Research On Correction Algorithm Of Grid Point Temperature Forecast Under Special Terrain In Sichuan

Due to the large topographic differences between the eastern and western parts of Sichuan,the terrain is high in the west and low in the east,and the terrain is complex,the model forecast effect of temperature needed to be improved.This study used the SCMOC model forecast products based on January 1,2016 to February 29,2020,the subjective forecast of SNFFD of Sichuan Meteorological Bureau and the objective forecast product of SPCO for 2m daily maximum precipitation,24 h cumulative precipitation,total cloud forecast and relative hygrometer wind direction wind speed and other forecast data studied the spatial distribution characteristics of the model forecast error in Sichuan.Combined with the site observation data of the same period,the correlation between model forecast errors and meteorological factors was analyzed using data such as relative humidity and precipitation,the errors of the station forecast results under different terrain classifications were analyzed,an error regression model was established,and the forecast was carried out Effect test analysis.The main conclusions are as follows:(1)From the perspective of the spatiotemporal changes in the temperature of the Tibetan Plateau and its eastern regions from 1960 to 2018,the temperature of the Tibetan Plateau was warming faster than that of Sichuan,and the warming trend was most significant in winter;(2)The temperature forecast error of the Sichuan region model showed obvious regional characteristics.On the whole,the forecast error of the western Sichuan model was greater than that of the eastern region.In the three model forecast(SCMOC,SNWFD and SPCO)results,the SPCO objective forecast of the Sichuan Meteorological Bureau had the smallest error,followed by the SNFFD subjective forecast,and the China Meteorological Administration SCMOC refined forecast had the worst result.The number of days with positive error is significantly less than the number of days with negative error under different forecast timeliness,indicating that the temperature forecast in Sichuan is mainly low;(3)The forecast error of 2m temperature in Sichuan area had a significant seasonal variation.The errors in summer and autumn are small,and the errors in spring and winter are large.The forecast error of the daily maximum temperature of2 m was greater than the forecast error of the daily minimum temperature,and as the forecast time increased,the forecast error gradually increased;(4)The three terrain factors of longitude,slope and altitude had a significant correlation with the model prediction errors.The errors decreased with increasing longitude and altitude,and increased with increasing slope.The error first decreased and then increased with the increase of the daily temperature difference,and it increases-decreases-increases with the increase of precipitation intensity.The average absolute error had a negative correlation with relative humidity.The greater the relative humidity,the smaller the model prediction error;(5)The error regression equations obtained under the SCMOC model had significantly improved the annual forecast results of the daily maximum and minimum temperatures,while the error regression equations obtained under the SNFFD and SPCO models had improved the model forecast errors under different terrains to varying degrees.Among them,the improvement effect of the prediction results was more obvious under the terrain such as flat dam,ridge and leeward slope.