基于NCEP CFSv2的西南地区气温和降水订正算法研究

doi:10.11755/j.issn.1006-7639-2026-03-0499

摘要/Abstract

摘要：

针对美国国家环境预报中心（National Centers for Environmental Prediction，NCEP）开发的第二代气候预报系统（Climate Forecast System Version 2，CFSv2）对西南地区次季节气温和降水预测存在较大系统偏差的问题，利用1999—2024年西南地区气象台观测资料和模式预测资料，采用非参数百分位映射法对模式日平均气温和降水量进行概率订正，并评估其在不同超前时效和不同时间尺度的预测结果。结果表明：模式能够较好模拟日平均气温近似正态分布特征，但普遍高估较小量级日平均气温和降水量发生概率，低估较大量级日平均气温概率。订正后，不同量级日平均气温和降水量的概率预测偏差得到有效改善，多年平均气温和降水量的空间分布误差及年际变化均方根误差显著降低，但对年际变化相关系数的提升有限。订正后，不同超前时效下日、候、旬和月尺度日平均气温预测技巧均有提高。对于日降水量，订正在日尺度上的效果存在一定不稳定性，但随着检验时间尺度增加，其预测技巧得到不同程度提升，并能够有效延长模式预测时效。

关键词: NCEP CFSv2, 概率订正, 检验, 西南地区

Abstract:

To address the substantial sub-seasonal forecast biases in temperature and precipitation over Southwest China produced by the NCEP CFSv2 model, this study uses station observations and model data from 1999 to 2024 to probabilistically correct the model’s daily mean temperature and precipitation using a nonparametric percentile mapping method. The results indicate that although the model can capture the approximately normal distribution of daily mean temperature, it generally overestimates the probability of low daily mean temperatures and light precipitation and underestimates the probability of high daily mean temperatures. After correction, the biases in the predicted probabilities across different magnitudes of daily mean temperature and precipitation are effectively reduced. The spatial biases in climatological mean temperature and precipitation, as well as the interannual root-mean-square errors, are significantly decreased, whereas the improvement in interannual correlation coefficients is limited. The correction enhances the predictive skill of daily mean temperature at daily, pentad, dekadal, and monthly time scales across various forecast lead times. For daily precipitation, the correction shows some instability at daily scale; however, its predictive skill improves to varying degrees with increasing verification time scales, effectively extending the model’s forecast lead time.

Key words: NCEP CFSv2, probabilistic correction, verification, Southwest China

中图分类号:

P456

吴遥, 唐红玉, 朱浩楠, 魏麟骁, 何慧根. 基于NCEP CFSv2的西南地区气温和降水订正算法研究[J]. 干旱气象, 2026, 44(3): 499-511.

WU Yao, TANG Hongyu, ZHU Haonan, WEI Linxiao, HE Huigen. Research on correction algorithm for temperature and precipitation in Southwest China based on NCEP CFSv2[J]. Journal of Arid Meteorology, 2026, 44(3): 499-511.

图/表 12

图1 西南地区1999—2023年各月观测和模式预测及订正的日平均气温概率密度

Fig.1 Probability density distributions of daily mean temperature observations， forecasts， and bias-corrected forecasts in each month over Southwest China during 1999-2023

图2 西南地区1999—2023年观测和模式预测及订正的日平均气温及偏差经1~44 d超前时效平均后的多年平均空间分布（单位：℃）

Fig.2 Spatial distributions of multi-year mean and bias of daily mean temperature observations， forecasts， and bias-corrected forecasts averaged over lead times of 1-44 days in Southwest China during 1999-2023 （Unit： ℃）

图3 西南地区1999—2023年1~44 d超前时效不同时间尺度模式预测及订正的日平均气温与观测值的空间相关系数及订正偏差多年平均的逐月分布

Fig.3 Multi-year mean monthly variations of spatial correlation coefficients of daily mean temperature forecasts and bias-corrected forecasts against observations and their differences for lead times of 1-44 days at different temporal scales over Southwest China during 1999-2023

图4 西南地区1999—2023年1~44 d超前时效不同时间尺度模式预测及订正的日平均气温与观测值的平均绝对误差及订正偏差多年平均的逐月变化（单位：℃）

Fig.4 Multi-year mean monthly variations of MAE of daily mean temperature forecasts and bias-corrected forecasts against observations and their differences for lead times of 1-44 days at different temporal scales over Southwest China during 1999-2023 （Unit： ℃）

图5 西南地区1999—2023年模式预测及订正的日平均气温与观测值的年际相关系数和均方根误差（单位：℃）经1~44 d超前时效平均后的空间分布

Fig.5 Spatial distribution of interannual correlation coefficients and RMSE （Unit： ℃） of daily mean temperature forecasts and bias-corrected forecasts against observations averaged for lead times of 1-44 days over Southwest China during 1999-2023

