Programme OS3e Regional and nationwide
scenarios 1 abstract 261
Climate and river runoff changes in West and Central Africa: past
variability and prediction of water resources for the 21st century.
Author(s): G. Mahé, J.E. Paturel, S. Ardoin, J.F.
Boyer, L. Casenave, A. Crès, A. Dezetter, C. Dieulin, S. Girard, M. Rescan, N. Rouché, E. Servat
HydroSciences Montpellier
Maison des Sciences de l’Eau
UM2 34095 Montpellier Cedex 5
France
gil.mahe@msem.univ-montp2.fr
Keyword(s): Climate, Runoff, West and Central Africa, Climatic variability,
River modelling, Water resources, GCM, HadCM3, FRIEND-AOC
Article:
Poster:
Session: OS3e Regional and nationwide
scenarios 1
Abstract The aim of this study is first to show how climate and river
runoff have changed over the last century in West and Central Africa, and secondly from this knowledge, in
combination with river modelling, to predict the water resources for the 21st century for more than 350 river basins
of West and Central Africa, obtained from the FRIEND-AOC database (UNESCO IHP).
The GR2M rainfall-
runoff conceptual model is calibrated and validated for each runoff series
at a monthly time step. Rainfall are
calculated from an IRD monthly gridded database, available on a website, PE are calculated from the CRU
database. The reservoir height is taken as the Water Holding Capacity, given by the FAO soil Map of the World. All
data are averaged for half degree squares. We simulate future runoff using rainfall and PE data derived from the
GCM outputs HadCM3, scenario A2, according to a procedure for estimating “realistic” rainfall and PE values from
the GCM outputs.
The results are presented for 3 time horizons: 2020, 2050 and 2080. In 2020 and 2050 one
observe an important variability over West Africa, except over three regions: the North-West (Senegal-Guinea-
Mauritania) and the North of the Congolese basin, where runoff decrease as soon as the 2020 horizon; and the
Chari basin where runoff increase. In 2080 runoff decrease everywhere in West and Central Africa, except over the
Chari basin. The seasonnal runoff dynamic, and particularly the flood peak occurrence, might be also modified by the
forthcoming climatic fluctuations predicted by the HadCM3 GCM.
The bias due the use of a single GCM and a
single scenario of greenhouse gases emission variability output data set is discussed, and we present a comparison
between 4 GCM outputs over West Africa.