为提升区域数值预报系统2 m气温预报性能,利用土壤温度和土壤湿度站点观测资料,对中国气象局陆面数据同化系统(CMA Land Data Assimilation System,CLDAS)陆面资料在浙江地区的精度进行评估,并将其融合应用于浙江省数值预报业务系统.结果表明:CLDAS土壤温度、土壤湿度产品相对于美国全球预报系统(Global Forecast System,GFS)分析场,与观测相比具有更小的均方根误差和更高的相关系数,在浙江省有较好的适用性.个例分析表明区域数值模式2 m气温预报对陆面资料变化较敏感,融合CLDAS地表温度、土壤温湿度实时分析产品的初始场,可持续影响到模式预报后期,主要通过地表感热、潜热通量直接影响气温变化.从均方根误差来看,与基于GFS分析场作为陆面初始场的区域模式预报相比,应用了CLDAS陆面资料的模式预报改进了13.6%.2021年7月阶段性应用结果显示,模式初始场融合CLDAS陆面资料后有效提高了浙江省2 m气温预报水平,融合后的预报改进效果夜间较白天明显,且晴热高温天气背景下较梅雨期、台风期改进更佳.高温天气预报评估进一步表明,CLDAS陆面资料的应用对浙江省高温事件预报有较好的改进,尤其对金衢盆地等高温区改进明显.
Research on improvement of temperature forecasts of the regional numerical prediction system using CLDAS land data
In order to improve the performance of the regional mesoscale numerical prediction system on temperature forecasts,the CLDAS(CMA Land Data Assimilation System)land data in Zhejiang Province is evaluated for its precision using observed soil tempera-ture data as well as soil moisture data,and then it is applied in Zhejiang numerical prediction system.The results are shown as follows:The CLDAS soil temperature and soil moisture products have smaller root mean squared errors and higher correlation coefficients com-pared to the observations than those products of the GFS(Global Forecast System)analysis field,and they have good applicability in Zheji-ang Province.The case study indicates that 2 m temperature predicted by the regional numerical model is sensitive to the change of land data.It exerts a sustainable influence on the model forecasts at a later stage through blending the CLDAS real-time surface temperature,soil temperature and soil moisture analysis products on the initial field.More specifically,the temperature changes are directly affected by the surface sensible heat flux and latent heat flux.Besides,the root mean squared error of the model prediction using the CLDAS data is lowered by 13.6%in comparison to that of the model prediction using the GFS analysis field as the initial land surface field.The phased application results in July 2021 show the model initial field blending the CLDAS data can effectively improve the forecast accuracy of the 2 m temperature in Zhejiang Province.Moreover,the results present a more favorable improvement at night than in the daytime,and more effective forecasts can be presented under dry-hot weather condition than tropical cyclones and Mei-Yu fronts periods.The evaluation of high temperature forecasts further indicates that the application of CLDAS land data has a good improvement effect on the predictions of high temperature events in Zhejiang Province,especially in high-temperature areas such as the Jinqu Basin.
soil temperaturesoil moisture2 m temperaturehigh temperaturenumerical weather prediction model
NETL
NSTL
万方数据
