HCO3 type waters, the concentration of sodium, potassium, magnesium and bicarbonate increase but Ca2+ enrichment is limited by an earlier saturation and precipitation of carbonates. The high TDS and the enrichment of sodium therefore testify that the thermal and the high TDS groundwaters have undergone a relatively pronounced degree of groundwater chemical evolution. High pH values are more often observed in the groundwaters of the basaltic plateau than in the high TDS Na-HCO3 groundwaters of the YTVL and LTG. High fluoride (F-) is observed in few water points issuing from acid volcanic rocks of the Quaternary acid volcanics in YTVL and in the groundwaters associated with thermal systems (eg. samples SK2, SK3, SK4, SK80, SK93, and SK102). The high fluoride in the groundwaters associated with acid volcanism has its source from leaching of fluoride bearing accessory minerals. Fluoride from leaching of acid volcanic rocks is a widely accepted explanation of high fluoride in the East African Rift Valley groundwaters (Darling et al., 1996; Gizaw, 1996, Chernet et al., 2001). Some rock forming minerals of acid volcanic rocks such as alkali amphiboles, alkali mica or accessory minerals such as apatite often contain F- associated with OH- groups of the minerals (Kilham and Hecky, 1973). Shallow unprotected springs and unprotected wells contain high nitrate (NO3-) and chloride (Cl-). The source of high nitrate in groundwater of the region is often attributed to anthropogenic pollution (agricultural or domestic waste) exacerbated by lack of well head or spring protection (McKenzie et al., 2001;Reimann et al., 2003 ). Despite the high altitude (mean altitude ~ 2500 masl), the low mean annual air temperature (~17 °C) in the basin, and the furthest distance of Ethiopian form the Atlantic moisture source, the cold groundwaters of the basin do not show commensurate δ18O depletion compared to modern meteoric waters of the Sahelian Africa. This confirms the previous observation made from the isotopic composition of East African rainfalls (Rozanski et al., 1996) and from the few groundwater isotope data across the Sahelian Africa (Joseph et al., 1992). 79 Unlike the Ethiopian Rift groundwaters and the groundwaters of shallow systems of the Sahel region, the groundwaters of the Blue Nile basin are charactersied by high (>15) deuterium excess. All the groundwaters plot above the Global Meteoric Water Line (GLWL) in δ18O vs. δD plot (figure 4). Lakes and rivers draining the lakes are enriched and they plot below the GMWL following a slope of 5.4. The Blue Nile River sampled at Khartoum (Farah et al., 2000) shows similar δ18O and δD compositions to the groundwaters of the Blue Nile basin. This reflects both a rapid water transfer time from the Northwestern Ethiopian Plateau to the Sudan and lack of strong en route evaporative effects. A clear difference exists between the δ18O and δD compositions of surface water originating from the Equatorial lakes region and the δ18O and δD compositions of meteoric waters of the Blue Nile River. The former shows isotopic enrichment and plot below the GMWL owing to evaporation in the equatorial lakes. This distinct signal has been used as a basis for groundwater tracing (Farah et al., 2000) in Central Sudan where the two hydrologic systems merge. The low temperature thermal waters (hypothermal waters) and the high TDS Na-HCO3 type waters of the LTG and the YTVL are characterized by relatively highly depleted δ18O compositions (figure 4 and 5). A tendency of depletion (- 0.1 ‰/100 m) of δ18O with altitude is observed in the low TDS cold groundwaters (figure 5). In general, the low TDS cold groundwaters in mountainous regions east and northeast of the Blue Nile basin are characterized by relatively depleted δ18O. The δ18O and δD compositions of the groundwaters are distributed around the average summer δ18O and δD composition of the Ethiopian rainfalls. The average δ18O of Ethiopian summer rainfall is -2.5 ‰ (Kebede et al., 2003). Some previous works (Gizaw 2002) indicate the presence of dissimilarity and imbalance between groundwaters and the annual average rainwater δ18O and δD compositions. In the Blue Nile basin the groundwaters δ18O and δD composition very well represents the average isotopic composition of Ethiopian summer rainfall as recorded at Addis Ababa IAEA station. Lack of influence of evaporative concentration of the isotopes in groundwaters and the similarity between the isotopic compositions of the groundwaters and that of the composition of summer rainfalls indicate that recharge occurs principally from summer rainfall. This rule out the importance of evaporative fractionation prior to recharge in affecting the isotope signals. The spring rainfall oxygen isotopic signature having δ18O>0 ‰ (Kebede et al., 2003) is not commonly observed in the groundwaters. This testifies that recharge to groundwaters takes place only from the summer rainfall ruling out the importance of the spring rainfall as a source of recharge. 80 Representative samples collected from the low TDS groundwaters from the basaltic plateau have appreciable amount of 3H: SK46 (5.8TU), SK47 (3.6TU), SK48 (5.8TU), SK52 (3.8TU) and SK92 (6.8TU). The thermal and the high TDS groundwaters from the YTVL and the LTG contain 3H concentration less than the detection limit: SK80 (0.7TU), SK93 (0.5TU), and SK102 (0.5TU). This reflects deeper circulation of groundwater and older ages of the high TDS Na-HCO3 groundwaters. 100 80 60 Average Summer Rainfall at Addis Ababa δD 40 20 0 * * Average March-April Rainfall at Addis Ababa -20 -40 -10 -8 -6 -4 -2 0 2 4 6 8 10 Lakes and rivers draining them GMWL -60 18 δ O Low TDS cold groundwaters high TDS Na-HCO3 waters from YTVL and LTG Figure 4. Isotope plot of the water samples compared to the global meteoric water line (GMWL). The waters which plot below the GMWL are mostly lake waters from the basin. The highly depleted δ18O composition of the high TDS Na-HCO3 springs of the YTVL and the LTG (figures 4 and 5) indicate that recharge must have taken place at higher altitude sources. However the δ18O of present day highest altitude cold springs are not as depleted as the δ18O of the high TDS waters (figure 5). This indicates that recharge of the high TDS waters took place under a probably colder climate than today. The absence of appreciable amount of tritium in high TDS waters also testifies lack of any modern day meteoric water mixing into them. These waters must have followed deeper circulation pathways before they emerge as low temperature thermal waters. 81 >? >@ >A ?B N # # # ( % 98 ' ! ( + 9: 8 # # # # # # # Bl # ue N il e riv # # ## # # # # # # no lca Vo # # # # # # # # d ie l Sh # # # # # ## # ## # # # # # # # # bo A#m# # n chi # Wo # # # # # ## # CB The Awa CB ! " # " $ %&'' ( )''' %*'' ( %&'' %)'' ( %*'' %+'' ( %)'' %''' ( %+'' + &'' ( %''' + *'' ( +&'' +)'' ( +*'' ++'' ( +)'' + ''' ( ++'' ,&'' ( +''' ,*' ' ( ,&'' ,)' ' ( ,*'' ,+ ' ' ( ,) ' ' ,'' ' ( ,+ ' ' &'' ( , ''' *'' ( & '' # # r Nil e Rive # B # ##Wanzaye# Andesa er # ( , 9: 8 ! ( , 9' ' (, -; . / " lue # ## # tem the Ethiopian Rift Sys ( ) 98 8 ! ( % 98 ' # ( + 9: 8 ! ( , 9: 8 # CD S # # E # ke Cho # # W # sh Basin and < 7 = 3 5 " $ 01 2$ " , & 3 4 5 67 ! " ( 8 9+ 8 ! ( ) 98 8 # # an a L. T Bu re CD AA # # s Di des # a # # # Woliso # # A A # # >? # >@ >A ?B 18 δ O of cold groundwaters and coldsprings elevation in masl 0 500 1000 1500 2000 2500 3000 3500 0 -2 -4 -6 Figure 5. Spatial and altitudinal variation of δ18O in groundwaters of the basin. The size of the symbols reflects the degree of depletion in δ18O. The average depletion of δ18O vs. altitude is -0.1‰/100m. The most depleted waters around lake Tana (Andesa and Wanzaye) and in the YTVL (Ambo, Woliso, Wonchi and Dedesa) are the high TDS low temperature thermal spring. 82 Recharge to the high TDS waters of the LTG most likely takes place around the Guna and Debretabor Shield volcanoes (north of the Blue Nile basin). The Choke shield volcano in the center of the Blue Nile basin is the principal site of recharge to the Bure (SK15) high TDS springs. The Wolliso (SK80) and the Ambo (SK102) high TDS thermal springs are most likely recharged around the highland midway between the two regions. The Wanzaye alkaline thermal springs (SK 19) show specific isotopic and chemical characteristics. These springs are characterized by the most depleted δ18O and δD isotopic compositions but they are also the most dilute with TDS less than 200mg/L. The depletion in the isotopic composition reflects the presence of groundwater which has been recharged under colder climate in the Miocene lacustrine deposits. Groundwaters in the Ethiopian Rift Valley east of the Blue Nile basin just out side the basin (figure 5) are relatively enriched in δ18O compared to the groundwaters of the bordering highland. This indicates lack of strong subsurface link between the plateau in the eastern part of the Blue Nile basin and the shallow Rift Valley aquifers. The relative enrichment of the Rift Valley waters is related to the importance of evaporative fractionation before recharge (Kebede et al., 2003). An important water body which may play a role as a recharge source for aquifers in the LTG is the Lake Tana. But its influence on the nearby groundwater is not evident from the isotope plots. Waters collected form southern, eastern and northern part of the lake do not show any sign of enrichment caused by mixing of lake water into them. This lack of groundwater outflow from the lake is the result of the lake-ward dipping blocks of rocks that favor groundwater flow to the lake than loss of lake water into the surrounding aquifers. 4.2. Statistical clusters and their correlation with hydrogeology Statistical classification of the hydrochemical data is used here to elaborate the forth going water geochemical types and associated chemical processes. We considered 11 variables (pH, Ca2+, Mg2+, K+, Na+, HCO3-, CO32-, Cl-, SO42-, TDS) to classify the 86 groundwater samples with complete chemical analysis (annex 1). In HCA the variables are log transformed and normalized so that each variable will have equal weight. Groups were selected visually from the dendrogram (figure 6) which is the output of the clustering. From the dendrogram, two major groups and nine subgroups were chosen using index of similarity = 0.25. This index of similarity was chosen because of the nine sub groups of waters that result were very clearly distinguishable in terms of their hydrogeological and geolocical variables. The two major groups are distinct by their TDS. Group I (deep 83 Group I TDS> 1000mg/L Group II TDS< 1000mg/L S K9 S K2 2 S K15 S K8 0 S K9 3 MP 31 S K10 2 S K3 S K19 M P 14 MP 32 M P 10 MP 9 M P 13 M P 11 M P 19 M P 16 M P 15 M P 12 S K5 6 MP 1 S K10 S K14 S K13 MP 29 M P 18 M P 17 MP 6 S K2 9 S K7 S K10 0 S K9 9 S K2 7 MP 21 S K8 4 MP 4 S K13 3 S K5 7 MP 23 S K3 5 S K9 5 S K2 8 S K4 8 S K12 MP 30 S K4 S K9 6 S K9 2 S K9 8 S K9 1 S K2 5 MP 5 S K3 1 S K3 8 S K3 0 MP 27 MP 26 MP 25 S K110 MP 22 S K2 3 S K2 6 MP 28 S K2 0 MP 2 S K10 8 S K10 7 S K8 3 S K8 1 S K13 4 S K116 MP 8 S K113 S K7 9 MP 24 MP 3 S K5 2 S K10 4 S K13 0 S K4 7 MP 20 MP 7 S K17 S K2 Members of the sub groups Sub group I: SK9 Sub group II: SK22 Sub group III: SK15 Sub group IV: SK80 Sub group V: SK3, SK102, MP31, SK93 Sub group VI: SK19 Sub group VII: 21 samples from SK7 to MP14 Sub group VIII: 16 samples from SK4 to SK100 Sub group IX: 38 samples from SK2 to SK96 0 1 2 3 4 5 In d e x o f s im ila rit y Figure 6. Dendrogram of the Q-mode hierarchical cluster analysis. The 'phenon line' is chosen at similarity index=0.25 to select nine subgroups. The left most subgroup is subgroup I and the right most ones are subgroup IX. The samples which belong to each subgroup are listed under the branches and in the box in the upper corner. 84 systems) waters have TDS greater than 1200 mg/L. Group II (shallow systems) waters have TDS less than 800 mg/L. The two groups have also distinct δ18O, δD, 3H and δ13C compositions. Group I waters are depleted in δ18O and δD, they are nearly tritium free, and they are enriched in δ13C. Group II waters are generally enriched in δ18O and δD, they contain appreciable amount of tritium and have depleted δ13C. Group I waters have five subgroups and Group II waters have four subgroups. Group I contains five fairly distinguishable subgroups. Group II have three distinguishable subgroups. The samples grouped under each subgroup and the average physico-chemical composition of each subgroup is presented in table 1 and figure 7. There is a good statistical coherence among the average subgroups, that is, the chemical composition of the subgroups can be clearly explained in terms of geologic history, aquifer type, and the human impact on water quality. Correlation of the average composition of the subgroups and the accompanying geological features is given in table 2. The ability of the statistical analysis to classify the groundwaters into these distinct categories of geological context helped us to gain additional insight on groundwater flow patterns and to conduct inverse geochemical modeling on the subgroups. Subgroup pH 7.67 I1 1 8.11 II 6.53 III1 7.75 IV1 6.78 V4 9.15 VI1 21 8.15 VII 16 6.99 VIII 6.88 IX38 TDS 3514 3922 4596 1099 1781 128 328 461 191 K+ 7.2 14.8 28.3 11.9 27.3 0.4 2.3 4.9 2.0 Mg2+ 149 336 462 20 21.2 0.02 8.9 17.7 7.9 Na+ 62.2 426 531 262 415 43.2 45 37.8 10.4 Ca2+ 670.8 1.2 22.9 1.6 29.8 0.6 28.9 52.4 25.2 HCO3 CO32442.9 4.1 3081 50 3500 1.8 732 2.7 1209 0.6 44 35 211.2 2.4 281.4 0.4 135.7 0.1 SO422142 5.1 18.4 0.1 27.2 2.1 14.8 9.3 1.3 Cl25.6 8.0 30.2 39.2 36.8 2.6 9.3 19.9 3.1 F1.25 0.06 0.06 26.6 3.1 0.39 0.24 0.26 0.30 NO39.2 0.04 1.5 2.3 11.3 0.04 4.69 37.1 5 SiO2 PCO2 SI-C SI-G SI-Ch 23.7 0.008 +1.42 +0.35 +0.15 113 0.021 0.00 -4.92 +0.82 124 0.900 -0.27 -3.13 +0.88 119 0.013 -0.67 -6.20 +0.85 128 0.189 -0.20 -2.58 +0.88 66.2 <0.001 -0.76 -5.07 +0.51 30.2 0.002 +0.53 -2.63 +0.24 62.0 0.029 -0.27 -2.62 +0.57 55.6 0.018 -0.97 -3.70 +0.52 3 H δ13C +1.0‰ +4.8‰ 0.7 -4.5‰ 0.5 +1.5‰ -15.5‰ -11.6‰ 6.8 Table 1. Mean values of chemistry of groundwater subgroups, numbers in the superscript indicate the number of members of the subgroup, 3 H and δ13C concentration is measured on selected representative samples of the subgroups. The saturation indices of calcite (SI-C) gypsum (SI-G) and chalcedony (SI-Ch) are also presented for the average group. 85 Cluster Subgroup I1 Subgroup II1 Subgroup III1 Subgroup IV1 Subgroup V4 Subgroup VI1 Subgroup VII21 Subgroup VIII16 Subgroup IX38 Geology The only cold spring from Mesozoic succession of the Blue Nile gorge. This spring emerge at a contact between evaporite beds and a thick sequence of limestone. A high TDS slightly alkaline pH cold spring from the Bure fault zone in the Ethiopian Plateau, no indication of thermal activity is observed in the region, the spring emerge at a contact between the Cenozoic trap basalt and the underlying Mesozoic sandstone. These springs have depleted δ18O. A high TDS, very low pH, thermal/cold spring from the LTG. Travertine deposits were observed at the issuing point of these springs. These springs have depleted δ18O. A near neutral pH, high TDS thermal spring in the southern part of YTVL. Underneath the area where these springs emerge the Mesozoic strata is very thin or absent. This spring is similar to sub group V waters but it contains lower PCO2. A low pH group of thermal springs in deeply faulted region of the YTVL, these subgroups are distinct from subgroup III water by their Mg content. While basalts are the dominant rock outcrop around subgroup III waters, quaternary acid volcanic rocks dominate the recharge region of subgroup V waters. Thick travertine and silica sinter deposits are ubiquitous around the issuing point of these springs. A hyperalkaline, very low TDS thermal spring east of the LTG. The geology of the area where this spring emerges is distinct by the presence of beds of late Miocene lignite, organic marls and Pliocene lacustrine deposits. There are three springs of this nature with in a radius of 5km. A Low TDS, relatively high pH springs and groundwaters from the basaltic plateau. The majority of these waters are deep wells from the basaltic plateau. These waters are saturated with respect to calcite. The relatively high pH and HCO3 and the saturation indices shows that these subgroups are derived from the subgroup IX waters by silicate hydrolysis. Low TDS, relatively high pH, high NO3 and high Cl cold springs and groundwater wells from basaltic plateau. These waters are affected by anthropogenic pollution because of lack of well-head or spring protection. Very low TDS, low pH, Ca-HCO3 type cold springs and shallow wells which may represent geochemically unevolved young meteoric waters from basaltic plateau. These waters are mainly collected from handdug wells and cold springs. These waters are under saturated with respect to calcite, aragonite and gypsum. Some members of these group contains appreciable amount of tritium testifying shallow circulation. Table 2. Description of the geologic characteristics accompanying the statistical subgroups. The HCA also shows that groundwaters which plot near each other in a simple piper plot (figures 3 and 7) may not be necessarily similar in their chemical evolution history. Waters of subgroups II, III, VII and VIII which plot near each other in the piper plot are statistically and geologically distinct. The same is true for waters of subgroups IV, V and VI. 86 100 0 V IV Mg 4 SO +K Na CO 3 VII CO 3+ H III II 0 10 0 VIII 0 10 0 IX g 0 +M Ca SO 4+ Cl+ NO 3 I 100 Ca 0 10 0 100 0 0 0 10 0 VI Cl+NO3 100 Figure 7. Piper plot of the subgroups, groundwater which plot near each other in a piper plot may not always be similar in their geochemical evolution history. The lines are some of the paths selected for the inverse geochemical modeling. The concentration of Ca2+ decreases generally from the more dilute groups to the high TDS groups (except in subgroup I). This may reflect precipitation of calcite along the flow path when the groundwater is transferred from the shallow to the deeper systems. Many of the ions (K+, Na+, Mg2+, HCO3-, F-, Cl-, SiO2) increase from the dilute systems to the high TDS systems. Exceptions to these are subgroup I and subgroup VI. The general increase in Mg2+, Na+, K+, SiO2 reflect the increased amount of hydrolysis of silicate minerals such as olivines, pyroxens, plagioclase, alkali feldspars. The pH generally decreases from the shallow to the deeper systems testifying additional input of CO2 gas. The concentration of NO3- is generally higher in the shallow systems than in the deep systems indicating a recent increase in anthropogenic pollution of the shallow groundwaters. Generally high TDS groundwaters from the YTVL contain the highest F- content. All the water groups are saturated or supersaturated with respect to silica. This reflects a relatively rapid hydrolysis of ferromagnesian minerals of the basaltic aquifers. The silica content however is higher in the high TDS systems because the high temperature of these groundwater systems increases the solubility of silica. 87 All the waters subgroups except subgroup I are under-saturated with respect to gypsum and anhydrite testifying that these minerals which abundantly exist in the sedimentary layers are not the main limiting factors of the compositions of SO42-. Only one cold spring sample (SK9) which emerges from the Mesozoic sedimentary layers is near saturation with respect to gypsum and anhydrite. The high TDS Na-HCO3 low temperature thermal springs are undersaturated with respect to the carbonate minerals (calcite, aragonite and dolomite), though deposition of these minerals are common around the springs. This may testify that disequilibrium (undersaturation) in these waters is caused by external input of CO2 gas from deeper sources which joins the groundwaters at shallower depths. Under this condition the groundwaters may not have enough time to dissolve more carbonate minerals to reach an equilibrium conditions. The statistical classification also shows that at least two geochemical types of groundwaters exist in the basaltic plateau. Waters of subgroup IX generally represent shallow circulation while those in sub group VII represent deeper circulation in basaltic plateau. The concentration of all the major elements (except SiO 2 and Ca2+) the pH and the TDS increase from subgroup IX to subgroup VII. The waters of subgroup VII are saturated with respect to silica and calcite while waters of subgroup IX are under saturated with respect to these minerals. These compositional differences implies the presence of at least two groundwater layers in the basaltic aquifers. Hydrolysis of volcanic minerals leads to increase in pH and the increase in the concentration of major elements when the water is transferred from the shallow basaltic aquifers to the deeper basaltic aquifers. Subgroup VIII represents anthropologically polluted members of the shallow groundwaters from the basaltic plateau. 4.3. Geochemical modeling and groundwater chemical evolution We conducted inverse geochemical modeling on the water subgroups that resulted from HCA. The average chemical composition of waters of subgroup IX was assumed to represent a pristine recharge area groundwater. Its composition is therefore taken as 'initial' water in the inverse geochemical modeling. The remaining subgroups were considered to be derived from this subgroup. Two exceptions are waters of subgroup VI and I. Subgroup VI has lower TDS and high pH which may not be possible to have its origin from the average of Subgroup IX under a logical assumption that the final waters in the inverse modeling have higher TDS than the initial waters. Field evidence shows that water of subgroup I directly get their recharge from rainfall without passing through chemical characteristics of waters of subgroup IX. Therefore we used the lowest TDS groundwater (SK 92) as the initial water to simulate the composition of subgroups VI and I. 88 Path SK92-Subgroup I Reaction Gypsum+Dolomite+Halite+NaX from ion exchange+CO2(g)→Ca-SO4water+Ca loss to ion exchange+ Chalcedony IX-Subgroup II Plagioclase+Olivene+Pyroxene+K-mica+CO2(g )→ Na-Mg-HCO3 water+Illite+Calcite+Fluorite IX-Subgroup III Plagioclase+ Olivine+Pyroxene+ K-mica+trace gypsum+ CO2(g) → Na-Mg-HCO3 water+Calcite+Illite+ trace fluorite Plagioclase+Pyroxene+K-mica+trace fluorite+CO2(g) → Na-HCO3 water+Chalcedony+Ca montmorilonite+trace Gypsum IX-Subgroup IV IX-Subgroup V Plagioclase+Pyroxene+K-feldespar+trace gypsum and fluorite → Na-HCO3 water Chalcedony+Calcite+Illite IX-Subgroup VII Plagioclase+Pyroxene+K-mica+trace gypsum +CO2(g) → Ca-Mg- HCO3 water+illite Table 3. Summary of inverse modelling for selected paths. The majority of the phases and thermodynamic data are taken from PHREEQC database (Parkhurst and Appelo, 1999) and from Kenoyer and Bowser (1992). The mineral phases were selected based on the saturation indices and the general mineralogical compositions of the rocks in the basin. The result of the inverse geochemical modeling (table 3) shows that, except in the evolution towards subgroup I, the hydrolysis of silicate minerals (principally feldspars and ferromagnesian minerals) without a major involvement of the sedimentary minerals (e.g. carbonates, evaporites) can satisfy the simulation. The major minerals that are required to dissolve were olivine, pyroxene, plagioclase, K-micas and Kfeldspars. Dissolution of gaseous CO2 is required in all cases. Removal of clay minerals such as illite or Camontmorillonite was required during groundwater transition in the aquifers. Precipitation of calcite or chalcedony or both were required in the models. Dissolution of gypsum, dissolution of dolomite and cation exchange (CaX ↔ NaX) were required to simulate the composition of subgroup I. While hydrolysis of olivines and pyroxenes were the principal reaction required to simulate group II and III waters, Plagioclase, K-feldspars and K-mica were the major phase required to simulate subgroups IV and V. 4.4. Further insight on groundwater chemical evolution: the role of CO2 4.4.1. Geochemical evidence-carbonate species One important observation that emerged from the carbonate species composition of the groundwaters and the HCA is the presence of four types of groundwater system as far as the role of CO 2 is concerned. This distinction is made based on the relation between pH, HCO3- and TDS (figure 8). The four systems represent different degree of involvement of CO 2 in the chemical evolution. The four systems can also be distinguished based on their δ13C. The four systems are, a) Subgroups III and V: a very low pH, high HCO3- and high TDS thermal springs; b) Subgroup II and IV: a near neutral pH, high HCO3- and high TDS groundwaters; c) Subgroup VI: a very high pH low TDS and low HCO3-; and, d) Subgroup IX, VIII, VII, and I: near neutral pH, low HCO3- and variable TDS. 89 Open system hydrolysis Open to metamorphic CO2' δ13C=+1‰ 10000 δ13C= +4.75‰ III δ13C= -4.5‰ 1000 HCO3 in mg/L II δ13C= +1.5‰ V IV I VIII 100 Open systme hydrolysis Open to lower amount deep CO2, VII δ13C= -11.6‰ IX VI δ13C= -15.5‰ Exchange with combusted organic matter 10 SK92 1 5 6 7 8 9 10 pH Figure 8. pH vs. HCO3 plot to show the role of external input of CO2 in chemical evolution of groundwaters, size of the circles reflects the TDS. Some d13C is also given. One important point about the chemistry of thermal groundwater springs of subgroups III, V is the presence of high partial pressure of CO2 and high HCO3-. Partial pressure of carbon dioxide as high as 0.9 atm (table 1) and concentration as high as 4000mg/L have been reported (BCEOM, 1999) in these springs. All the waters with these characteristics emerge as hot springs exclusively in the LTG and in the YTVL. The high partial pressure of CO2 coupled with the high mineralization, the low pH, the relatively high temperature, and the fact that they emerge along the deep grabens and associated lineaments testify that the system is open to an external input of CO2 from deeper sources. The relatively depleted δ18O composition of these waters and the very low tritium content reveals a recharge source at a high altitude far from the emanation point of the springs. These in turn reflect a deep circulation of groundwater and long subsurface residence time. The source of CO2 may be a direct source from the mantle along the deep faults or from metamorphic decarbonation of the underlying sedimentary sequence by heat from a magma chamber- the same magma chamber which has led to the eruption of the Quaternary volcanoes in the YTVL and the formation of Quaternary basalts in the LTG. Two exceptions of the high TDS Na-HCO3 waters that may not be completely explained by the above model are the Bure cold springs (subgroup II) and the Wolliso thermal springs (subgroup IV). These two waters evolve with lower amount of CO2 involvement compared to the subgroup III and V waters. While lack of heating from below restricts the decarbonation and major influx of CO2 from the Mesozoic sediments in the Bure cold spring area, lack of or thinning of Mesozoic formation in southern part of YTVL restricts major influx of CO 2 from deeper sources. 