Skill and usefulness of regional seasonal forecasts for adoption to climate change for agricultural production in Tanzania

Abstract

The world climate is changing and sub-Saharan Africa experiences it through climate variability. Since the majority of people in the region depend on agriculture for their livelihood, the predictability of rainfall is very critical in reducing their vulnerability and seizing opportunities. This study aimed at verifying skill and usefulness of the forecasts from two drought monitoring centres-GHACOF and SARCOF-by looking at accuracy, bias and skill in 16 synoptic stations in Tanzania. The results have indicated that seasonal forecasts by the two centres have similar patterns of accuracy. Both have better accuracy in the northern parts of Tanzania compared to the southern parts during October·November-December (OND) period. During January-February- March for SARCOF and March·April·May for GHACOF forecasts show better accuracy indices in the southern parts compared to the northern parts of the country. The usefulness of both forecasts is still very low because the accuracy levels are below 0.8. The forecasts are not showing much bias, i.e., they are not over· or under· forecast rainfall amounts. However, the forecasting skill for both centres is very low (majority of stations with HSS < 0.2). This study recommends the two centres, together with national meteorological services, to improve accuracy and skill of seasonal rainfall forecasts. The seasonal forecasts should be disseminated widely to users, including farmers. However, a caution should be provided to farmers as the accuracies of the seasonal forecasts are seasonal, location and forecast dependent.