Calibration of GFS model by using real time weather data for water requirement forecasting of vegetables in Chtouka region

In the ultimate goal to forecast weather parameters in the agriculture field and precisely in predicting water requirements, a software was developed based on Global forecasting system. To assess its reliability, data of six weather stations were used in the Chtouka region to provide measured weather parameters and allow, thus, software calibration. Air temperature, air humidity, global radiation, evapotranspiration, precipitation and wind speed/direction are studied as weather parameters. Based on statistical comparisons using normalized root mean square error (NRSE), absolute percentage error (MAPE) and mean bias error (MBE), the software was revealed efficient to provide reliable temperature, global radiation and evapotranspiration forecasting but remains unable to predict relative humidity, wind speed/direction and precipitation in an accurate way.