Predictive Skill of African Easterly Waves in the ECMWF Subseasonal-to-Seasonal Reforecasts

This study evaluates the subseasonal predictive skill of African easterly waves (AEWs) in the ECMWF reforecasts during the West African monsoon season (July-September) from 1995 to 2014. There are two wave tracks over West Africa based on the relative location to the African Easterly Jet (AEJ): northern waves and southern waves. The ECMWF model realistically represents the propagations and the horizonal and vertical structures of AEWs along the two tracks out to 16-day forecast lead time, but under-predicts the intensity of the northern waves and overpredicts the southern waves. The positive biases in the southern waves are associated with stronger monsoonal westerlies and stronger deep convection, which is coincident with a northward shift of the AEJ. The negative biases in the northern waves are linked to a weaker AEJ, which results from a cold bias in the Sahara and a weaker meridional surface thermal contrast. The biases suggest possible deficiencies in the model convection, surface, and planetary boundary layer schemes.

The predictive skill of the weekly AEW activity was examined using an AEW index defined based on the first two principal components of the 2.5-9-day bandpass-filtered meridional winds at 850-hPa. The model shows some skill in predicting the active and inactive weeks of the AEW activity in the week-1 reforecasts, but the skill deteriorates quickly with the forecast lead times. The impacts of some large-scale climate modes such as the El Niño–Southern Oscillation (ENSO) and the Madden-Julian Oscillation (MJO) on the predictive skill were investigated.