CMIP6 HighResMIP模式对青藏高原东侧极端降水的模拟评估及热动力效应  

Assessment of extreme precipitation simulations over the eastern Qinghai-Xizang Plateau using CMIP6 HighResMIP models:thermodynamic and dynamic contributions

作　　者：王康宁 陈权亮 葛非 林芷叶 WANG Kangning;CHEN Quanliang;GE Fei;LIN Zhiye(School of Atmospheric Sciences/Plateau Atmosphere and Environment Key Laboratory of Sichuan Province/Chengdu Plain Urban Meteorology and Environment Observation and Research Station of Sichuan Province/Sichuan Meteorological Disaster Prediction and Early Warning Engineering Laboratory,Chengdu University of Information Technology,Chengdu 610225,China;Heavy Rain and Drought-Flood Disasters in Plateau and Basin Key Laboratory of Sichuan Province,Chengdu 610027,China)

机构地区：[1]成都信息工程大学大气科学学院/高原大气与环境四川省重点实验室/成都平原城市气象与环境四川省野外科学观测研究站/四川省气象灾害预测预警工程实验室,四川成都610225 [2]高原与盆地暴雨旱涝灾害四川省重点实验室,四川成都610072

出　　处：《大气科学学报》2025年第2期240-254,共15页Transactions of Atmospheric Sciences

基　　金：国家自然科学基金项目(U2442210,42375047);四川省自然科学基金项目(2024NSFTD0017,2024NSFSC0064);高原与盆地暴雨旱涝灾害四川省重点实验室开放研究基金项目(SZKT202304)。

摘　　要：青藏高原东侧(the eastern slope of the Qinghai-Xizang Plateau,ESQP)因其复杂多变的地形和显著的地势梯度,一直是气候模式模拟挑战性极高的地区。本研究基于第六次国际耦合模式比较计划中的高分辨率模式比较计划(CMIP6 HighResMIP),利用高分辨率(high-resolution,HR)模式及其对应的低分辨率(low-resolution,LR)模式,对青藏高原东侧地区降水模拟进行了评估。结果表明,LR与HR模式均能较好地重现青藏高原东侧地区年降水的空间分布,且均反映出从西北向东南递增的降水模态。对比之下,HR模式在降低青藏高原东侧地区降水模拟偏差方面较LR模式展现了显著的优势。多模式集合(multi-model ensemble,MME)的分析结果揭示,年均降水量的模拟偏差由LR模式的1.05 mm/d减少至HR模式的0.96 mm/d。在模拟极端降水事件方面,HR模式也显著优于LR模式,这一提高在四川盆地尤为突出。采用强降水天数(R10mm)和单日最大降水量(Rx1day)作为评价极端降水模拟的指标,相较于LR模式,HR模式的R10mm相对误差下降6%,Rx1day相对误差降低5%。最后,利用物理尺度诊断方程分析了LR和HR模式在极端降水模拟中的热力、动力效应,讨论了HR模式模拟性能提升的原因,发现更高分辨率的模式能够更为真实地再现大气环流和水汽条件。量化分析表明,HR模式在动力效应提升方面有91%的相对贡献,在热力效应提升方面有8%的相对贡献。这表明,模式分辨率的提升对于青藏高原东侧地区降水模拟准确性是一个至关重要的因素。The eastern slope of the Qinghai-Xizang Plateau(ESQP)features complex topography and substantial elevation gradients,making it a region of intricate climate dynamics.This study assesses the ability of the Coupled Model Intercomparison Project Phase 6(CMIP6)High-Resolution Model Intercomparison Project(HighResMIP)models to simulate precipitation over the EQSP,with a focus on comparing high-resolution(HR)and low-resolution(LR)models in capturing both annual and extreme precipitation events.Using data from 1985 to 2014,including historical simulations and observational datasets,the study evaluates extreme precipitation indices such as R10mm(heavy precipitation days)and Rx1day(maximum 1-day precipitation).Results show that both HR and LR models successfully capture the general spatial distribution of annual precipitation,which increases from northwest to southeast across the region.However,HR models exhibit significant improvements over LR models in reducing biases and improving simulation accuracy.Specifically,the annual precipitation bias is reduced from 1.05 mm·d^(-1)in LR models to 0.96 mm·d^(-1)in HR models,demonstrating the benefits of increased resolution in minimizing simulation errors.For extreme precipitation events,HR models outperform LR models in both occurrence and intensity representation.The R10mm index shows a 6%reduction in relative error for HR models compared to LR models,while the Rx1day index exhibits a 5%improvement in HR model performance.These improvements are particularly notable in the Sichuan Basin,a region historically challenging to simulate due to its complex terrain and variable moisture conditions.Further analysis investigates the thermodynamic and dynamic contributions to differences between HR and LR model simulations using a physical scaling diagnostic equation.Results indicate that dynamic effects account for 91%of the improvements observed in HR models,while thermodynamic effects contribute only 8%.This suggests that higher resolution enhances the representation of atmospheric

关 键 词：青藏高原 CMIP6 High-ResMIP 极端降水 物理尺度诊断方程 

分 类 号：P426.6[天文地球—大气科学及气象学]