Adaptive Weather Forecasting using Local Meteorological Information

In general, meteorological parameters such as temperature, rain and global radiation are important for agricultural systems. Anticipating on future conditions is most often needed in these systems. Weather forecasts then become of substantial importance. As weather forecasts are subject to uncertainties, there is a need in minimising the uncertainties. In this paper, a framework is presented in which local weather forecasts are updated using local measurements. Kalman filtering is used for this purpose as assimilation technique. This method is compared and combined with diurnal bias correction. It is shown that the standard deviation of the forecast error can be reduced up to 6 h ahead for temperature, up to 31 h ahead for wind speed, and up to 3 h for global radiation using local measurements. Combining the method with diurnal bias correction leads to a further increase in performance in terms of both bias and standard deviation.