90 Waters of subgroup VI shows a unique characteristics. They have very high pH, extremely low amount of PCO2 and very low TDS. They are associated with faulted area in the eastern part of the LTG. The isotopic signature of these waters indicates that they have meteoric origin. Their depleted δ18O relative to nearby groundwater bodies may indicate that recharge takes place at high altitude. Unlike the other thermal springs in the Blue Nile basin, + - this water contains an extremely low amount of dissolved carbon dioxide. Na , HCO3 and SiO2 dominate the chemistry. Ca2+ and Mg2+ are extremely low. One of the most plausible explanations for the existence of these hyperalkaline, very low TDS thermal springs is that these waters interact with a pre-metamorphosed Miocene lacustrine lignite and mud rock beds imbedded between the trap series basalt. This metamorphosed zone could act as major zone of sink in CO2, Ca2+ and Mg2+ and an increase in Na+, K+, pH, and SiO2 leading to this unique characteristic. These kinds of hyperalkaline very low TDS waters are not uncommon. Clark et al.(1994) have reported the presence of a similar type of thermal springs in an area characterized by a similar geologic condition (high heat flow and covered by metamorphosed organoclastic deposits). The remaining subgroups (Subgroup VII, VIII, IX, I) evolve under a relatively closed system silicate hydrolysis and dissolution reactions. Under such a condition, the initial H+ produced by the reaction between soil CO 2 and the infiltrating water will be consumed by the silicate hydrolysis reaction. Because of lack of additional CO2 from deeper sources both the pH and HCO3- increase along the evolution direction SK92→IX→VII or along SK92→I until saturation is reached with respect to carbonate minerals. The relatively deeper systems of the basaltic plateau (subgroup VII) are more closed to external input of CO2 than the shallower systems of the basaltic plateau (subgroup IX). 4.4.2. Carbon-13 evidence Carbon-13 were measured in six samples each one representative of different groundwater subgroups. The δ13C content of the high TDS Na-HCO3 groundwaters varies between -4.2 and +6.3‰ PDB. These compositions are more enriched than the δ13C compositions of mantle carbon dioxide which varies between -3 and -8‰ (Hoefs, 1997). This enriched range of δ13C is most likely the result of interaction of groundwaters with CO2 from metamorphic decarbonation of carbonate rocks beneath the YVTL and the LTG. This is most plausible mechanism because the δ13C of carbon dioxide from carbonate rocks ranges between -4 and +4% (Craig, 1963). This confirms the geochemical evidence from pH, HCO3 and TDS of the influence of deep CO2 from deeper metamorphic decarbonation sources on the high TDS waters. Waters of subgroup III and V are relatively more 91 depleted in δ13C than waters of subgroup II and IV confirming the hypotheses from the carbonate species compositions that relatively lower amount of deep CO2 is involved in the geochemistry of these springs. The Jiga cold springs (SK 103) representative of the shallow Ca-Mg-HCO3 type groundwaters show a δ13C of 11.6‰ reflecting soil CO2 as a principal source of carbonate species in the shallow groundwaters. Cold groundwaters from basaltic aquifers around Addis Ababa (on the southern water divide of the Blue Nile Basin) show a δ13C ranging between -4 and -12‰ (Gizaw, 2002). These testify the dominant source of CO2 in the shallow groundwaters of the basaltic plateau is soil carbon dioxide. The Wanzaye thermal springs (subgroup VI) are the most depleted in both δ13C (δ13C = -15.5‰) and δ18O (5.4‰) contents. The very high pH of these springs, the extremely low PCO2 , its Na-HCO3 nature, and the depletion in δ18O shows that these springs represent evolved groundwater systems which have undergone significant degree of water-rock interaction. The extreme depletion of δ13C of these springs confirms the hypothesis from geochemical evidences that they must have interacted with organic matter at depth. These reflect that the Wonzye thermal springs interact with organic matter of the Miocene organic rich sediments in the LTG. 5. Summary and conclusions The Hierarchical Classification Analysis elaborately classified the groundwaters of the upper Blue Nile basin into two major groups and 9 subgroups. The advantage of the method was that the subgroups were objective and a clear geohydrological patterns were recognized. The nine subgroups show different degree of water rock interaction, subsurface residence time, aquifer composition, influence of CO2, and exposure to anthropogenic pollution etc. In a poorly known hydrogeological system the exercise of associating the results of the cluster with geo-hydrological conditions facilitates the understanding of the groundwater flow systems in the basin. Traditionally, it was thought that two groundwater layers (shallow/deep; fresh/saline; unconfined/confined) exist in volcanic aquifers of the region (Chernet, 1982; Chernet, 1990). However, the Hierarchical Classification Analysis and the geochemical approach show that more groundwater flow patterns can be distinguished in the region adding more understanding to the previous knowledge of the Ethiopian plateau volcanic hydrogeology. 92 Schematic groundwater flow pattern Wolisothermal spring δ13C=+6.3‰ δ13C=-4.2‰; δ18O=-5.3‰ * andCraterLake Wonchi thermal springs δ13C=+1.5‰; δ18O=-4.4‰ sinter associatedtravertineandsilica Ambofault Ambothermal springs and YTVL (60 km) * * CO2 ( D) CO2 (D) CO2 (M) CO2 ( M) Heat and CO 2( M) source Precambrian basement * Mesozoic Sandstone &Limestone Cenozoic trap basalt Quaternary trachyte & rhyolite volcanics Magmatic CO2 CO2 from decarbonation CO2 rich thermal & high TDS springs Schematic groundwater flow paths The Lake Tana Graben: Eastern Subgraben Debretabor Coldspring δ 18O =-3.2 ‰ * Wonzaye low TDS NaHCO3 thermal springs δ 13 C=-16‰ ; δ 18O=-5.2 ‰, pH =9.4 * Lake Tana δ 18 O=+4.5 ‰ * Heat source δ18O=-2.25 ‰ δ13C=+1‰; δ18O=-3.5 ‰ Gish Abay Coldsprings Lake Tana block faults δ13C=-12‰; δ18O=-1. 4 ‰ Recent lacustrine and alluvial sediments Groundwater flow paths Jiga Coldsprings Quaternary to recent basaltic flows Locally metamorphosed Miocene organoclastics Bure Baguna cold springs Cenozoic trap basalt and associated volcanics Choke Mountain Chain * * * Mesozoic Sandstone and Limestone Source of recharge and main geochemical processes 1. Recharge at high altitude by CaHCO3 type waters, major residence in acid volcanic rocks, open system hydrolysis of silicate minerals, lowering of pH by addition of CO2 from metamorphic decarbonation of the underlying Mesozoic sediments or from direct CO 2 input form deeper sources along the fault zone, release of CO2 on emergence or before emergence and deposition of travertine and silica sinter, waters with Na-HCO3 characteristics as final composition. 2. Recharge at around Guna Shield volcanoes, flow of Ca-HCO3 type water through a metamorphosed zone of the Miocene lignite beds and sink of CO2, exchange of Ca for Na and K in the zone, formation of high pH and high SiO2 water plume, emergence of Na-HCO3 water plume at low TDS and high pH. 3. Recharge at high altitude around shield volcanoes of the central part of the Blue Nile basin, major subsurface residence of the water in the basaltic aquifer, closed system dissolution of silicate minerals without major addition of CO2 from depth, emergence of the water at a contact between the Mesozoic and the Cenozoic as Na-Mg-HCO3 water at above neutral pH. Cenozoic trap basalt & associated acid volcanics Groundwater flowpath Cenozoic trap basalt and associated volcanics Mesozoic Sandstone and Limestone Groundwater flow paths δ18O=+3.3 ‰ δ13C=+4.8‰; δ18O=-4.96‰ * Abay River * Andesa highTDS springs δ18O=-2.73‰ Bahrdar twonwell * δ18O=-2.2‰ Bahrdar water spully spring The Lake Tana graben: south Quaternary to recent basaltic flows Precambrian basement 4. Recharge by Ca-HCO3 waters around the Choke shield volcanoes in the center of the basin, major residence of the water in basaltic aquifer, open system silicate hydrolysis by CO2 influx from metamorphic decarbonation of the Mesozoic sediments or from a direct input of CO2, deposition of travertine and silica sinter upon emergence, and formation of low pH, Na-Mg-HCO3 waters. CO2 from decarbonation Magmatic CO2 Figure 9. Schematic sections showing origin and evolution of selected groundwater systems/springs. 1-Thermal springs in the YTVL, 2Wonzaye thermal springs in LTG, 3-Bure cold springs on the plateau, 4-Andesa thermal spring in LTG. 93 The geochemical, isotopic, stratigraphic, structural data and the Hierarchical Classification Analysis helped to schematize geo-hydrological characteristics of important zones in the Blue Nile basin and to gain for the first time a general picture of the groundwater circulation and its chemical evolution in the basin. Conceptual models that schematize the major geochemical processes and recharge source for selected groundwater systems are presented in figure 9. In the basaltic plateau recharge is rapid, groundwater circulation is shallow and the waters are characterized by low TDS. Two structurally deformed zones, namely the LTG and the YTVL plays a major role in favoring the existence of regional and probably deeper groundwater flows. The geologic processes which formed the two structural basins resulted in juxtaposition of the Mesozoic sediments with the Cenozoic volcanic cover and this promotes the presence of artesian springs at the contact between the Mesozoic sediment and Cenozoic volcanic cover. Furthermore carbon dioxide from deeper sources (CO2 produced by metamorphic decarbonation of the Mesozoic sediments underlying the basaltic trap) along the deformed zones influences geochemical evolution of the high TDS thermal groundwaters of the YTVL and the LTG. The CO2 gas from depth promotes acid - hydrolysis of the volcanic covers, which explain generally low pH and high HCO3 of these groundwaters. Despite the similarity in climatic conditions and the general similarity in Mesozoic lithology, the hydrogeological characteristics of the Blue Nile basin are different from Hydrogeology of Sahelian Africa. The majority of groundwaters of the Blue Nile basin are often highly flushed, young, low TDS groundwaters with rapid recharge through fractured rocks. These characteristics are mainly the result of uplifting and erosional fragmentation of the aquifers of the Blue Nile basin. In the shallow sedimentary aquifers of the Sahel, evaporation prior to recharge seems an important hydrologic process (Sonntag et al., 1982; Dodo and Zuppi, 1997). This process seems unimportant in the upper Blue Nile basin. The classical sedimentary basin aquifer (Intercalaire aquifers) that underlie the majority of the Sahel and Northern African countries (Sudan, Chad, Senegal, Mali, Niger, Tunisia, Egypt etc) containing the late Pleistocene or early Holocene groundwaters (Sonntag et al., 1982; Andrews et al., 1994; Edmunds et al., 2003; Dabous and Osmond 2001) is represented by uplifted Mesozoic sediments which are not accessible to groundwater circulation due to the thick (often greater than 1km) basaltic cover. In some places such as the LTG and the YTVL, however the Mesozoic sediments play an indirect role in influencing the hydrogeochemistry of the groundwaters by supplying CO 2 for water-rock interaction. 94 The important role that CO2 from deeper sources plays in groundwater chemical evolution is a widely accepted model in the East African Rift System also. In general, the δ13C content of the high TDS groundwaters of the Blue Nile basin is more enriched than the δ13C content of groundwaters of the Ethiopian Rift Valley, the δ13C of the later as documented in Darling et al., 1996 and Craig et al., 1977. This reflects that CO 2 from decarbonation of marine carbonates is more important in the Blue Nile basin high TDS thermal groundwaters than in the Rift Valley groundwaters. Many previous models (Darling et al., 1996; Gizaw, 2002) consider the source of CO2 in thermal groundwaters of the Central Ethiopian Rift or that of the Addis Ababa region is the mantle and the influence of Metamorphic CO2 is non existent or minor. Structural evidence coupled by geochemical data shows that the YTVL is an extensive east - west zone that intersects the Ethiopian Rift Valley. The influence of the Mesozoic sedimentary layers as a source of CO2 in thermal groundwaters may not be minor beneath Addis Ababa and the Main Ethiopian Rift Valley as it was thought before. The information obtained and the major conclusions made in this study would help to select future targets of detailed groundwater resources assessment programs. The schematic diagrams may be used to select suitable sites for groundwater resources development. The recently flourishing Ethiopian Gaseous Soda Spring bottling plants may find the schematic diagrams and the water quality data very useful. The isotope and geochemical data from the Blue Nile basin also allowed us in an independent work (Kebede et al., 2003) to trace the subsurface hydrogeologic link between the Northwestern Ethiopian Plateau and the Ethiopian Rift Valley. Acknowledgements: This research was funded partly by the French Ministry of Foreign Affairs through its office at the French Embassy in Ethiopia. The first author would like to thank the Ethiopian American Foundation for financial support for travel within Ethiopian during the collection of water samples. The Department of Geology and Geophysics, Addis Ababa University provided field vehicle and other field logistics. This work would not have been completed without the isotope analysis provided by The International Atomic Energy Agency through its TC project (ETH/8/007). Thanks also to Drs Dereje Ayalew, Tesfaye Korme and Balemwal Atnafu for their constructive comments and discussions during the start of this work. This paper is part of the first author's dissertation research. Last but not least we would like to thank one anonymous reviewer, Prof. M. Edmunds and Prof. Yves Tardy for constructive comments which helped us bring this article to publishable quality. 95 References Abate, B., Koberl, C., Buchanan, C.P., Korner, W., 1996. Petrography and geochemistry of basaltic and rhyodacitic rocks from Lake Tana and the Gimjabet-Kosober areas (North Central Ethiopia). J. African Earth Sci. 26, 119-134. Abebe, T., Mazzarini, F., Innocenti, F., Manetti, P., 1998. The Yerer-Tullu Wellel volcanotectonic lineament:a transtensional structure in central Ethiopia and the associated magmatic activity. J. African Earth Sci. 26, 135 150. Abebe, T., 2000. Geological limitations of a geothermal system in a continental Rift zone: example the Ethiopian Rift Valley; Proceedings World Geothermal Congress 2000; Kyushu - Tohoku, Japan, May 28 - June 10, 2000. Adams, S., Tredoux, G., Harris, C., Titus, R., Pietersen, K., 2001. Hydrochemical characteristics of aquifers near Sutherland in the Western Karoo, South Africa. J. Hydrol. 24, 91– 103. Alberto, W.D., Del Pilar, D.M., Valeria, A.M., Fabiana, P.S., Cecilia, H.A., De Los Angeles, B.M., 2001. Pattern recognition techniques for the evaluation of spatial and temporal variations in water quality. A case study: Suquý`a River Basin (Co´rdoba-Argentina). Water Resear. 35, 2881–2894. Andrews, J.N., Fontes, J.-Ch., Aranyossy, J.-F., Dodo, A., Edmunds, W.M., Joseph, A., Travi, Y., 1994. The evolution of alkaline groundwaters in the Continental Intercalaire aquifer of the Irhazer Plain, Niger. Water Resour. Res. 30, 45–61. Asrat, A., Barbey, P., Gleizes, G., 2001. The Precambrian geology of Ethiopia: a review. African Geosci. Rev. 18, 271–288. Assefa, G., 1991. Lithostratigraphy and Environment of deposition of the Late Jurassic-Early Cretaceous sequence of the central part of northwestern plateau, Ethiopia. Neues Jahrbuch Geologie Paleontologie Abhandubglen. 182, 255-284. Ayenew, T. 1998. The Hydrogeological system of the lake district basin, central Main Ethiopian Rift. Phd thesis, ITC publication Number 64, the Netherlands, 200p. Barbieri, P., Adami, G., Favretto, A.,Lutman, A., Avoscan,W., Reisenhofer, E., 2001. Robust cluster analysis for detecting physico-chemical typologies of freshwater from wells of the plain of Friuli (northeastern Italy). Analytica Chimica Acta. 440,161–170. BCEOM, 1999. Abay River Basin integrated master plan, main report, Ministry of Water Resources, Addis Ababa. Chernet, T., 1982. Hydrogeology of the Lakes Region, Ethiopia. Ministry of Mines and Energy, Addis Ababa. Chernet, T., 1990. Hydrogeology of Ethiopia and water resources development, Ministry of Mines and Energy report, 157p. Chernet , T., Travi, Y., Valles, V., 2001. Mechanism of degradation of the quality of natural water in the lakes region of the Ethiopian Rift Valley. Water Reser.35, 2819–2832. Chorowiz, J., Collet, B., Bonavia, F., Mohr P., Parrot, J-F., Korme, T., 1998. The Tana basin, Ethiopia. Intraplateau uplift, Rifting and subsidence. Tectonophys. 295, 351-367. Clark, D.I., Dayal, R., Khoury , H.N., 1994. The Maqarin (Jordan) natural analogue for 14C attenuation in cementitious barriers. Waste Manag. 14, 467-477. Craig, H., 1963. The isotopic geochemistry of water and carbon in geothermal areas, In: Tongiorgi, E.(Ed). Nuclear geology on geothermal areas, CNR, Pisa, pp. 17-23. Craig, H., Lupton J.E., Horowiff, R.M., 1977. Isotope Geochemistry and Hydrology of geothermal waters in the Ethiopian rift valley. Scripps Institute of Oceanography, University of California report, 160p. Dabous, A.A., Osmond J.K., 2001. Uranium isotopic study of artesian and pluvial contributions to the Nubian Aquifer, Western Desert, Egypt. J. 243, 242- 253. Darling, G., Gizaw, B., Arusei, M., 1996. Lake-groundwater relationships and fluid-rock interaction in the East African Rift Valley: isotopic evidence. J. African Earth Sci. 22, 423-430. Darling, WG., 1996. The Geochemistry of fluid processes in the eastern branch of the east African rift system, Ph.D thesis, British Geological Survey, UK, 235p. Dodo, A., Zuppi G.M., 1997. Groundwater flow study in the Bilma-Djado Basin (Niger) by means of environmental isotopes. Earth and Planetary Sci. 325, 845-852. Edmunds ,M., Smedley, P., 2000 Residence time indicators in groundwater: the East Midlands Triassic sandstone aquifer. Appl. Geochem. 15, 737-752. Edmunds, W.M., Guendouz, A.H., Mamou, A., Moulla, A., Shand, P., Zouari, K., 2003. Groundwater evolution in the Continental Intercalaire aquifer of southern Algeria and Tunisia: trace element and isotopic indicators. Appl. Geochem. 18, 805–822. Farah, E.A. Mustafa, E.M.A., Kumai, H., 2000. Sources of groundwater recharge at the confluence of the Niles, Sudan. Env. Geol. 39, 667-672. Feseha, M., 2002. Sequence stratigraphy, petrography, and geochronology of the Chilga Rift basin sediments, northwest Ethiopia. Ph.D. dissertation, University of Texas at Austin, 330p. 96 Gizaw, B. 1996. The origin of high bicarbonate and fluoride concentrations in waters of the Main Ethiopian Rift Valley, East African Rift system. J. African Earth Sci. 2,391-402. Gizaw, B., 2002. Hydrochemical and Environmental Investigation of the Addis Ababa Region, Ethiopia. Ph. D dissertation, Faculty of Earth and Environmental Sciences Ludwig-Maximilians-University of Munich, 157p. Güler, C., Thyne, D.G., McCray, E.J., Turner, A.K., 2002. Evaluation of graphical and multivariate statistical methods for classification of water chemistry data. Hydrogeol. J. 10, 455-474. Güler, C., Thyne, D.G., 2004. Hydrologic and geologic factors controlling surface and groundwater chemistry in Indian Wells-Owens Valley area, southeastern California, USA. J.Hydrol. 285, 177-198. Hidalgo, M.C-L., Cruz-Sanjulian, J., 2001. Groundwater composition, hydrochemical evolution and mass transfer in a regional detrital aquifer Baza basin, southern Spain. Appl. Geochem. 16, 745-758. Hoefs, J., 1997. Stable isotope Geochemistry. Springer, Berlin, 210p. Joseph, A., Frangi, P., Aranyossy, J.F., 1992. Isotopic composition of Meteoric water and groundwater in the Sahelo-Sudanese Zone. J. Geophys. Res. 97, 7543-7551. Kebede, T., Koeberl, C., Koller, F., 1999. Geology, geochemistry and petrogenesis of intrusive rocks of the Wallagga area, western Ethiopia. J. African Earth Sci. 29, 715-734. Kebede, S., Travi, Y., Alemayehu, T., Ayenew, T., Aggarwal, P., 2003. Tracing sources of recharge to groundwaters in the Ethiopian Rift and bordering plateau: Isotopic evidence. Paper presented at the Fourth International Conference on Isotope for Groundwater Management, Vienna, IAEA. Kenoyer, G.J., Bowser, C.J., 1992. Groundwater chemical evolution in a sandy aquifer in Northern Wisconsin. 2. Reaction Modeling. Water Resour. Res. 28, 591-600. Kilham, P., Hecky, R.E., 1973. Fluoride: Geochemical and ecological significance in east African waters and sediments. Limnol. Oceanog. 18, 932-945. McKenzie, J., Siegel, D., Patterson, W., McKenzie, J., 2001. A geochemical survey of spring water from the main Ethiopian Rift Valley, southern Ethiopia: implication for well head protection. Hydrogeol. J. 9, 265-272. Meng, S.X., Maynard, J.B., 2001. Use of statistical analysis to formulate conceptual models of geochemical behaviour: water chemical data from the Botucatu aquifer in Sa˜o Paulo state, Brazil. J. Hydrol. 250, 78–97. Parkhurst, D.L., Appelo, C.A.J., 1999. User’s guide to PHREEQC (ver. 2)—A computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations. US Geological Survey. WaterResources Investigation Report, pp. 99–4259. Pik, R., Deniel, C., Coulon, C, Yirgu, G., Hofmann, C., Ayalew, D., 1998. The northwestern Ethiopian Plateau flood basalts: Classification and spatial distribution of magma types. J. Volcanol. and Geothermal Res. 81, 91111. Plummer, L.N., Parkhurst, D.L., Thorstenson, D.C., 1983. Development of reaction models for ground-water systems. Geochim. Cosmochim. Acta. 47, 665–686. Plummer, L., Busby, J., Lee, R., Hanshaw, B., 1990. Geochemical modelling of the Madison aquifer in parts of Montana, Wyoming, and South Dakota. Water Resour. Res. 26, 1981–2014. Reimann, C., Bjorvatn,K., Frengstad, B., Melaku, Z., Tekle-Haimanot, R., Siewers, U., 2003. Drinking water quality in the Ethiopian section of the East African Rift Valley, part I: data and health aspects. The Sci. Tot. Env. 31, 65-80. Rozanski, K., Araguas-Araguas, L., Gonfiantini, R., 1996. Isotope patterns of precipitaion in the East African Region. In: Johnson, T.C., Odada, E. (Eds). The Climatology, Palaeoclimatoloy, Paleoecology of the East African Lakes, Gordon and Breach,Toronto, pp. 79-93. Schoell, M., Faber, E., 1976. Survey on the isotopic composition of waters from NE Africa. Geologisches Jahrbuch. 17, 197-213. Sonntag, C., Thorweihe, U., Rudolph, J., 1982. Isotopenuntersuchungen zur Bildungsgeschichte Saharischer Paläowässer. Geomethod. 7,55-78. Swanson, S.K., Bahr, J.M., Schwar, M.T., Potter, K.W.,2001. Two-way cluster analysis of geochemical data to constrain spring source waters. Chem. Geol. 179, 73–91. UN,1989. Ethiopia. In: Groundwater in Eastern, Central and Southern Africa. Natural Resources/Water Series No. 19, United Nations, New York, pp. 84-95. Varsanyi, I., Kovacs, L., 1997. Chemical evolution of groundwater in the River Danube deposites in the southern part of Pannonian Basin (Hungary). Appl. Geochem. 12, 625-636. Wang, Y., Ma, T., Luo, Z., 2001. Geostatistical and geochemical analysis of surface water leakage into groundwater on a regional scale : a case study in the Liulin karst system, northwestern China. J. Hydrol. 246, 223-234. Yemane, K., Bonnefille, R., Huges, F., 1985. Palaeoclimatic and tectonic implications of Neogene microflora from the Northwestern Ethiopian highlands. Nature , 318, 653-656. 