Maize yield forecasting for Zimbabwe farming sectors using satellite rainfall estimates

Southern Africa rainfall station network is suffering from an unfortunate serious decline while climate-related food insecurity is worsening. In the current work, we demonstrate the possibility of exploiting the complementary roles that remote sensing, modeling, and geospatial data analysis can play in forecasting maize yield using data for the growing seasons from 1996/1997 to 2003/2004. Satellite-derived point-specific rainfall estimates were input into a crop water balance model to calculate the Water Requirement Satisfaction Index (WRSI). When these WRSI values were regressed with historical yield data, the results showed that relatively high skill yield forecasts can be made even when the crops are at their early stages of growth and in areas with sparse or without any ground rainfall measurements. Inferences about the yield at national level and small-scale commercial farming sector (SSCF) sector can be made at confidence levels above 99% from the second dekad of February. However, the most unstable models are those for the communal farming sectors whose inferences for yield forecast can only be made above the 95% confidence level from the end of February, after having recovered from a state of complete breakdown two dekads earlier. The large-scale commercial farming (LSCF) sector has generally the weakest fitting, but it is usable from the first dekad of February to the end of the rainy season. Validation of the national yield models using independent data set shows that an early estimation of maize yield is quite feasible by the use of the WRSI.