浙江地区CMA-WSP2.0地面太阳辐射预报产品的检验评估  

Test and Evaluation of CMA-WSP2.0 Surface Solar Radiation Forecast Products in Zhejiang Region

作　　者：王伟 周悦 潘娅英[1] 张加易 Wang Wei;Zhou Yue;Pan Yaying;Zhang Jiayi(Zhejiang Meteorological Service Center,Hangzhou 310059,China)

机构地区：[1]浙江省气象服务中心,杭州310059

出　　处：《科技通报》2025年第4期1-9,29,共10页Bulletin of Science and Technology

基　　金：浙江省自然科学基金联合基金项目(LZJMY23D050001);浙江省基础公益研究计划(LY24D050002);浙江省气象局重点项目(2023ZD11);浙江省气象局重点项目(2020ZD15)。

摘　　要：本文对中国气象局气象服务平台2.0版本的地面太阳辐射预报效果进行检验评估。基于杭州站和洪家站2023年7月至2024年6月的地面太阳辐射观测资料,利用评价指标对模式的辐射预报开展全年、逐月、分季节、日变化、分云量、分雨量评估。总体来看,杭州站14 d预报与观测的相关系数在0.67~0.86,洪家站在0.67~0.85,均随预报时效延长而递减;1~5 d相关系数均大于0.8,预报效果较好,6~14 d相关系数均小于0.8,预报效果较差。针对逐月预报,模式对11—12月的预报效果最好,11月预测与观测相关系数达0.947,6—8月的预报效果最差。对于日变化评估,模式在中间时段(10:30—15:30)预报效果相对较差,其他时段相对较好。此外,模式对地面太阳辐射的预报存在较强的季节性,5 d预报效果冬天最好且稳定,平均相关系数达0.91,夏季5 d预报效果最差且差异较大,春秋季节5 d预报效果相对较好。不同云量条件下,阴天预报效果最好,多云最差,晴天较好;降水天气的预报效果较差,且随着雨量的增加预报准确性总体递减,但天气过程的持续稳定发展有利于模式预报。总体来说模式1~5 d地面太阳辐射预报效果较好,6~14 d较差,不同月份、季节、时段、云量、雨量下预报效果有所差异,未来仍需开展模式产品订正优化以提高其本地化应用效果。In order to test and evaluate the surface solar radiation forecast effect of China Meteorological Administration Weather Service Platform(CMA-WSP) version 2.0.Based on the surface solar radiation observation data of Hangzhou Station and Hongjia Station from July 2023 to June 2024,the annual overall,monthly,seasonal,diurnal,different cloud cover and rainfall of the model radiation forecast were evaluated using evaluating indexes.In general,the correlation coefficients of 1-14 d forecast and observation ranged from 0.67 to 0.86 at Hangzhou station and from 0.67 to 0.85 at Hongjia Station,which decreased with the forecasting time;The correlation coefficients of 1~5 d were all >0.8,indicating a good forecasting effect,the correlation coefficients of 6-14 d were all <0.8,indicating a poor forecasting effect.In the monthly evaluation results,the model has the best prediction effect from November to December,the correlation coefficient between prediction and observation in November is 0.947,and the prediction effect in June to August is the worst.In the diurnal variation evaluation results,the prediction effect of the model is relatively poor in the middle period(10:30 to 15:30),but relatively good in other periods.There is a strong seasonality in the forecast of surface solar radiation,the 5 d forecast effect is the best and stable in winter,with an average correlation coefficient of 0.91,the 5 d forecast effect is the worst and the difference is large in summer,and the 5 d forecast effect is relatively good in spring and autumn.Under different cloud cover conditions,overcast is the best,cloudy day is the worst,sunny day is better;The forecast effect of precipitation weather is poor,and the forecast accuracy decreases with the increase of rainfall,but the steady development of weather process is conducive to the model forecast.In general,the prediction effect of surface solar radiation on 1-5 d of the model is better than that on 6-14 d of the model,the prediction effect varies with different months,seasons,time period

关 键 词：CMA-WSP2.0 地面太阳辐射 评估检验 准确性 浙江 

分 类 号：P456[天文地球—大气科学及气象学]