97 L CS TS DW CS CS CS CS CS CS DW CS L CS R TS DW R TS DW L CS DW CS CS DW CS CS HDW CS L CS L CS HDW DW DW DW AR CS CS DW R R R CS HDW R CS CS CS HDW TS DW HDW DW Locality Wonchi Wonchi Wonchi Ambo-Awaro Duber Tibtab Abay Bridge Kurar,Dejen Kurkur Amanuel Bure Bure Baguna Zengena Zengena Gilgel Abay Wanzaye Wonzaye Abay Fall Andasa Kunzila Lake Tana Bahir Dar Bahir Dar Debretabor Tirtira Nefas Mewcha Workaye Boya Kebero Meda Sanka Hayk Gidirso Hardibo Angolela Menechera Chacha Wonto Sheno Dire Termaber Debre Sina Shoa Robit Jeweha river Atayae Borkena Tosa, Dessie Korke Gimba Harbou Chefa Fontanina Awash-Jima Woliso Goro Wolkite Locality Wolkite Code SK1 SK2 SK3 SK4 SK7 SK9 SK10 SK11 SK12 SK13 SK14 SK15 SK16 SK17 SK18 SK19 SK20 SK21 SK22 SK23 SK24 SK25 SK26 SK27 SK28 SK29 SK30 SK31 SK32 SK33 SK34 SK35 SK36 SK38 SK39 SK40 SK41 SK42 SK43 SK46 SK47 SK48 SK49 SK50 SK51 SK52 SK53 SK54 SK55 SK56 SK57 SK79 SK80 SK81 SK82 Code SK83 UTM N 971492 971401 969696 991200 1044859 1112265 1114488 1118000 1131748 1164821 1182942 1184475 1206900 1207789 1258000 1302829 1302829 1270000 1274056 1313579 1282130 1281517 1281000 1311755 1299941 1295397 1300708 1302315 1260251 1272746 1253500 1254013 1245000 1065581 1062598 1052374 1044425 1036471 1012616 1088521 1091179 1105060 1116152 1142751 1225635 1230501 1235974 1287421 1277463 1259198 1212945 963344 943265 928389 915480 UTM N 915860 UTM E 377438 376889 376889 378500 486203 415243 410962 406888 381952 336873 287178 288084 277200 276609 286100 355132 355100 345800 334411 285598 324600 312595 324500 392610 421690 439865 472879 506903 546632 550039 576000 573000 583000 547696 546632 551175 542088 530728 485155 580555 582739 598835 606940 605757 579914 568448 574863 573588 571397 568923 584265 418884 387830 375286 365553 UTM E 369097 δ18O 4.8 4.7 -1.9 -3.8 -2.9 -2.5 -1.2 -2.4 -1.5 -2.0 -3.5 4.9 -1.9 -1.0 -5.2 -1.1 3.0 -5.0 -2.1 3.3 -2.2 -2.7 -3.2 -4.3 -4.6 -3.4 -3.0 -3.9 -4.6 7.1 -1.4 5.9 -4.0 -3.8 -3.6 -3.2 -3.7 -2.5 -4.2 -4.1 -1.6 -0.9 -1.0 -0.5 -2.8 -1.8 -0.3 -1.5 -2.0 -1.1 -2.9 -5.3 -2.7 -1.1 δ18O -2.7 δD 34.6 33.4 23.6 -4.8 -10.2 -6.3 2.6 4.8 0.3 6.3 3.9 -4.7 37.2 -2.4 5.7 -21.5 2.2 25.9 -20.5 -0.8 30.3 1.9 -3.8 -2.6 -12.3 -12.1 -7.1 -7.0 -16.4 -15.7 46.4 -1.9 37.5 -15.5 -15.2 -15.1 -14.1 -10.7 2.2 -15.2 -14.6 -1.2 0.9 -1.1 1.2 -9.8 -1.8 5.5 -1.5 -4.5 0.1 -10.0 -25.7 -8.9 3.8 δD -8.1 δ13C 6.3 -6.6 1.0 pH PCO2 7.6 7.1 7.1 8.2 7.7 8.0 0.004 0.066 0.051 0.001 0.008 0.003 6.7 8.4 8.4 8.1 0.042 0.000 0.001 0.022 7.0 0.010 -15.7 9.2 6.5 0.000 0.028 4.8 6.5 6.7 0.907 0.039 6.1 6.4 5.8 6.8 7.9 6.5 6.7 0.111 0.040 0.102 0.014 0.002 0.022 0.010 6.9 0.065 6.4 0.051 7.4 7.3 0.004 0.017 7.5 0.008 8.0 7.5 7.6 7.8 6.8 0.002 0.016 0.004 0.013 0.027 pH 7.2 PCO2 0.011 -4.2 δ13C T 15.5 15.7 32.8 32.0 21.0 22.4 17.6 23.1 23.3 21.2 21.6 24 21.1 18.9 20.9 40.5 24.2 24.4 25.2 24.4 19.2 24.2 23.9 18.2 15.6 16.6 18.5 14.4 13.3 16.7 22.1 22.4 24.5 17.1 16.4 19.6 18.1 24.1 21.6 18.0 17.0 24.0 25.0 24.0 18.0 19.0 20.0 25.0 23.0 24.0 22.0 22.5 30.2 31.1 22.3 T TDS K Mg Na Ca HCO3 CO3 SO4 Cl F 309 1419 909 224 3538 725 5.7 19.8 12.0 1.8 7.2 3.0 0.3 1.3 30.4 6.1 149.0 38.7 69.5 348.4 76.6 17.5 62.2 53.5 2.2 4.1 93.0 26.3 670.8 81.7 163.5 856.4 564.9 134.2 442.9 244.0 0.4 0.8 0.5 1.1 4.1 1.8 0.7 0.1 10.1 1.9 2142.0 178.3 3.5 40.1 32.0 3.0 25.6 21.0 6.6 6.3 1.1 0.3 1.3 0.5 505 75 350 4035 4.8 0.1 1.4 14.8 21.6 2.5 15.5 336.0 11.0 2.3 7.0 426.0 74.2 6.7 37.7 1.2 201.3 29.3 213.5 3080.5 0.1 0.3 3.2 50.1 34.9 0.1 1.3 5.1 24.0 0.5 1.6 8.0 0.1 0.1 NO3 SiO2 RE SI C SI D SI G 13.2 2.6 9.2 59.2 56.8 141.8 75.5 29.0 23.7 43.6 2% 1% 1% 4% -4% 2% -1.23 -0.88 0.24 0.33 1.14 0.74 -3.5 -2.5 -0.2 -0.2 1.43 0.95 -5 -6.2 -2.5 -3.6 -0.1 -1.3 47.4 22.9 64.1 113.1 0% -3% -4% -6% -0.61 -0.66 0.89 -0.19 -2 -2 1.19 1.87 -2 -5.6 -3.6 -5.2 86.2 10.7 4.3 189 2.4 9.7 5.7 14.1 97.6 0.1 0.8 1.1 195 130 0.4 1.8 0.0 6.4 43.2 5.4 0.6 15.3 44.0 87.0 35.0 0.0 2.1 2.4 2.6 2.4 0.4 0.2 10.7 47.1 0% -1.2 -2.8 -4.2 66.2 9.3 -3% 0% -0.77 -1.7 -3.2 -4 -5.1 -3.6 4719 325 28.3 1.1 461.6 14.8 530.8 10.1 22.9 31.2 3500.0 200.0 1.8 0.1 18.4 0.8 30.2 3.6 0.2 1.5 123.7 63.7 4% -3% -0.5 -0.87 0.1 -2.3 -3.5 -3.9 231 192 179 310 234 191 111 1.2 0.4 0.9 11.0 2.2 0.6 0.8 10.8 11.3 9.0 8.1 5.5 7.5 3.8 8.4 5.9 6.0 14.8 25.8 4.5 3.1 18.6 16.6 23.0 37.0 23.3 27.0 14.5 130.0 98.0 60.0 85.0 145.0 62.0 45.0 0.0 0.0 0.0 0.0 0.7 0.0 0.0 0.8 5.0 5.4 18.9 2.9 2.4 1.7 1.0 4.4 21.2 27.6 3.9 9.0 9.5 8.9 9.6 32.7 43.0 2.8 46.8 12.6 51.4 41.0 20.8 64.9 21.3 31.4 20.4 -3% 1% -1% 3% 3% 2% -2% -1.89 -1.74 -2.44 -1.1 0.07 -1.67 -1.83 -4.2 -3.9 -5.5 -3.1 -0.7 -4.1 -4.4 -4.1 -3.3 -3.1 -2.4 -3.4 -3.4 -3.8 697 1.7 24.4 31.4 92.4 450.0 0.2 7.4 17.2 7.4 64.8 1% -0.03 -0.9 -2.6 237 1.5 8.9 6.0 29.4 120.0 0.0 3.0 7.6 19.5 40.8 -1% -1.44 -3.6 -3.3 115 523 1.0 3.9 3.8 14.8 5.2 47.0 15.0 53.1 85.4 329.4 0.1 0.4 0.9 42.7 0.9 22.6 0.1 0.8 2.7 8.0 -7% -8% -0.88 0.08 -2.6 -0.6 -4 -2 291 0.4 7.9 16.6 40.2 212.5 0.4 4.4 4.0 0.2 4.2 -5% -0.03 -1 -3 394 706 204 1079 242 1.2 0.8 2.9 11.9 8.6 10.7 38.0 3.8 4.9 75.2 73.5 16.3 262.3 18.2 16.9 54.7 26.2 1.6 28.9 230.0 480.0 149.0 732.0 176.9 1.4 1.0 0.3 2.7 0.1 39.3 15.9 3.2 0.1 3.2 10.3 12.8 2.0 39.2 1.5 0.5 8.6 29.3 1% 0% -4% -4% -3% 0.15 0.4 -0.17 -0.71 -0.84 -0.1 0.44 -1.4 -0.4 -2.7 -2.5 -2.5 -3.3 -6.5 -3.3 0.4 26.6 0.0 2.3 0.0 TDS 241 K 7.6 Mg 5.2 Na 19.4 Ca 24.2 HCO3 181.0 CO3 0.2 SO4 1.0 Cl 0.8 F 1.0 NO3 0.5 RE -7% SI C -0.51 SI D -1.9 SI G -3.9 0.0 0.1 0.3 0.0 SiO2 98 DW CS HDW CS HDW DW TS HDW HDW CS HDW DW CS HDW CS TS R CS DW L DW CS DW CS HDW HDW DW L CS CS R L L L L R TS DW DW R CS CS DW CS R CS DW CS L L L HDW HDW CS DW DW DW DW Abelti Sokoru Lalo Agaro Bedele Bedele Dedesa Dedesa Arjo Kewisa Nekempt Nekempt Bako Ijaji Gedo Ambo Ginchi Jiga Kuarit Zengena Mandura Mandura Jawi Jawi Jawi Gilgel Beles Chagni Tirba Tirba Zengena Gilgel Abay Tana Tana Tana Tana Gumarat Wanzaye Wanzaye Woreta Arno Ferenjua Azezo Azezo Infraz Tisabay Kolit-Adet Adet town GishAbay Gudera Tach Bahir Lay Bahir Mertulemariam Debre Work Hayk Kalu Metekel Wuchale Locality Yilmandens SK84 SK85 SK88 SK89 SK91 SK92 SK93 SK94 SK95 SK96 SK97 SK98 SK99 SK100 SK 101 SK102 SK103 SK104 SK105 SK106 SK107 SK108 SK109 SK110 SK111 SK112 SK113 SK114 SK115 SK116 SK117 SK118 SK119 SK120 SK121 SK122 SK123 SK124 SK125 SK126 SK127 SK128 SK129 SK130 SK131 SK132 SK133 SK134 SK135 SK136 SK137 SK138 SK139 SKsp MP1 MP2 MP3 Code MP4 903928 876830 862353 868826 934826 935494 955235 960766 969576 990014 1005486 1005787 1008101 994776 997452 992850 997941 1040679 1040673 1206900 1227800 1227800 1244790 1280000 1280000 1234960 1214430 1197500 1196000 1207789 1258000 1282130 1282130 1282130 1282130 1302829 1302829 1302829 1334584 1346000 1362000 1387515 1380000 1350000 1270000 1260000 1245100 1215213 1204773 1210560 1210500 1200330 1204590 1250300 1223920 1214430 1246326 UTM N 1230922 342170 325020 248457 233855 208762 209397 210783 214923 222936 223120 231602 232505 286359 316217 328735 373740 404143 372269 322672 277200 218600 218600 213000 225000 225000 209860 231768 264250 265000 276609 286100 324600 324600 324600 324600 355132 355132 355100 361140 355700 348000 328546 334563 351500 345800 328100 332000 305120 306061 421140 421121 415000 368857 576000 553726 231680 567729 UTM E 328472 -1.8 -1.1 -0.9 -0.8 -1.3 -2.0 -3.3 -0.7 -1.7 -1.8 -0.9 -1.6 -1.4 -1.3 -2.7 -4.4 -0.9 -1.4 -2.1 5.7 -1.9 -1.4 -0.5 -1.2 -1.4 -0.6 -0.8 6.3 -0.7 -1.5 -2.3 4.9 4.9 4.9 4.9 -1.1 -4.7 -2.2 -5.4 -1.3 -1.3 -1.2 -0.1 -0.6 1.6 -1.6 -2.3 -2.3 2.8 7.7 8.4 -2.1 -2.8 -1.5 δ18O -2.4 5.4 6.7 8.3 5.5 1.5 -9.8 8.6 5.2 2.2 9.8 5.4 5.7 4.8 -4.0 -20.3 6.0 1.8 -2.1 39.5 0.0 3.5 9.4 5.1 2.9 8.5 5.1 44.6 8.1 4.7 -1.7 37.9 38.0 37.8 37.7 8.9 -20.7 -2.6 -27.2 6.3 4.1 3.4 15.7 8.3 22.9 1.0 -5.0 0.4 32.2 48.3 56.9 -3.3 -5.9 -1.7 δD 7.3 0.014 6.1 5.6 6.8 0.039 0.020 0.265 7.9 5.8 0.000 0.037 6.3 6.1 6.0 0.023 0.091 0.037 1.5 6.5 0.360 -11.6 7.3 0.008 7.1 6.7 0.026 0.058 6.6 0.046 7.6 0.004 7.7 0.001 23.0 22.0 20.7 20.3 18.8 19.0 28.5 20.8 16.8 18.5 17.8 17.1 19.3 17.4 15.1 36.0 22.6 22.4 19.6 372 6.4 14.8 12.4 56.0 245.0 0.3 0.1 11.4 0.6 24.6 1% -0.03 -0.9 -4.9 70 16 2614 1.7 0.2 30.2 1.4 1.4 14.1 1.9 0.6 728.0 11.5 1.5 16.5 46.0 7.0 1650.0 0.0 0.0 0.9 0.8 0.3 95.5 1.2 1.1 36.7 0.1 4.6 5.3 4.2 37.6 -5% 0% 5% -2.48 -4.67 -0.48 -6.1 -9.6 -1.2 -4.2 -5.5 -2.5 106 39 11.0 1.1 3.0 1.6 11.6 1.4 8.8 5.8 6.0 21.0 0.0 8.0 0.4 21.9 0.9 0.1 35.3 6.5 1% 3% -1.76 -3.42 -4.2 -7.6 -3.3 -4.7 65 258 249 1.5 3.0 7.0 2.7 7.9 7.6 1.6 19.0 20.6 10.3 32.0 26.8 41.0 96.0 36.6 0.0 0.0 0.0 0.1 4.2 1.2 2.3 25.2 26.8 0.1 0.2 0.1 5.0 70.4 122.4 1% -6% -5% -2.41 -1.87 -2.39 -5.6 -4.5 -5.5 -5.4 -3.1 -3.8 1581 31.5 29.8 268.1 45.7 1172.0 0.3 0.1 27.9 0.9 4.8 -8% -0.36 -1.1 -5.2 207 1.4 13.4 7.3 23.0 158.6 0.2 0.1 0.9 0.1 2.2 -1% -0.52 -1.5 -4.8 20.2 22.2 23.1 21.0 22.1 28.6 23.0 18.0 384 362 1.7 1.2 10.5 12.4 7.4 8.5 71.0 63.0 290.0 271.0 0.2 0.1 1.8 4.0 0.3 1.2 0.2 0.2 0.7 0.8 1% 0% -0.06 -0.52 -1.2 -1.9 -3.2 -2.9 204 1.2 9.1 8.4 24.5 160.0 0.0 0.3 0.3 0.1 0.5 -6% -1.22 -3.1 -4.3 158 0.9 3.8 17.9 17.4 116.0 0.2 1.2 1.0 0.2 0% -0.51 -1.9 -3.9 17.2 88 2.5 8.2 5.7 1.5 59.6 0.2 0.6 1.1 9.1 -5% -1.69 -2.9 -5.2 0.007 32.4 20.0 22.1 19.9 19.9 21.7 19.2 21.2 141 1.5 1.5 9.0 21.4 103.0 0.1 1.4 2.3 1.0 -5% -0.8 -3 -3.7 0.024 0.001 18.4 20.2 16.1 461 68 1.6 1.0 18.3 2.5 21.6 1.9 65.7 8.6 281.0 45.0 0.2 0.1 0.1 0.9 13.6 0.6 58.8 7.1 -2% -9% -0.11 -1.06 -1 -2.9 -4.9 -4.2 600 512 263 TDS 396 1.5 1.3 0.7 K 1.1 22.4 34.0 10.0 Mg 17.5 37.4 29.0 6.9 Na 25.5 80.2 62.4 47.2 Ca 46.5 439.2 375.0 183.0 HCO3 244.0 4.8 0.4 0.4 CO3 0.5 0.1 2.0 0.1 SO4 0.1 14.2 4.0 3.0 Cl 11.3 0.0 4.0 12.0 NO3 49.0 -2% 6% 3% RE -2% 1.25 0.09 0.12 SI C 0.11 1.75 -0.3 -0.6 SI D -0.4 -4.8 -3.3 -4.9 SI G -5 20.2 19.8 19.6 20.6 7.2 7.1 7.7 0.1 20.6 16.5 16.0 17.4 δ13C 8.2 7.2 7.6 pH 7.5 0.003 0.024 0.005 PCO2 0.008 T 0.0 0.1 0.1 F 0.4 SiO2 32.8 99 DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW DW HDW DW DW DW DW Dembia Didessa Debretabor Gidakiremu Goma Gondar Gondar BH3 Gondar BH2 Gondar BH4 Gondar BH5 Debre Markos Debre Markos 12 Debre Markos 4 Debre Mardos 5 Debre Markos Debre Markos 8 Debre Markos 6 Chagni BH4 Mankush BH4 Mota BH4 JabiTehinin JabiTehinin JabiTehinin JabiTehinin Dangila BahrDar Ambo Guder MP5 MP6 MP7 MP8 MP9 MP10 MP11 MP12 MP13 MP14 MP15 MP16 MP17 MP18 MP19 MP20 MP21 MP22 MP23 MP24 MP25 MP26 MP27 MP28 MP29 MP30 MP31 MP32 1349950 960766 1311750 1076886 834629 1387900 1382400 1386300 1387000 1388000 1137100 1138500 1137100 1141301 1141301 1142701 1141301 1214100 1263129 1226721 1204500 1204000 1204100 1204316 1270131 1281000 992880 992866 269459 214900 392680 198042 238652 334563 334470 334400 329990 329980 359080 359080 357680 364681 361881 360480 360480 231460 153232 336082 265400 265258 265258 265258 275060 324990 373660 360480 6.7 7.9 7.0 7.7 8.2 8.2 8.2 8.2 8.2 8.6 8.3 8.2 7.8 7.9 8.3 7.1 7.3 6.8 6.9 7.6 6.5 6.3 6.6 6.5 8.0 7.8 6.7 8.4 0.026 0.002 0.015 0.002 0.001 0.001 0.002 0.002 0.002 0.001 0.001 0.001 0.003 0.002 0.001 0.017 0.013 0.019 0.068 0.006 0.043 0.104 0.059 0.019 0.002 0.010 0.228 0.001 188 195 199 135 287 309 426 348 462 369 247 255 239 244 223 282 358 134 730 319 160 258 273 82 259 893 1652 322 3.0 2.8 4.5 1.7 6.6 5.0 0.7 0.7 1.2 1.3 1.4 1.5 2.0 1.4 1.8 2.6 5.0 0.7 1.7 1.1 2.0 3.2 2.5 1.2 1.4 7.3 27.5 6.5 3.7 4.4 7.8 5.8 4.0 5.0 1.0 10.0 1.0 8.8 11.0 7.1 6.1 8.3 1.9 15.6 13.6 7.8 22.5 10.0 7.8 13.6 12.7 2.0 4.9 31.1 39.4 5.8 13.0 14.5 18.0 6.0 40.8 40.8 110.0 43.0 120.0 78.2 19.0 26.0 27.2 19.0 60.8 9.5 28.9 7.4 45.0 11.0 6.7 8.8 9.0 4.6 48.4 160.5 314.0 35.0 28.9 26.4 20.0 22.4 25.7 25.7 4.8 32.0 5.5 12.8 33.3 28.1 24.0 30.1 3.2 38.5 40.1 13.6 108.6 54.4 21.6 35.2 39.2 12.0 13.2 25.7 52.9 46.4 122.0 142.0 143.0 96.0 170.8 195.2 293.0 244.0 317.2 244.0 164.7 183.0 164.7 177.0 146.4 207.4 256.2 97.6 533.0 229.4 114.7 195.2 207.4 58.0 183.0 634.4 1159.0 194.0 0.0 0.6 0.1 0.3 1.6 1.8 2.7 2.3 2.9 12.0 1.9 1.7 0.6 0.8 1.6 0.2 0.3 0.0 0.3 0.6 0.0 0.0 0.1 0.0 1.0 2.8 0.5 3.1 0.1 0.1 0.1 0.1 24.0 23.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 1.0 1.0 1.0 1.0 0.5 0.1 0.1 13.0 24.0 17.1 4.0 4.0 2.0 12.8 11.3 14.0 14.0 14.0 10.6 10.6 7.1 7.1 7.1 7.1 7.1 14.2 4.0 12.0 7.0 1.0 1.0 1.0 3.6 7.1 24.1 42.5 6.0 0.1 0.1 0.1 0.1 0.3 0.5 0.0 0.0 0.0 0.4 0.4 0.0 0.4 0.5 0.4 0.2 0.0 0.0 0.0 0.0 0.2 0.2 0.2 0.0 0.1 0.5 0.5 0.0 0.0 0.0 1.8 0.7 0.4 0.3 0.0 1.7 0.0 1.0 4.8 0.2 6.4 0.0 0.0 1.1 0.1 2.6 7.0 4.8 4.8 0.1 0.0 0.1 0.1 6.4 2.6 1.3 39.0 21.0 13.0 21.0 53.0 25.0 25.0 25.0 24.0 19.5 30.0 52.3 160.0 18.0 -2% -2% 1% 7% -2% -6% -2% -2% 1% 1% 4% -2% -2% -2% 6% 1% -1% -3% 0% 0% 1% 1% 0% -3% 1% -1% 1% 6% -1.13 0.05 -0.93 -0.35 0.43 0.49 -0.07 0.67 0.01 0.67 0.63 0.51 0 0.23 -0.4 -0.42 -0.12 -1.49 0.14 0.27 -1.52 -1.27 -0.93 -1.97 -0.01 0.49 -0.13 0.94 -3.4 -0.9 -2.5 -1.5 -0.1 0.06 -1 0.63 -0.9 0.98 0.58 0.23 -0.8 -0.3 -1.2 -1.4 -0.9 -3.4 -0.6 -0.4 -3.7 -3.2 -2.6 -4.9 -0.7 0.86 -0.6 0.77 -5.1 -5.1 -5.2 -5.2 -2.5 -2.5 -5.9 -5.1 -5.9 -5.5 -5.1 -5.1 -5.2 -5.1 -6 -5 -5 -5.4 -4.7 -3.6 -3.9 -3.7 -3.7 -4.4 -5.4 -5.4 -2.8 -2.3 Annex I. Water chemical and isotope data. Major elements and silica in mg/L and isotopes of water in per mil (‰) calibrated against V-SMOW, δ13C in per mil calibrated against PDB. RE is reaction error. Saturation indices are also given for calcite (SIC), dolomite (SID) and Gypsum (SIG). Samples with code MP1-MP32 are from BCEOM, 1999. The coordinates are in UTM. 100 Introduction 18 Les isotopes stables de l’eau (δ O, δD) ont été largement utilisés comme méthode alternative dans les études hydrologique et paléohydrologique sur les lacs. Il s’agissait en particulier de tracer de manière qualitative l’écoulement souterrain autour des lacs, d’estimer le taux d’évaporation ou de quantifier les flux souterrains. Le traçage des interactions lacs-eaux souterraines utilise le fait que suite aux pertes par évaporation, l’eau des lacs est enrichie par rapport aux précipitations actuelles et aux eaux souterraines adjacentes. Les eaux souterraines amont et celles qui sont partiellement ou complètement alimentées par de l’eau lacustre peuvent être identifiées en considérant leur enrichissement isotopique. Ceci peut être mis en évidence par un simple diagramme δ18O-δD ou par une carte de répartition spatiale des teneurs isotopiques autour des lacs. Il sera alors possible d’identifier les directions de l’écoulement souterrain autour du lac. L’estimation quantitative des flux souterrains repose sur la possibilité, en combinant le bilan hydrologique et les équations du bilan isotopique, d’obtenir deux équations et deux composants du bilan hydrologique inconnus. Une solution mathématique est alors possible pour séparer les flux souterrains inconnus. Dans les cas où ces flux sont négligeables ou pré-déterminés par des approches physiques, la méthode a été utilisée de manière alternative pour estimer l’évaporation en surface. Pour la reconstruction paléoclimatique à partir des sédiments lacustres on utilise la capacité de certains sédiments (eg. carbonates, cellulose, silice des diatomées) de conserver une trace isotopique de l’eau du lac dans laquelle ils se sont formés. Ces archives paléo-isotopiques peuvent être interprétées en termes de paléo-hydrologie (paléo-précipitation, paléo-évaporation/flux entrant, paléo-temperature etc.) du lac. L’applicabilité de la méthode isotopique pour tracer les interactions lacs-eaux souterraines, pour quantifier le bilan hydrologique ou pour interpréter les archives paléoclimatiques, n’est cependant pas universelle et dépend des conditions spécifiques du site. Ces conditions comprennent : le climat, l’hydrographie, la salinité, la taille du bassin, etc. En prenant en considération les conditions spécifiques de l’Ethiopie, l’appendice permet : 101 • De montrer les avantages et les contraintes d’utilisation de la méthode du bilan sous les conditions de l’Ethiopie, • De calculer une droite d’évaporation locale hypothétique (DELC) et la composition hypothétique des lacs de rapports évaporation/flux d’entrée donnés sous les conditions météorologiques d’Ethiopie, • De comparer la composition isotopique de quelques lacs éthiopiens sélectionnés avec la composition isotopique calculée, • D’interpréter la comparaison en termes de bilan hydrologique des lacs et les facteurs non climatiques qui font dévier les compositions isotopiques par rapport aux valeurs hypothétiques calculées. • De recommander les techniques de comparaison entre la composition isotopique mesurée et la Droite d’Evaporation Locale d’Ethiopie comme un outil d’estimation rapide de l’Hydrologie des lacs peu connus de la région. La méthode dV = I −Q−E L’équation du bilan hydrologique: dt Vdδ + δdV = Iδ I − Qδ L − Eδ E dt Et l’équation du bilan isotopique: Sous des conditions de régime permanent peuvent être combinées pour relier les compositions isotopiques à l’Hydrologie des Lacs. L’application de cette méthode n’est pas simple du fait de la difficulté à estimer la composition isotopique de l’eau en train de s’évaporer (δE). δE Le modèle Craig et Gordon (1965) (α δ = ) − hδ A − ε (1 − h + 10 −3 ∆ε ) est largement utilisé pour estimer * L la composition isotopique du flux d’évaporation. Combinées ces équations peuvent s’écrire sous la forme : h δ * − δ SS 1 Ι = = x Ε 1 − h δ SS − δ I (1) 102 Où h est l’humidité, x le rapport évaporation/flux d’entrées (E/I), δ* = (hδA+ε)/ (h-10-3ε) ≈ δA+ε/h est la composition isotopique limite qui correspond à l’enrichissement isotopique qui pourrait être approchée par les eaux résiduelles d’un volume d’eau se réduisant sous l’effet de l’évaporation, δSS la composition isotopique en oxygène ou en deutérium en régime permanent, du lac, δI est la valeur δ18O ou δD de l’eau d’alimentation du lac. La Figure 1 utilise l’équation (1) pour calibrer les 'relations composition isotopique – caractéristiques hydrologiques' de référence sous les conditions de l’Ethiopie 7 . Ensuite, la composition isotopique mesurée des lacs est comparée avec la Droite d’Evaporation Locale modélisée et la composition hypothétique des lacs. Cette approche donne une possibilité pratique, peu onéreuse et simple de classification des lacs en différentes catégories hydrologiques sur la base de leurs compositions en δ18O et δD. Ces différentes catégories comprennent : les lacs à flux de sortie dominant, à évaporation dominante, à diminution de volume ou eau souterraine dominantes. Le Tableau 1 indique la manière d’utiliser la figure pour interpréter les compositions isotopiques des lacs éthiopiens en rapport avec leur hydrologie. 90 80 70 60 δ D SS 50 Shrinking Lakes 40 30 Evaporation dominated Lakes 20 10 0 Outflow dominated lakes -10 -6 -4 -2 0 2 4 6 δ OSS x= 5 8 10 12 14 16 18 x=0 x=0.25 x=0.5 x=0.75 x=1 x= 10 δ∗ Figure 1. DEL calculée pour différentes valeurs de x sous les conditions hydroclimatiques de l’Ethiopie. Sous les conditions qui prévalent en Ethiopie et avec la composition des eaux météoriques d’Ethiopie, la DEL est définie par la droite δD=5.3δ18O + 6.2. La composition isotopique mesurée des lacs peut être comparée avec le graphe pour visualiser et examiner rapidement les rapports entre évaporation et flux d’entrée. 7 Sous les conditions éthiopiennes les estimations sont: -2.5 et -5 ‰ pour la composition en δ18O et δD des flux d’entrée; la composition isotopique limite est estimée respectivement à 14 et 82 pour δ18O et δD, l’humidité est estimée à 0.6 103 δ18O δD x (E/I) Classe de lac <4.5‰ <27‰ < 0.5 Le flux de sortie prédomine sur l’évaporation. La majeure partie des pertes correspond au flux de sortie 4.5 27 Flux de sortie égal à l’évaporation. Flux d’entrée également réparti entre évaporation et flux de sortie. 4.56.0 27-36 0.5-0.75 7.2 >7.2 42 >42 1 Doivent être calculés en tenant compte des données spécifiques du site 14 82 " >14 >82 " autre autre x δ18O ≠ ≠x δD 0.5 L’évaporation dépasse légèrement le flux de sortie, ou une part importante du flux d’entrée est perdue par évaporation. Lac terminal Lacs avec effets hydrographiques ou connectés. Composition isotopique pas entièrement fonction de leur hydrologie mais aussi des caractéristiques de leur basin versant • Lacs avec des mares en amont • Lacs avec un rapport surface du basin, surface du lac important • Lacs remplis par des lacs amont Composition maximum pour un lac à bassin simple Petits lacs avec un flux d’entrée amont évaporé à l’amont, à leur extension la plus faible Lacs salés affectés par l’effet des sels dissous Quelques exemples de cette étude Lac Tana après saisons humides Petits réservoirs artificiels (Koka, Dire). Mares saisonnières après saison humide Lac Tana à l’échelle annuelle Lac Wonchi Lac Zengena, Lac Tirba Lac Tana en saison sèche, Lac Hora Lacs Gamari, Hayk, Abijata, Lac Gamari Lac Afrera Table 1. Relation entre la composition isotopique des lacs en régime permanent et leurs caractéristiques hydrologiques sous les conditions de l’Ethiopie. Quelques exemples sont également indiqués. Quelques résultats 1. La comparaison entre les compositions isotopiques mesurées et estimées, des lacs actuels, montre qu’en plus des facteurs climatiques tels que l’humidité, la composition isotopique de la vapeur ambiante, et la température de l’air, un certain nombre d’autres facteurs non climatiques comme la présence de volumes d’eau à l’amont, le rapport des surfaces entre le lac et son bassin versant, la salinité des lacs, la variation saisonnière de la composition isotopique du flux entrant des eaux météoriques et les caractéristiques hydrologiques des lacs, joue un rôle significatif sur le 104 régime isotopique des lacs éthiopiens (table 1). Ceci est un élément important pour choisir un lac adapté aux études paléo-hydrologiques et pour la gestion des ressources en eau. 2. En utilisant les valeurs moyennes de la composition isotopique des lacs, les données pluviométriques et sur l’écoulement de surface, et les estimations de l’évaporation à partir d’autres méthodes physiques, l’index de la perte par évaporation est calculé pour des lacs à hydrographie simple. Le Tableau 2 donne une première estimation du bilan hydrologique de ces lacs. Les résultats de l’estimation du bilan isotopique montrent que l’eau souterraine joue un rôle significatif dans l’hydrologie des lacs éthiopiens. Lac δ18Omoy x I E P+RI RO GI GO 0 Remarques Hora 7.2 1.00 1710 1710 1170 0 560 31% 0% Index lac terminal Bishoftu 6.9 0.92 1846 1710 997 0 849 46% 136 Babogaya 6.7 0.87 2043 1710 1022 0 1022 50% 333 16% Awassa 6.9 0.81 2168 1750 1860 0 308 14% 418 19% δ18O I= -1.5‰ Hardibo 7.1 0.85 2011 1710 1912 0 100 5% 300 15% δ18O I= -1.5‰ Langano 6.6 0.82 2213 1860 0 Ziway 6.2 0.74 2508 1860 2094 320 400 16% 328 13% Zengena 4.9 0.60 2754 1300 2200 0 500 18% 1450 53% Tana 4.8 0.52 2827 1478 2610 1113 217 8% 137 8% 5% Table 2. Bilan isotopique annuel en isotope de l’oxygène de quelques lacs sélectionnés pour leur hydrographie simple, le composant eau souterraine est également reporté en pourcentage du flux entrant total ou de la perte totale par évaporation. Pour le lac Awassa le bilan isotopique est estimé en supposant un δ18OI de -1.5‰. 3. Des deux isotopes de l’eau, l’oxygène est le plus largement utilisé pour l’étude des bilans isotopiques des lacs, et on considère qu’il donne de meilleurs résultats (eg. Rozanski et al., 2004 parmi beaucoup d’autres exemples). Les lacs éthiopiens montrent également quelques désaccords entre les bilans effectués avec l’oxygène et avec le deutérium (figure 2). Ce relatif désaccord a beaucoup d’implications sur les bilans isotopiques des lacs éthiopiens et de beaucoup d’autres lacs: • Comme cela est démontré de manière théorique dans l’appendice 3 le modèle d’évaporation de Craig et Gordon pose un problème théorique pour estimer la composition isotopique du flux d’évaporation. • Le bilan isotopique est plus performant pour indiquer rapidement l’état hydrologique des lacs que pour obtenir de manière précise les rapports E/I. • Les lacs étendus et les petits lacs se comportent différemment et l’application du modèle de Craig et Gordon ou le bilan isotopique nécessitent en général des adaptations et des modifications spécifiques au site. 105 • Sous les conditions actuelles en Ethiopie, plutôt que d’utiliser l’approche isotopique pour estimer l’évaporation depuis les lacs, il est plus judicieux de calculer les flux d’eau souterraine sur leur pourtour. L’estimation du bilan isotopique en oxygène (table 1) est plus suggestive et préliminaire que définitive. • 3 E/I(δ D) 2 1 Lake Tana Small area lakes Salt lakes 0 0 1 E/I ( δ 18 O ) 2 3 Figure 2. Comparaison entre les rapports E/I déterminés sur les lacs éthiopiens à partir des isotopes de l’oxygène et de l’hydrogène. Les deux ne donnent pas le même résultat sur les lacs de faible surface, sur les lacs salés et sur les grands lacs. 106 Submitted to Chemical Geology Application of stable isotopes of water in hydrological studies of Lakes: cases from Ethiopia Seifu Kebede1,2* , Yves Travi1 1 Laboratory of Hydrogeology, University of Avignon, 33 Rue Louis Pasteur, 84000, Avignon, France 2 Department of Geology and Geophysics, Addis Ababa University, POBox: 1176, Addis Ababa, Ethiopia 1. Background Stable isotopes of water (δ18O, δD) were widely used as an alternative method of lake hydrological and paleo-hydrological studies. The method has been used in qualitatively tracing groundwater flow around lakes, in estimating open water evaporation rate or in quantifying groundwater flux around lakes. Lake Groundwater interaction tracing around lakes uses the advantage that, owing to evaporative water loss, lakes are enriched with respect to the present day rainfall and with respect to groundwaters adjacent to them. Up gradient groundwaters and groundwaters that are partially or completely fed by lake water can then be identified based on the degree of their isotopic enrichments. Direction of groundwater flow around lakes can be inferred from a simple δ18O-δD plot or from a spatial plot of these isotopes around lakes. Quantitative estimation of groundwater flux around lakes uses the advantage that by combining the water budget and the isotope budget equations we will be left with two equations and two water budget components. This allows a mathematical solution to separate the unknown groundwater fluxes. This allows the net groundwater flow that can not be determined from simple hydrologic equation to be separated into inflowing and outflowing components. In cases where groundwater flux around a lake is negligible or predetermined by physical approaches, the method was used as an alternative method of estimating open water evaporation. Paleo-climate reconstruction from lake sediments use the advantage that some sediments (eg. carbonates, cellulose, diatom silica) register the paleo isotopic composition of the paleo-lake water in which they form. The paleo-isotope record can then be interpreted in terms of paleo * Correspondence to: S. Kebede, email: [email protected] 107 hydrology (paleo precipitation, paleo evaporation to inflow flux, paleo temperature etc.) of the lake. Although stable isotopes of water were widely used as an alternative method of water balance studies of lakes, the method is not widely used in Ethiopian Lakes hydrological studies. Some previous studies used the method to trace the direction of groundwater flow around lakes (Darling et al., 1996; Ayenew, 1998; Chernet et al., 1998; Kebede et al., 2002). This part first calculates a reference local evaporation line with which the isotopic compositions of the lakes in the region is compared. From the comparison one can infer the the present day hydrology of the lakes. One can also get how exactly we can interprate the paleo isotope records from the lakes in terms of paleo climate and hydrological variations. The departure of the lakes isotopic composition from the reference line will also provide useful information (eg. selecting a suitable lake for paleoclimate study, understanding how the present day lakes behave etc) for paleo climate studies that use lake sediments as proxies. In these regard this chapter does the following, • Presents the isotopic characteristics of selected Ethiopian Lakes • Briefly describes the theoretical background of the isotope balance method and computes a reference hypothetical local evaporation line. • Discusses the conveniences and limitations of the isotope balance method in hydrological study of Ethiopian lakes • Uses isotope method to trace groundwater flow direction around lakes • Compares the measured isotopic composition of Ethiopian lakes with the calculated LEL • Compares the results of oxygen isotope balance method with that of hydrogen isotope balance method • Discusses the application of the isotope balance method in computing the evaporation to inflow ratios of the selected Ethiopian lakes 2. Isotopic characteristics of the Lakes 2.1. Location of the lakes and their general isotopic composition The studied Lakes (figure 1) include 1) Lakes in the Northwestern Ethiopian Plateau (NWP) namely Lake Tana and the Plateau Crater Lakes (Zengena, Tirba and Wonchi); 2) Lakes in the Main Ethiopian Rift (MER) namely the Ziway-Shalla Lakes, the Bishoftu Crater Lakes (BCL), the Awassa Lake; 3) Lakes in the Afar Depression namely the Afrera Salt Lake and the Gamari Flood Plain Lakes. The other lakes include the Hayk-Hardibo Volcanic Crater Lakes. 