基于机器学习模型的青藏高原日降水数据的订正研究

doi:10.7522/j.issn.1000-0240.2017.0065

冰川冻土 ›› 2017, Vol. 39 ›› Issue (3): 583-592.doi: 10.7522/j.issn.1000-0240.2017.0065

基于机器学习模型的青藏高原日降水数据的订正研究

陈浩¹, 宁忱¹, 南卓铜², 王玉丹³, 吴小波³, 赵林³

1. 宝鸡文理学院地理与环境学院, 陕西宝鸡 721013;
2. 南京师范大学地理科学学院, 江苏南京 210023;
3. 中国科学院西北生态环境资源研究院, 甘肃兰州 730000

收稿日期:2016-12-20 修回日期:2017-02-10 出版日期:2017-06-25 发布日期:2017-09-09
通讯作者: 南卓铜,E-mail:nanzt@njnu.edu.cn E-mail:nanzt@njnu.edu.cn
作者简介:陈浩(1978-),男,陕西宝鸡人,讲师,2015年在中国科学院寒区旱区环境与工程研究所获博士学位,从事降水与气象灾害研究.E-mail:chenhao@bjwlxy.cn
基金资助:
国家自然科学基金项目（41471059）；陕西省科技统筹创新计划项目（2016KTCL03-17）；陕西省教育厅重点实验室项目（16JS006）资助

Correction of the daily precipitation data over the Tibetan Plateau with machine learning models

CHEN Hao¹, NING Chen¹, NAN Zhuotong², WANG Yudan³, WU Xiaobo³, ZHAO Lin³

1. School of Geography and Environment, Baoji University of Science and Art, Baoji 721013, Shaanxi, China;
2. School of Geography Science, Nanjing Normal University, Nanjing 210023, China;
3. Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou 730000, China

Received:2016-12-20 Revised:2017-02-10 Online:2017-06-25 Published:2017-09-09

摘要/Abstract

摘要： 选择了5种机器学习模型，即k最近邻方法（KNN）、多元自回归样条方法（MARS）、支持向量机（SVM）、多项对数线性模型（MLM）和人工神经网络（ANN），利用海拔、相对湿度、坡向、植被、风速、气温和坡度等因子订正ITPCAS和CMORPH两种常用的青藏高原日降水数据集。五折交叉验证表明，KNN的订正精度最高。在三个验证站点（唐古拉、西大滩和五道梁）的误差分析，以及对青藏高原年降水量的空间分析均表明，KNN对CMORPH的订正效果显著，对ITPCAS在局部区域有一定订正效果，ITPCAS及其订正值的降水空间分布准确度高于CMORPH的订正值。主成分分析法表明降水订正是气象和环境因子综合作用的结果。

关键词: 机器学习模型, 降水数据, 订正, 青藏高原

Abstract: In this paper, five machine learning models, namely k-nearest neighbor (KNN), multivariate adaptive regression splines (MARS), support vector machine (SVM), multinomial log-linear models (MLM) and artificial neural networks (ANN), are selected to correct two commonly used precipitation datasets, ITPCAS (Institute of Tibetan Plateau Research, Chinese Academy of Sciences) and CMORPH (climate prediction center morphing technique), over the Tibetan Plateau by establishing the relationship between daily precipitation and environmental data (elevation, slope, aspect, vegetation), as well as meteorological factors (air temperature, humidity, wind speed). The 5-fold cross validation shows that the KNN has the highest accuracy. The error analysis over the Tanggula, Xidatan and Wudaoliang Stations and the spatial analysis on annual precipitation over the plateau show that the KNN model can significantly correct the CMORPH over the plateau and the correction on the ITPCAS is significant locally. The KNN-corrected CMORPH has lower accuracy than the two ITPCAS precipitation. Principal component analysis indicates that the correction is the comprehensive effects of both environmental and meteorological factors.

Key words: machine learning model, precipitation data, correction, Tibetan Plateau

中图分类号:

P407

陈浩, 宁忱, 南卓铜, 王玉丹, 吴小波, 赵林. 基于机器学习模型的青藏高原日降水数据的订正研究[J]. 冰川冻土, 2017, 39(3): 583-592.

CHEN Hao, NING Chen, NAN Zhuotong, WANG Yudan, WU Xiaobo, ZHAO Lin. Correction of the daily precipitation data over the Tibetan Plateau with machine learning models[J]. JOURNAL OF GLACIOLOGY AND GEOCRYOLOGY, 2017, 39(3): 583-592.

