基于天气尺度环流评估全球气候模式对长江中游夏季极端降水的模拟

Extreme rainfall (ER) is a meteorological disaster occurring frequently over the middle reaches of Yangtze River during summer and the projection of its future changes is of great socio-economic importance. Global climate models from the Coupled Model Intercomparison Project are the primary tools to project future changes under various radiative forcing scenarios. But uncertainties and discrepancies are still found in the projected changes in ER at regional scales. Therefore, it is essential to evaluate models' representations of the characteristics of precipitation extremes in historical record and select those models with better performances. Many previous studies of model evaluation directly focused on the statistical characteristics of modeled rainfall or the composite circulation features associated with ER. Given the variety of ER-producing circulation patterns and the diverse response of different circulation patterns to radiative forcing changes in the atmosphere, the circulation used to define evaluation metrics should be classified into distinct patterns and the relationship with ER should be robust. Our earlier work has identified three synoptic-scale circulation patterns driving ER formation. In this study, the statistical characteristics of these patterns and rainfall in models, as well as the model intrinsic patterns associated with modeled ER will be investigated, which can be used to understand whether the models are producing ER through similar synoptic-scale drivers as observed and select a subset of models that are dynamically well constrained. The confidence we place on the projected future changes will be improved if this subset of models are used in generating projection ensembles. The results of this study will also allow us to track biases of simulated extreme rainfall to provide recommendations for model improvements.

极端降水是长江中游夏季常见的气象灾害，对其未来变化的预估有重要的社会经济意义。全球气候模式是极端降水未来预估的首要工具，但其在区域尺度上仍存在不确定性和模式间分歧，这就需要我们评估模式对历史时期极端降水的模拟以选择更可靠的模式。以往的模式评估直接基于极端降水或与其有关的平均环流。但由于导致极端降水的天气尺度环流的多样性以及不同类型环流对辐射强迫响应的差异，作为模式评估标准的环流需要进一步分型和检验。本项目将基于前期研究得到的导致长江中游夏季极端降水的三类典型天气尺度环流，通过分析模式中这些环流与极端降水的误差，以及检验模式自身典型环流等方式，评估模式能否再现这种“环流-降水”关系，即模式产生极端降水的环流是否与观测的类似。这将帮助我们选择“动力上”对极端降水模拟好的模式；用这些模式作为集合成员可以增加未来预估结果的可信度。此外，该研究也将帮助我们理解模式的主要误差来源，为模式改进提供参考。

极端降水是长江中游夏季常见的气象灾害，评估模式对极端降水的模拟性能对于理解和订正模拟结果、改进模式十分重要。但从动力过程上评估模式产生极端降水的研究目前仍很缺乏。本项目基于产生极端降水的典型天气尺度环流评估了全球模式，发现模式对典型天气尺度环流的模拟误差可以导致极端降水的模拟误差，并分析了其可能的误差来源是局地反馈和气候态背景场。项目圆满完成了预期研究目标，主要研究结论如下：.1）典型天气尺度环流在模式中的统计特征及其误差.基于观测中导致长江中游夏季极端降水的典型天气尺度环流评估CMIP5模式，发现大部分模式可以在误差上再现观测中“环流—降水”的关系，即大部分模式低估了能够导致长江中游夏季极端降水的典型天气环流的出现频次，进而低估了长江中游夏季极端降水的出现频次，也因此低估了长江中游夏季平均降水。.2）模式自身的典型天气尺度环流.基于模式自身导致极端降水的典型天气尺度环流评估模式，发现大部分模式可以再现观测中“环流—降水”的关系，即模式自身的典型天气环流与观测中的典型天气环流相似。部分模式自身的典型天气环流与观测中的差异较大且与模拟降水关系较弱，所以虽然这些模式可能再现了合理的降水分布，但其产生极端降水的方式在天气动力上是不合理的。由于不同天气环流对辐射强迫变化的响应不同，仅依据降水评估挑选的模式的未来预估仍有很大不确定性。.3）误差来源.通过新的梅雨锋锋生方程和分段位涡反演发现长江中游夏季极端降水的模拟偏差可能通过非绝热作用反馈到梅雨锋环流，从而导致极端降水发生后的天气环流的偏差；通过分析能量转换发现模式气候态背景场的偏差可能影响急流附近天气扰动的发展，导致下游典型天气环流的模拟偏差。