Verification Of Medium-Range Weather Forecasts Over Tanzania

ABSTRACT

There are several weather forecasting systems within and outside Tanzania issuing medium-range rainfall forecasts. However, few attempts have been made to evaluate the accuracy of these forecasts. Hence the major objective of this study was to evaluate some selected medium-range rainfall forecasts and forecasting systems issuing those forecasts. Based on the above objective, the study could firstly evaluate medium-range rainfall forecasts in selected bimodal and unimodal areas of Tanzania; secondly identify the most important verification indicators on rainfall forecast accuracy relevant to agriculture and lastly compare forecasted rainfall amounts by Accuweather with actual rainfall amounts as recorded by TMA. The datasets used include 10-day rainfall forecast generated by two forecasting systems in Africa namely; the Tanzania Meteorological Agency (TMA), the Intergovernmental Authority on Development (IGAD) Climate Prediction and Application Centre (ICPAC) formally known as Drought Monitoring Centre (DMC)-Nairobi. Others included the National Centre for environment Prediction (NCEP) in the USA, and State College Accuweather Company also in the USA, and observed 10-day accumulations of precipitation of five selected stations from TMA. The method used to investigate the overall and specific objectives included the contingency table that shows the frequency of “yes” and “no” forecasts and occurrences. In the same way, categorical statistics skill measures were computed from the elements in the contingency table to describe particular aspects of forecast performance. Results from Accuweather forecast products appeared to perform better in the short rainy season (OND) than in the long rainy season (MAM) with respect to accuracy. Similarly, the ICPAC forecasts products appeared to perform better in the OND than in MAM season. Generally on the other hand, all the three forecasting systems (NCEP, Accuweather and ICPAC) all forecasts were worse than the reference forecast based on the HSS negative skill scores obtained for all rainy seasons.