4
modèle a été appliquée sur l'ensemble de la série pour simuler les écoulements. Cette
opération a permis d'étendre les séries de débits mensuels de trois bassins (Bafing Makana,
Dakka Saidou et Sokotoro) de 1923 à 2005. Globalement, le modèle reproduit d'une manière
satisfaisante la forme des hydrogrammes observés. Les débits de pointe calculés sont bien
situés dans le temps, même s'ils sont parfois sous-estimés (particulièrement pour le bassin de
Bafing Makana pour la période allant de 1981 à 1985) ou surestimés. Les étiages, par contre,
sont parfaitement reconstitués et on observe une bonne superposition des hydrogrammes des
débits observés et calculés. En raison de la variété des bassins (taille, physiographie, états de
surface et caractéristiques des sols), qui leur donne une sensibilité différente aux différents
termes du climat, les impacts d'un potentiel changement climatique sont variables d'un bassin
à l'autre. Ainsi, par rapport aux données observées, les scénarios utilisés prévoient sur nos
bassins des variations d'écoulement moyen annuel qui dépendent étroitement des prévisions
de précipitation de chaque modèle climatique.
MOTS-CLÉS : haut bassin, fleuve Sénégal, modélisation pluie-débit, ressource en eau,
modèle climatique, OMVS, Sénégal.
ABSTRACT
Today water resources management is crucial at various scales (local, regional, national
and international). This management is encouraged by the extreme hydrological events
(droughts or floods) that can have dramatic consequences on human, economic and political
aspects. Appropriately managing a resource requires its evaluation. However, over the last
years, there was a tremendous decrease in hydrological monitoring capacities in the riverine
countries of the Senegal River to produce sufficiently good information to meet end-users
requirements on the main river and its tributaries. Thus there are some gaps in the knowledge
of water resources and its seasonal variations in the upper Guinean basin due to discontinuous
bservation series. o
In this context, the main objective of this thesis was to simulate missing hydrological data
using the GR2M model, especially in the upper Guinean basin. To this end, a physiographic
and climatic characterization of the upper basin was performed. The pluviometric regime was
characterized using a statistical analysis of annual, monthly and daily rainfall, with a spatial
analysis of results. The produced maps provide a visualisation and synthesis tool, not only at
the local scale but at the scale of the studied zone. Then the GR2M hydrological model was
calibrated and validated on a reference period, which served for infilling gaps in monthly flow
time series for the Bafing Makana, Dakka Saïdou and Sokotoro catchments over the 1923-
2005 period. On average, the model satisfactorily reproduces the shape of the observed
ydrographs. h
Last, we evaluated the potential impacts of climate change on the evolution of water
resources on the upper basin using the outputs of four climate models (CSMK3, HADCM3,
MPEH5 and NCPCM) from the 2007 IPPC runs based on the A2 SRES scenario. The impacts
of climate change on flows differ between catchments due to different sensitivities to the
various climatic variables. On the Bafing Makana catchment, an increase of flow is predicted
by the year 2030 with then a decrease by the years 2060 and 2090 when using the CSMK3
and HADCM3 models. A continuous decrease is predicted until 2090 when using the MPEF5
model. Only low variations are predicted with the NCPCM model. At the Dakka Saidou
station, the four models yield a decrease of flows for the three time horizons (2030, 2060 and
2090). The Sokotoro catchment is different from the two previous ones. On this catchment, an