ECMWF高分辨率数值模式对广安地区气温预报性能检验分析

为了解ECMWF高分辨率数值预报模式(以下简称“EC”)对广安地区气温的预报性能，提高预报质量。利用EC气温预报产品，对2015～2017年广安地区最低(高)气温进行预报性能检验。结果表明:EC模式预报最低气温，正确率较高；预报最高气温，正确率波动大，随月份呈明显的“V”型变化，盛夏7、8月最低。预报误差随时效延长，略有增大；最低气温误差小于最高气温误差；最低气温误差各月无明显差异，最高气温误差在盛夏7、8月最大。最低气温预报效果区域差异不明显；最高气温预报效果受地形影响较大。根据订正指标，进行气温订正预报，可有效提升预报正确率。EC模式预报高温时的最高气温偏小，经过订正后，各站各时效正确率均明显提高，正确率提升20.6～91.3%，具有较高的参考价值。 

Verification and Analysis on the Temperature Forecast Performance of ECMWF High Resolution Numerical Model in Guang'an

To better understand and improve the forecast performance of the daily minimum and maximum temperatures in Guang’an city, Using ECMWF high-resolution numerical forecasting model (hereinafter referred to as EC) temperature forecasting products to test the daily minimum and maximum temperatures in 2015-2017.The results show that the EC model predicts the minimum temperature and the correct rate is higher. The maximum temperature is predicted, and the correct rate fluctuates greatly. It shows an obvious “V” type change with the month, and it is the lowest in July and August in midsummer. The forecast error is extended by time and slightly increased; the minimum temperature error is less than the maximum temperature error; the minimum temperature error has no significant difference in each month, and the maximum temperature error is the largest in July and August in midsummer. The regional difference in the effect of the minimum temperature forecast is not obvious; the effect of the maximum temperature forecast is greatly affected by the terrain. According to the revised indicators, the temperature correction forecast can effectively improve the forecast accuracy rate. The EC model predicts that the maximum temperature is too small at high temperature. After correction, the correct time rate of each station is obviously improved, and the correct rate is increased by 20.6 to 91.3%, which has a high reference value.