基于GWR模型的中国中东部降水与海拔高度关系特征分析

利用中国国家级地面气象站逐时降水资料,采用地理加权回归(Geographically Weighted Regression,GWR)模型系统分析了中国中东部暖季降水与海拔高度的关系,并将二者关系作为一种客观标准,评估了ECMWF-IFS模式对2017年暖季降水的预报能力。主要结论如下:(1)总体来看,中国中东部降水频率(强度)随海拔高度升高而增加(减小),二者在不同地区的贡献程度不同导致降水量与海拔高度关系的区域差异显著。(2)通过对比午后短时和夜间长时降水事件与海拔高度关系的差异,发现午后短时降水事件的降水量主要随海拔高度升高而增加,且以降水频率与海拔高度关系的贡献为主。而夜间长时降水事件的降水量与海拔高度关系的区域一致性较差。相较于午后短时降水事件,夜间长时降水事件中有更多站点表现出降水量随海拔高度升高而减小的特征,在大地形周边陡峭地形处的站点所表现出的此种差异较东部孤立地形处更加显著。(3)根据ECMWF-IFS模式的评估结果,模式能够较好地刻画出中国中东部2017年暖季降水气候态的空间分布特征,且与观测具有较大的空间相关系数。但从降水与海拔高度关系来看,观测与模式的空间相关性偏弱。此外,模式能够表现出降水强度(频率)主要随海拔高度升高而减小(增加)的特征,但绝大多数站点在模式中的降水强度(频率)与海拔高度的负(正)回归关系要弱(强)于观测结果。

Analysis of the relationship between precipitation and altitude over central and eastern China based on the Geographically Weighted Regression Model

The relationship between warm season precipitation and altitude over central and eastern China was studied using hourly rain-gauge data from national surface stations of China based on the Geographically Weighted Regression Model.Using the relationship between the two as an objective criterion,the ECMWF-IFS model’s performance in predicting precipitation during the warm season in 2017 was also investigated.The main conclusions are as follows.(1)On the whole,the precipitation frequency(intensity)increases(decreases)with altitude over central and eastern China.The significant regional differences in the relationship between precipitation amount and altitude result from the distinct contribution of frequency and intensity in different regions.(2)By comparing the differences in the relationships between different kinds of precipitation events and altitude,it was found that the precipitation amount of short-afternoon precipitation events mainly increases with altitude,which comes from the contribution of the relationship between precipitation frequency and altitude.Meanwhile,the regional consistency of the relationship between precipitation amount of long-nighttime precipitation events and altitude is poor.Compared to short-afternoon precipitation events,there are more stations where precipitation amount decreases significantly with altitude in long-nighttime precipitation events.This kind of difference is more pronounced in the stations at the steep terrain around large-scale topography than the stations at the eastern isolated topography.(3)According to the evaluation of the ECMWF-IFS model,it can be seen that the model can well forecast the spatial distribution of the precipitation climate state characteristics during warm season in 2017 over central and eastern China,with high spatial correlation coefficient with observation.But in terms of the prediction of the relationship between precipitation and altitude,the spatial correlation between observation and model forecast is weakened.Furthermore,the model can show the characteristic that precipitation intensity(frequency)mainly decreases(increases)with the increasing altitude.However,the negative(positive)regression relationship between precipitation intensity(frequency)and altitude is weaker(stronger)in model than that in observation.