图6 西南地区1999—2023年不同月份观测和模式预测及订正日降水量概率密度分布

Fig.6 Probability density distributions of daily precipitation observations， forecasts， and bias-corrected forecasts in each month over Southwest China during 1999-2023

图7 西南地区1999—2023年观测和模式预测及订正的日降水量及偏差经1~44 d超前时效平均后的多年平均空间分布（单位：mm）

Fig.7 Spatial distributions of multi-year mean and bias of daily precipitation observations， forecasts， and bias-corrected forecasts averaged over lead times of 1-44 days in Southwest China during 1999-2023 （Unit： mm）

图8 西南地区1999—2023年1~44 d超前时效不同时间尺度模式预测及订正的日降水量与观测值的空间相关系数及订正偏差多年平均的逐月分布

Fig.8 Multi-year mean monthly variations of spatial correlation coefficients of daily precipitation forecasts and bias-corrected forecasts against observations and their differences for lead times of 1-44 days at different temporal scales over Southwest China during 1999-2023

图9 西南地区1999—2023年1~44 d超前时效不同时间尺度模式预测及订正的日降水量与观测值的平均绝对误差及订正偏差多年平均的逐月变化（单位：mm）

Fig.9 Multi-year mean monthly variations of MAE of daily precipitation forecasts and bias-corrected forecasts against observations and their differences for lead times of 1-44 days at different temporal scales over Southwest China during 1999-2023 （Unit： mm）

图10 西南地区1999—2023年模式预测及订正的日降水量与观测值的年际相关系数和均方根误差（单位：mm）经1~44 d超前时效平均后的空间分布

Fig.10 Spatial distribution of interannual correlation coefficients and RMSE （Unit： mm） of daily precipitation forecasts and bias-corrected forecasts against observations averaged for lead times of 1-44 days over Southwest China during 1999-2023

图11 西南地区2024年1~44 d超前时效日尺度模式预测及订正的日平均气温与观测值的空间相关系数和平均绝对误差（单位：℃）及其订正偏差的逐月变化

Fig.11 Monthly variations of spatial correlation coefficients and MAE （Unit： °C） of daily mean temperature forecasts and bias-corrected forecasts against observations and their differences for lead times of 1-44 days at daily scale over Southwest China in 2024

图12 西南地区2024年1~44 d超前时效日尺度模式预测及订正的日降水量与观测值的空间相关系数和平均绝对误差（单位：mm）及其订正偏差的逐月变化

Fig.12 Monthly variations of spatial correlation coefficients and MAE （Unit： mm） of daily precipitation forecasts and bias-corrected forecasts against observations and their differences for lead times of 1-44 days at daily scale over Southwest China in 2024

