概率密度分布法在“一带一路”地区极端月气温评估和预估中的应用

“一带一路”地区人口众多,气候类型复杂,亟待加强区域气候变化风险的认识。文中将该区分成10个区域,基于第五次耦合模式比较计划（CMIP5）中的31个全球模式模拟结果,应用概率密度分布（PDF）方法评估历史阶段（1986—2005年）各模式模拟暖月和冷月气温的能力,挑选并建立较优模式集合,用以预估21世纪中叶（2041—2060年）和21世纪末（2081—2100年）的极端月气温。结果表明,模式对观测中冷月气温距平PDF的模拟水平整体较暖月高。与多模式平均以及中位值相比,较优模式集合方法更适于极端暖/冷月气温的评估。在中等排放RCP4.5情景下,与低纬度地区相比,较优模式模拟中高纬地区未来极端暖/冷月气温的增温幅度的不确定性范围较大。21世纪中叶和21世纪末较优模式模拟的极端暖月气温在地中海增幅整体最大,东南亚增幅整体最小。对较优模式集合预估的极端冷月气温而言,无论是21世纪中叶还是世纪末,北欧增幅整体最大,东南亚增幅整体最小。

Evaluation and projections of monthly temperature extremes over the Belt and Road region based on PDF-adjusted method

The Belt and Road region has widespread population and complex climate types, of which the risks of climate change are yet to be recognized. In this study, we investigated the simulations of the annual extreme monthly temperatures by 31 models of the Coupled Model Intercomparison Project Phase 5 (CMIP5) for 10 sub-regions. The performances of the models for the historical period (1986-2005) were assessed by comparing the simulated and observed probability density functions (PDFs) of the warm/cold month temperatures. A “better-model ensemble” was applied to project extreme monthly temperatures in the middle (2041-2060) and the end (2081-2100) of the 21st century. The models tend to perform relatively better for the cold month than for the warm month temperatures over the historical period. Compared to all-model ensemble/the median, the “better-model ensemble” is more suitable for evaluation of extreme monthly temperatures. As to the changes in extreme warm/cold month temperatures projected under the RCP4.5 scenarios, the inter-model uncertainty is larger over high latitudes than over low latitudes via the “better-model ensemble”. The changes of monthly warm extremes show the largest increase in the Mediterranean Basin and the smallest in Southeast Asia in both the middle of the end of the century. The changes of monthly cold extremes show the largest warming in Northern Europe and the smallest warming in Southeast Asia in both the middle and the end of the century.