108 Isotopic characteristics of the lakes water and the surrounding groundwaters have been measured since 1970s. The average δ18O and δD compositions of the Lakes along with of topographic, hydrographic, geochemical characteristics are presented in table 1 and figure 2. The lakes, in addition to playing a critical role in water and aquatic resources in Ethiopia, they contain useful sediment archives for paleo-climate reconstruction. All the lakes are isotopically enriched (table 1, figure 2) compared to the composition of modern precipitation and groundwaters adjacent to them. This reflects effects of evaporative isotope enrichment. Despite comparable evaporation amount and ambient environmental conditions (such as humidity, isotopic composition of ambient vapor) the closed Lakes in Ethiopia are enriched than the other closed lakes of Tropical East Africa. The relative enrichment of the Ethiopian lakes reflects the relative enrichment of their inflows (rainfall waters and groundwaters). The relative enrichment of the Ethiopian rainfalls compared to the East African rainfalls is a known phenomenon (Rozanski et al., 1996). 109 Figure 1. Location map showing the studied Lakes. The depression that divides the Southeastern plateau and the North western plateau is the Ethiopian Rift Valley. 110 Plateau Crater Lakes Ziway Shalla Lakes Bishoftu Crater Lakes Afar Depression Hayk-Hardibo Lakes Tana Awass Tirba Wonchi Ziway LanganoAbijata Shalla Hora Zengena a Lakes Gamari Afrera Hayk Hardibo >80 88.2 64 70 22.8 16 0 2.9 2.3 3 8.1 7.13 BabogayBishoftu a Max. Depth 14 23 148 Mean Depth 9 10 78 9 47 10 266 40 65 89 70 30 4 20 5 103 17.5 38 55 92 0.5 4.4 442 241 159 329 1.03 0.58 0.93 28000 900 39 249 1700 4800 800 34000 18 22 52 4.3 14 0.4 2 14.1 7.7 61 10.4 0.61 0.51 0.42 6 7 8 40 2 Lake A(Km2 3156 Volume 153 <2 (106m3) AC / AL 18 Mean δ O Mean δ D 4.76 17 8 6.9 34.717 43.78 3 5.0 1 5.6 37.23 39.5 4.1 2 9 11 6.2 6.6 8.4 7.55 292 42.86 47.17 55.59 54.43 7.1 ∞ 2 6.4 10 6.7 6.9 15.5 37.88 4040 42.92 92.5 27.910 513 493 2.3 0.8 1.5 10 181 0.8 0.5 1 TDS (g/L) 0.13 0.77 0.08 0.1 0.22 0.58 1.8 34.6 23.7 E mm/yr 1500 1400 1300 130 1300 1780 1780 1780 1780 1710 1710 1710 3000 3000 1500 1500 1 8 1 1 1,2, 3,6,7 1,2, 3,6,7 1,2, 3,6,7 1,2, 3,6,7 9, 10 9,10 9,10 1, 5 4 1 1 0 Source of 1 data Table 1. Morphometric and physicochemical characteristics of the selected lakes, numbers in superscript are the number of isotope data used to calculate the average composition. Source of data: 1.This work 2. Craig et al., 1977; 3. Chernet, 1998; 4. Gonfiantini et al., 1973; 5. Hailemichael et al., 2002; 6. Ayenew, 1998; 7. Ayalew, 2003; 8. IAEA TC project unpublished reports 9. Kebede et al., 2002; 10. Kebede, 1999. 111 The δ18O composition of the Ethiopian Lakes varies between +4 ‰ and +16 ‰ (figure 2b). The best fit regression line that connects the isotopic composition of the Ethiopian Lakes (The Local Evaporation Line, LEL) intersects the Addis Ababa Local Meteoric Water Line (AALMWL) at -2.5‰ δ18O and -5‰δD. This is similar to the isotopic composition of the weighted average summer rainfall at Addis Ababa GNIP station. This reflects the major input to the lakes comes from the summer rainfalls (June-July-August-September) than from the spring rainfalls (March-April) which have more enriched compositions (0‰ in δ18O and 12‰ in δD). Generally lakes in the NWP are depleted that the MER and the Afar lakes. Despite the extremely high evaporation, Lake Afrera does not show commensurate isotope enrichment. The shallow seasonal lakes at the terminus of the Awash River are the most enriched lakes in the region. 200 150 δD 100 Ethiopian Groundwaters LEL 50 Lake Afrera 0 -50 -10 -5 0 5 10 15 20 -100 18 δ O Ethiopian Lakes Victoria Malawi Rukwa, Tanzania Spring rain average Kenay, Mountain Lake Naro Moro Nivasha, Kenya Rift Tanganyika Summer rain average AA LMWL 112 100 90 80 δD 70 60 50 40 30 20 2 4 6 8 10 12 14 16 18 18 δ O BCL Tana PCL Afrera Gamari Awassa Ziway Langano Abijata Shalla Hayk-Hardibo Figure 2a.and b. General trend in isotopic composition of Ethiopian lakes with respect to the Addis Ababa summer and spring rainfall isotopic compositions and the AALMWL. Comparison is also made with the average compositions of other East African Lakes (the majority of which are closed basin lakes). Data on other East African Lakes are from Begonzini et al. 2001; Gonfiantini, 1986; Rickettes, 1996; Arusei et al., 2003. The lower graph shows the δ18O-δD plot of Ethiopian lakes. 2.2. Isotopic and hydrologic characteristics of each Lake Lake Tana Tana is a shallow lake situated on the Northwestern Ethiopian Plateau. By surface area it is the largest lake (3200 km2) in Ethiopia. Direct precipitation on the lake accounts for about 50 % of total surface water inflow. Evaporation from the lake is estimated at 1500 mm/year. The lake is drained by the Abay River which contributes 7% of the Blue Nile flow at the Ethio-Sudan border. Seasonal stratification is weak owing to its shallow depth (Wudneh, 1998). Lake Tana is the most depleted of the studied Lakes (table 1, figure 3). Though it is not a pronounced variation as in high latitude lakes, the lake water shows seasonal variation in δ18O (figure 2b). The most depleted lake water (figure 3) is observed in October corresponding to the maximum lake level following the major rainy season (JuneSeptember). The lake water reaches its maximum isotopic enrichment in June which is the end of the long dry season and the start of the rainy season. The physical hydrology and a simplified water balance model of lake Tana is given in appendix 2. 113 S # 0.5 W S # E S 4.0 # S 4.4 S # 4.3 2.8 S # 4.4 #S#S # 3.9 S 1.7 4.0 #S 4.3 # S S # 4.3 S # S # 4.4 4.1 S# S # 4.2 3.7 S # -0.3 S # 5.2 S # 4.6 J ^ _ ` L a H O J M J T L b HO M N H c d M H J a W V e[ W W V eV W V W F e\ F e\ W V S # V W Ve\ S # Ve\ W E S # E W Ee\ # S E e\ W ] S # ] W ]e\ S # ] e\ W X # S X W Xe\ S # X e\ W \ e\ P Q#SR L S L T M U VE W VX VV W VE Y W VV Z W Y [ W Z \ W [ ] W \ E W ] F W E S # N 4.0 S# 3.7 # S # #S S 3.9 # S # # S S 4.4 -1.6 S # 3.6 S#3.7 3.6 # S 3.6 -1.1 S # # S S # 3.5 3.6 # S EF F 3.7 # S S # EF G HI J K L M L NO 10 18 δ O 8 6 4 2 0 N D J F M A M J J A S O N D J F M A M J J A S Figure 3. Monthly (November 2001 to September 2003) and spatial (September 2003) variation in δ18O of Lake Tana. Spatial δ18O sampling conducted in September 2003 shows that the lake is isotopically homogeneous (figure 3). The most depleted value is observed near the surface water inflow zones in the southern bay area. Generally the western and northern part of the lake shows the most enriched δ18O and the southern portion of the lake including bay area shows relatively depleted values. The southern bay area is slightly depleted compared to the northern part of the lake owing to low water residence time in the bay area. The bay area has shorter water residence time because the major inflow and the outflow zone are located in this area and the northern part of the lake water body is blocked by water inflows in the southern zone. The Bishoftu Crater Lakes The Bishoftu Crater Lakes are a group of five permanent lakes located on the western escarpment of the MER. Three of these lakes (from north to south: Babogaya, Hora and Bishoftu) are considered in this study. The lakes are roughly circular in shape, with areas between 0.6 and 1 km2. They have no perennial inlets or outlets, and they are presumably fed by groundwater (Kebede et al., 2002). Their open water evaporation is estimated at 1710mm using the Penman method (Kebede, 1999). The lakes 114 stratify briefly during the summer wet season (Prosser et al., 1968). Monthly and depth wise monitoring of the isotopic composition of Lake Babogaya (Kebede et al., 2002) shows no pronounced depth or seasonal variation in δ18O (figure 4) except the slight depletion (0.5 ‰) of the lake surface water in August 1999 due to dilution by rainfall . 8 7 δ 18 O 6 5 4 3 2 N D J F M A M J J A S O N D J F M A Month since January 1998 1 Meter 10Meter 60Meter Figure 4. Seasonal variation in δ18O of Lake Babogaya at three different depths; data from Kebede, 1999. Lake Hora, one of the lakes with the highest salinity (2.3 g/L) and with the most enriched δ18O is of particular interest. Chloride mass balance and physical hydrogeological evidence shows that the lake is not affected by a major groundwater outflow (Kebede et al., 2002). Furthermore the lake is located under an average meteorological condition of Ethiopia making the lake ideal index lake. This lake has been used as an index lake from which the isotopic composition of ambient vapor of the central and northern sector of Ethiopia has been previously estimated (Kebede et al., 2002). The Ziway-Shalla Lakes The Ziway-Shalla lakes are a chain of four large lakes (from north to south: Ziway, Abijata, Langano, Shalla). The lakes are located in the central sector of the Main Ethiopian Rift (MER). The lake basin is bounded from east and west by relatively humid plateau which is the source of the major rivers feeding the lakes. The lakes are characterized by a wide range of hydrological and bathymetric characteristics (table 1; figure 3). No present day surface outflow exists from the basin. Lakes Ziway and Langano are fed by rivers draining the escarpment bounding the rift valley. Both these lakes feed Lake Abijata via their surface water outflows. Lake Shalla is fed by two rift valley rivers and presumably by groundwater. Open water evaporation is estimated at 1800mm in the region (ValletCoulomb et al., 2000). Seasonal stratification is absent in lakes Abijata, Langano and Ziway owing to there shallow depths (Baxter et al., 1965). The two upstream lakes Ziway and Langano are relatively isotopically depleted compared to the downstream lakes Abijata and Shalla (table 1). Lake Ziway shows the highest seasonal δ18O isotope variation compared to Lakes Abijata and Shalla. This reflects that the seasonal variation in the isotopic composition of Lake Ziway inflows is subdued in the lake 115 before the lake overflows into downstream Lake Abijata. Despite its highest depth (266m) Lake Shalla does not show any major depth δ18O variations (Chernet, 1998). Lake Awassa Lake Awassa is located within a caldera south of the Ziway-Shalla Lakes. It is a closed-basin lake with no surface water outflow. Despite this the lake is dilute with salinity less than 1000 mg/L. Previous investigators (Darling et al., 1996; Telford et al., 1999, Wood and Talling, 1988) attribute the low salinity of this closed lake to groundwater seepage out of the lake. The lake is fed both by numerous streams on the northern and western slopes of the catchment. These streams drain swamps on the north-east side before they reach the lake. The lake is polymictic (Zinabu and Taylor, 1989), a suitable condition for the existence of an isotopically well mixed lake. Lake water isotope analyses conducted at different times since 1977 show δ18O compositions ranging between +5.4‰ and +8.2‰. The Afar Lakes In the Afar depression numerous high salinity lakes exist. Two of these are the Gamari flood plain lakes and the Afrera salt lake. The Gamari flood plain lakes receive their inflows from the Awash River. The Awash River before it reaches the Gamari flood plain is subject to evaporation in swamps and shallow lakes. The Gamari flood plain lakes are the most enriched lakes in Ethiopia (table 1). During the wet season the δ18O is +12‰. In the dry season its δ18O isotopic composition reaches +16‰. Lake Afrera is located in the northern part of the Afar depression at an altitude of 70 meters below sea level. It has an area of 70km2 and its depth is more than 80 meters. Its salinity is extremely high (158g/l). The major component of the lake salinity is due to high NaCl and Gypsum (Gonfiantini et al., 1973). The lake is principally fed by thermal and cold-springs and evaporation is the major means of water loss. Lake Afrera has a unique isotopic composition among Ethiopian lakes. Despite the location of the Lake under extremely high evaporation environment its δ18O and δD compositions are not high (table 1, figure 3). Other lakes There are few other lakes for which water isotope data is measured (table 1). This includes the Hayk Hardibo Lakes, and the Ethiopian Plateau Crater Lakes which include Lakes Zengena, Tirba and Wonchi. The Hayk - Hardibo lakes are volcanic lakes in northern sector of Ethiopia. Lake Hayk the downstream lake receives part of its water inflows from Lake Hardibo (table 1). The lake basin is currently closed lacking any surface water outflow. Although the lakes are closed, the salinities of the lakes are relatively low (table 1). There is a 2‰ difference in δ18O composition of the dry season and wet season lake water of Hayk. Lake Hayk is enriched than Lake Hardibo owing to its terminal position. The Crater lakes Zengena, Tirba and Wonchi show nearly similar isotopic enrichment. Their 116 δ18O composition varies between +4.1 ‰and and +5.9‰ (table 1). Two of these Lakes Zengena and Tirba are groundwater fed. There is no surface water inlet or outlet from the two lakes. Lake Wonchi has larger catchments than the other two lakes and it has both surface water outflows and inflows. 3. The isotope balance method The underlying principle behind the use of isotopes of water in quantifying water flux around lakes is based on the fact that using hydrologic budget (equation 1) and isotopic budget (equation 2) of a lake there will two equations and two water budget components that can not be directly measured, namely groundwater inflow and outflow. This allows simultaneous solution of the unknown components. In cases where the groundwater flux around a lake is predetermined or negligible the method can be used to estimate the rate of evaporation (a water budget component that is difficult to estimate accurately). dV = I −Q− E dt (1) Vdδ + δdV = Iδ I − Qδ L − Eδ E dt (2) Where, dV/dt is the rate of change of volume; I is the total inflow (on lake precipitation+ runoff from the catchment + groundwater inflow); Q is total outflow (river outflow + groundwater outflow); E is evaporation from lake surface; δ is the isotopic composition of each flux with which it is associated; δQ=δL for a well mixed lake. ρB= ρι,B ι, Turbulent sub-layer Diffusive sub-layer Saturated sub-layer ε=ε∗+∆ε Non fractionating boundry ρA≠≠ ρι,A ι, ∆ε ε∗ Lake/Liquid Figure 5. A simplified representation of the Craig and Gordon evaporation model. Evaporation involves an equilibrium fractionation (ε∗) between the liquid layer and the saturated sub-layer; and a kinetic fractionation (∆ε) between the saturated sub-layer and the diffusive sublayer. In the diffusive sub-layer(A) the transport resistance of the isotopic species(ρi,A) and the normal species of water (ρA) are not equa leading to kinetic fractionation of isotopes of waterl. The turbulent mixing zone(B) is non-fractionating (ρi,B=ρB). The application of this method is not straightforward because of difficulties in estimating the isotopic composition of evaporating water (δE). Evaporation involves both equilibrium and kinetic fractionation of isotopes of water. Therefore, the isotopic composition of evaporating water is not measurable from the equilibrium liquid-vapor fractionation factor alone. In addition, the isotopic composition of 117 evaporating water cannot be easily measured. Craig and Gordon (1965) used a linear resistance model (figure 5) to derive equation 3 as a means of estimating the isotopic composition of the instantaneous net evaporation flux. δE = (α δ ) − hδ A − ε (1 − h + 10 −3 ∆ε ) * L (3) Where α*(=1/α) is the equilibrium liquid-vapor isotope fractionation at the surface temperature of evaporating surface (α is equal to 1.0098 for δ18O at 25°C). δA is the isotopic composition of ambient moisture unaffected by vapor from the lake, and h is humidity normalized to surface temperature of the evaporating surface; ε is total isotopic separation factor which is the sum of the equilibrium (ε*) and the kinetic (∆ε) separation factors. ε* is a convenient way to express fractionation factors (α* orα) in per mil (‰) so that ε*=1000(1-α*). The kinetic fractionation factor is given by Craig and Gordon (1965) and Gat (1995) and is expressed as: ρ ρ i, A ρ i,B ρ ∆ε = (1 − h) A ( − 1) + B ( − 1) ∑ ρ ρ A ∑ ρ ρ B Where the terms in the bracket ( ρ ) are the transport resistance of the isotopic species in the diffusive sub-layer (A) and in the turbulent sub layer (B). Because the turbulent sub-layer is not fractionating ρi,B ρi, A −1 −1 ρB equals zero. The transport resistance in the diffusive sub layer ρ A is obtained from the relative differences between the molecular diffusivities of the normal and isotopic species of water in air (Merlivat, 1978; Merlivat and Jouzel, 1979) and is determined to be Ck=14.3‰ for δ18O and 12.5‰ for δD (Gonfiantini, 1986). Various assumptions are often made to combine equations 1 to 3 to use them in lake hydrological studies. Before assumptions were made however detailed site-specific knowledge of the lake system is required. These different site specific assumptions allows equations 1 to 3 to be combined in many different ways resulting in multiple ways of solving lake isotope water budgets (Gibson, 2002). The first such assumption is whether the lake is in a steady (isotopic and hydrologic) or non steady state condition. Isotopic and hydrologic steady state conditions were assumed in various previous studies (Zuber, 1983; Dincer, 1968, Krabbenhof et al., 1990b; Yehdegho et al., 1997; Kebede et al., 2002). Non steady state systems such as artificial reservoirs which have not yet reached isotopic steady state, ephemeral desiccating lakes in arid climate, and seasonal Sub Arctic Lakes were respectively studied by Zimmerman (1979), Fontes and Gonfiantini (1967) and Gibson (2002). 118 For a lake in hydrologic steady state but not in isotopic steady state, equations 1-3 can be rearranged as follows: dδ L = −[(1 + mx )δ L − δ I − xδ *] I V dt Where m= (4) h − ε / 1000 1 − h + ∆ε / 1000 , x is the Evaporation to inflow rations (E/I), δ* = (hδA+ε)/ (h-10-3ε) ≈ δA+ε/h is the limiting isotopic composition which is the isotopic enrichment value that would be approached by residual waters of a shrinking water body exposed to evaporation. For lakes with hydrologic steady state but not in isotopic steady state the solution of equation 4 leads to equation 5 (Gonfiantini, 1986; Gibson, 2002). δL = δ SS − 123 steady − state − part (δ SS − δ O ) exp − (1 + mx ) It V3 44444 14444 42 transient − part (5) Where the steady state lake isotopic composition (δSS) is given as: δ SS = [ δ I + mxδ * 1 + mx ] or 1 Ι h δ * − δ SS = = x Ε 1 − h δ SS − δ I (6) The isotopic composition of a constant volume lake that has departed from its initial isotopic steady state or an artificial lake that is just filled will evolve towards its steady state composition following a characteristic curve (according to equation 5) as demonstrated in figure 6. The trend of the curves depends on the local climatic (δa, h, δI) and E/I (x) ratio. To draw figure 6 from equation 5, the input parameters in equation 5 were taken from isotope and meteorological data in the region8. The isotopic evolution characteristics of a temporally monitored lake can then be compared with the curves (figure 6). The lake's E/I ratio can then be estimated by the 'best fit' method as in Zimmermann (1979) or Gibson (2002). If isotopic and hydrologic steady state conditions (dV/dt=dδ/dt=0) are reasonably assumed equation 1 to 3 can be reduced to equation 6. Equation 6 can then be used directly to estimate the x if the other input parameters (h, δa, δ*) were known. A simple practical way of gaining rapid hydrological information of lakes under steady state condition is to compare their isotopic compositions with a calculated reference hypothetical Local Evaporation Line. 8 At least four estimates of the isotopic composition of ambient moisture exist for Ethiopia. An earlier estimate (Gonfiantini et al., 1973) was based on the index lake method using Lake Afrera (A salt lake in Afar depression). Recently Kebede et al., 2002 used Lake Hora, a shallow crater lake at 1900 m altitude. The two estimates give a δ18OAcomposition of -12% under humidity of 60%. More recently vapour samples were taken from Central Ethiopia at altitude of 2400 masl. The result shows that the average annual δ18OA composition is -12‰ confirming the previous estimates from the Index Lake method. Assuming also an equilibrium fractionation between the ambient moisture and the summer rainfall isotopic compositions of the Addis Ababa rainfall the δA= δP/α-ε* gives the vapour isotopic composition of -12.3‰. The δDAcomposition is estimated at -87‰ using the index lake method. 119 20 15 18 δ OL 10 5 0 0 200 400 600 800 1000 1200 1400 1600 1800 2000 -5 t δ∗ x=0.25 x=0.5 x=0.75 x=1 x=5 x=10 90 80 70 60 δDL 50 40 30 20 10 0 -10 0 500 1000 1500 2000 t δ∗ x=0.25 x=0.5 x=0.75 x=1.0 x=5 x=10 Figure 6. Schematic representation of evolution of δ18O (upper graph) and δD (lower graph) composition of lakes with different x towards isotopic steady state for constant volume lakes and towards the limiting isotopic composition (δ*) for shrinking lakes. The figure uses the average isotopic composition of ambient vapor estimated at δAO ≈ -12‰ and δD ≈ -87‰. These compositions are the average isotopic composition of the ambient vapor estimated for the Ethiopian region by Gonfiantini et al. (1973) and Kebede et al. (2002). The isotopic composition of inflow used in the figure (δIO ≈-2.5‰, δID ≈-5‰) is estimated from the intersections of the measured LEL and the Addis Ababa Local Meteoric Water Line (δD = 8δ18O+ 15). The composition of the intersection points is also similar to the average isotopic composition of the summer rainfall of the Ethiopian highland. From meteorological data across the study region the humidity is estimated at 60%. Figure 7 uses equation 6 to estimate the δ18O and δD compositions of hypothetical lakes with given x and to calculate hypothetical Local Evaporation Line (CLEL). The best fit line that connects the isotopic composition of hypothetical lakes with different x is defined by: δD=5.3δ18O + 6.2 (Figure 7). The comparison of the measured isotopic composition with the calculated isotopic composition of hypothetical lakes of a given x can be used as a basis for a simple visual estimation of lakes water budget index (x) and as a basis of comparison of hydrology of lakes. Thus lakes can be classified in to evaporation dominated, groundwater dominated, shrinking or terminal lakes by visual inspection. Deviations of measure isotopic compositions from hypothetical values also provide useful information about the factors that influence the isotope regime of lakes. This in turn has a wide ranging implication in the suitability of using δ18O and δD isotopes in paleo-climate reconstructions from Lakes or in selecting a suitable lake for retrieving paleo isotope studies. 120 90 80 70 60 δDSS 50 Shrinking Lakes 40 30 Evaporation dominated Lakes 20 10 0 Outflow dominated lakes -10 -6 -4 -2 0 2 4 6 δ OSS x= 5 8 10 12 14 16 18 x=0 x=0.25 x=0.5 x=0.75 x=1 x= 10 δ∗ Figure 7 . Calculated LEL from various x values under Ethiopian hydro-climatic conditions. The slope of the line that connects the steady state isotopic compositions is 5.3. The limiting δ18O composition is about 14‰, the δ18O composition of a terminal Ethiopian lake is estimated at 7.1 ‰. The limiting deuterium isotopic composition (δD) is estimated at 81‰. Under a condition of prevailing climatic and isotopic composition of meteoric waters of Ethiopia, the C LEL is defined by the line δD=5.3δ18O + 6.2. 4. Discussion 4.1. Constraints of the isotope balance method: the Ethiopian Lakes perspective Introduction of two unknowns in the Craig and Gordon model: One of the difficulties in fully exploiting the isotope method in precisely estimating the isotopic composition of evaporating water is the difficulty of measuring the temperature of evaporating surface to which the humidity has to be normalized in equation 3 or 9. This parameter is rarely integrated into the Ethiopian meteorological or hydrological monitoring networks. In absence of this parameter, often assumption was made such that the surface temperature is equal to the local air temperature or normalized humidity equals the atmospheric humidity (as in figure 7). This assumption however significantly underestimates/influences the rate of evaporation when evaporation is estimated from physically based models as it is already demonstrated in the Ziway Shalla Lakes region (Vallet-Coulomb et al., 2001). The assumption surface temperature equals the air temperature obviously affects the estimates of the isotopic composition of evaporating water. The isotopic composition of evaporating water obtained from this approach is extremely sensitive to a small variation in humidity (Turner et al., 1996; Kebede et al., 2002). Therefore the strict suitability of equation 3 or 9 in precisely estimating water flux around lakes requires a precisely determined surface temperature to which the humidity has to be normalized. Another difficulty of using the Craig and Gordon (1965) model lies in the introduction of δA and the lack of a suitable and direct means of measurement of this variable. Several attempts have been made 121 to overcome these limitation by estimating δA by using index lake with known water balance (Dincer, 1968; Gat and Levy, 1978; Zuber, 1983; Kebede et al., 2002), or using small pans which are kept in thermal equilibrium with the lake (Allison et al., 1979; Allison and Leany, 1982; Gibson et al., 1999) or using measurement of the isotopic composition of the ambient vapor samples collected by Cryostat (Gibson et al., 1999). Under a humid climate, a condition under which equilibrium can be assumed to exist between isotopic composition of local precipitation (δP) and the isotopic composition of ambient moisture, δA can also be estimated from the equation δA= δP/α-ε* (Craig and Gordon, 1965; Gibson et al., 1993; Krabbenhoft et al., 1990b; Yehdegho et al., 1997). Dependence of δE and E: This problem is a universal problem of the isotope balance method of estimating evaporation which is not specific to the Ethiopian Lakes. The dependence of the isotopic composition of evaporating vapor on conditions, rate and amount of evaporation is a well known phenomenon (Craig and Gordon, 1965; He and Smith, 1999). Since δ E → f (E ) , the isotope method of estimating evaporation is a simple conversion of the subject of dealing with ordinary water to the subject of dealing with isotopic species of water. This is true because the starting relation from which equation 3 is derived has the form of equation 7. Equation 7 is fundamentally identical to equation 8, a physically based evaporation rate estimation method (mass transfer method) that considers evaporation as a diffusive process. E = Γ (e s − e a ) E = (bo + bVa)(es - ea ) (7) (8) While we deal with the vapor pressure deficit (es-ea), nature of the smoothness evaporation surface (Va*) and the molecular (ρM) and turbulent(ρT) resistances of isotopic species of water via Γ, which has Γ= the form Va * ρ M + ρ T ( Craig and Gordon, 1965; Merlivat and Jouzel, 1979; Gonfiantini, 1986), or via equation 3 in estimating the isotopic composition of evaporating isotopic species of water, we deal with vapor pressure deficit and wind-speed (Va) as well as kinetic constants (bo and b1: empirical constants that depend chiefly on the height at which wind speed and air vapor pressure are measured) in estimating evaporation. In dealing with lake evaporation estimate, the isotopic method cannot be considered theoretically independent from other physical approaches. In addition, estimating evaporation by the isotope balance method (equations 1, 2, 3) requires additional knowledge of the other water budget components (I and Q). The accuracy of estimate for evaporation by isotope method therefore depends on the accuracies of the estimates on I, Q and δE. Only some of the parameters required in estimating the isotopic composition of evaporating flux would have be enough to estimate lake evaporation. 