T.G. Doeswijk and K.J. Keesman (The Netherlands)
weather forecast, Kalman filter, forecast error, bias, standard deviation
For controlling agricultural systems, weather forecasts can be of substantial importance. Studies have shown that fore cast errors can be reduced in terms of bias and standard de viation using forecasts and meteorological measurements from one specific meteorological station. For agricultural systems usually the forecasts of the nearest meteorological station are used whereas measurements are taken from the systems location. The objective of this study is to evaluate the reduction of the forecast error for a specific agricultural system. Three weather variables , that are most relevant for greenhouse systems are studied: temperature, wind speed, and global radiation. Two procedures are used consecu tively: diurnal bias correction and local adaptive forecast ing. For each of the variables both bias and standard devi ation were reduced. In general, if local measurements are reliable, forecast errors can be reduced considerably.
Important Links:
Go Back
