一种订正格点温度资料的途径

A Method for Correcting Grid Temperature Data

摘要: 基于欧洲中期天气预报中心ECMWF高分辨率模式24 h预报场资料，从2016年6月1日~2019年9月30日2 m逐3 h最高、最低温度资料中提取日最高、最低温度格点预报值，选取模式网格产品中最靠近武威市区域内6个站点的格点值代替站点的最高、最低温度预报值。检验评估了模式产品输出和回归模型预报订正能力，发现模式输出的最高、最低温度预报准确率川区优于山区，特别是最低温度预报准确率山区误差较大；经过回归模型订正后，区域内6个站点的最高温度预报准确率均 ≥ 75.1%，川区和浅山区站点的最低温度预报准确率 ≥ 71%，提高幅度较大，但深山区的天祝、乌鞘岭的最低温度预报准确率仅为65.1%、55.2%，订正效果不显著。深入分析误差产生的原因，从局地温度变化影响因子出发，构造时空误差订正方法，有效提高了高海拔地区最低温度的预报订正能力，深山区站点的最低温度预报准确率平均提高8.8%~12%，预报绝对误差平均减小了0.5℃；业务试运行中各站最低温度预报准确率达71.7%~85.2%。构造的时空误差订正方法，按照温度年变化分上升区间、下降区间建立预报模型，使建立的预报模型更为客观；在温度订正方法上加入了温度个别变化对局地温度变化影响因素，使得构造的温度订正模型更趋合理；该方法计算简单，便于在格点极端温度预报订正上开展释用，为格点极端温度预报订正提供了一种新的思路和方法。

Abstract: Based on the 24 h forecast field data of the ECMWF high resolution model of the European Centre for Medium-Range Weather Forecasts, the daily maximum and minimum temperature grid point forecast values are extracted from the 2 m every three hour maximum and minimum temperature data from June 1, 2016 to September 30, 2019. The maximum and minimum temperature of grid points closet to the stations in Wuwei City were selected as the value of the stations. By testing and evaluating the forecast correcting capability between the model product output and the regression method, it was found that maximum and minimum temperature forecast accuracy rate of the model output in plain area was better than that in the mountainous area, especially the error of minimum temperature in the mountainous area was larger; after the regression model correcting, the maximum temperature forecast accuracy rate of the 6 stations in the region was all ≥75.1%, and the minimum temperature forecast accuracy rate of the stations in plain areas and shallow mountainous areas was ≥71%, which is a significant increase, but the minimum temperature forecast accuracy rate of Tianzhu and Wushaoling in the deep mountainous area was only 65.1%, 55.2%, respec-tively, the correction effect is not significant. Indepth analysis of the causes of errors, starting from the local temperature change influencing factors, constructing a spatiotemporal error projection correction method, effectively improved the forecast correction ability of the minimum temperature in high-altitude areas, and the minimum temperature accuracy rate of the stations in deep mountainous areas increased by an average of 8.8%~12%, the absolute forecast error has been re-duced by 0.5˚C on average; the minimum temperature forecast accuracy rate of each station in the operational trial operation reached 71.7%~85.2%. Spatio-temporal error correction method is established according to the rising and falling intervals of annual temperature change, which makes the prediction model more objective. The influence factors of individual temperature changes on local temperature changes are added to the temperature correction method, which makes the constructed temperature correction model more reasonable. The method is simple in calculation and easy to be applied in the prediction and correction of extreme temperature in grid, which provides a new idea and method for the prediction of extreme temperature in grid.