The East African Paradox
What this page covers
The most important honest caveat in this wiki. A persistent CMIP6 bias affects future projections of March-May rainfall over the Horn of Africa. The observational record shows drying; the models project wetting; the mechanism is understood; no model subset resolves it.
The observed signal
To be written. CHIRPS / ERA5 evidence of multi-decadal long-rains drying over Ethiopia, Somalia, Kenya. Sources: Funk et al. (CHIRPS); the recent IPCC AR6 WGI assessment of historical change; Wainwright et al. 2019 on the shorter-not-weaker rains framing.
The modelled signal
To be written. CMIP6 ensemble projects wetting under SSPs. Park et al. 2023 evaluation of long-rains representation; Schwarzwald et al. 2024 confirms persistence in CMIP6.
The mechanism
To be written. Equatorial Pacific SST biases propagating through Walker Circulation. Sources: Park et al. 2023 (Climate Dynamics); the broader literature on Pacific-Indian Ocean teleconnections to East African rainfall.
Affected geography
To be written. Ethiopia, Somalia, Kenya, Djibouti, Eritrea, South Sudan, northern Tanzania, northern Uganda. The MAM long-rains season specifically. Other seasons and other variables (temperature) less affected.
What this means in practice
To be written.
- Observational “Recent Changes” view: correct; the drying trend is real.
- Future-projection view for MAM rainfall over named countries: treat magnitude with extra caution; the direction itself may be wrong.
- No CMIP6 subset (AFR-13, AFR-EAF, AFR-8, FULL-18) resolves the paradox. This is structural across CMIP6.
How to handle it in a climate rationale
To be written. Recommended framing language:
“The observational record shows substantial drying. Climate models project wetting. The scientific consensus is that the models are biased over this region. Adaptation planning should account for continued drying as a credible scenario.”
Cite Schwarzwald 2024 + Park 2023 + Wainwright 2019 in proposals.
What CMIP7 might bring
To be written. The mechanism (Pacific SST biases) is a major target of CMIP7 development. Whether CMIP7 resolves the paradox will be one of the headline tests of the new model generation. Cross-link to CMIP7 + CORDEX-Africa.