122 Under a condition where isotope monitoring is not yet part of the hydrologic monitoring networks the isotope method may not be therefore justifiably used in evaporation estimation. However, the method is often a suitable way in estimating groundwater flux around lakes, particularly in Ethiopia where hydrological measurements are rarely available. Interdependent δE parameters: In estimating open water evaporation by the mass transfer approach, since the vapor pressure of the evaporating surface is determined from measured surface temperature by the nonlinear e relation s = 611 . exp( 17.3Ts ) Ts + 237.3 , and since wind speed and vapor pressure differences may be correlated, one cannot assume that the approach (equation 8) will give the correct time averaged rate of evaporation when time averaged temperatures and wind speeds are used as independent variables. This approach gives a reasonably accurate evaporation estimate only for measurements integrated over a maximum of one week (Dingman, 1994). Mass transfer approach of estimating evaporation significantly underestimates the rate of evaporation in many instances within arid regions of Ethiopia (Ayenew, 1998; Kebede, 1999). Likewise, while dealing with estimating the isotopic composition of evaporating water (via equation 3) the phenomenon of interdependence of parameters is re-reflected via the kinetic fractionation factor (∆ε). The kinetic fractionation factor depends on the humidity deficit (1-h) and the transport resistances both of which in turn are ρ A ρ i, A ρ i,B ρ − 1) + B ( − 1) → f (1 − h) ( ρ ρA ∑ ρ ρ B interdependent. That is ∑ . Equation (3) can give a better estimate only for instantaneous evaporation rate at the time when measurements are made on humidity, temperature and transport resistances. Likewise equation 6 gives a better estimate of the E/I for input parameters integrated over short period as already recommends by Yehdegho et al., 1997. 4.2. Methodological conveniences of the isotope balance method in Ethiopian Lakes The Index lake method: The 'index lake method' is a suitable alternative and found to be more promising to indirectly infer δA and h (Dincer, 1968; Gat and Levy, 1978; Zuber, 1983; Kebede et al., 2002). A homogenous terminal lake with known hydrology often serves as an index lake. For a terminal lake with virtually no outflow an assumption can be made so that δE = δI (Dincer, 1968) and δA can then be estimated from equation 3. By assuming that there is no remarkable variation in δA over the region, the δA obtained from the index terminal lake can then be used to derive the δE for other lakes under a similar climatic condition. The presence of many closed basin lakes with wide differences in hydrology, bathymetry, local topography, chemistry, stratification and wind protection in Ethiopia provides an opportunity to select a suitable index terminal lake. Two previous studies (Kebede et al., 2002; and Gonfiantini et al., 1973) used Lake Hora and Lake Afrera as index terminal lakes to estimate the isotopic composition of ambient vapor. The two values are similar to each other 123 and to other estimates in Tropical Africa and to measurements made recently on moisture samples from central Ethiopia9 . The similarity in vapor isotopic value estimated from different approaches reflects the validity of using the index lake method in estimating the isotopic composition of ambient vapor. Lake stratification: One assumption often made in using equation 2 and 3 is that the lake is well mixed so that δQ=δL and that δL be reasonably estimated. In stratified lakes the volumes of the epilimnion and hypolimnon should be known and if the two parts have distinct isotopic compositions, their respective composition should be separately sampled to arrive at a reasonable isotope water budget estimate (Gat, 1995). In lakes with complex/changing stratification regime a monthly monitoring of the isotopic composition is needed. In tropical lakes, seasonal changes in temperature and insolation are not as great as in temperate climate. Tropical lakes often have a simple stratification pattern. Deep tropical lakes are generally permanently stratified (Hecky, 2000, Kilham and Kilham, 1989). Lake Shalla is such an example. In such lakes therefore a long term monitoring of the lake isotopic composition may not be strictly required. However, both the epilimnion and the hypolimnion column should be sampled during the time of sampling. In shallow tropical lakes, patterns in stratification and mixing are closely related to local climate and topography and often they frequently mix because of a marked diurnal variation in temperature and wind speed. The above arguments show that lakes to which isotope method can be suitably applied can be easily selected based on their bathymetry and local climate. Steady state conditions and trends of Local Evaporation Line: The steady state condition is not always attained in highly seasonal systems (Gibson, 2002). This is because of strong seasonal variability in isotopic composition of inflows in high latitude regions. Summer and Winter Local Evaporation Lines (LEL) follows different trends in high latitude region. Unlike the temperate and high latitude lakes, the δA‰ -10.5 -11.8 9 Region Afar , Ethiopia NWP, Ethiopia Source of data Gonfiantini et al., 1973 Kebede et al.,2001 Method The index lake method The index lake method -12 Tanzania Bergonzani et al., 2001 The index lake method -9 Ghana Turner et al., 1996 Subjective -12 NWP, Ethiopia This work, see figure below Direct vapor sampling (figure below) Estimates of vapor isotopic composition over Ethiopia derived from the index lake method assuming humidity of 60% . Data from Tropical Africa and Western Africa is included for comparison. Data from direct measurement of isotopic composition of ambient moisture is also included. The vapor sample is collected at Addis Ababa (altitude 2400 masl, mean annual temperature 17°C) at height of 5m above ground surface. 0,0 19.04-20.04.2003 17.05-18.05.2003 14.06-15.06.2003 23.07-24.07.2003 25.08-26.08.2003 15.11.2003 16.01.2004 16.03.2004 -2,0 -4,0 -6,0 -8,0 -10,0 -12,0 -14,0 -16,0 A two years ambient vapor isotope data measured on moisture samples extracted from the air by cryostat between years 2003 and 2004. The average δ18O A composition is -12‰. The seasonal variation in the isotopic composition reflects the seasonal variation in sources of moisture. The departures from the major trends are the result of local meteorological processes. 124 tropical lakes are characterized by little seasonal changes in isotopic composition of ambient moisture and in the isotopic composition of their inflows. In tropical regions the seasonal variation in isotopic composition of rainfall is not as pronounced as this variation in high latitude regions (Rozanski et al., 1993) therefore lakes often have a single input regime and follow a single LEL. This simplifies the need of seasonal monitoring of isotopic composition of lake the input parameters in equation 6 in tropical lakes. 4.3. Lake-groundwater relation tracing Where networks of groundwater monitoring wells are unavailable, tracing groundwater relation using physical groundwater flow study may not be practical. In such cases stable isotopes become very important tools to obtain information on lake-groundwater relationships and tracing groundwater flow directions around lakes (Krabbenhoft et al., 1990a; Krabbenhoft et al., 1994; Yehdegho et al., 1997). The method has been previously used (Craig et al., 1977;Darling et al., 1996; Kebede et al., 2002) to indirectly detect the influence of East African Lakes on adjacent groundwater bodies. The method uses the advantage that lakes are highly enriched and groundwaters which are influenced by lake water show isotopic enrichment compared to pristine groundwaters which are not influenced by lakes. Aquifer and groundwater wells which show isotopic enrichment were observed around the Bishoftu Crater Lakes, the Ziway Shalla Lakes and around Lake Awassa. The spatial δ18O isotopic composition plot and the δ18O-δD plot of groundwaters around Lake Tana show the presence of two groundwater types (figure 8). Low temperature thermal waters from the Eastern part of Lake Tana basin are the most depleted waters. This is because of high altitude recharge of the waters or/and recharge under a colder climate conditions. The shallow groundwters are relatively enriched compared to the deep groundwaters but they plot close to the present day rainfall isotopic composition of the region ruling out the effects of evaporative enrichment or the role of lake water mixing into them. Both the spatial δ18O plot around Lake Tana and the δ18O-δD plot show no evidence lake water influence on adjacent groundwater bodies. Groundwaters around the BCLs show two types of isotopic distribution (figure 9). Those groundwaters in the up gradient of the lakes (North, Northwest, and Northeast sector) are isotopically depleted and those groundwaters in the southern sector of the region are enriched. This reflects a southward groundwater flow from north passing through the Lakes. The δ18O-δD plot shows that some groundwater wells contain as high as 50% of lake water component. 125 uz p g t f jg kg u| u | w v | S v | w } | S } | w x | S x | w | | S g q r i q j q rl m g q ri N S W # E # S S # S S S S S S S S #S S SS el Gilg A b ay S S # S S S S S S S S S S S S S o i k w u z r m qr l m r k $ $ S T $ S $ â # S # # S Aba T $ S â $ T yO u tfl ow # $ # # T $ w } | w w v v | # # # w x | w w } | w vv | w w u w u w x { { # f g h i j g k g l g m n o p i m qrs p g t uv w u x uu w uv y w uu z w y { w z | w { } w | v w } ~ w v â # # w | | w w x | # 80 δD 40 0 -40 -6 -4 Lake Tana -2 δ18 O 0 Groundwaters 2 4 Deep thermal groundwaters 6 8 Addis Ababa MWL Figure 8. Spatial and δ18O-δD plot of groundwaters and lakes around Lake Tana Basin. The most depleted water are thermal groundwaters. Isotope data is taken from this work. The isotope water samples were collected between 2001 and 2004. 126 -2.9 +6.7 +2.1 -1.4 -2.9 -2.8 -0.1 -3.0 -3.4 -2.9 +7.2 -1.3 -0.5 +0.6 +0.9 +6.9 -1.4 80 60 δD 40 20 0 -4 -2 0 2 4 6 8 -20 18 δ 18Ο Lakes Downgradient groundwaters Upgradient groundwaters AALMWL Figures 9. δ18O spatial plot (above) of groundwaters around the Bishoftu Crater Lakes and δ18O-δD plot of groundwaters and lake waters. The arrows indicate general direction of groundwater flow inferred from δ18O distribution. The groundwaters in the central and southern sector of the region shows isotopic enrichment while groundwaters in northern, north western and northeastern sectors are isotopically depleted. Data from Kebede, 1999 and Kebede et al., 2002. The groundwater flow direction inferred from the isotope distribution also follows the general topography of the region. 127 Spatial plot of δ18O in groundwater around the Ziway Shalla Lakes (figure 10) shows that the groundwaters in between the lakes are relatively enriched relative to groundwaters in highland areas bordering the lakes basin. This was previously (Darling et al., 1996; Ayenew, 1998) interpreted as the influence of the Ziway-Shalla Lakes hydraulic connection to the adjacent groundwater bodies. Z S # ( â S # â ( ( â((( â S # â ( â â ( â â ( â NWP â â Z â â # # â S # ( A L $ â â â # ( ( SEP S # ( # ( ( ( $ S # â ââ(( ( ( â (ââ â â â â Legend O -4 to -6 ‰ Θ -2 to -4 ‰ ● 0 to -2‰ 2 to 0 ‰ ▲ ● >2‰ Figure 10. Spatial plot of δ18O cold groundwaters in and around the Ziway Shalla Lake basin. The lakes from North to South are Ziway (Z), Abiyata (A), Langano (L) and Shalla (S). Groundwaters north of Lake Langano, between Lake Langano and Shalla, between Ziway and Abijata are enriched owing to lake water mixing into them. Piezometric map around the lake also supports this hypothesis. Isotope data from Ayalew, 1998; Darling 1996;and Craig, et al., 1977. 128 420000 425000 430000 435000 440000 445000 450000 455000 795000 460000 465000 795000 S # $ $ 790000 790000 $ S # 785000 S # 785000 Ú d d d â $ d 780000 d 780000 d d S # $ d d $ S # d 775000 Ñd Ú 775000 $ Ú $ $ S # $ S # 770000 ¯ ° ± ² ¤ ³ ¨ © £ ´¢ ¦ µ ´¶ · ¦ ´¸ ³ ª ¡ ¢ ¢ ¡ ª ® ª ª ¹ â ª ¹ ª ª« S # ª« ª ª ¨ $ ª ¨ ª ¬ S # ¨ ª « Ú « ª ¹ Ñ ¹ d ¡ ¢ º¢ £ ¤ ¥ ¦ § ¨© ª « ¬ ¨ ª ¨ © ¨® ª ¨ ¨« ª ¨ ® ¨¬ ª ¨ « ©ª ¨¬ ª © ®ª «ª ® ¬ª « S # $ # S 770000 S # S # 765000 765000 S # 420000 425000 430000 435000 440000 445000 450000 455000 460000 465000 100 80 60 δD 40 20 0 -6 -4 -2 0 2 4 6 8 10 -20 -40 18 δ O Awassa Groundwaters AAMWL Figure 11. δ18O map of groundwater and δ18O-δD plot of Lake Awassa water and groundwaters of the basin. The arrows in the spatial map indicate direction of groundwater flow inferred from Piezometric map (Geremew, 2000). Isotope data is taken from IAEA TC Project ETH/8/006. Lake Awassa bathymetry is given in m. Spatial δ18O and δ18O-δD plot around the lake shows that the southern, northern and eastern part of the basin are characterized by depleted groundwaters while in the southwest and west of the groundwaters are enriched (figure 11). From isotopic point of view, this reflects major westward and southwestward groundwater seepage out of the lake. Piezometric map around the lake (Geremew, 2000) however 129 shows the presence of southwestward and a northward groundwater seepage following a very narrow zone. Groundwaters in western part of the Lake Awassa though they are enriched they are not the result of lake water mixing into them. The enrichment seems to be the result of evaporative fractionation prior to recharge of the vertically infiltrating water. Evaporative enrichment prior to recharge and lake water mixing with groundwaters may have similar isotopic characteristics in arid regions with high evaporation. Therefore care should be taken in tracing the origin of groundwaters around lakes. 4.4. Comparison of the lakes isotopic composition with the CLEL: index of the hydrology of the Lakes Figure 12 compares the isotopic composition of hypothetical steady state lakes of given x with the measured isotopic composition of the selected Lakes of Ethiopia. Table 2 a way of interpreting the comparison between the calculated and the measured isotopic compositions. The figure shows a good general fit between the CLEL and the trends in the measured isotopic composition of the studied lakes. The figure demonstrates that there are generally three groups of lakes. The first groups are those lakes which plot at the extreme enriched end of the CLEL (Lake Gamari). These are groups of saline flood plain lakes whose isotopic composition is closer to the limiting isotopic composition. This kind of isotopic characteristics is attained only for shrinking lakes with very large evaporation to inflow ratios. The second group of lakes is those which plot below the LEL. Lake Afrera, a highly saline lake in Afar depression is an example of such lakes. As the salinity of the lake exceeds the salinity of the ocean it is believed that the dissolved salt effect play a role in the isotopic evolution of this lake (Gonfiantini et al., 1973). The third group of lakes is those which plot near or on the LEL showing a wide range of compositional values. Among the third groups Lake Tana has slightly different characteristics. The δ18O and δD of Lake Tana plot slightly above the CLEL showing slightly higher slope. This reflects evaporation under higher humidity condition than that has been estimated at 60% in estimating the CLEL. The fact that under higher humidity conditions the LEL has higher slope is a well known phenomena (Gonfiantini, 1986). Since Lake Tana is large (~3200km2) evaporation from the lake surface could result in humidity buildup above the lake leading to higher slope. From the graph it can be inferred that many of the Ethiopian rift valley lakes (The Ziway-Shalla Lakes, Lake Awassa, the BCLs, and the Hayk-Hardibo Lakes) are evaporation dominated (0.75<x<1) and the Ethiopian Plateau Lakes (Lake Tana and the PCL) are lakes with important outflow (0.5<x<0.75). The Ethiopian Plateau Lakes including lake Tana shows isotopic composition characteristic of a lake 130 whose net inflow is equally shared between evaporation and outflow (x=0.5). Greater than 75% lake inflows are lost to evaporation in many of the Ethiopian Rift Lakes. 100 The Limi ting isotopic composition 80 δD 60 Lakes wi th upstream evaporated water infl ow 'or lakes infl uenced by 'chain of lake effect' Terminal Lakes Evaporation domi nated Lakes 40 Lakes with important outflow Lakes with equal Evap. and Ouftflow 20 0 -20 -4 -2 0 2 4 6 8 10 12 14 16 18 δ 18O x=0 x=0.25 x=0.5 x=0.75 x=1 x= 5 x= 10 δ∗ BCL Tana PCL Afrera Gamari Aw assa Ziw ay Langano Abijata Shalla Hayk-Hardibo Figure 12. Comparison of the modeled isotopic compositions of hypothetical lakes of given x with the isotopic composition of the Ethiopian lakes to infer their x values and a basis of comparison of lakes hydrology and factors that influence the isotope regime of the lakes . δ18O δD x (E/I) Lake Class Some examples from this study <4.5‰ <27‰ < 0.5 Outflow dominates evaporation, Lake Tana during wet seasons Major loss of inflow is to outflow Small artificial reservoirs like Koka, Dire 4.5 27 4.5-6.0 27-36 6.0-7.2 36-42 7.2 >7.2 42 >42 14 82 >14 >82 other other 0.5 Outflow equals evaporation, Inflow equally partitioned between evaporation and outflow 0.5-0.75 Evaporation slightly dominates outflow loss, or the important part of inflow is lost to evaporation 0.75-1 Evaporation dominated Lake, Evaporation is the dominant water loss, minor outflow 1 Terminal lake Should be computed Lakes with hydrogrphic or lake connectivity based on site specific effects. Isotopic composition of the lakes not entirely the function of their hydrology but also data that of their catchment characteristics • Lake with upstream swamps • Lakes with large catchement to lake area ratio • Lakes fed by upstream lakes " The maximum attainable composition for lake with simple catchement " Small lakes with upstream evaporated water inflow at their shrinking stage x δ18O ≠ ≠x δD Salt Lakes affected by dissolved salt effect Lake Tana on annual scale Lake Wonchi Lake Zengena, Lake Tirba Lake Tana during dry season, Laks Langano, Babogaya Ziway, Shalla, Lake Hora Lakes Gamari, Abijata, Hayk, Lake Gamari Lake Afrera Table 2. The relation between isotopic compositions of steady state lakes and their E/I Ethiopian conditions. Lakes which lack any surface water outlet like lakes Babogaya and Bishoftu (two of the Bishoftu Crater Lakes) plot on a position where outflow is possible (x<1). Therefore groundwater loss should 131 be present in these lakes. The presence of groundwater leakage from these lakes is also a possible reason for their relative freshness (TDS less than 1800mg/L). Isotopic enrichment greater than the expected values (isotopic composition often exceeding or closer to the composition of a theoretical terminal lake) is observed in some of the constant volume lakes such as Awassa Abijata, Gamari and Hayk. Hydrographically these lakes are characterized either by very large catchement to lake area ratios or by the presence of upstream swamps or the presence of upstream lakes that feeds them with pre-evaporated waters. The enriched isotopic composition of Lake Awassa compared to the terminal lake composition reflects the input of pre-evaporated and therefore isotopically enriched water from the swamps- the swamps which contribute the major water inflow to Lake Awassa. The intersection between the LEL and the LMWL (figure 11) indicates that the initial isotopic composition of the inflow water to Lake Awassa is greater than the assumed isotopic composition of the inflows (δ18OI= -2.5‰ and δD = -5‰) by about 1‰ in δ18O . The same phenomenon of high enrichment regardless of the expected value is observed in Lake Abijata. This is because the lake receives enriched water from Lakes Ziway and Langano- the two principal feeders of Lake Abijata. The same is true for Lake Gamari which is more enriched than the limiting isotopic composition of the Ethiopian region in the dry season. This is because the lake receives its major inflow from the Awash River which is enriched already by en route evaporation before it overflows its bank to form the flood plain lake Gamari. Figure 12 shows that the isotopic composition of lakes with large catchment area to lake area ratio (eg. Lake Gamari, Lake Abijata, Lake Awassa, Lake Ziway, Lake Shalla) departs from the predicted values often exceeding the isotopic composition of a hypothetical terminal lake. The departure of the lakes isotopic composition from the expected values in large catchement lakes shows the importance of catchement size in influencing the isotopic composition of lakes. In lakes having large catchement to lake area ratios, there is a likelihood of evaporative enrichment with in the basin which influences the isotopic composition of lake inflows. In such lakes, the isotopic composition of the lakes may not always reflect the isotopic composition of rainfalls. This suggests care should be taken in interpreting the paleo δ18O and δD records from lakes which are characterized by the presence of upstream lakes or swamps. 4.5. Estimates of groundwater flux around the lakes 132 Although the strict precision of the groundwater flux computed using equation 6 depends on the accuracy of the input parameters and the time step (weekly, monthly, yearly) over which the input parameters are integrated, the equation provides a preliminary annual δ18O based isotope water budget estimates of some of the lakes (table 3). The oxygen isotope water budget is computed only for lakes with simple hydrography. That is lakes whose contribution of upstream evaporated water input can be estimated. The results in the table are estimated using the average values of the δ18O composition of lakes, the rainfall and runoff data, and evaporation estimates from the Penman method. Groundwater flux is considered negligible in many instances around Ethiopian Lakes (eg. VonDamm and Edmound, 1984). The results of the isotope budget estimates show that it plays a significant role in the hydrology of Ethiopian Lakes. The groundwater component of Lake Tana is about 7% of the total water flux around the lake. Lake Ziway and Langano the upstream lakes of the Ziway-Shalla Lake basin are characterized by evaporation dominated hydrology with low groundwater inflow component. Previous estimates based on chloride balance shows that groundwater outflow from Lake Awassa represents 7% of the total water loss (Telford, 1999). The isotopic approach shows an estimate significantly higher than this value. Lake δ18Oav x I E P+RI RO GI 0 560 31% GO Remarks e Hora 7.2 1.00 1710 1710 1170 0 0% Index terminal lake Bishoftu 6.9 0.92 1846 1710 997 0 849 46% 136 8% Babogaya 6.7 0.87 2043 1710 1022 0 1022 50% 333 16% Awassa 6.9 0.81 2168 1750 1860 0 308 14% 418 19% δ18O I= -1.5‰ Hardibo 7.1 0.85 2011 1710 1912 0 100 5% 300 15% δ18O I= -1.5‰ Ziway 6.2 0.74 2508 1860 2094 320 400 16% 328 13% Zengena 4.9 0.60 2754 1300 2200 0 500 18% 1450 53% Tana 4.8 0.52 2827 1478 2610 1113 217 8% 137 5% Table 3. Annual isotope budget of selected lakes of simple hydrography in mm per unit mm area of the lakes, groundwater components are also reported in percentage of total inflow or total water loss. For Lakes Awassa and Hardibo the isotope budget is estimated assuming δ18OI of -1.5‰. The rainfall, runoff and evaporation are estimated from hydrological and meteorological data around the lake basins. 4.6. Comparison of oxygen and deuterium budgets 133 Among the two stable isotopes of water, in lake balance studies oxygen is widely used and considered to give good results than deuterium budget (Rozanski et al., 2004). There is some apparent disagreement between the oxygen isotope budget and the deuterium isotope budgets (figure 13) in the Ethiopian lake also. The Lakes which show disagreement are a) those lakes with very large catchment area, b) lakes with very small area, and c) saline lakes. The cause of such disagreement is still the subject of scientific debate (Rozanski et al., 2004). Although the causes for such apparent disagreement can be attributed to local hydrological conditions, the limitation of the Craig and Gordon model in fully describing the evaporative fractionation of the deuterium isotope can not be ruled out. The departure of lake Tana from the 1:1 trend may be related to the size of the lake which results in humidity buildup over the lake surface. This causes an apparent high humidity and reduction in the kinetic fractionation effects and thus higher slope. The departure of the Afrera lakes from the 1:1 trend is caused by the salinity of the lake which reduces the activity of water and low apparent humidity. 