CWRF模式在中国夏季极端降水模拟的误差订正

Bias Correction of Summer Extreme Precipitation Simulated by CWRF Model over China

利用1980—2015年6—8月我国逐日降水观测数据评估CWRF模式(Climate-Weather Research and Forecasting model)多种参数化方案对我国夏季日降水的模拟能力,并考察累积概率变换偏差订正法(CDFt)的订正效果通过将广义帕累托分布(GPD)引入到偏差订正模型中,提出针对极端降水的累积概率变换偏差订正法(XCDFt),检验和评估其对极端降水订正的适用性。结果显示:CWRF模式微物理过程选用Morrison-aerosol参数化方案组合对我国降水场的模拟较好,CDFt订正效果良好;XCDFt偏差订正模型能够较好地提取模式建模与验证时期变化信号,订正后相比订正前与观测极端降水的概率分布更为接近;经过XCDFt订正后华南、华中和华北地区20年一遇的极端降水重现水平较模拟值更接近观测值,可为CWRF模式提高极端降水的业务预测水平提供参考。

Accurate forecast of extreme precipitation plays an important role in guiding the national economy and people’s livelihood. The newly developed Climate-Weather Research and Forecasting model(CWRF) is applied to the operational forecasting experiment of China National Climate Center. It provides valuable scientific basis for improving the operational prediction for extreme precipitation.CWRF integrates a comprehensive ensemble of alternate parameterization schemes for each of key physical processes,including surface(land,ocean),planetary boundary layer,cumulus(deep,shallow),microphysics, cloud, aerosol, and radiation. This facilitates the use of an optimized physics ensemble approach to improve weather and climate prediction.Daily precipitation data simulated by CWRF from June to August during 1980-2015 and the observation by China Meteorological Administration are used to evaluate the performance of various parameterization schemes, and cumulative distribution function transform(CDFt) correction method is introduced.Based on the CDFt, the probability bias correction model XCDFt is proposed for extreme rainfall by introducing generalized Pareto distribution(GPD) and results are assessed. It shows that Morrison-aerosol parameterization scheme of CWRF model can simulate the spatial distribution of extreme precipitation better as well as correct daily precipitation by CDFt.Simulated results of Morrison-aerosol for daily precipitation threshold and super-threshold samples in summer in North China, Central China and East China are similar to those observed in this scheme. In Changsha, Jinan, Nanjing, and Nanning, GPD characterizes the distribution of each extreme precipitation well. It shows that XCDFt can preserve the CDF form of the observed calibrated precipitation and acquire the small change from the calibration to validations. XCDFt can improve the consistency between model simulation and observation in regional extreme precipitation recurrence levels. In North China, Central China and South China, after model simulation correction by XCDFt, the 20-year recurrence interval of extreme precipitation are closer to the observation, which shows that the revised data are more reliable.Error correction can only be used as a supplementary means to improve extreme precipitation prediction, though. The precision description of model physical process and the improvement of model resolution are the key to improve extreme precipitation prediction level.