Abstract:
Over the past decades, large variations of precipitation were observed in Africa, which
often led to dramatic consequences for local society and economy. To avoid such
disasters in the future, it is crucial to better anticipate the expected changes, especially
in the current context of climate change and population growth. To this date, however,
projections of precipitation over Africa are still associated with very large uncertainties.
To better understand how this uncertainty can be reduced, this study uses an advanced
Bayesian analysis of variance (ANOVA) method to characterize, for the first time in the
regional climate projections of CORDEX-AFRICA, the different sources of uncertainty
associated with the projections of precipitation over Africa.
By 2090, the ensemble mean precipitation is projected to increase over the Horn of Africa
from September to May and over the eastern Sahel and Guinea Coast from June to
November. It is projected to decrease over the northern coast and southern Africa all year
long, over western Sahel from March to August, and over the Sahel and Guinea Coast
from March to May. Most of these projections however are not robust, i.e., the magnitude
of change is smaller than the associated uncertainty. Over time, the relative contribution
of internal variability (excluding interannual variability) to total uncertainty is moderate
and quickly falls below 10%. By 2090, it is found that over the Horn of Africa, northern
coast, southern Africa, and Sahel, most of the uncertainty results from a large dispersion
across the driving Global Climate Models (in particular MIROC, CSIRO, CCCma, and
IPSL), whereas over the tropics and parts of eastern Africa, most of the uncertainty results
from a large dispersion across Regional Climate Models (in particular CLMcom).