3 E/I(δ D) 2 1 Lake Tana Small are a lake s Salt lake s 0 0 1 E/I (δ δ 18 O ) 2 3 Figure 13. Comparison between E/I ratios determined for Ethiopian Lakes based on the oxygen and hydrogen isotopes. 5. General remarks The isotope method is found to be a useful approach in rapidly or visually understanding/quantifying the hydrological regime (particularly the Evaporation/Inflow ratio) of poorly known lakes of Ethiopia. In some hydrogrphically simple lakes, the method has been used to separate groundwater flux around lakes into its inflowing and out flowing components. The method provides a suitable basis to rapidly classifying lakes into shrinking lakes, evaporation dominated lakes, and outflow dominated lakes, groundwater lake categories. This makes the isotope method a simple approach to obtain rapid hydrological information once an isotope hydrological regime (δI, h, δA) of the lakes region under study is determined and a theoretical local evaporation line (CLEL) is calibrate. 134 It is demonstrated that the isotopic composition of the lakes of the Ethiopian region is influenced not only by the climate but also by their hydrological characteristics, by the hydrological characteristics of their catchement (eg. presence of upstream swamps and marshes, catchment to lake area ratio, presence of upstream lake, etc) and by their salinity. This has a wide ranging implication in lake management and in the interpretation of the paleo-δ18O and paleo-δD proxies. Given the departure of the isotopic composition of lakes from the predicted E/I ratios, quantitative reconstruction of past climate record should be made taking into account the site-specific hydrographic conditions. It is observed that in highly saline lakes further evaporation may not always lead to further enrichment of δ18O. Therefore the application of dual δ18O and δD tracers to lake sediment archives should be used to trace changes in paleo-slope of the evaporation line to provide a basis for examining past evaporative enrichment estimates. It is also shown that the isotopic composition of the lakes with large catchments is influenced by the characteristics of their catchments. Therefore paleo-climate reconstruction form lakes having large catchement size should consider paleo-catchemnt characteristics. Improvements on the accuracies of the isotopic compositions of ambient vapor (δA), the normalized humidity (h) and the isotopic composition of inflow (δI) of the region will help to quantitatively determine the water budget components with improved accuracy and to refine the reference CLEL. Overall, the isotope method provides rapid lake hydrological information in poorly known systems such as the Ethiopian Lakes. The calibration curve (reference graph) between the present day evaporation to inflow flux of lakes and their respective isotopic composition (figure 7), after some, refinement can be extended on paleo- δ18O and δD sediment records to get quantitative paleohydrology. The extension of the reference graph to interpret the paleo-records however is possible only on a time scale where the input parameters (h, δA, δI) can be reasonably assumed similar to the present day conditions. References Allison, G.B., Brown, R.M. and Fritz, P., 1979. Evaluation of water balance parameters from isotopic measurements in evaporation pans. In: isotopes in lake studies. International Atomic Energy Agency (IAEA), Vienna, pp: 21-32. Allison, G.B. and Leaney, F.W., 1982. Estimation of isotopic exchange parameters, using constant-feed pans. Journal of Hydrology, 55: 151-161. Arusei, M.K., Sanga, J.K. and Tole, M.P., 2003. Outflow of Lake Naivasha based on stable isotope composition of groundwater. 2nd Gengen geothermal conference, April 7-9, 2003, Nairobi, Kenya Ayenew, T. 1998. The Hydrogeological system of the lake district basin, central main Ethiopian Rift. Ph.D. Thesis, ITC publication Number 64, Netherlands. Ayalew, E., 2003. Application of stable isotopes in the study of lake dynamics in Ziway-Shala basin, Ethiopia. M.Sc. Thesis, Addis Ababa University, Ethiopia. Baxter, R.M., Prosser, M.V., Talling, J.F. and Wood, R.B., 1965. Stratification in Tropical African Lakes at moderate altitudes(1,500 to 2,000 m). Limnology Oceanography, 10: 510-520. Bergonzini, L., Gibert, E., Winckel, A. and Merdaci, O., 2001. Bilans hydrologique et isotopiques (18O et 2H) du lac Massoko, Tanzanie. Quantification des échanges lac–eaux souterraines Earth and Planetary Sciences 333:617–623 135 Chernet, T., 1998. Etude des Mechanismes de mineralisation en fluorure et elements associes de la region des lacs du rift Ethiopien. Ph.D. Thesis, Avignon, France. Craig, H. and Gordon,L.I., 1965. Deuterium and oxygen 18 variations in the ocean and marine atmosphere. In: E. Tingiorgi (editor.), Stable Isotopes in Oceanographic Studies and Paleotemperatures. Lab. Geologica Nucleare, Pisa, pp. 9-130. Craig, H., Lupton, J.E. and Horowiff, R.M., 1977. Isotope Geochemistry and Hydrology of geothermal waters in the Ethiopian rift valley. Technical report, Scripps Institute of Oceanography University of California, USA. Darling, W.G., Allen, D.J. and Armannsson, H. 1990. Indirect detection of outflow from a Rift valley lakes. Journal of Hydrology, 113: 297-305. Darling W.G., Gizaw, B. and Arusei K.M., 1996. Lake-groundwater relationships and fluid rock interaction in the East African Rift Valley: isotopic evidence. Journal of African Earth Sciences, 22: 423-431. Darling, W.G., 1996. The Geochemistry of fluid processes in the Eastern branch of the East African rift System. Phd Thesis, British Geological Survey, UK. Dincer, T., 1968. The use of oxygen-18 and deuterium concentrations in the water balance of lakes. Water Resources Research, 4:1289-1305. Dingman, S.L., 1994. Physical Hydrology. Prentice-Hall, New Jersey, 575p. Fontes, J.-Ch. and Gonfiantini, R., 1967. Comportment isotopidque au cours de l'evaporation de deux basins sahariens. Earth and Planetary Sciences Letters, 3: 258-266. Gat, J.R. and Levy, Y., 1978. Isotope hydrology of island sabkhas in the Bardawil area, Sinai. Limnology Oceanograpy, 23: 841-850. Gat, J.R., 1995. Stable isotopes of fresh and saline lakes. In: A. Lerman, D. Imboden and J.R, Gat (Editors), Lakes, Springer Verlag, New York, pp. 139-165. Geremew, Z., 2000. Engineering geological Study of Awassa Town and Its Surroundings, M.Sc. Thesis, Addis Ababa University, Ethiopia. Gibson J.J., 2002. Short-term evaporation and water budget comparisons in shallow Arctic lakes using nonsteady isotope mass balance. Journal of Hydrology, 264: 242–261. Gibson, J.J., Edwards, T.W.D. and Prowse, T.D., 1999. Pan-derived isotopic composition of water vapour and its variability in northern Canada. Journal of Hydrology, 217: 55–74. Gibson, J.J., Edwards, T.W.D., Bursey, G.G. and Prowse, T.D., 1993. Estimating evaporation using stable isotopes: quantitative results and sensitivity analysis for two catchments in northern Canada. Nordic Hydrology, 24:79–94. Gonfiantini, R., Borsi, S., Ferrara, G. and Panichi, C., 1973. Isotopic composition of waters from the Danakil Depression (Ethiopia). Earth and Planetary Science Letters 18: 13-21. Gonfiantini, R., 1986. Environmental isotopes in lake studies. In: P. Fritz and J.-Ch. Fontes (Editors), Handbook of Environmental Isotope Geochemistry. Elsevier, New York, Vol. 13, pp. 113-168. Hailemichael, M., Aronson, J., Samuel, S., Tevesz, M. and Carter, J., 2002. δ18O in mollusk shells from Pliocene Lake Hadar and modern Ethiopian Lakes: implications for history of the Ethiopian monsoon. Palaeogeogrphy Palaeoclimatoloty Paleoecology, 186: 81-99. He, H. and Smith R.B., 1999. An advective-diffusive evaporation-condensation model. Journal of Geophysical research, 104: 18619-18630. Hecky, R.E., 2000. A biological comparison of Lakes Superior and Malawi and the limnological consequences of an endless summer. Aquatic Ecosystem Health and Management, 3: 23-33. IAEA TC project ETH006,2002. An ongoing project conducted by International Atomic Energy Agency, the Ethiopian Science and Technology Commission and the Ethiopian Geological Surveys, Various unpublished and expert visit reports, Ethiopian Geological Survey, Addis Ababa, Ethiopia. Kebede, S., 1999. The Hydrology and Hydrochemistry of Bishoftu Crater Lakes. Hydrological and Isotope Modeling. M.Sc. Thesis, Addis Ababa University, Ethiopia. Kebede, S., Lamb,H., Telford,R., Leng, M. and Umer, M., 2002. Lake-Groundwater relationships, oxygen isotope balance and climate sensitivity of the Bishoftu Crater Lakes, Ethiopia. Advances in Global Change Research, 12: 261-275. Kilham, P., Kilham, S.S., 1989. Endless summer: Internal loading processes dominate nutrient cycling in tropical lakes. Fresh Water Biology, 23: 379-389. Krabbenhoft, D.P., Anderson, M.P. and Bowser, C.J., 1990a. Estimating Groundwater Exchange With Lakes, 2, Calibration of a Three-Dimensional, Solute Transport Model to a Stable Isotope Plume. Water Resources Research, 26: 2455-2462. Krabbenhoft, D.P., Bowser, C.J., Anderson, M.P. and Valley, J.W., 1990b. Estimating Groundwater Exchange With Lakes, 1, The Stable Isotope Mass Balance Method. Water Resources Researrch, 26: 2445-2453. Krabbenhoft, D.P., Bowser, C.J., Kendall, C. and Gat, J.R., 1994. Use of Oxygen-18 and Deuterium to Assess the Hydrology of Ground-water/Lake Systems. In: L.A. Baker (Editor), Environmental Chemistry of Lakes and Reservoirs. American Chemical Society, Washington, D.C., pp. 67-90. 136 Merlivat, L., 1978. Molecular diffusivities of H216O, HD16O and H218O in gasses. Journal of Chemical Physics, 69: 2864-2871. Merlivat L. and Jouzel, J., 1979. Global Climate interpretation of the deuterium oxygen 18 relationship for precipitation. Jouranal of Geophysical Research, 84: 5029-5033. Prosser, M.V., Wood, R.B. and Baxter, R.M. 1968. The Bishoftu Crater Lakes: A bathymetric and chemical study. Arch.Hydrobiology, 65, 309-324. Rozanski, K., Araguas-Araguas, L. and Gonfiantini, R., 1993. Isotopic Pattern of modern global precipitation. In: P.K. Swart, K.C. Lohmann, J.A. McKenzie, and S. Savin (Editors), Climate change in continental isotopic records. AGU monograph, 78: 1-36. Rozanski, K., Araguas-Araguas, L. and Gonfiantini, R., 1996. Isotope patterns of precipitaion in the East African Region. In: Johnson, T.C., Odada, E. (Eds). The Climatology, Palaeoclimatoloy, Paleoecology of the East African Lakes, Gordon and Breach,Toronto, pp. 79-93. Rozanski, K., Gonfiantini, R., Gourcy, L., Paredes, M., Campos, J. and Rojas, R., 2004. Isotope studies of Lake Titicaca- an overview, International Workshop on the application of isitope techniques in Hydrological and Environmental studies. UNESCO, Paris, France, September 6-8, 2004. Saxena, R.K., Jaedicke, C. and Lundin, L.-C., 1999. Comparison of Lake evaporation estimated by isotope massbalance, bulk aerodynamic and Bowen retio methods. Physics and Chemistry of Earth (B). 24:851-859. Telford, R.J. and Lamb, H.F. 1999 Groundwater mediated response to Holocene climatic change recorded by the diatom stratigraphy of an Ethiopian crater lake. Quaternary Research 52, 63–75. Turner, B.F., Gardner, L.J., Sharp, W.E. and Blood, E.R., 1996. The geochemistry of lake Bosumtiwi, a hydrologically closed basin in the humid zone of tropical Ghana. Limnology and Oceanography 41: 1415-1424. Vallet-Coulomb, C., Legesse, D., Gasse, F., Travi, Y. and Chernet, T., 2000. Lake evaporation estimates in Tropical Africa (Lake-Ziway, Ethiopia). Journal of Hydrology, 245: 1-18. Von Damm, K.L., and Edmond, J.M., 1984. Reverse weathering in the closed basin lakes of the Ethiopian Rift. American Journal of Sciences, 284: 835-862. Wood, R.B. and Talling, J.F., 1998. Chemical and algal relationships in a salinity series of Ethiopian inland waters. Hydrobiologia 158: 29-67. Wudneh, T., 1998. Biology and management of fish stocks in Bahir Dar Gulf, Lake Tana, Ethiopia. Ph.D. Thesis, Wageningen University, The Netherlands. Yehdegho, B., Rozanski, K., Zojer, H. and Stichler W., 1997. Interaction of dredging lakes with the adjacent groundwater field: an isotope study. Journal of Hydrology, 192: 247-270. Zimmermann, U., 1979. Determination by stable isotopes of underground inflow and outflow and evaporation of young artificial groundwater lakes. In: isotopes in lake studies. International Atomic Energy Agency (IAEA), Vienna, pp: 87-94. Zinabu, G.-M. and Taylor, W.D., 1989. Seasonality and spatial variation in abundance, biomass and activity of heterotrophic bacterioplankton in relation to some biotic and abiotic variables in an Ethiopian rift-valley lake (Awassa). Freshwater Biology, 22: 355-368. Zuber, 1983. On the environmental isotope method for determining the water balance components of some lakes. Journal of Hydrology, 61: 409-427. 137 PARTIE V Synthèse et Perspectives 138 Remarques Générales Ce travail essaye de donner un large aperçu du régime isotopique des eaux météoriques en Ethiopie et des facteurs météorologiques qui contrôlent ou influencent leurs variations spatio-temporelles. On utilise ensuite ces acquis pour ; 1) examiner le mécanisme de la recharge des eaux souterraines dans trois secteurs d’Ethiopie (Le Plateau Nord Ouest, Le Rift Ethiopien Principal et l’Afar); 2) examiner les ressources en eau souterraine du Bassin du Nil Bleu ; et, 3) calculer le bilan isotopique de quelques lacs sélectionnés. Le régime isotopique des eaux météoriques en Ethiopie est influencé par l’interaction d’une variété de processus à l’échelle globale ou locale. La comparaison de la variation temporelle de la composition en δ18O et δD des pluies avec les processus météorologiques montre que le principal facteur influençant la variation saisonnière de la composition isotopique des eaux météoriques d’Ethiopie est le déplacement saisonnier de la ZITC et les modifications de l’origine de l’humidité qui l’accompagnent. Le signal isotopique montre aussi l’influence de la vapeur continentale issue de l’évapotranspiration sur le régime isotopique des pluies. Une fois que l’humidité de la mousson a atteint le Plateau Ethiopien, le soulèvement orographique, l’évaporation sur le versant sous le vent et les activités convectives jouent un rôle important sur la variation spatiale de la composition isotopique des eaux météoriques en Ethiopie. Le travail montre que, avec une bonne connaissance des conditions géologique, géochimique et météorologique, les isotopes de l’eau peuvent fournir une information rapide, peu coûteuse et même performante sur les ressources en eau de surface et souterraine. Trois exemples le démontrent; a) le traçage des relations entre d’une part les eaux souterraines du Plateau Nord Ouest et le RiFt Ethiopien principal ou la Dépression de l’Afar b) l’examen des mécanismes de recharge, de l’origine des eaux souterraines et de leurs conditions d’écoulement, et l’identification des zones intéressantes du point de vue hydrogéologique dans le Bassin du Nil Bleu; et, c) des modèles isotopiques sur l’Hydrologie de quelques lacs éthiopiens sélectionnés. Du fait de leur position sous le vent par rapport aux sources d’humidité, le Rift Ethiopien Principal et la Dépression de l’Afar, sont caractérisés par un déficit général en humidité et la rareté des eaux libres. Néanmoins, ces régions renferment beaucoup de systèmes géothermiques et de lacs . Une bonne compréhension des sources de recharge des eaux géothermales et des eaux souterraines de ces régions arides est un problème critique pour estimer les possibilités de développement de la ressource géothermale et du point de vue de la gestion des eaux souterraines. Le traçage isotopique montre l’influence des eaux souterraines infiltrées sur le Plateau Nord Ouest alors que la source eau souterraine diminue lorsqu’on se déplace de l’aquifère près de l’escarpement vers la plaine de l’Afar. Bien que l’écoulement de sub-surface du Plateau Nord Ouest vers la Dépression de l’Afar paraisse d’importance mineure, l’infiltration depuis les cours d’eau pérennes et 139 temporaires, marqués par l’évaporation, près de l’escarpement semble une importante source de recharge dans l’Afar. Dans le Rift Ethiopien principal, cependant, le marquage isotopique montre que beaucoup de sources thermales ou de puits profonds possèdent le signal isotopique des eaux des plateaux adjacents. En considérant plus attentivement le climat, la topographie et la géologie des deux secteurs du rift (Rift Ethiopien Principal et Afar) et du plateau on s’aperçoit que a) la proximité du rift du plateau humide; b) les hauteurs de pluie du plateau adjacent; et c) la présence de structures régionales reliant le rift et le plateau, jouent un rôle majeur dans le transfert des eaux souterraines. L’absence de quantité significative d’eau de « type plateau » dans la dépression de l’Afar est liée à la relative faiblesse des pluies sur sa bordure et l’éloignement très important de la vallée. Dans le Rift Ethiopien Principal les bords des plateaux sont relativement humides et proches de la vallée. Les isotopes de l’eau et la Géochimie des éléments en solution ont prouvé qu’ils constituaient un outil puissant d’investigation des ressources en eau souterraine du bassin du Nil Bleu. Le bassin abrite 44% de la population de l’Ethiopie, et l’eau souterraine représente la principale source d’alimentation en eau potable. La connaissance pratique des conditions de recharge, de l’écoulement souterrain et de son évolution géochimique, et de l’origine des eaux souterraines est encore faible. Les isotopes de l’eau associés aux données géochimiques montrent qu’au moins trois systèmes géochimiques d’eau souterraine existent dans le Bassin du Nil bleu. Il s’agit des eaux souterraines de la zone du linéament volcanique de Yerer Tulu Welel, des eaux souterraines du Graben du Lac Tana et des eaux souterraines extérieures à ces deux zones. Dans la zone extérieure aux deux bassins structuraux les principaux aquifères se trouvent dans les basaltes du Cénozoïque. Dans les régions structuralement déformées, les sédiments du Mésozoïque sont directement ou indirectement accessibles. Dans ces deux zones, la diffusion de CO2 profond et son influence sur la composition chimique des eaux souterraines constitue un des processus majeur. Le gaz, soit issu de la « cuisson » des sédiments mésozoïques, soit directement depuis le manteau joue un rôle prépondérant dans l’évolution chimique des eaux souterraines fortement minéralisées et de type Na-HCO3. Dans ces deux régions, les structures favorisent également l’existence d’écoulements souterrains. Les isotopes de l’eau ont prouvé qu’ils étaient aussi un bon outil pour obtenir une information hydrologique rapide sur les lacs. Ils ont au moins deux applications dans les études de l’eau des lacs. Il s’agit des études qualitatives sur l’interaction eaux souterraines – eaux de surface et de la détermination quantitative du composant eau souterraine ou du flux d’évaporation au niveau des lacs. Dans les lacs d’Ethiopie, dont l’hydrologie est peu connue (quelques uns d’entre eux ne possèdent encore aucune donnée de base) les isotopes peuvent être utilisés pour obtenir une information 140 préliminaire rapide. Ceci nécessite de connaître l’humidité, la composition isotopique de la vapeur ambiante et de la pluie mobilisable pour le ruissellement et la recharge des eaux souterraines. Une fois que ces variables sont déterminées pour une région donnée, il est possible de calculer ce que 'devrait être' la composition isotopique des lacs correspondant à un rapport prédéfini ou hypothétique entre l’évaporation et le flux d’entrée (E/I). La comparaison du régime isotopique des lacs hydrologiquement mal connus avec les lacs hypothétiques est une technique rapide pour quantifier l’hydrologie du lac. La technique peut aussi être utilisée de manière pertinente sur les paléo-archives pour interpréter les paléo-δ18O en terme de flux d’eau. Sur la base de cette approche les lacs ont été classés en lacs à évaporation dominante, à flux de sortie dominant, à réduction de volume etc.. Pour quelques lacs au fonctionnement hydrologique simple le flux d’eau souterraine a été estimé. Généralement l’influence des eaux souterraines sur l’hydrologie des lacs est variable d’un système lacustre à un autre. Le lac Tana, par exemple, est le moins influencé par les apports et les pertes en eau souterraine. L’étude isotopique des lacs permet aussi de discuter la spécificité des lacs éthiopiens et de la nécessité de connaître de manière détaillée l’hydrographie spécifique du site avant d’utiliser les isotopes pour quantifier les flux d’eau. Les lacs renferment des archives sédimentaires parmi lesquelles certaines sont varvées. La calibration entre le rapport évaporation/flux d’entrée (E/I) actuel et la composition des lacs éthiopiens montre que le régime isotopique des lacs est également influencé par l’hydrographie, le climat ou l’hydrologie. Ceci implique la nécessité de considérer la spécificité du site lorsqu’on interprète les δ18O et δD des lacs. La calibration montre également qu’il y a au moins trois facteurs non climatiques qui influencent le régime isotopique des lacs : 1) la salinité – impliquant une forte évaporation ne conduit pas toujours à un enrichissement isotopique dans les lacs très salés ; 2) la taille du bassin versant (effet hydrographique ou effet d’inter-connection) plus la taille du bassin versant est grande plus la probabilité est grande pour que l’enrichissement par évaporation du flux entrant soit fort; et 3) La présence de lacs à l’amont (effet de lac interconnectés). Les compositions en δ18O et δD enregistrées dans les sédiments lacustres ne reflètent pas toujours le seul signal climatique. Généralement, dans les études paléoclimatiques en Ethiopie ou ailleurs (sédiments) les lacs terminaux sont considérés comme le système le plus sensible et idéal (Street Perrot et Harrison, 1985; Lamb et al., 2001) si on prend garde, dans les lacs des zones volcaniques, à tenir compte de l’influence des entrées d’eau géothermales (Telford et Lamb, 1999). L’étude isotopique des lacs démontre que quelques uns des lacs terminaux d’Ethiopie (eg Abijata) sont influencés par les lacs situés à l’amont et qui ont des bassins étendus. Ceci leur donne une composition isotopique qui s’écarte des valeurs prédites du fait des effets hydrographiques ou de lacs interconnectés. C’est pourquoi, une reconstruction paléoclimatique à partir de ce type de lac nécessite de prendre en plus en compte la géométrie du paléo-lac et la paléo-hydrographie. 141 Un bref résumé de quelques résultats majeurs de ce travail figure dans l’encadré 4 (appendice 1). Perspectives En rapport avec les zones et les sujets examinés dans cette thèse il existe une grande variété de perspectives concernant l’utilisation future des outils de l’hydrologie isotopique et de la géochimie. On peut citer les exemples suivant, Signaux isotopiques et processus météorologiques locaux: En général, lorsqu’on utilise la méthode isotopique, la première étape devrait être la détermination du signal isotopique des eaux qui rechargent le système à étudier. Ceci nécessite d’appréhender clairement la variation spatiale des isotopes et leur contrôle météorologique. Notre travail émet l’hypothèse de l’importance de la vapeur recyclée sur le régime isotopique des eaux en Ethiopie. Mais la compréhension précise des activités convectives associées à l’orographie et son influence sur le régime isotopique, la connaissance de la taille et du type de surface terrestre qui influence le régime isotopique (eg: bassin continental végétalisé à large échelle ou humidité du sol locale ou volumes d’eau libre et mares), et les arguments en faveur de l’influence de la ré-évaporation depuis la surface du sol sur les pluies, sont encore vagues. Les prélèvements préliminaires de la vapeur atmosphérique et le suivi de la chimie des pluies sont apparus comme des outils appropriés pour déterminer les sources de vapeur. La comparaison de la composition isotopique de la pluie à l’échelle événementielle, de la composition isotopique de la vapeur et de la chimie des pluies avec les processus météorologiques correspondant devrait être un des futurs objectifs. Pour comprendre vraiment à quelle échelle la surface du sol peut avoir un impact sur le régime isotopique, il faut avoir une vision régionale des paramètres de surface de la région traversée par la mousson avant d’atteindre les reliefs éthiopiens. Transfert des eaux souterraines du Plateau vers le Rift : Dans beaucoup de régions montagneuses qui présentent des dépôts alluviaux au pied des escarpements, les approches chimique et isotopique on démontré leur efficacité pour identifier les sources de la recharge dans la vallée alluviale. Le massif montagneux et le versant de la montagne sont les deux principales sources de recharge des plaines qui bordent les escarpements faillés. Depuis le massif montagneux la recharge se fait par infiltration sur les zones hautes tandis que sur le versant elle se fait à partir des pertes sur les axes d’écoulements et les crues dans les zones basses. L’importance de la recharge depuis les massifs montagneux dépend de la hauteur de pluie, de la position de l’aquifère, de la proximité de la plaine alluviale et de la présence de zones favorables au transfert (comme les grabens ou les failles transversales). Dans ce travail, on montre que les eaux ayant subi un fractionnement par évaporation constituent la principale source de recharge des eaux souterraines dans l’Afar. Le transfert du Plateau Nord ouest vers la Dépression de l’Afar ne 142 correspond pas à une relation simple entre zone de recharge et zone d’écoulement comme dans les bassins sédimentaires. L’importance des montagnes bordant le rift dépend de la présence de structures appropriées. La région autour d’Addis Abeba où le Linéament de Yeler Tulu Welel (un vieux rift est-ouest) traverse le Rift Ethiopien Principal constitue une de ces zones à structure favorable. Ceci crée un bon 'Seuil Hydrogéologique)' qui permet le transfert du Plateau Nord Ouest humide vers la vallée du rift (figure 1). La zone riche en eau souterraine entre Addis Ababa Wolisso, Debrezeyit, Akaki, Majo, Koka etc semble être en relation avec ce seuil hydrogéologique. La zone d’intersection correspond aussi à un contrôle majeur du drainage de l’Awash supérieur. L’encadré 5 (appendice 1) donne un bref résumé de l’hydrogéologie isotopique de la région d’Addis Abeba (à l’intersection du Rift Ethiopien Principal et du Linéament de Yeler Tulu Welel). Un examen plus attentif des structures géologiques le long et à travers les escarpements bordant la vallée du rift, associé à des investigations géochimiques et isotopiques devrait donner de bonnes informations sur les circulations en eau souterraine entre le plateau et le rift. Les grabens marginaux tels que « Raya Valley », « Atayé Valley », « Kombolcha Valley », « Kobo valley » et le système alluvial de Diredawwa devraient aussi être examinés attentivement du point de vue structural et isotopique. Comprendre les mécanismes de transfert de l’eau souterraine des montagnes vers les vallées alluviales et les plaines est aussi d’un intérêt scientifique majeur (hydrogéologie ou hydrogéochimie) là où les isotopes et l’hydrochimie ont été abondamment utilisés. La grande variété géologique, tectonique et climatique le long de l’escarpement du Rift Ethiopien Principal et de l’Afar fait de ces derniers un laboratoire expérimental de terrain pour étudier les transferts d’eau souterraine entre les montagnes et les vallées. 