Evaluate the effectiveness of local rainfall forecast application for farmers making decision

Abstract

In Niger rainfall forecasts are provided at large scale and by administrative region which does not support farmer for strategic decisions on crop management and risk reduction strategies. This study evaluate effectiveness of two local rainfall forecasts Application (Accuweather and Weather channel) for agricultural decision making in the regions of Zinder and Tahoua in Niger. The applications were installed in the smartphone of 10 fields technicians based in 10 communes, respectively. Results showed that the best forecast effectiveness was 67% with Accuweather against 73% for weather channel and with rainfall occurrence above 91% and 81%, respectively. For both applications, the higher is forecast, and the greater is the rainfall occurrence. We infer that forecast percentage threshold is important for users for decision making. For example, with a forecast above 70% indicates high rainfall occurrence. In this context, farmers can trust forecasts and schedule their agricultural activities. However, there was a maximum number of forecasts within 21-30% interval with an effectiveness below 25%. Here, it needs not trust the forecasts of these two applications for making agricultural decisions. In this study, many farmers were interested in using these applications. At this stage, it is highly recommended to repeat the experimentation while focusing on acceptance threshold, which could better serve farmers in their decision making.