The impact of climate change on rainfall extremes over Northeast Africa

Africa is highly vulnerable to rainfall variability and thus to climate change. The highest
impact of this variability is through extremes, such as floods and long periods of
severe drought. In this study, these extremes in rainfall have been investigated.
For comparison, rainfall extremes over Northeast Africa, which includes Ethiopia
and Sudan, have been analyzed in the ERA-40 reanalysis dataset. The analysis showed
three clear signals during rainfall extremes. (1) A negative connection with Sea Surface
Temperatures (SSTs) in the equatorial Pacific Ocean, i.e. ENSO. (2) A positive
connection with the Tropical Easterly Jet (TEJ) at 200 hPa, which overlies the region,
and (3) a positive connection with moisture advection over central Africa, southwest
of the region.
The second part of the study focuses on the projected changes under global warming
of extreme precipitation events in a model ensemble of the ECHAM5/MPI-OM
GCM. In the model, the future climate period of 2051-2100 is compared with the
present climate between 1951 and 2000. The rain season, from May to October, is
projected to shift towards later in the season. In terms of rainfall extremes, for both
tails the magnitude is projected to decrease in June and July, and increase from August
to October.
The climate signals found in the ERA-40 dataset are analyzed on their changes
under global warming, and if they can explain the projected changes in rainfall. The
strongest connection is found for the moisture advection southwest of the region, from
which changes in the 10% largest extremes coincide quite well with the projected
changes of wet extremes. Similar results are found for the TEJ, although the correlation
is weaker.
The connection with the dry extremes was less evident and could not completely
be address to either the TEJ or to changes in the moisture advection. Furthermore,
changes of the ENSO signal where not found in the model, but both the present and the
future climate showed a significant and negative correlation between ENSO SSTs and
rainfall in the region.