WRF模式中初始土壤温湿度对华北冬季近地面要素预报的影响
| 1. 河北省气象与生态环境重点实验室,河北 石家庄 050021; |
网络出版日期: 2025-12-31
基金资助
河北省气象局青年基金项目(22ky20);河北省重点研发计划项目(22375404D);河北省自然科学基金项目(D2025304001);中国气象科学研究院基本科研业务费专项基金项目(2023Z004,2023Z015)
Impact of Initial Soil Temperature and Moisture on the Temperature and Relative Humidity at 2 m in the WRF Model over North China in Winter
Online published: 2025-12-31
基于WRF模式,研究了来自CMA-GFS分析场、CLDAS融合实况产品两种土壤温、湿度初始场对2023年 12月 10-25日华北地区近地面温、湿要素预报效果的影响,并分别讨论了初始土壤偏冷/偏暖、偏干/偏湿对2 m气温的影响差异。结果表明:(1)在预报时长21 h之前,以CLDAS为土壤温湿度初始场预报的相对湿度/2 m气温效果较差;在21 h之后,CLDAS初始土壤温湿度对相对湿度/2 m气温的预报效果更好,均方根误差(RMSE)最多降低8. 3%/10%。(2)以CLDAS为土壤温湿度初始场时,由于模式初始场中,大气和土壤温、湿度的数据来源不同,大气和土壤温、湿度通过更多地表向大气输送的感热、潜热通量进行热调整,在21 h达到平衡,故在21 h之后相对湿度/2 m气温的预报效果转优。(3)空间上,在山西南部、河北中南部及其以南地区,CLDAS土壤温湿度为初值预报的相对湿度负偏差更小、日最高气温暖偏差更小、日最低气温暖偏差更大,在河南的预报效果更好,其相对湿度偏差降低了8%,日最高气温偏差减少了1. 5 ℃;在山西北部、河北北部及其以北地区,CLDAS土壤温湿度初值预报的相对湿度负偏差更大,预报的日最高、最低气温均更优。(4)当初始土壤偏湿、偏冷时,2 m气温的预报效果最好,几乎接近于真实气温;当初始土壤偏湿、偏暖时,土壤温湿度初值对2 m气温的预报效果影响较小,2 m气温预报效果整体欠佳。相比土壤温度,土壤湿度初值对2 m气温预报影响更大,当初始土壤偏干时,对地表热通量的影响最大,感热通量更大,潜热通量更小,日最低气温的预报效果更好。
张 琳, 张卫红, 尹金方, 丁明虎 . WRF模式中初始土壤温湿度对华北冬季近地面要素预报的影响 [J]. 高原气象, 0 : 1 . DOI: 10.7522/j.issn.1000-0534.2025.00058
The effects of initial soil temperature and moisture were explored by performing a series of 84-hour numerical simulations form 10 to 25 December 2023,using the WRF model with soil temperature and moisture initialized with CMA-GFS forecasts and the China Meteorological Administration Land Data Assimilation Sys‐ tem(CLDAS)analysis respectively. It showed that the relative humidity(RH2m)and the temperature(T2m)at 2 m were poorly during the first 21-hour integration and getting better during the following time integration ini‐ tialized with CLDAS than that with CMA-GFS,with a maximum root mean square error(RMSE)decrease of 8. 3% and 10% respectively. Further diagonosis indicated that the first 21-hour integration usually a accompanied with more sensible and latent heat fluxes in case of the soil temperature and moisture initialized with CLDAS. It means that the model would take a longer spin-up time during which any satisfiable forecasted T2m and RH2m might not be gotten,due to the in consistanece in initial conditions which came from the different source datasets containing some more reliable variables although. Spatially,when the model was initialized with CLDAS,it rep‐ resented much more favourable with smaller negative deviation of RH2m and positive deviation of the daily max‐ imum T2m,which maximally decreased by 8% and 1. 5 °C in Henan respectively,and unfavourable with larger positive deviation of the daily minimum T2m in the southern region of Shanxi and central Hebei. Whereas,it rep‐ resented pretty well performance of daily maximum and minimum T2m and bad performance of RH2m with larg‐ er negative deviation in the northern region of northern Shanxi and northern Hebei. Among groups of initial soil conditions,the moister and colder initial soil group had the best forecasting performance of T2m which is tightly closed to the observation,and the moister and warmer initial soil group has the worst although the T2m differ‐ ence between the simulations initialized with CLDAS and CMA-GFS is small. Compared with the initial soil tem‐ perature,the initial soil moisture has a greater impact on T2m. The drier initial soil would lead to a better fore‐ casting performance of daily minimum T2m,accompanied with larger sensible heat flux and smaller latent heat flux in the integration.
/
| 〈 |
|
〉 |