143 nR A l if t-M uto -L a ER n Mai n E Sodo thio pia i Diji bo u t pr ess ion De rpme nt Esca tern Eas h t u So S ern a st h -E t u o E S t A .Ú M W (S EP ) a Pl Ú KÚ Aw asaofel e Ú ½ À ½ ÅÆ ÌÍÎÏ ËÐ Ð Ì ÎÐ Ñ Ò Ð Ñ ÎÑ Ç Ï ÍÏ ÎÒ ÏÑÒÍ N Af ar rginal grabens & Alluvial Valleys Esca rpment, Ma L gan o BC » ¼½ ¾ ¿ À Á à ÁÃ Ä ÇÈ É Ê ËÈ ËÈ Ì Í ÎÏ Ê ËÐ Ð Ì ÎÐ Ê Ñ Ò Ð Ñ ÎÑ Ê ÇÏ ÍÏ ÎÒ Ê ea u * Yerer Tulu Welel Lineament Zone-YTVL A KE A.A GNIP Ú S NW P North-W est e rn u Bl eN iver ile R hl and Hi g Wol l o ra be n Pla tea u- Me t ek el L ow la n d a m Plat e Goj au l alo D L ak G e T a na Ó¿ Ô½ Õ ¿ à ¿ ar i a O ga de la ow nL nd Mo yale È ÉÈÈ Ö Á¼ Â Ä ½ À ½ Å Æ Le seuil hydrogéologique de LYTV Figure 1. Quelques zones cibles pour des investigations futures Le Linéament de Yeler Tulu Welel et le Graben du lac Tana: La déformation du Plateau Nord Ouest fournit des conditions favorables lorsque les sédiments du Mésozoïque sont directement ou indirectement disponibles pour influencer la circulation et la chimie de l’eau souterraine. Le Linéament de Yeler Tulu Welel et le Graben du Lac Tana correspondent à deux zones de ce type. Les observations géochimiques et isotopiques montrent que les chambres magmatiques et le CO2 profond, issu soit du magma soit de la décarbonatation des sédiments du Mésozoïque, influencent la formation des systèmes thermaux riches en gaz. Les observations géologique et géophysique montrent que à la fois les sédiments du Mésozoïque et la chaleur sont présents dans les niveaux profonds de la zone du Linéament de Yeler Tulu Welel. Un échantillonnage de gaz devrait être entrepris dans ces deux zones sur les eaux thermales fortement minéralisées pour comprendre leur évolution hydrochimique. Dans le Graben du Lac Tana le rôle des sédiments lacustres anciens sur la dynamique des eaux souterraines est mal connu. Au cours de notre dernière campagne de terrain (Avril 2004) on a pu observer de l’eau saumâtre dans le Graben du Lac Tana. Un examen plus attentif de l’origine de ces saumures par rapport à la géologie de la région et de leur impact sur l’hydrochimie du lac Tana pourrait constituer un autre sujet de recherche future. Isotopes dans l’étude des lacs: Au début de ce travail, un des objectifs était de calculer le bilan isotopique du lac Tana. On a pu observer, par la suite, que l’approche isotopique est très pertinente 144 pour comparer l’hydrologie des lacs et obtenir une information hydrologique rapide, plutôt que pour déterminer avec précision les flux d’évaporation des lacs. Toutefois, la méthode reste un moyen approprié d’étude de la dynamique des eaux souterraines autour des lacs. Dans ce travail, on essaye de calibrer la relation entre la composition isotopique de quelques lacs sélectionnés et leur hydrologie. Le modèle développé ici peut être utilisé comme référence pour obtenir une information rapide sur l’hydrologie d’autres lacs éthiopiens, à partir seulement de leur composition isotopique. La précision du modèle développé dans ce travail dépend de la précision de ses paramètres d’entrée, comme la composition isotopique de la vapeur ambiante, du flux d’entrée des lacs et de la température et de l’humidité autour des lacs. Une des futures exigences, si l’on veut extrapoler avec précision le modèle à d’autres lacs pas encore étudiés, sera d’améliorer la qualité de ses paramètres (δI, hN, δA)(voir l’annexe jointe à la fin du mémoire de thèse). Pour déterminer la composition isotopique de la vapeur ambiante nous avons utilisé une méthode simple appelée 'the index lake method'. La détermination de la composition isotopique de la vapeur ambiante peut être faite expérimentalement avec les bacs d’évaporation. Cette approche pourrait être un autre sujet d’étude de l’hydrologie isotopique des lacs. Par ailleurs, un examen approfondi du modèle de Craig et Gordon (1965) montre qu’il est important de l’adapter aux conditions climatiques locales. L’adaptation aux conditions locales d’évaporation des lacs éthiopiens pour estimer précisément leur bilan constitue une des questions scientifiques à résoudre. 145 Intérêt de ce travail pour le Bassin du Nil Bleu : quelques éléments Les ressources en eau du Bassin du Nil constituent le sujet majeur de discorde ou de coopération entre les 10 pays riverains d’Afrique. Le Plateau Ethiopien représente la principale source du débit du Nil (70%). Bien qu’il y ait des différences dans la manière de concevoir les accords et la gestion, entre les pays (certains mettant en avant la nécessité d’acquérir de nouvelles données, d’autres considérant que les données disponibles sont suffisantes pour s’occuper du partage des ressources en eau), la partie III de ce travail peut être utilisée comme une plate forme pour l’évaluation future des ressources en eau du Bassin du Nil Bleu en Ethiopie. La partie IV de cette thèse fournit une méthode utile, simple et fiable pour quantifier les pertes par évaporation depuis les surfaces d’eau libre et les marais. Le modèle isotopique d’évaporation développé pour l’Ethiopie (partie IV) peut être extrapolé, pour ce genre d’investigations, sur des étendues d’eau comme le lac Victoria, les marais Sudd au Soudan et les principaux réservoirs artificiels. Le modèle peut aussi être utilisé pour préciser les flux d’eau et de soluté, et l’interaction entre les différents compartiments de grands marécages comme ceux de Sudd au soudan. Ceci comprend la détermination du degré d’interconnexion entre différents secteurs de l’ensemble marécageux, qui aide, à son tour, à comprendre le fonctionnement écologique et la migration des solutés ou des polluants dans l’étendue marécageuse ou lacustre. Le données de l’hydrologie isotopique (partie I et III) de cette thèse montre que les eaux météoriques des reliefs éthiopiens dans la région des sources du Nil Bleu ont une composition isotopique (oxygène, hydrogène et excès en deutérium) qui est différente de celle des eaux météoriques issues de l’Afrique équatoriale. Près de la zone de confluence, les eaux issues du plateau éthiopien et les eaux issues de la région des lacs équatoriaux ont conservé leur composition isotopique distincte. Cette distinction peut être utilisée comme un moyen de quantification de l’eau issue des deux fleuves, au Soudan, à l’aval de Khartoum. Par ailleurs, cette différence dans le signal isotopique des eaux météoriques des deux régions peut être utilisée pour tracer l’importance relative des deux ensembles d’eau dans la recharge des aquifères situés près de la confluence et au delà. Dans l’ensemble, l’hydrologie isotopique peut être intégrée à la gestion des ressources en eau et dans les programmes d’étude comme le « Nile Basin Initiative » et fournir une approche indépendante pour l’investigation des ressources en eau Recommandations 1. Si on considère la rareté des données hydrogéologiques, la topographie compliquée et la complexité du climat, en Ethiopie, étudier les ressources en eau en utilisant uniquement l’approche physique ne paraît pas judicieux. Dans cette optique, l’utilisation des outils de l’hydrogéochimie comme les isotopes et les éléments chimiques dissous peut apporter une information complémentaire aux approches physiques. 2. L’approche géochimique dans l’étude des ressources en eau peut tirer avantage du fait que la détermination de la qualité des eaux (éléments en trace et majeurs) est nécessaire (au moins au niveau de la police des eaux en Ethiopie) avant toute utilisation pour la consommation. Cette opération fournit une information très utile pour l’hydrogéologie (comme cela est montré dans la partie III de ce travail). 146 Dans cette optique, on peut recommander de renforcer l’action légale de détermination de la qualité de l’eau pour non seulement vérifier sa qualité mais aussi en termes de valorisation de la connaissance hydrogéologique et des ressources en eau de la région. 3. Les différences de mécanismes de recharge dans le Plateau Nord Ouest, le Rift Ethiopien Principal et la Dépression de l’Afar montrent que l’isotope de l’oxygène utilisé tout seul ne donne pas toujours une information suffisante. L’excès en deutérium est un outil pertinent pour comprendre les processus hydrogéologiques. On peut recommander l’utilisation conjointe des isotopes de l’oxygène et de l’hydrogène pour obtenir une information complète. 4. Au minimum, à des conditions économiques courantes, on peut faire des mesures simples comme le pH, la conductivité électrique et l’alcalinité sur le terrain, sur les eaux souterraines et autres types d’eau. Ces éléments peuvent fournir une quantité d’informations sur les aquifères et l’origine des eaux souterraines. La Partie III de ce travail montre que ces trois paramètres, mesurables à peu de frais, ont pu être utilisés pour déterminer l’origine des eaux souterraines dans le Linéament de Yeler Tulu Welel et le Graben du Lac Tana. Dans ce contexte, on recommande la mesure de ces paramètres simples et peu coûteux pour chaque type d’eau. 5. Apparemment aucune donnée isotopique n’est disponible sur les zones basses du Nord Ouest, le plateau sud est et les zones basses qui l’entourent, et le Nord de l’Ethiopie. Obtenir des données isotopiques sur ces régions, en plus de leur intérêt pour l’étude de leurs ressources en eau, pourrait aider à mieux comprendre la relation entre les variables météorologiques et l’évolution spatiale des teneurs en isotopes. Ceci pourrait aussi renforcer les études météorologiques dans la région. Globalement, les futures études d’hydrologie isotopique peuvent suivre deux voies séparées mais interdépendantes. Il s’agit premièrement de comprendre la variation spatiale et temporelle des isotopes de l’eau et leur relation avec les paramètres météorologiques, et deuxièmement d’appliquer cette connaissance aux disciplines concernées (ressources en eau souterraine, hydrologie des lacs, paléoclimat, ingéniérie, etc.). Limites de ce travail On ne peut pas prétendre que toutes les données utilisées dans ce travail (en particulier celles extraites de travaux antérieurs) sont de très bonne qualité. L’utilisation des isotopes pour l’étude des ressources en eau en Ethiopie a débuté au début des années 1970. Lorsqu’on utilise des données provenant de laboratoires différents et correspondant à des époques différentes on introduit inévitablement des incertitudes. Pour présenter la variation spatiale on a besoin des coordonnées géographiques. En l’absence de cette information, nous avons essayé de trouver la position la plus proche possible à partir des cartes et des localités citées dans les différents rapports. References: Lamb, A., Leng, M., Lamb, H., Telford, R., Umer, M., 2002. Climatic and non climatic effects on the δ18O and δD compositions o lake Awassa, Ethiopia, during the last 6.5Ka. 147 Street Perrot, F.A., and Harrison, S.P., 1985. Lake levels and climate reconstruction. In: Hecht, A.,D. (ed) Paleoclimatic analysis and modeling. Wielly, Newyork , pp 291-240. Telford, R., J and Lamb, H.F., 1999. Groundwater mediated response to Holocene climatic change recorded by the diatom stratigraphy of an Ethiopian crater lake. Quaternary Research 52: 63-75. 148 ANNEXES 149 Box 1: Short Note on rainfall geochemistry at Addis Ababa In addition to being indicators of sources of moisture, rainwater chemical composition is important from various perspectives. The constituent of solutes in rainfall plays significant role as nutrient loading of freshwater lakes. Some constituents of rain water chemical composition (such as Cl and Br) can be used as an index of evaporation prior to groundwater recharge or as an indicator of rate of recharge in arid regions (Cook et al., 1992). Since the Br/Cl depends partly on rainfall and evaporation conditions and because the ratio remains constant in many rock types, paleo-climate at the time of recharge can be inferred from the Br/Cl ratio in groundwaters (Edmunds et al., 1992). The objective of this short note is to present a one year biweekly monitored rain water chemical composition of Addis Ababa. With the objective of addressing sources of rainfall and sources of solutes in the rainfall, a bulk biweekly accumulated summer rainfalls of the year 2003 at Addis Ababa were analyzed for major ions, The δ18O and d excess composition of the rainfall waters are also presented with the chemical data(table B1). The two principal sources of solutes in rainwaters are the Sea Salt Component (eg gypsum, halite, and marine earosols) and the Land Surface Components (desert dust, forest fires, pollution from cities and agricultural lands etc). Code RNSK1 RNSK2 RNSK3 RNSK4 RNSK5 RNSK6 RNSK7 RNSK8 RNSK9 RNSK10 RNSK11 RNSK12 Na/Cl 0,58 0,40 0,42 0,45 0,44 0,49 0,47 0,43 0,46 0,43 1,07 1,01 δ O 0,34 -2,08 -3,29 -3,77 -5,42 -6,32 -5,33 -5,03 -2,69 -2,14 -1,39 -6,50 18 δD dexcess 17,20 14,5 0,50 17,1 -8,10 18,2 -11,20 19,0 -27,70 15,7 -34,10 16,5 -24,60 18,0 -22,50 17,7 -4,70 16,8 2,30 19,4 5,00 30,2 -21,80 16,1 - Cl 0,90 0,38 0,32 0,28 0,21 0,16 0,15 0,29 0,38 0,52 0,52 0,54 -- SO4 3,58 2,39 2,27 1,86 1,81 1,29 1,23 1,47 2,38 2,65 0,48 1,64 ++ Ca 4,64 2,63 1,63 1,18 0,98 0,66 0,59 0,72 1,44 1,39 0,20 0,86 ++ Mg 0,40 0,24 0,13 0,13 0,04 0,07 0,00 0,03 0,06 0,23 0,03 0,17 + Na 0,52 0,15 0,13 0,13 0,09 0,08 0,07 0,12 0,18 0,23 0,56 0,55 + K SiO2 0,20 1,23 0,11 0,65 0,11 0,72 0,12 0,36 0,10 0,33 0,12 0,25 0,11 0,26 0,13 0,29 0,07 0,27 0,16 0,50 0,42 0,12 0,34 0,17 Date 7/6/03 -25/6/03 25/6/03 -11/7/03 11/7/03 -18/7/03 18/7/03 -25/7/03 25/7/03 -1/8/03 1/8/03 -8/8/03 8/8/03 -19/8/03 19/6/03 -26/8/03 26/8/03 -2/9/03 2/9/03 -10/9/03 16/09/2003 21/09/2003 Table B1. Major ion composition in mg/L of summer rainfall at Addis Ababa. 150 The result (table B1) shows clear monthly chemical and isotopic variations. The start (June) and end (September) of the monsoon are marked by high solute loads. The rainwater chemistry is dominated by Ca++ and SO2-4. Both elements are often regarded as solutes form continental sources. Except in the convective rainfalls of September 16 and 21, the ratio of Na/Cl nearly remains constant and close to the marine value (Na/Cl=0.5) in the Addis Ababa rainfall. This testifies that Na and Cl in the monsoon rains have dominantly their source from marine origin. The deviation of Na/Cl ratio from the marine value in September may be related to the fact that origin of rainfall in September over Ethiopian plateau is principally influenced by local convection following the weakening of the summer monsoon. Compared to the Sahel rains reported by Goni et al., (2001) the Addis Ababa rains show lower amount of solute load. However, the ratio of Na/Cl in rains of both regions maintains their oceanic value. These suggest that the two regions have similar moisture sources but different rain bearing moisture trajectories and different degree of recyling history. The dilution of the Ethiopian rainfalls generally is the result of Ethiopia's location at furthest distance from the oceanic sources (mainly that of the Atlantic) and lower amount of dust involvement along the moisture trajectory. The continuous montly dilution of rain water chemical composition at Addis Ababa is accompanied by a continuous depletion in δ18O and an increase in d-excess (figure B1). The months with the highest rainfall (July and August) are characterized by the most depleted δ18O and the most dilute chemistry. While isotopic depletion is caused by the saturation of the atmosphere and decrease in rain evaporation, the continuous dilution of the chemistry is related to washing out and cleaning of atmospheric dust by rainfall in the proceeding months. The ratio of ions (particularly Na/Cl) of the single heavy local storm rainfall on the September 16 2003 compared to that of the preceding summer months is most likely related to difference in origin of Rainfall. The heavy rainfall of the date is characterized by the highest d-excess and δ18O enrichment. Field observation during rainfall shows that the rainfall happened in during the noon in an arealy localized place. The rain was mainly in solid form (characterstics of September rain in the region). The highest d excess and the enriched δ18O in the rainfall reflect the nature of rainfall and the origin of the moisture. Gonfiantini et al., (2001) states that during formation solid precipitation isotope equilibrium may not be attained and the condensate may have higher d excess. Although more data is still required the general enrichement of the Addis Ababa rainfall in Septermber and their high d excess value may be related to the rainfall formation mechanism. In September the monsoon is retreating and the cloud cover is lower than the preceeding months. This allows heating of the ground during the morning and heavy local convective rains in the afternoon and the evenings. 151 21/09/2003 16/09/2003 2/9/03 -10/9/03 26/8/03 -2/9/03 19/6/03 -26/8/03 8/8/03 -19/8/03 1/8/03 -8/8/03 25/7/03 -1/8/03 18/7/03 -25/7/03 11/7/03 -18/7/03 25/6/03 -11/7/03 7/6/03 -25/6/03 5 35 30 18 Concentration in mg/L andδ O 25 0 dexess 20 SO4-- 15 Ca++ -5 Mg++ Na+ 10 K+ SiO2 d18O 5 Cldexcess -10 0 Figure B1. Temporal variation in major ions chemistry of bulk summer rainfall of at Addis Ababa. The short length of record and lack of data from other rainfall stations do not allow to fully associating the seasonal variation in rainfall chemistry and the corresponding δ18O and d-excess variations to land surface vapor feed back. But it is clear that land surface properties at least play an important role in influencing the chemistry of Ethiopian rainfalls. Potentially the rainfall chemical compositions can be suitably used in clarifying the origin of different moisture masses in Ethiopia. References Cook, P.G., Edmunds, W.M., and Gaye, C.B. 1992. Estimating paleorecharge and paleo-climate from unsaturated zone profiles. Water Resources Research 28: 2721-2731. Edmunds, M., Darling, G., Kinniburgh, G., Kotoub S., and Mahgoub, S. 1992. Sources of recharge at Abu Delaig, Sudan, Journal of Hydrology,131:1-24. Gonfiantini, R., Roche, M-A.,Olivry, J-C., Fontes, J-Ch., and Zuppi, G.M., 2001. The altitude effect on the isotopic composition of tropical rains, Chemical Geology, 181:147–167. Goni, I.B. Fellman, E., and Edmunds W.M., 2001. Rainfall geochemistry in the Sahel region of northern Nigeria Atmospheric Environment 35: 4331-4339. 152 Box: 2. Tritium content of Addis Ababa rains Dating of groundwater mean residence time using tritium is either done qualitatively or quantitatively. The presence of appreciable amount of tritium may indicate young groundwater age while its absence may indicate old groundwaters. To quantitatively determine groundwater ages however lumped parameter models (black box models) are widely used (Zuber, 1986 ). The lumped parameter models require the time series of tritium input, the time series of tritium in monitored groundwaters and the selection of site specific, suitable, and carefully chosen transfer functions. In Ethiopia so far temporally monitored groundwater tritium data is apparently unavailable. So the use of lumped parameter models in dating groundwater in Ethiopia is not feasible until the tritium monitoring were made. Initiation of monitoring program requires the understanding of the tritium input functions. Determination of tritium input function can provide the likely range of groundwater tritium content at a given time. Tritium concentration in Addis Ababa rainfall has been measured intermittently since late 1960's. In late 60s and early 70's the concentration was higher than today owing to the emission of nuclear bomb tritium. The concentration has remained stable and returned to its natural-cosmogenic concentration since 1980s. The natural concentration varies seasonally (figure B2-1). The highest tritium concentration occurs in the summer (JJAS) when the sun is in the northern hemisphere. The average modern day concentration of tritium in the summer rainfall (rainfall that is available for groundwater recharge) is estimated at 10TU. 14 12 3 H in TU 10 8 6 4 2 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure B2-1. Seasonal variation in mean cosmogentic tritium in Addis Ababa rainfall (1984 to present) Because of missing tritium data during the early 1960s and between the late 1970s and early 1980s in Addis Ababa rainfall, the measured rainfall tritium data alone cannot be used as a complete input function for groundwater mean residence time estimation in Ethiopia. Figure B2-2 uses the Doney et al., (1992) method (equation 1) to fill the missing data. The factors (C, f) are geography and time dependent factors of Addis Ababa taken from Doney's data base, t is the year for with the bomb tritium has to be estimated. Figure B2-2 shows the calculated tritium fits well with the measured ones for the bomb tritium (1960-1974) while it overestimates the natural tritium content. TU = [ f1c1 (t ) + f 2 c2 (t )] 1442443 equation 1 400 350 300 Estimated input function 3H (TU) 250 200 Measured tritium 150 100 50 j-98 j-96 j-94 j-92 j-90 j-88 j-86 j-84 j-82 j-80 j-78 j-76 j-74 j-72 j-70 j-68 j-66 j-64 j-62 j-60 0 Figure B2-2. Annual variation in tritium concentration in Addis Ababa rainfall. The high values in 1960s and 1970s is caused be nuclear weapon tests. The upper curve is tritium concentration in rainfall estimated using the Doney et al (1992) method Using estimated tritium input for the bomb part (1960-1984) and the measured tritium (1984 to present) for natural tritium, table B1 qualitatively estimate the mean residence time of Ethiopian groundwaters corresponding to a given range of tritium concentration. These estimates assume a 'piston flow' type groundwater circulation with no accumulation near recharge region or with no body of stagnant water in the system. It should be noted that these values will change with time, for example in 2013 the groundwaters recharged during bomb, and the rainfall tritium will have nearly similar composition. Distinction between the two may not be therefore possible after 2013. However ancient and modern groundwater will still be distinguished using tritium. No unique mean residence time can be determined for tritium content of groundwater ranging between 0.6TU and 15TU until tritium monitoring and lumped parameter models were used. Concentration (tu) Water residence time and origin <0.6 TU Exclusively prebomb recharge 0.6 to 15 TU No unique age can be until models were used >20TU Exclusively bomb tritium > 15 TU Dominantly bomb tritium (this value will be >10 TU in 2013) 153 Table B2-1. Concentration of tritium in Ethiopian groundwaters and proposed interpretation of residence time for sampled in 2003. Box 3: Note on isotope hydrogeology of Lake Tana basin: origin of groundwaters & lake-groundwater interaction The Lake Tana basin is identified hydrogeologically interesting zone (part III) where, unlike many other region of the upper Blue Nile basin, complex lithologic, tectonic, and geothermal activities play a significant role in controlling groundwater flow, its subsurface residence time and its geochemical evolution. The main geologic features of the basin are: subsidence and block fault formation, accumulation of Miocene organic rich sedimentary rocks, Plieo-Pleistocene basic volcanism and late quaternary volcanic activity and recent lacustrine and alluvial deposits. What isotopes of water tell us about the hydrogeology of the basin? The isotope plot 18O-D plot (figure B3-1) of groundwater around the lakes shows that there are at least two types of groundwater in the basin. These are, 1) the most isotopically depleted groundwaters including the Wanzayé thermal spring, Gurambayé thermal springs, the Andesa high TDS spring and the Amora Gedel brackish oozes, and 2) the isotopically relatively enriched groundwaters mainly issuing as cold springs or abstracted from the basaltic rocks of the region. The Lake Tana waters plot below the LWML showing the influence of evaporative enrichment of the heavy isotopes of water. Another notable future is that the most isotopically depleted waters in the basin are much more depleted than the highest altitude (just on the eastern basin divide) present day meteoric waters. This suggests the most depleted groundwaters are recharged under a colder climate or it indicates the depleted groundwater have longer flow path. The spatial isotope plot of the samples (figure B3-2) shows that the most depleted waters are located in the Eastern part of the basin. Around the issuing point of the springs fault blocks are a common future. 