The Sahel is the transitional semi-arid region of North Africa south of the Sahara desert between 12.5°N and 17.5°N. Its inhabitants are sensitive to rainfall amounts in the wet season (typically July to September), with droughts causing significant hardship. Rainfall is very variable from year to year and decade to decade. The Soudan is the belt immediately to the south of the Sahel between 12.5°N and 10°N. While the wet season is wetter on average than in the Sahel, it also has considerable variability. The Guinea Coast is a humid tropical region south of 10°N encompassing the countries of Cote d'Ivoire, Ghana, Togo, Benin and western Nigeria. Forecasts of the most likely of five rainfall categories (very wet, wet, average, dry or very dry) are produced using dynamical forecast models. Each of these categories has occurred in one-fifth of past years.
Issued July 2013
Our updated forecast for the July-August-September 2013 wet season indicates that the most likely category is 'wet' in the Sahel , 'wet' in the Soudan and 'dry' or 'very dry' in the Guinea Coast region. For a detailed look at the 2013 prediction, see the updated forecast for JAS 2013 (PDF, 306 kB) .
The forecast for July-August-September 2012 (PDF, 473 kB) favoured rainfall in the 'dry' category in the Sahel whilst the observed rainfall was in the 'very wet' category. The forecast for the Soudan indicated the 'dry' category as most likely whilst the observed rainfall was in the 'wet' category. The forecast for the Guinea Coast favoured the 'dry' category whilst the observed rainfall was in the 'dry' category. Therefore the forecast for the Guinea Coast region provided good guidance whilst the forecasts for the more northern regions were misleading.
The table below compares observed outcomes with our past forecasts. Overall, the correlations between the predicted and observed category are 0.10 for the Sahel, 0.13 for the Soudan, and 0.29 for the Guinea Coast. In general, these do not represent a statistically significant level of skill. Our current forecast methods do demonstrate skill, however, being considerably better than earlier methods when tested in forecasts of past years.
|1992||dry||dry||v. dry||avg||v. dry||dry|
|1993||v. dry or dry||dry||v. dry||v. dry||dry or avg||avg|
|1994||v. dry or dry||wet||dry or v. dry||avg||dry or avg||avg|
|1996||dry or avg||v. dry||avg||v. dry||v. wet||avg|
|1997||dry or avg||v. dry||dry or avg||v. dry||dry||v. dry|
|1998||v. dry||avg||v. dry||avg||wet or v. wet||dry|
|1999||wet||v. wet||wet||wet||v. wet||wet|
|2003||v. wet||wet||wet||wet||v. dry||avg|
|2004||v. wet||avg||avg||dry||v. dry||dry|
|2005||v. wet||avg||avg||dry||v. dry||dry|
|2006||avg or wet||avg||dry or avg||avg||dry or avg||dry|
|2007||dry or avg||wet||dry or avg||avg||dry or avg||wet|
|2009||dry or avg||wet||dry or avg||avg||avg or wet||dry|
|2010||avg||wet||dry||wet||dry or v. dry||v. wet|
|2011||wet||avg||dry||avg||dry or v. dry||dry|
Last updated: 30 January 2014