African easterly waves (AEWs) are the primary synoptic-scale weather feature found in sub-Saharan Africa during boreal summer. Many studies have used observations and idealized models to highlight processes associated with the movement and growth of AEWs, yet there have been few studies documenting the performance of operational ensemble prediction systems (EPSs) for these phenomena. Here, the predictability of AEWs in ECMWF EPS forecasts is assessed during two periods of enhanced AEW activity (July--September 2007--2009 and 2011--2013). Specifically, AEW predictability is analyzed through three main areas of focus: 1) verifying AEW position and intensity forecasts, and assessing their relation to convective errors; 2) evaluating environmental features associated with AEW intensity uncertainty; and 3) understanding AEW intensification differences for individual forecasts.

The forecast verification evaluates ECMWF EPS AEW position and intensity forecasts against an average of four operational analyses. During 2007--2009, AEW position forecasts were mainly under-dispersive and characterized by a slow bias, while intensity forecasts were characterized by an over-intensification bias, yet the ensemble-mean errors generally matched the forecast uncertainty. Although 2011--2013 position forecasts were still under-dispersive with a slow bias, the ensemble-mean error is smaller than 2007--2009. In addition, the 2011--2013 intensity forecasts were over-dispersive, and had a negligible intensity bias. Forecasts from 2007--2009 were characterized by higher precipitation in the AEW trough center and high correlations between divergence errors and intensity errors, suggesting the intensity bias is associated with errors in convection. By contrast, forecasts from 2011--2013 have smaller precipitation biases than 2007--2009 and exhibit a weaker correlation between divergence errors and intensity errors, suggesting a weaker connection between AEW forecast errors and convective errors.

The AEW intensity uncertainty is evaluated through the ensemble forecast standard deviation (SD). Both periods exhibit different AEW intensity SD growth rates during the first 48 h, yet are similar thereafter. Forecasts from both periods are stratified based on the 72-h AEW intensity SD to evaluate hypotheses for how different processes contribute to large forecast SD. Forecasts with large SD are characterized by higher relative humidity values downstream of the AEW trough, which is associated with higher precipitation and precipitation SD, suggesting that uncertainty associated with diabatic processes could be associated with large AEW intensity SD. Previous work has suggested that water vapor and convection over Africa is modulated by the phases of convectively coupled equatorial waves and longitude. The 2007--2009 period is characterized by larger intensity SD values in eastern Africa while western Africa is associated with lower intensity SD values; however, few differences exist during 2011--2013. Moreover, AEW intensity SD is generally not a function of convectively coupled equatorial wave phases, suggesting that large-scale convective environment changes are not a major contributor to AEW intensity predictability.

The individual case study component analyzes 6 different cases that exhibit large 72-h AEW intensity SD. For each individual case, differences between two subsets consisting of 10 strong and 10 weak ensemble members are also used to determine whether barotropic, baroclinic, and/or diabatic processes are associated with creating intensity differences. Overall, the high intensity subsets were associated with more intense convection, which can result in larger baroclinic energy conversions and subsequent AEW intensification. This suggests that small-scale convective differences are generally associated with creating intensity differences.