80 Lake T ana waters 60 Groundwaters from volcanic rocks δD 40 Depleted groundwaters 20 0 -20 -40 -6 -4 -2 0 δ18 O 2 4 6 8 Figure B3-1: Isotope plot of groundwaters and Lake Tana N ó æëð á Øå öý ×÷ Û Ø ÜØ þæ ÷æí÷ íè ç ÷í ç ÷ í è î ÷í î ÷ í è é ÷í S é ÷ í è í ÷í S ù Øú ö â û ãüÚ âô Û âãÝ õ Þ Ø âãÚ ô ð ñà ò Ú Ü è æ ë ó ãô Þ âãÝ õ Þ ãö Ü è í ÷ í è è é ÷ø í $ è é ÷ ø í è è î ÷í â è î ÷ í è è ç ÷ç í T è ç÷çí è è æ $ # è æ è é ÷ì ì ×Ø Ù Ú Û Ø Ü Ø Ý Ø Þ ßà á Ú Þ âãä á Ø å æç è æ é ææ è æç ê è ææ ëè ê ìè ë íè ì îè í çè î ïè ç ÿ S W # E S # S S S # S S S S S S S S #S S S S # S # T $ S $# S S # $ S S S S el Gi lg A b ay S S S S S S S S S S S Aba T y $ S $ â â # Ou tflo w # # # # ÿ # # # T $ # Figure B3-2: Spatial isotope map of the Lake Tana Basin. What geochemistry tells us about the hydrogeology of the basin? The isotope data identifies at least two type of groundwaters in the basin. However a closer look at the geochemistry and carbon-13 isotope data shows the isotopically the most depleted groundwaters reflect a quasi complex subsurface geology. With in short space in Eastern part of the basin, the groundwaters change from mineralised, gas rich, low pH groundwaters ( eg. Andesa spring), to highly alkaline pH >9, yet very low TDS (<200mg/L), Na-HCO3 type thermal springs (eg. Wanzayé and Gurambayé) and to brackish (EC= 12 000 mg/L) neutral pH groundwaters (eg. Amora Gedel). The δ13C composition of these groundwaters is also different. The Andesa springs are the most enriched (δ13C>0‰) reflecting the involvement of deep seated CO2 (magmatic or gas from de-carbonation of carbonate 154 rocks due to geothermal heating). The Wanzayé thermal springs are the most depleted ( δ13C=-16‰). These depletion reflects the involvement of carbon dioxide from organic matter. The most plausible source of CO2 in the Wanzayé and Gurambaye thermal springs is the old organic rich lacustrine sediments. This could also explain the low mineralization in the waters. The brackish groundwates of the Amora Gedel may reflect the presence of old meteoric waters trapped in the geologic structures. The relatively enriched groundwaters of the basin shows a simple geochemistry. They are Ca-Mg-HCO3 type, have low TDS. These groundwater shows circulation within the basaltic rocks of the Tana graben. How much is groundwater loss from lake Tana? The isotope map shows that at least the groundwaters in BahrDar Town and in the western part of the basin (Kunzila Town) do not show the evidence of lake water mixing into them. Although major conclusion is can not be made until complete data is available around the lake (particularly in the western part), the isotope map suggests there is little groundwater seepage loss from Lake Tana. Isotope budget calculation shows the groundwater loss represent about 5% of total water loss. Similar evidence shows groundwater inflow represent about 7% of total water inflow. The lake ward dipping block faults favour more groundwater inflow towards the lake Tana than out of the Lake. What is the future target of this study or recommendations of this study? The origin of isotopically depleted groundwaters is some what intriguing. This requires a clear understanding of the subsurface geology. Particularly the extent of the Miocene or the Mesozoic sediments underneath the Lake Tana graben should be understood to clearly understand the hydrogeology of the basin. Geophysics and tectonic study would reveal what type of rocks exist underneath the Lake and the basin. The impact of the deep seated gas, and the impact of mixing of the brackish waters on the water chemistry of the lake and on that of the shallow groundwaters should also be investigated. 155 Box 4: Summary of the major results of this work Origin of Isotopic composition of Ethiopian meteoric waters: Rainfall isotope seasonality mirrors climate seasonality which in turn is controlled by the ITCZ drift Altitude effect on the windward face of the NWP: -0.1‰/100 meters Pseudo altitude effect on the leeward face: irregular but -0.14‰/100 meters There is no direct relationship between spatial variation in rainfall amount and spatial variation in isotopic composition of meteoric waters There is good relation between spatial variation in rainfall isotopic composition and spatial variation in mean annual temperature: δ18O‰ = 0.2(T°C)-6.5 The isotopic composition of Ethiopian meteoric waters shares common characterstics with the Sahel rains in JJAS and with the East African regions in MA The Ethiopian meteoric waters are enriched than meteoric waters of the Sahel and the Eastern Africa. This is reflected in rains, groundwaters, and lakes in the region There is a clear west-east depletion in δ18O of meteoric waters from the western lowland of Ethiopia to the central Wollo highland this follows the direction of the summer monsoon flow over the plateau, these rules out the major importance of the tropical or African Easterly Jet in influencing the isotope regime of the Ethiopian rains Recharge mechanism and groundwater transfer in NWP, MER and Afar Depression The degree of evaporative recharge increases as one moves from the NWP to MER and the Afar The summar rain is the principal source of recharge in the NWP There is a good balance between the isotopic composition of groundwaters and the isotopic composition of rainfall in the NWP The influence of plateau type water (groundwater from the plateau) is higher in the MER than in the Afar depression. The configuration of aquifers between NWP and the Afar do not seem to a simple recharge area discharge area configuration it is rather short discontenous flow paths with variable source of recharge and time of recharge Main source of recharge in Afar seems evaporatively (incomple evaporation) fractionated floods and wadies draining the escarpments Infiltration rate is rapid and degree of evaporative fractionation is low in the NWP The presence of old groundwaters recharged under colder climatic condition can not be ruled out in Afar Depression Groundwater resources of the Blue Nile Basin Two structurally deformed basin with hydrogeogeochemically distinct history exists in the Blue Nile Basin these are the YVTL and the LTG Thermal and high TDS groundwaters in the two zone is the outcome of interaction between heat, decarbonation of CO2 from Mesozoic sediments and leaching of volcanic covers 156 About 9 groups of groundwaters can be distinguished based on their flow history, residence time, recharge, and degree of antropogenic influence. This furher classifies the groundwaters of volcanic aquifers of the Blue Nile basin which are traditionally treated as deep/shallow, freash/saline, confined/unconfined etc. Groundwater recharge takes place rapidly through fractures, recharge is selective, flow paths are short in the basaltic plateau Physically approach shows a total groundwater flow of at least 1.4x109m3/year in the Blue Nile Basin Unlike the Mesozoic sedimentary aquifer of the Sahel which contains paleo groundwaters, the Mesozoic sediments of the Blue Nile basin are only locally available for groundwater circulation. However in places like the YTVL , Addis Ababa, and the LTG the sediments release CO2 and influce the chemical composition of major NaHCO3 springs and influence indirectly the groundwater chemical evolution. Isotopes in Lakes Hydrology The work proposes the following reference local Evaporation Line and a table to compare and rapidly investicate the isotopic composition of other lakes 90 80 70 60 50 δDSS Shrinking Lakes 40 30 Evaporation dominated Lakes 20 10 0 Outflow dominated lakes -10 -6 -4 x=0 -2 x=0.25 0 x=0.5 2 x=0.75 4 δ1 8 OSS x=1 x= 5 6 8 x= 10 10 12 14 16 δ∗ Figure B4. Calculated LEL for various Evaporation to Inflow ratios (x) under Ethiopian hydro-climatic conditions. The limiting δ18O composition (δ*, the maximum enrichement under Ethiopian present day climate )is about 14‰ in δ18O. δ18O composition of a terminal Ethiopian lake is estimated at 7.2 ‰. The limiting deuterium isotopic composition (δD) is estimated at 81‰. Under a condition of prevailing climatic and isotopic composition of meteoric waters of Ethiopia, the C LEL is defined by the line δD=5.3δ18O + 6.2. Comparision of isotopic composition of lakes in the region with this curve would help to infer the evaporation to inflow ratios of the lakes. The graph can be interpreted as in table below. δ18O <4.5‰ δD <27‰ x (E/I) < 0.5 4.5 27 0.5 4.5-6.0 27-36 0.5-0.75 6.0-7.2 36-42 0.75-1 7.2 42 1 Lake Class Outflow dominates evaporation, Major loss of inflow is to outflow Some examples from this study Lake Tana during wet seasons Small artificial reservoirs like Koka, Dire Seasonal swamps after wet season Outflow equals evaporation, Lake Tana on annual scale Inflow equally partitioned between evaporation Lake Wonchi Lake Zengena, Lake Tirba and outflow Evaporation slightly dominates outflow loss, or the Lake Tana during dry season, important part of inflow is lost to evaporation Evaporation dominated Lake, Laks Langano, Ziway, Shalla, Evaporation is the dominant water loss, minor Babogaya outflow Terminal lake Lake Hora 157 Should be computed Lakes with hydrogrphic or lake connectivity Lakes Gamari, Abijata, Hayk, based on site specific effects. Isotopic composition of the lakes not data entirely the function of their hydrology but also that of their catchment characteristics • Lake with upstream swamps • Lakes with large catchement to lake area ratio • Lakes fed by with upstream lakes 14 82 " The maximum attainable composition for lake with simple catchement >14 >82 " Small lakes with upstream evaporated water Lake Gamari inflow at their shrinking stage x δ18O ≠ ≠x δD other other Salt Lakes affected by dissolved salt effect Lake Afrera Table B4. Reference table to interprate measured annual average isotopic composition of Ethiopian lakes, examples are included >7.2 >42 The isotope lake balance method The hydrogen balance method and the oxygen balance method of lake hydrological computation do not always give similar result in Ethiopian Lakes (this is a well known problem in isotope balance study). This work proposes why and the need of local adaptation of the isotope existing isotope balance method Box 5: Preliminary note on isotope hydrogeology of Addis Ababa Geology: The geology of Addis Ababa is the result of the intersection of two major tectonic features and cenozoic to quaternary volcanism. The two tectonic features are the the Yerer Tullu Welele Volcanic Lineament (YTVL) and the western margin of the Main Ethiopian Rift (MER). The YTVL ia an east-west running old fault zone. Abebe et al. (1998) elaborately described the origin of the structures in YTVL, their evolution and their importance in controlling the origin of quaternary volcanics in the region. Hydrogeologically, the structures at the junction zone open the plateau towards the rift making 'the YTVL hydrogeologic switch'. switch' In places the EW faults act as a barrier to the North-South groundwater flow around Addis Ababa resulting in the emergence of productive thermal springs in the central Addis Ababa. The hydrogeology of this region is of critical importance because the Addis Ababa City (population >3000000) gets part of its water supply from groundwater aquifers (Akaki well field in southern suburb of Addis Ababa) which is part of the intersection zone. Recharge rate estimation in the region have been the subject of many previous studies (Gizaw, 2002 and references therein). Many previous studies assume groundwater flow path follow the topography (AAWSA et al., 2002) and it is generally from north to south. Previous isotopic evidence shows the volcanic aquifers of the region are highly variable and complex (Travi, 1998). None of the previous works show the importance of the East-West structure in controlling the groundwater flow and its chemical evolution. Furthermore because of its proximity to the MER, the hydrogeology of Addis Ababa is conceptualised interms of the MER hydrogeology. This short note uses isotopic and geochemical evidence to show groundwater flow pattern and groundwater origin around Addis Ababa (along the Entoto - Akaki transect). It attempts to show the importance of the East-West structures related to the YTVL in influencing groundwater flow patterns and its chemical evolution. Isotopic and geochemical evidence of groundwater occurrence and flow: flow Geochemical and isotopic evidence show there are at least four recognizable groundwater occurrence between Entoto and Akaki. The four compartment of groundwaters are identified based on tritium content, δ18O and the geographic distribution of the samples as shown in figure B5. The following major observations obtained from the figure characterises the groundwater flow and occurrence in the region. • Around the Entoto ridge (a) both very high tritium and very low tritium (pre-bomb10) waters are observed • In the central transect of the town two types of groundwater exist. These are 1) the 18O enriched but tritium rich waters (c) and 2) the 18O depleted but low tritium waters (b). The Filwuha thermal springs the the most depleted of the 18O depleted low tritium waters of type b. • The Filwuha thermal springs (b) are depleted than the modern day or the pre-bomb waters of Entoto • Some of the Filwuha springs although they are the most depleted they contains some tritium (~3TU) 10 Exclusively pre bomb recharge refers tritium content less than 0.6 TU if measured in 2003 in Addis Ababa 158 • • Excluding the Filwuha thermal springs which are the most depleted, the low tritium containing waters in central transect (b) of the town have similar δ18O composition to the pre-bomb or post bomb Entoto groundwaters The low tritium and 18O depleted waters of the central transect of the region (b) are rarely observed in the low lying plain beyond Kaliti (d). Likewise the 18O depleted water of Entoto type are rarely observed in the southern sector of Addis Ababa. Generally water around the Akaki well field (d) have similar δ18O composition to that of c type waters. But the low tritium waters of d are more similar in δ18O to the zone c waters. Most high tritium containing waters of zone d waters generally contain also enriched δ18O Origin of groundwaters around Addis Ababa • • • • • • • Groundwaters of the region are not well mixed. At least four compartments vertical/horizontal can be identified from isotopic evidence The presence of pre-bomb type groudwaters in Entoto indicates entrapment or slow flow of groundwaters in rhyolitic Entoto ridge. This shows the low transmisivity of the ridge. This is consistent with the aquifer transmisivity distribution of Addis Ababa (AAWSA, 2000) determined from physical approaches. Entoto type meteoric waters (low δ18O) are not observed in groundwater wells or springs beyond the Kaliti slope break, therefore the Entoto ridge is not the major recharge source to groundwater wells beyond Kaliti including the Akaki wellfield. The flow of Entoto type water underneath the wellfield can not be ruled out until data were obtained. However, low tritium waters in the central sector ( b) which have similar δ18O as that of the Entoto waters are most likely recharged by infiltration at Entoto. The similarity in their 18O content shows that the groundwaters of b in central sector of Addis Ababa seems to be the main recharge water/zone for the low tritium groundwaters beyond the Kaliti break. The difference in in tritium is related to radioactive decay (ageing of the water) as it moves from the central sector of the transect to the lowland. The high tritium and 18O enriched groundwaters beyond Kaliti are most likely recharged locally by vertical infiltration. The enrichement in 18O and the high tritium indicate local vertical recharge. The source of tritium in tritium containing thermal waters of Addis Ababa is local mixing with shallow tritium containing waters of the sector (c) upon ascent. Origin of the Filwuha thermal springs Chemically and isotopically (δ18O, δD and δ13C) the Filwuha thermal springs (b) are distinct from adjacent cold temperature groundwaters. As already noted by Gizaw (2002) the springs are characterised by high TDS, high Na and bicarbonate. Furthermore the springs are enriched in carbon-13. They are the most depleted in oxygen 18 and deuterium. All these observations indicate that the waters are characterised by intensive degree of rock water interaction relative to the cold groundwaters. Moreover, on Entoto plateau the present day δ18O content of groundwaters is 2‰ more enriched than the Filwuha springs. These suggest: • The present day recharge on Entoto ridge is not the major source of recharge for the Filwuha springs • The Filwuha springs follow deeper circulation path and were recharged under a relatively colder climate than today but they were recharged at Entoto or eleswhere • The enrichment in carbon-13 is caused by CO2 coming from mantle or from the metamorphic de-carbonation of the mesozoic sediment buried at deeper levels At least three derivative questions follow these suggestions: 1. What is the source of CO2? and 2) What is the structure that controls the diffusion of the deep seated CO2 towards the surface? 3) what is the cause of depletion of the 18O? Eighty kilometres west of the Filwuha springs (Ambo-Wonchi-Wolisso) there are many similar springs with similar geochemistry, δ18O and carbon 13 characteristics. Part III of this work associate the origin of those springs to the interaction between heat, Mesozoic sediments, and groundwater circulating in volcanic cover. The east west oriented Entoto ridge/fault (northern boundry of the YTVL) can be followed up to Ambo or beyond (Abebe et al., 1998). Although the surface expression is not prominent around Addis Ababa as that of the MER faults, the YTVL is generally a zone with intensive faulting. The lithology of the YTVL is characterised by volcanic rocks underlain by Mesozoic sediments and Precambrian basement. Geothermal heating is also associated with quaternary volcanics of the YTVL. Addis Ababa shares the geologic characterstics of both the MER and the YTVL. Geophysical evidence also shows that the Mesozoic sediments do exist beneath Addis Ababa at about 2km (Momo, T. personal communication). This reflects the Mesozoic sediments can be an important sources of CO2 around Addis Ababa as they are in the western part of the YTVL. Although the surfacial expression of the EW faults is not very clear or laterally discontinous due to cross cutting with the MER faults, the same EW structure and associated faults seems to be responsible for the compartmentalisation groundwater flow patterns around Addis Ababa. The emergence of 159 Filwuha springs follow this localy prominent east west fault. The EW faults also act as a condiute for the deep seated CO2 as well as barrier for the North-South groundwater flow. Aquifer vulnerability map vs tritium distribution: distribution Tritium composition of groundwaters can indirectly tell the vulnerability of aquifers to pollution. The assumption is that tritium containing groundwaters are recently recharged and have a direct connection to the surface. Low tritium waters represent those flow lines which follow deeper circulation pathways or those with longer subsurface residence time (and likelyhood of pollutant attenuation). There is a likely hood of high vulnerability near recharge zones. The tritium distribution shows groundwater aquifers type b, which contains the highest tritium, is located in the central sector of Addis Ababa. These aquifers are the most vulnerable. The deeper aquifers of the Akaki plain and some region of the Entoto ridge are the leaset vulnerable. Although some slight differences are observed around the Entoto ridge, evidence from tritium is consistence with the GIS based aquifer vulnerability map (Alemayehu, 2004) of Addis Ababa. The recharge role of Entoto ridge: The role of mountains in recharging adjacent groundwater bodies (alluvial fills or mountain front aquifers) is not as simple as one can imagine as 'mountains are recharge sources and valleys or adjacent low lands are discharge areas'. The role of mountains in recharging adjacent aquifers depends on the permiability of the mountain block, the rainfall condition on the mountains, the presence of favorable geologic sturcures, and the proximity of the mountains to the the aquifers. Worldwide, the mechanism of groundwater transfer from mountains to valleys, the groundwater flow path identification, geochemical evolution along these flow paths are subjects of scientific investigations. Noble gases, isotopes, geochemistry were widely used in these studies. Part II of this thesis shows the mechanism of plateau-rift groundwater transfer at regional scale. The Entoto- Akaki transect can be used as a downscaled example of mountain-valley groundwater transfer mechanism in the region. Figures B5 shows that the majority of groundwater wells and springs at the foot of the Entoto mountain have higher tritium content than the majority of springs emerging within the mountain. This testifies that the Entoto mountain is not feeding the wells and springs at its foot. One would imagine that the infiltration at Entoto follows deeper flow paths and is not detecable in shalower aquifers at the foot of the mountain. However, deeper aquifers furhter south of the Entoto ridge contains 18O enriched waters than the Entoto ridge springs or wells. The groundwaters in the southern sector of Addis Ababa have similar 18O content to aquifers in the central part of the transect. Thus main recharge source for the aquifers at furthest distance (Akaki) from Entoto seems the central sector of the slope (foot of Entoto mountain) or local vertical recharge for those 18Oenriched (shallow) aquifers. Where is the Entoto type water gone? The Entoto type water either follow still deeper flow paths or emerges following the EastWest fault barrier as thermal spring or as low tritium groundwater at deeper levels of the aquifers. Generally the Entoto Akaki transect shows the presence of recharge continum and compartmentalisation of groundwater flow paths. This entails the need of site specific conceptualisation of mountain-valley groundwater transfer mechanisms before delinating the bounderies for groundwater flow modelling. Previous groundwater modelling of the Akaki well field delinated the Entoto ridge as a major boundery with inflow. However isotopic evidence shows the hydrogeologic boundry of the Akaki well field is the area starting from the foot of Entoto ridge. General conclusions The Entoto-Akaki groundwater transfer is not a simple continuous flow that mixes along the flow line. At least for compartments exist. Waters of the Entoto ridge emerges in the central sector of Addis Ababa. Then infiltration in the central sector of Addis Ababa recharges the groundwaters down the Kaliti slope break. It can be generally said that the principal source of inflow to wells in the Akaki well field comes from recharge that takes place in the central transect of the slope after the Entoto slope break and from local vertical recharge. The groundwater flow pattern is schematically shown in figure B5. This work suggests the need of a closer look at the E-W structures to get improved picture of groundwater flow in the region. The isotope and geochemical interpretation would help better define boundary conditions to refine the groundwater modelling of the Akaki well field. Futher isotope and geochemical data in eastern and western suburbs of Addis Ababa is required to refine the schemiatic diagram. References Alemayehu, T., Ayenew, T., Legesse, D., Tadesse, Y., Waltnigus, S., Mohammed, N., 2003. Groundwater vulnerability mapping of the addis ababa water supply aquifers, Ethiopia, UNEP/UNESCO/UN-HABITAT/ECA report, 80pp. Abebe, T., Mazzarini, F., Innocenti, F., Manetti, P., 1998. The Yerer-Tullu Wellel volcanotectonic lineament:a transtensional structure in central Ethiopia and the associated magmatic activity. J. African Earth Sci. 26, 135 150. AAWSA, BCEOM, SEURECA and TROPICS (2000) Addis Ababa water supply project - stage III A groundwater - Phase II, modelling of Akaki well field, V1, Draft main report. Addis Ababa Water and Sewerage Authority, Addis Ababa, Ethiopia. Gizaw, B., 2002. Hydrochemical and Environmental Investigation of the Addis Ababa Region, Ethiopia. Ph. D dissertation, Faculty of Earth and Environmental Sciences Ludwig-Maximilians-University of Munich, 157p. Travi Y., 1998. Isotope techniques for water resources management, International Atomic Energy Agency, Field Mission Report, ETH/8/006, report number RU-8303 160 980000 T $ T $ $$ T T$ T T $ T $ T $ 990000 T$T $ T $ $ T $$T $T T T $ T $ T $ N T $ 470000 480000 490000 Zone of alternating prebomb inflow & modern recharge 980000 990000 c Examples of prebom b recharge -1 Kaliti Akaki Lowland -2 δ 11 88 Ο -5 1010000 * ? ? Examples of post bomb/recentrecharge mixing of thermal waters withW E shallow groundwaters CO2 -6 470000 480000 490000 ; 1 > & *% % 4 2: 0 8 4 4 25 6 8 500000 Entoto1030000 1020000 a Kaliti slope break -7 Mesozoic sediments GW flow direction Heat source t ul fa b pre bomb thermal waters Filwuha thermal springs Heat source CO2 Filwuha slope break L ? ? ? -4 460000 d C G M IH N @ A D E E ? F O EC @ @ J I? K B P @ -3 450000 E S Zone recent recharge with entrapped post bomb waters 1000000 K 0 970000 Filwuha ts 1 960000 500000 ? ? @ @ A A D @ E BC F @ 460000 970000 970000 450000 T $ $ T 970000 970000 T $ 4 29 5 9 4 4 2: 0 8 & < = / * T $ T $ W T $ 4 25 6 8 4 4 20 0 7 T $ T $ T$T $ 4 20 0 7 4 4 2; ; 5 1000000 T $ T $ 980000 T $ T $ T $ T $ T $ 980000 T $ T $ T $ T $ 0 0 5 7 1 6 7 1 6 27 21 6 7 1 7 1 1 3 T $ T $ T $ 0 0 0 : 9 8 : 9 8 2: 29 28 : 9 8 : 9 8 : 8 7 T$T $ T $ T $T $ 980000 $ T T $ T $ $$ T T T $ T$T $T $$T$ T T$ T $ T$ $ T$T T $ T $ T $ T $ T $ 990000 T $ T $ 990000 990000 T $ T $ T $ T $ 500000 T $ T $ T $ T $ T $ T $ T $ T $ T $T $T $ $ T T$ $ T$T$T$T$T$T$T$T$T$T $T$T T $ 490000 1000000 T $ 480000 T $ $$ T TT $ T $ T T$ $ T$ $ T$ T T$T$T $ T $ T $ T $ T $ T $ 1000000 1000000 T $ T $ 470000 T $ T$ $ T T$ $ T $ T T $ T $ 460000 ! " " $# % & 4; ' 2 ; 5 4 4 26 ; T $ T $ T $ $ T T $ T $ E S W 450000 *() + , 0 0 0 0 0 0 %# 5 : 9 8 7 1 & 5 : 9 8 7 1 25 2: 29 28 27 21 . 5 : 9 8 7 1 , 5 : 9 8 7 1 5 : 8 7 1 3 / * 4 4 4 4 4 4 500000 0 0 0 6 ; 0 6 ; 0 6 2; 20 4 ; 0 ; 0 0 ; 0 ; ; 4 4 2; 0 0 ; 0 5 ; 0 5 ; 20 25 0 5 0 5 0 5 490000 1010000 480000 T $ $$ T T 1010000 470000 N 1010000 460000 1010000 450000 GW flow direction Figure B5 .Spatial tritium and oxygen 18 variation in groundwaters of Addis Ababa (the upper figures). Latitudinal variation in δ18O and tritium content of groundwaters around Addis Ababa along Entoto Akaki transect (lower left), data from Gizaw 2002, IAEA project (AWSSA) and this work. Tritium concentration varies between 0.1TU (the smallest circle) and 20TU (the largest circle). The figure in the lower right corner shows a schematic diagram of groundwater flow patterns along N-S transect . The EW fault in the central sector of the region acts as a major barrier to the N-S flowing groundwaters. Carbon dioxide from decarbonation of the Mesozoic sediments is the principal catalyst for the hydrolysis reaction that produced the Filwuha high TDS high HCO3 waters. The groundwater flow at deeper levels, its continuity and influence on rift groundwater flow is not clear. Along the east west Filwuha fault barrier the groundwater inflow from west parallel to the fault can not be ruled out. 161 162