不同偏差校正法对GCM降水数据的应用效果分析  

Evaluations of Different Bias Correction Methods on the GCM Precipitation Data

作　　者：李昕潼 李占玲[1] 韩孺村 LI Xintong;LI Zhanling;HAN Rucun(School of Water Resources and Environment,China University of Geosciences,Beijing 100083,China)

机构地区：[1]中国地质大学(北京)水资源与环境学院,北京100083

出　　处：《水文》2023年第3期93-100,117,共9页Journal of China Hydrology

基　　金：水文水资源与水利工程科学国家重点实验室“一带一路”水与可持续发展科技基金资助项目(2021nkms03)。

摘　　要：大气环流模式(GCM)是获取时空连续气候数据的主要手段之一,但由于其在区域尺度上的输出数据往往存在一定偏差,因此在使用前需要进行校正。以黑河流域为研究区,提取了CMIP6中的BCC-CSM2-MR模式的降水数据,采用线性标度法(LS)、经验分位数映射法(EQM)和Gamma分位数映射法(GQM)对1985—2014年逐月降水数据进行偏差校正,选择标准化均方根误差(NRMSE)、平均绝对误差(MAE)、皮尔逊相关系数(R)和Sen斜率(Sen’s slope)为评价指标,分析校正前后评价指标的变化以及不同校正方法的差异和有效性。结果表明:(1)LS法校正后降水数据总量指标值有明显改进,NRMSE、MAE和均值误差明显减小,然而由于该方法没有考虑数据在频率上的差异性,因此,LS校正后的降水数据在频率上仍存在一定偏差;(2)经过EQM法和GQM法校正后,NRMSE、MAE指标均有所改进,频率指标方面校正效果更优,校正后的降水频率空间分布与变化趋势更加符合实测数据;另外,通过分析各指标,EQM方法在本研究区的适用性更强。General Circulation Model(GCM)can provide spatial-temporal continuous meteorological data in the studies of climate change,while such data will inevitably contain some biases.Bias correction are quite necessary for such data before use.Taking the Heihe River Basin as the study area,this study mainly compared three methods in bias correcting the monthly precipitation data from the BCC-CSM2-MR model in CMIP6.The methods were Linear Scale(IS)method,Empirical Quantile Mapping(EQM)method and Gamma Quantile Mapping(GQM)method.Four evaluation indicators were included,the Normalized Root Mean Square Eror(NRMSE),Mean Absolute Error(MAE),Pearson correlation cofficient(R)and Sen's slope.The results showed that:(1)IS method showed good performance in improving the total indicators of precipitation data,and reducing the NRMSE,MAE and mean error,while poor performance in improving the frequency indicators;(2)EQM and GQM methods also showed good performance in improving the NRMSE and MAE,although the overall effect was not as good as that of LS method,they were greatly improved compared with those before correction.In terms of the frequency indicators,the results of these two methods were obviously better,and the spatial distributions and variation trends of precipitation frequency after correction were more consistent with the measured data.In addition,EQM has stronger applicability in this study area comparatively.

关 键 词：GCM 偏差校正 经验分位数映射法 线性标度法 Gamma分位数映射法 

分 类 号：P338[天文地球—水文科学] TV125[水利工程—水文学及水资源]