青藏高原极端气温的动力降尺度模拟

Dynamic Downscaling Simulation of Extreme Temperature Indices over the Qinghai-Xizang Plateau

利用ERA-Interim再分析资料作为边界条件,基于耦合陆面模式Noah-MP的区域气候模式WRF在东亚区域进行了动力降尺度模拟(简称WRF2),对比格点观测资料,评估了动力降尺度对青藏高原极端气温指数的模拟能力,在此控制试验基础上,分别将WRF的陆面模式替换为Noah LSM,边界条件替换为CCSM4,进行了两组敏感性试验(分别是WRF1和WRF3),通过与控制试验的比较,分析了边界条件和陆面模式对极端气温指数模拟的影响。结果表明,WRF2能较好地模拟青藏高原极端气温指数气候态的空间分布,但存在一定的冷偏差;受边界条件影响WRF3模拟的极端气温指数的气候倾向率存在负偏差。同时,尽管采用不同的边界条件,耦合相同陆面过程的两次数值试验对极端气温空间分布的模拟能力相似,相比WRF2,WRF1表现出更强的冷偏差;但边界条件对极端气温指数气候倾向率的影响大于陆面模式,WRF3模拟的极端气温指数气候倾向率与观测结果更为接近。

To produce high-resolution climate simulations over Qinghai-Xizang Plateau（QXP） for studying extreme temperature indices,a dynamic downscaling experiment（WRF2 for short） has been performed over East Asia based on WRF coupled with Noah-MP,forced by ERA-Interim（ERA-Int）.Compared with grid observation data,the performance of ERA_Noah-MP in simulating extreme temperature indices over the QXP has been evaluated.To investigate the impacts of the land surface model and the boundary conditions on the simulated extreme temperature indices.Different land surface models（Noah LSM） and boundary conditions（CCSM） are used to run two other dynamic downscaling experiments（WRFl,WRF3 for short）,which have the same model configurations with WRF2 except for the land surface models and the boundary conditions.The results show that the basic features of the spatial patterns of all extreme temperature indices over the QXP can be finely reproduced by WRF2.The centered pattern correlation coefficients of WRF2 compared to observation for most extreme temperature indices are above 0.77 and the simulation of the length of growing season performs best.But simulations for extreme temperature indices suffer from systematic biases to a certain extent,On the whole,a cold bias exists in the simulations of WRF2,especially in winter.WRF2 can simulate the observed inter-annual variations well,and can reproduce the warm trend of extreme temperature indices.However,due to the impacts of the boundary conditions,climatic tendency rates of extreme temperature indices for WRF2 exist negative deviation.In spite of being forced by the different boundary conditions,the two experiments coupled with the same land surface model reproduce the similar spatial distributions of extreme temperature indices under current model configurations;between two land surface models tested,WRFl produced significantly larger cold bias.This difference between two experiments is caused by the different reproduction between two Land surface models for snow cover and snow melting process,bare soil albedo over QXP,which lead to a higher surface albedo in Noah LSM compared with Noah-MP.However,the boundary condition has more influence on climatic tendency rates of extreme temperature indices compared with the land surface model,and thus the climatic tendency rates of WRF3 are close to observation.