The quality of precipitation forecasting has a very important role in optimizing irrigation of field crops. Forecasting timing and amount of rainfall accurately has a significant impact on water-efficient irrigation. The total amount of rainfall, its intensity and its distribution influences soil moisture which has a major impact on plant growth and development. Great variability of rainfall is one of the main characteristics of the Polish climate. A high level of verifiability of weather forecasting is the basic and required condition. The aim of the analyses was to evaluate the quality of rainfall forecasts using the Unified Model (UM) and the COAMPS (Coupled Ocean/Atmosphere Mesoscale Prediction System) mesoscale numerical models. The accuracy of the forecasts (rainfall quantity and time) was compared with the data obtained by automatic meteorological stations for Skierniewice (51掳57鈥睳, 20掳08鈥睧), Biala Rzadowa (51掳15鈥睳, 18掳27鈥睧) and Bialousy (53掳24鈥睳, 23掳13鈥睧). The verification covered the period from April to October 2010 and 2011. For the rain quantity, mean absolute error (calculated for the sum of daily rainfall) and mean relative error (for the sum of the rainfall for the period of April-October) were calculated. The forecast accuracy indicator (%) was determined for the evaluation of accuracy of the beginning of rainfall forecast. The presented evaluation of the forecast rainfall showed a significant space and time diversity between forecast and measured values. On the basis of the evaluation carried out in the seasons of 2010 and 2011, the most credible forecasts were obtained using the UM model. The quality of forecasting using the UM model seems to be high enough to apply it as an operational model in Decision Support Systems.CT VII International Symposium on Irrigation of Horticultural Crops, Geisenheim, Germany.
展开▼
