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 D ES 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 Professor Université d’ Oxford, Center for Water Research Rapporteur J.L. Michelot CR, HDR Université de Paris Sud, LHGI Rapporteur Y.Travi Professeur Université d'Avignon, Labo. Hydrogéologie Directeur de Thèse T. Alemayehu Asso. Professor Addis Ababa University, Geology Department Examinateur K. Rozanski Professor University of Krakow, Dept. Nuclear & Env. Physics Examinateur P. Aggarwal Doctor Head, Isotope Hydrology Section, IAEA Examinateur B. Blavoux Professeur Université d' Avignon, Labo. Hydrogéologie Examinateur

2

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

3

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

4

INTRODUCTION GENERALEINTRODUCTION GENERALEINTRODUCTION GENERALEINTRODUCTION GENERALE

i.i.i.i. Objectifs et démarche utilisésObjectifs et démarche utilisésObjectifs et démarche utilisésObjectifs 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

5

(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 hydrologie-

isotope-climat sur quelques lacs actuels sélectionnés est une approche pratique pour examiner d’autres

lacs ou réservoirs peu connus.

6

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)ii. Approche (méthodologie)ii. Approche (méthodologie)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.

Isotopehydrogeology

Modernlake isotope

water balance

Modern isotopeclimate

relations

Paleoclimate

modern

calibration/modeling

Gw tracing

Composition of input signal

Lake groundwater link

Modern

calibration/modelling

proxy

Input signal

proxy

7

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.iii.iii.iii. Organisation de la thèse Organisation de la thèse Organisation de la thèse 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 3H, le Carbone -13, et l’hydrochimie précisent les contrôles sur l’évolution chimique des eaux

8

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

9

iv. Localisation de la zone d’étude, toponymie, régionsiv. Localisation de la zone d’étude, toponymie, régionsiv. Localisation de la zone d’étude, toponymie, régionsiv. 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. v. v. v. Définitions, symboles, notationsDéfinitions, symboles, notationsDéfinitions, symboles, notationsDéfinitions, 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 16O (abondant) et 18O (rare). Ces isotopes se combinent pour former quatre types

d’eau : H216O (le plus abondant), D2

16O, H218O et D2

18O (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.

( ) 1000*1)16/18(

)16/18( .18

−=

SMOW

Ech

OO

OOpourmilleOδ et ( ) 1000*1

)/(

)/( .

−=

SMOW

Ech

HD

HDpourmilleDδ

10

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

11

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

NWP- The North Western Ethiopian Plateau SEP- The South Eastern Ethiopian Plateau MER- The Main Ethiopian Rift GNIP- Global Network for Isotopes in Precipitation of the IAEA BCL- Bishoftu Crater Lakes YTVL- The Yerer Tulu Welel Volcanic Lineament LTG- The Lake Tana Graben JJAS- June-July-August-September CLEL- 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.

12

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.

13

PARTIE I

Composition isotopique des eaux météoriques en Ethiopie

(Plateau NW, Afar et Rift Ethiopien Principal)

14

Introduction

Les concentrations en δ18O et δD des eaux météoriques en Ethiopie, ainsi que leur variations spatio-

temporelles 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).

15

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

16

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).

17

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.

18

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 level1. The isotopic

composition of rains associated with the Indian and Atlantic Ocean monsoons are relatively well-

studied (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.

19

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 δ18O-

δD 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

20

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.

21

-5

0

5

10

15

20

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

δ18 Ο

an

d d

-exc

ess

0

50

100

150

200

250

300

350

400

Ra

infa

ll in

mm

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.

-10

-8

-6

-4

-2

0

2

4

6

8

Jan Feb Mar Apr May June Jul Aug Sept Oct Nov Dec

δδ δδ18 0

Figure 4. Monthly Maximum-Minimum-Mean plot of the δ18O ‰ of the Addis Ababa rainfalls (1965-2002). Highly variable δ18O content

is observed in June and October which marks the northward and south ward passage of the ITCZ over central part of Ethiopia.

22

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 Africa3 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.

23

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 the East African Rainfall stations include: Ndola, Dar es Salam, Kampala, Harare, Antananarivo, Entebbe 4'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.

24

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

25

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

26

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 range5.

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

A.

B

-60

-40

-20

0

20

40

60

80

100

-10 -8 -6 -4 -2 0 2 4 6 8 10

δδδδ18O

δδ δδD

Low TDS cold groundwaters high TDS Na-HCO3 waters from YTVL and LTG Lakes and rivers draining them GMWL

*

Ave rage Summer Rainfal lat Addis Ababa

*Ave rage March-Apri l Rainfal lat Addis Ababa

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.

27

• 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 northeast-

southwest 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

28

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.

29

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 evapo-

transpiration (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,

30

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

31

-8

-6

-4

-2

0

2

4

6

8

j-65

j-67

j-69

j-71

j-73

j-75

j-77

j-79

j-81

j-83

j-85

j-87

j-89

j-91

j-93

j-95

j-97

j-99

j-01

δδ δδ18 O

Automn Spring May Summer

Figure 6. Long-term variation in amount weighted δ18O of Addis Ababa precipitation.

-10

-5

0

5

10

15

20

jan

v-6

5

jan

v-6

7

jan

v-6

9

jan

v-7

1

jan

v-7

3

jan

v-7

5

jan

v-7

7

jan

v-7

9

jan

v-8

1

jan

v-8

3

jan

v-8

5

jan

v-8

7

jan

v-8

9

jan

v-9

1

jan

v-9

3

jan

v-9

5

jan

v-9

7

jan

v-9

9

jan

v-0

1

D e

xces

s in

spr

ing

and

sum

mer

ra

infa

lls

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.

32

• 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.

6The 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.

33

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