地学前缘 ›› 2023, Vol. 30 ›› Issue (3): 515-528.DOI: 10.13745/j.esf.sf.2022.12.60

• 非主题来稿选登 • 上一篇    下一篇

基于CMIP6多模式的和田河流域未来气候变化预估

何朝飞1,2(), 骆成彦3, 陈伏龙1,2,*(), 龙爱华1,4, 唐豪1,2   

  1. 1.石河子大学 水利建筑工程学院, 新疆 石河子 832000
    2.生态水利工程兵团寒旱区重点实验室, 新疆 石河子 832000
    3.新疆水利水电勘测设计研究院有限公司, 新疆 乌鲁木齐 830000
    4.中国水利水电科学研究院 流域水循环模拟与调控国家重点实验室, 北京 100038
  • 收稿日期:2022-10-18 修回日期:2023-01-10 出版日期:2023-05-25 发布日期:2023-04-27
  • 通讯作者: *陈伏龙(1978—),男,教授,博士生导师,主要从事水文学及水资源问题研究工作。E-mail: cfl103@shzu.edu.cn
  • 作者简介:何朝飞(1992—),男,博士研究生,主要从事水文学及水资源问题研究。E-mail: 530731798@qq.com
  • 基金资助:
    国家自然科学基金项目(52169005);国家自然科学基金项目(51769029);南疆重点产业创新发展支撑计划项目(2022DB024)

CMIP6 multi-model prediction of future climate change in the Hotan River Basin

HE Chaofei1,2(), LUO Chengyan3, CHEN Fulong1,2,*(), LONG Aihua1,4, TANG Hao1,2   

  1. 1. College of Water Conservancy & Architectural Engineering, Shihezi University, Shihezi 832000, China
    2. Key Laboratory of Cold and Arid Regions, Eco-Hydraulic Engineering of Xinjiang Production & Construction Corps, Shihezi 832000, China
    3. Xinjiang Water Resources and Hydropower Survey and Design Institute Co., Ltd., ürümqi 830000, China
    4. State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin, China Institute of Water Resources and Hydropower Research, Beijing 100038, China
  • Received:2022-10-18 Revised:2023-01-10 Online:2023-05-25 Published:2023-04-27

摘要:

气候系统模式是对历史和未来气候模拟最广泛有效的工具,但存在一定的不足和局限性,使其无法直接用以预估未来气候变化。本文采用基于分位数映射的日偏差校正(DBC)、多模式集合(MME)平均和基于皮尔逊r相关系数的加权集合(r-MME)平均方法,以1971—2000年为基准期,评估6种气候模式在和田河流域的适用性;运用r-MME方法对未来SSP1-2.6、SSP2-4.5和SSP5-8.5情景下各模式的偏差校正结果进行集合,分析未来近期(2021—2050年)、远期(2061—2090年)日最高、最低气温以及降水的时空演变特征。结果表明:基于DBC的r-MME融合方法能够综合考虑各模式的优势,可大幅提高气候模式模拟的精度,年均最高气温、最低气温和降水量与实测序列的相关系数分别达到0.918、0.821和0.878;3种情景下的气温和降水均呈现增加趋势,其中低强迫SSP1-2.6情景下的增幅最小,远期年最高气温、最低气温和降水的平均增量分别为2.830、2.523 ℃和46.412 mm,高强迫SSP5-8.5情景下的增幅最大,远期年最高气温、最低气温和降水的平均增量为5.697、6.452 ℃和93.206 mm;同时,未来温差的逐渐减少使得流域 “暖湿化”现象更加明显。该研究结果可为未来气候变化下和田河流域经济发展和生态文明建设,以及水资源预测提供重要的理论依据。

关键词: 和田河流域, CMIP6, 模式评估, 偏差校正, 气候变化

Abstract:

Climate system models are the most extensive and effective tool for simulating historical and future climate, but they cannot be used directly to predict future climate change due to certain deficiencies and limitations of the models. In this study, taking 1971-2000 as base period, six climate models were evaluated using quantile-based daily bias correction (DBC), multi-model ensemble (MME) and correlation coefficient-based weighted multi-model ensemble (r-MME) methods to assess their applicability to predicting future climate change in the Hotan River Basin. In addition, r-MME was used to aggregate deviation-correction factors for the six models under three futuristic scenarios, SSP1-2.6, SSP2-4.5 and SSP5-8.5, to predict the near future (2021-2050) and long-term (2061-2090) spatial/temporal trends in daily maximum and minimum temperatures and precipitation. Results show that combining r-MME and DBC can comprehensively utilize each model’s advantages and greatly improve simulation accuracy, as correlation coefficients of 0.918, 0.821 and 0.878 were achieved, respectively, in prediction of maximum/minimum annual average temperatures and annual average precipitation. These three climate indices all showed an upward trend, under the three scenarios, where the low-forcing SSP1-2.6 had the smallest long-term increments of 2.830 ℃, 2.523 ℃ and 46.412 mm, respectively, and the high-forcing SSP5-8.5 had the largest increments of 5.697 ℃, 6.452 ℃ and 93.206 mm, respectively; besides, the decrease of temperature difference between maximum and minimum temperatures in the future will result in more obvious warming and humidification in the basin. The research results can provide an important theoretical basis for the economic and heathy-ecosystem developments in as well as water resources prediction for the Hotan River basin under future climate change.

Key words: Hotan River Basin, CMIP6, model evaluation, deviation correction, climate change

中图分类号: