基于CMIP6模式和ERA5再分析资料的青藏高原降雪的时空特征研究

Spatio-Temporal Characterization of Snowfall over the Qinghai-Xizang Plateau Based on CMIP6 Models and ERA5 Reanalysis Data

摘要: 为进一步增强对青藏高原降雪机制和影响因素的认识，丰富对高原气候系统的理解，本文基于4个国内的CMIP6 (Coupled Model Intercomparison Project in Phase 6)模式模拟资料以及欧洲中期天气预报中心(ECMWF)的ERA5再分析资料，对青藏高原降雪的时空变化特征进行分析，并评估青藏高原在当前气候情景下的降雪变化。结果表明：(1) 4个模式和ERA5模拟的降雪空间分布存在不同程度的差异，除BCC-ESM1外，其他3个模式和ERA5模拟出在高原东南部都存在降雪大值区，而BCC-ESM1模式在模拟高原西部地区时出现了大值中心；(2) 从空间分布趋势来看，BCC-CSM2-MR、FGOALS-f3-L、FIO-ESM-2-0模拟出高原大部分地区降雪存在减少趋势，但FIO-ESM-2-0模拟出降雪在高原东南部减少更为显著。ERA5模拟出高原东部和西部地区降雪均在增加。从时间上看，仅有FIO-ESM-2-0模拟的降雪呈现较为明显的减少特征；(3) 在春季，FIO-ESM-2-0模拟的降雪年际变化没有显著增减趋势；BCC-CSM2-MR、BCC-ESM1和FGOALS-f3-L模拟的均为弱的增加趋势。而在冬季，BCC-CSM2-MR和BCC-ESM1模拟的降雪无明显增减趋势，仅有FIO-ESM-2-0模拟的呈明显减少趋势；(4) 4个模式中青藏高原降雪年内变化均表现为“双峰”型，峰值分别在春季和秋季，且春季大于秋季。ERA5的年内变化趋势与模式一致，但变化相比更平缓。

Abstract: To further enhance the understanding of the snowfall mechanisms and influencing factors over the Qinghai-Xizang Plateau, enriching the comprehension of the plateau’s climate system, this study analyzes the spatiotemporal characteristics of snowfall over the TP based on simulation data from four domestic CMIP6 (Coupled Model Intercomparison Project Phase 6) models and the ERA5 reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF). This paper also assesses the impact of the current climate scenario on snowfall over the TP. The results indicate: (1) There are varying degrees of differences in the spatial distribution of snowfall simulated by the four models and ERA5. Except for BCC-ESM1, the other three models and ERA5 all display a high-value snowfall area in the southeastern part of the plateau, while BCC-ESM1 shows a high-value center in the western part of the plateau; (2) In terms of spatial distribution trends, BCC-CSM2-MR, FGOALS-f3-L, and FIO-ESM-2-0 simulate a decreasing trend in snowfall over most areas of the plateau, but FIO-ESM-2-0 shows a more significant reduction in the southeastern part. ERA5 simulates an increase in snowfall in the eastern and western parts of the plateau. Temporally, only FIO-ESM-2-0 exhibits a more pronounced decreasing characteristic in snowfall; (3) In spring, the interannual variability of snowfall simulated by FIO-ESM-2-0 shows no significant increasing or decreasing trend; BCC-CSM2-MR, BCC-ESM1, and FGOALS-f3-L all simulate a weak increasing trend. In winter, BCC-CSM2-MR and BCC-ESM1 simulate no significant trend in snowfall, with only FIO-ESM-2-0 showing a clear decreasing trend; (4) The inter-annual variation of snowfall over the TP in the four models all exhibit a “double-peak” pattern, with peaks in spring and autumn, and the spring peak is greater than autumn peaks. The inter-annual variation trend of ERA5 is consistent with the models, but shows a smoother variation.