Abstract:
In this paper, a comprehensive correction technique by combining moving-biweight correction method with successive correction method for spatial error is proposed to correct and analyze the bias of 2 m maximum and minimum temperature forecast of ECMWF (European Centre for Medium-Range Weather Forecasts) high resolution model within 24-168 h forecast time lengths during the period from 1 May 2016 to 1 May 2017. The main conclusions are as follows. (1) The 2 m maximum and minimum temperature forecasts of ECMWF model are obviously lower, on average, than the observation in Jiangxi Pro-vince. It is feasible to correct the bias of model temperature considering that the spatial distribution of ECMWF model temperature shows a significant systematic deviation which is stable at different forecast times. (2) The moving-biweight correction method having longer moving correction period has a better effect on the model temperature, which is ideal for the 20 d forecast. By combining the successive correction method for spatial error, the quality of temperature forecasting with the moving-biweight correction method could be further improved, even though the effect of moving-biweight correction method might not be satisfactory during the seasonal transition period. (3) The temperature accuracy shows that the quality of temperature forecasting has been significantly improved after bias correction conducted by the comprehensive correction technique by combining moving-biweight correction method with successive correction method for spatial error. After bias correction, the maximum temperature accuracies of forecast errors ≤2℃ in 24, 48, 72 h forecasts are greatly increased from 0.59, 0.55, 0.52 to 0.75, 0.68, 0.62, respectively, and the minimum temperature accuracies are increased from 0.84, 0.83, 0.82 to 0.89, 0.87, 0.85, respectively. After bias correction, the maximum and minimum temperature accuracies of the 72 h forecast are even greater than that of the 24 h forecast before correction. In general, the deviation of the temperature forecasting is effectively reduced by this proposed comprehensive correction technique, even with a more uniform spatial distribution. (4) For the alpine station, the maximum and minimum temperatures after correction are basically consistent with the observation. In addition, the successive correction method for spatial error has a positive correction effect on temperature correction for the reason that the deviation is within ±1℃, but shows a certain negative correlation with that of moving-biweight correction method. This comprehensive correction technique has been successfully applied to the objective forecasting operational system of fine meteorological elements in Jiangxi Province.