Last week Ousmane Ndiaye, a graduate student at DEES and a graduate research assistant at the IRI, gave his thesis defense to a packed house in Monell. The presentation, entitled “Predictability of the Sahel Climate: Seasonal Sahel Rainfall and Onset over Senegal,” considered issues of rainy-season predictability in Ndiaye’s home country. It also earned him his PhD.
Ndiaye’s investigation centered on two main questions. First, he explored methods to predict the overall quality of the rainy season in Sahel. This is particularly important in West Africa, where short and variable rains have a dramatic effect on livelihoods.
Over the course of his study, Ndiaye found that tropical Atlantic winds are a good proxy for Sahel rainfall at seasonal to multi-decadal timescales. He also found that the National Centers for Climate Prediction’s Climate Forecast System (CFS) was able to accurately characterize the rainy season, even at long (~6 months) lead times around the spring predictability barrier. According to Ndiaye, the ability of CFS to accurately predict Sahel’s rainy season is related to the extent to which CFS accurately characterizes sea surface temperatures (SST) in the eastern tropical Pacific.
As a result of Ndiaye work, it now seems possible to generate skillful, reliable seasonal forecasts of Sahel’s July-to-September rainy season as early as April.
Importantly, Ndiaye’s research also considered the predictability of rainy season onset over Senegal. This is important especially to smallholder farmers whose scarce resources make it particularly important for them to avoid getting fooled by false starts and plant only when the rainy season truly begins.
Ndiaye’s research revealed that there are in fact two onset regions in Senegal and that these regions follow the influence of the Intertropical Convergence Zone. Ndiaye also found that the predictability of northern Senegal’s onset date is tied to global SST, while the predictability of southern Senegal’s onset date is related to the tropical Atlantic SST dipole. This information gives forecasters a way to predict onset date in both regions, which was previously not possible.
Ndiaye’s research contributes to broad our understanding of West African climate. It also provides information that is immediately useful to farmers, water managers, and health workers in Senegal. Ndiaye will continue to bridge the divide between climate and society within the IRI until the end of the year. After that, he will join the Senegalese National Weather Service, where his research skills will be used to for the benefit of the local population and for institutions including Agrhymet and the Red Cross. His presentation is available here.
How reliable are such forecasts? It seems to me most weather forecasts are not accurate.
On the seasonal time scale (not weather) the performance expressed in term of correlation is r=0.6 base on past records (comparing observation and model outcome). Our confidence is much higher during some years (ENSO, Tropical Atlantic Dipole) than others (no signal).
We have to understand with such long lead-time (forecasting the seasonal outcome months in advance) it’s very difficult to hit high score due to uncertainties : in our observing system, lack of representing the whole climate system processes, …
The way we have to see this type of forecast is, it’s better than taking decision without any prior information, its impact on decision making in this region of the world, and this is the best we can do (at the moment).
Salaam,
Well done Ousmane, I hope your results will be used to reinforce the seasonal predictions over West Africa. Also your decision to go back home is encouraging, may Allah (SWT) assist you and your family for this endeavor.
Regards,