参考文献 35

[1]	陈子凡, 王磊, 李谢辉, 等, 2022. 西南地区极端降水时空变化特征及其与强ENSO事件的关系[J]. 高原气象, 41(3):604-616.
[2]	郭准, 吴春强, 周天军, 等, 2011. LASG/IAP和 BCC大气环流模式模拟的云辐射强迫之比较[J]. 大气科学, 35(4):739-752.
[3]	郝民, 龚建东, 田伟红, 等, 2018. L波段探空仪湿度资料偏差订正及同化试验[J]. 应用气象学报, 29(5):559-570.
[4]	黄珊, 杨扬, 王含嘉, 等, 2020. 中国西南地区地表感热和潜热通量时空变化特征[J]. 干旱气象, 38(4):601-611.
[5]	李俊, 杜钧, 陈超君, 2014. 降水偏差订正的频率(或面积)匹配方法介绍和分析[J]. 气象, 40(5):580-588.
[6]	李维京, 2012. 现代气候业务[M]. 北京: 气象出版社:202-315.
[7]	刘佳, 徐金霞, 马振峰, 等, 2014. 第二代月动力延伸预报产品对西南汛期降水的预报检验[J]. 高原气象, 33(6):1468-1 479.
[8]	刘雅静, 韦志刚, 陈广宇, 等, 2021. NCEP CFSv2对北半球平流层极涡边界年代际变化特征模拟的评估[J]. 高原气象, 40(5):1012-1 023.
[9]	卢新玉, 魏鸣, 王秀琴, 2017. TRMM月降水量产品在新疆地区的订正[J]. 应用气象学报, 28(3):379-384.
[10]	穆穆, 陈博宇, 周菲凡, 等, 2011. 气象预报的方法与不确定性[J]. 气象, 37(1):1-13.
[11]	孙靖, 程光光, 张小玲, 2015. 一种改进的数值预报降水偏差订正方法及应用[J]. 应用气象学报, 26(2):173-184.
[12]	王娟怀, 李清泉, 汪方, 等, 2021. 基于DERF2.0的华南前汛期降水订正[J]. 应用气象学报, 32(1):115-128.
[13]	吴遥, 蒋兴文, 唐红玉, 2020. 集合样本数影响月平均500 hPa高度场预测技巧分析[J]. 热带气象学报, 36(3):360-369.
[14]	吴遥, 唐红玉, 蒋兴文, 等, 2024. NCEP CFSv2模式对中国西南月尺度降水预测能力评估[J]. 高原山地气象研究, 44(3):67-75.
[15]	晏红明, 李清泉, 孙丞虎, 等, 2013. 中国西南区域雨季开始和结束日期划分标准的研究[J]. 大气科学, 37(5):1111-1 128.
[16]	袁良, 谭江红, 闫彩霞, 等, 2024. 概率匹配订正法在湖北襄阳地区降水预报中的应用[J]. 沙漠与绿洲气象, 18(2):92-98.
[17]	章大全, 陈丽娟, 2016. 基于DERF2.0的月平均温度概率订正预报[J]. 大气科学, 40(5):1 022-1 032.
[18]	赵琳娜, 马清云, 杨贵名, 等, 2008. 2008年初我国低温雨雪冰冻对重点行业的影响及致灾成因分析[J]. 气候与环境研究, 13(4):556-566.
[19]	郑凤琴, 钟利华, 罗小莉, 等, 2020. 利用天气关键区相似法预测西江流域6月延伸期强降水天气过程[J]. 干旱气象, 38(3):490-496.
[20]	郑志海, 2013. 月动力延伸预报研究进展回顾[J]. 气象科技进展, 3(1):25-30.
[21]	BETTS A K, HONG S Y, PAN H L, 1996. Comparison of NCEP-NCAR reanalysis with 1987 FIFE data[J]. Monthly Weather Review, 124(7): 1 480-1 498.
[22]	CANE M A, ZEBIAK S E, DOLAN S C, 1986. Experimental forecasts of El Niño[J]. Nature, 321(6073): 827-832.
[23]	COLLINS M, 2002. Climate predictability on interannual to decadal time scales: The initial value problem[J]. Climate Dynamics, 19: 671-692.
[24]	DHANYA C T, VILLARINI G, 2017. On the inherent predictability of precipitation across the United States[J]. Theoretical and Applied Climatology, 133: 1 035-1 050.
[25]	ENTEKHABI D, RODRIGUEZ-ITURBE I, CASTELLI F, 1996. Mutual interaction of soil moisture state and atmospheric processes[J]. Journal of Hydrology, 184(1/2): 3-17.
[26]	GAO J M, SANG Y H, 2017. Identification and estimation of landslide debris flow disaster risk in primary and middle school campuses in a mountainous area of Southwest China[J]. International Journal of Disaster Risk Reduction, 25: 60-71.
[27]	JIANG X W, YANG S, LI Y Q, 2013. Seasonal-to-interannual prediction of the Asian summer monsoon in the NCEP climate forecast system version 2[J]. Journal of Climate, 26(11): 3 708-3 727.
[28]	LIU W, SUN J, GONG J, 2019. Assessment of the WRF model’s ability to simulate different precipitation types[J]. Journal of Geophysical Research: Atmospheres, 124(8): 4 224-4 239.
[29]	MA Z F, LIU J, ZHANG S Q, et al, 2013. Observed climate changes in southwest China during 1961-2010[J]. Advances in Climate Change Research, 4(1): 30-40.
[30]	PANOFSKY H A, BRIER G W, 1968. Some applications of statistics to meteorology[M]. University Park, Pa: The Pennsylvania State University Press: 224-225.
[31]	PIANI C, HAERTER J O, COPPOLA E, 2010. Statistical bias correction for daily precipitation in regional climate models over Europe[J]. Theoretical and Applied Climatology, 99 (1/2): 187-192.
[32]	POLADE S D, PIERCE D W, CAYAN D R, et al, 2014. The key role of dry days in changing regional climate and precipitation regimes[J]. Scientific Reports, 4: 4364. DOI:10.1038/srep04364.
[33]	RAISANEN J, RATY O, 2013. Projections of daily mean temperature variability in the future: Cross-validation tests with ENSEMBLES regional climate simulations[J]. Climate Dynamics, 41(5/6): 1 553-1 568.
[34]	SAHA S, MOORTHI S, WU X R, et al, 2014. The NCEP climate forecast system version 2[J]. Journal of Climate, 27(6): 2 185-2 208.
[35]	SHIN C S, HUANG B, ZHU J, et al, 2019. Improved seasonal predictive skill and enhanced predictability of the Asian summer monsoon rainfall following ENSO events in NCEP CFSv2 hindcasts[J]. Climate Dynamics, 52(5): 3 079-3 098.