Abstract:

In order to improve correction ability and forecasting level of intelligent grid. Based on the slice data of Gansu Province of objective guidance product from Central Meteorological Observatory of China and daily grid temperature data from Chinese Land Data Assimilation System Version 2.0 (CLDAS-V2.0) of CMA, the maximum and minimum temperature of 0.05°×0.05° grid points in the eastern Hexi Corridor (101.0°E-104.5°E, 36.0°N-40.0°N) were corrected, tested and evaluated by using Kalman filtering method and sliding training correction method. The results are as follows: (1) For seasonal comparison, the mean absolute errors of maximum and minimum temperature of Kalman filter and sliding training correction products were both smaller than objective guidance product at all seasons, and all values were less than 2.00 ℃. The forecast accuracy of maximum and minimum temperature of Kalman filter and sliding training correction products were greater than 70% at all seasons. which the maximum temperature was 6%-13% higher and the minimum temperature was 8%-24% higher. (2) For spatial comparison, the mean absolute errors of the maximum and minimum temperature of Kalman filter and sliding training correction products were 1.00-2.00 ℃, but greater than 2.00 ℃ in a few areas. The forecast accuracy of maximum (minimum) temperature of Kalman filter and sliding training correction products were greater than 70% (60%-70%) in most areas, and greater than 80%(70%) in a few areas. (3) As a whole, the correction skills of maximum and minimum temperature of Kalman filter and sliding training correction products were basically positive, and were greater than 0.300 in a few seasons and a few areas. It showed that the two correction methods have good prediction and correction ability, which can provide certain technical support for the future temperature forecasting operations.

Key words: intelligent grid, Kalman filtering, sliding training correction, temperature correction forecast, test and evaluated

摘要：

为提高智能网格的订正能力及预报水平，基于中央台客观指导产品的甘肃省切片数据和中国气象局陆面数据同化系统（Chinese Land Data Assimilation System Version 2.0，CLDAS-V2.0）日网格实况产品，采用卡尔曼滤波和滑动训练订正两种方法，对河西走廊东部地区（101.0°E—104.5°E，36.0°N—40.0°N）0.05°×0.05°格点最高、最低气温进行订正、检验和评估。结果表明：（1）季节对比，卡尔曼滤波和滑动训练订正产品对四季最高、最低气温的平均绝对误差均小于中央台客观指导产品，均小于2.00 ℃；卡尔曼滤波和滑动训练订正产品对四季最高、最低气温的预报准确率均大于70%，其中最高气温偏高6%~13%，最低气温偏高8%~24%。（2）空间对比，卡尔曼滤波和滑动训练订正产品对最高、最低气温的平均绝对误差绝大部分地区在1.00~2.00 ℃，个别地区大于2.00 ℃；卡尔曼滤波和滑动训练订正产品对最高（最低）气温的预报准确率大部分地区大于70%（60%~70%），个别地区大于80%（70%）。（3）总体上，卡尔曼滤波和滑动训练订正产品对最高、最低气温订正技巧基本为正技巧，个别季节和部分地区订正技巧大于0.300。说明两种订正方法具有较好的订正预报能力，可为今后的温度预报业务提供一定的技术支持。

关键词: 智能网格, 卡尔曼滤波, 滑动训练订正, 温度订正预报, 检验和评估