参考文献

[1] Wu Guoxiong, Wang Jun, Liu Xin, et al. Numerical modeling of the influence of Eurasian orography on the atmospheric circulation in different season[J]. Acta Meteorologica Sinica, 2005, 63(5):603-612.[吴国雄, 王军, 刘新, 等. 欧亚地形对不同季节大气环流影响的数值模拟研究[J]. 气象学报, 2005, 63(5):603-612.]
[2] Wang Tongmei, Wu Guoxiong, Wan Rijin. Role of the Tibetan Plateau thermal forcing in the summer climate patterns over subtropical Asia[J]. Plateau Meteorology, 2008, 27(1):1-9.[王同美, 吴国雄, 万日金. 青藏高原的热力和动力作用对亚洲季风区环流的影响[J]. 高原气象, 2008, 27(1):1-9.]
[3] Moulin L, Gaume E, Obled C. Uncertainties on mean areal precipitation:assessment and impact on streamflow simulations[J]. Hydrology and Earth System Sciences, 2009, 13(2):99-114.
[4] He Hongyan, Guo Zhihua, Xiao Wenfa. Review on spatial interpolation techniques of rainfall[J]. Chinese Journal of Ecology, 2005, 24(10):1187-1191.[何红艳, 郭志华, 肖文发. 降水空间插值技术的研究进展[J]. 生态学杂志, 2005, 24(10):1187-1191.]
[5] Shen Yan, Xiong Anyuan, Wang Ying, et al. Performance of high-resolution satellite precipitation products over China[J/OL]. Journal of Geophysical Research:Atmospheres, 2010, 115(D2)[2016-01-30]. http://onlinelibrary.wiley.com/doi/10.1029/2009JD012097/full.
[6] Xie Pingping, Xiong Anyuan. A conceptual model for constructing high-resolution gauge-satellite merged precipitation analyses[J/OL]. Journal of Geophysical Research:Atmospheres, 2011, 116(D21)[2016-01-30]. http://onlinelibrary.wiley.com/doi/10.1029/2011JD016118/full.
[7] Han Zhenyu, Zhou Tianjun. Assessing the quality of APHRODITE high-resolution daily precipitation dataset over contiguous China[J]. Chinese Journal of Atmospheric Sciences, 2012, 36(2):361-373.[韩振宇, 周天军. APHRODITE高分辨率逐日降水资料在中国大陆地区的适用性[J]. 大气科学, 2012, 36(2):361-373.]
[8] Liu Shiwei, Wu Jinkui, Zhang Wenchun, et al. Comparison analysis of sampling methods to estimate the regional precipitation based on Kriging interpolation method:a case study in Gansu Province[J]. Journal of Glaciology and Geocryology, 2015, 37(3):650-657.[刘世伟, 吴锦奎, 张文春, 等. 基于克里金插值估算区域降水量的抽样方法对比分析:以甘肃省为例[J]. 冰川冻土, 2015, 37(3):650-657.]
[9] Qi Wenwen, Zhang Baiping, Pang Yu, et al. TRMM-data-based spatial and seasonal patterns of precipitation in the Qinghai-Tibet Plateau[J]. Scientia Geographica Sinica, 2013, 33(8):999-1005.[齐文文, 张百平, 庞宇, 等. 基于TRMM数据的青藏高原降水的空间和季节分布特征[J]. 地理科学, 2013, 33(8):999-1005.]
[10] Xu Shiguang, Niu Zheng, Shen Yan, et al. A research into the characters of CMORPH remote sensing precipitation error in China[J]. Remote Sensing Technology and Application, 2014, 29(2):189-194.[许时光, 牛铮, 沈艳, 等. CMORPH卫星降水数据在中国区域的误差特征研究[J]. 遥感技术与应用, 2014, 29(2):189-194.]
[11] Xu Shiguang, Niu Zheng, Shen Yan, et al. Evaluation and modification of CMORPH multi-satellite precipitation estimates in summer over Tibetan Plateau[J]. Remote Sensing Information, 2015, 30(1):71-76.[许时光, 牛铮, 沈艳, 等. CMORPH对青藏高原地区夏季降水的模拟精度研究与修正[J]. 遥感信息, 2015, 30(1):71-76.]
[12] Li Qiong, Yang Meixue, Wan Guoning, et al. Analysis of the accuracy of TRMM 3B43 precipitation data in the source region of the Yellow River[J]. Journal of Glaciology and Geocryology, 2016, 38(3):620-633.[李琼, 杨梅学, 万国宁, 等. TRMM 3B43降水数据在黄河源区的适用性评价[J]. 冰川冻土, 2016, 38(3):620-633.]
[13] Yu Jingjing, Shen Yan, Pan Yang, et al. Improvement of satellite-based precipitation estimates over China based on probability density function matching method[J]. Journal of Applied Meteorological Science, 2013, 24(5):544-553.[宇婧婧, 沈艳, 潘旸, 等. 概率密度匹配法对中国区域卫星降水资料的改进[J]. 应用气象学报, 2013, 24(5):544-553.]
[14] Chen Yingying, Yang Kun, He Jie, et al. Improving land surface temperature modeling for dry land of China[J/OL]. Journal of Geophysical Research:Atmospheres, 2011, 116(D20)[2016-01-30]. http://onlinelibrary.wiley.com/doi/10.1029/2011JD015921/full.
[15] Guo Donglin, Wang Huijun. Simulation of permafrost and seasonally frozen ground conditions on the Tibetan Plateau, 1981-2010[J]. Journal of Geophysical Research:Atmospheres, 2013, 118(11):5216-5230.
[16] Xue Baolin, Wang Lei, Yang Kun, et al. Modeling the land surface water and energy cycles of a mesoscale watershed in the central Tibetan Plateau during summer with a distributed hydrological model[J]. Journal of Geophysical Research:Atmospheres, 2013, 118(16):8857-8868.
[17] Kan Baoyun, Su Fengge, Tong Kai, et al. Analysis of the applicability of four precipitation datasets in the upper reaches of the Yarkant River, the Karakorum[J]. Journal of Glaciology and Geocryology, 2013, 35(3):710-722.[阚宝云, 苏凤阁, 童凯, 等. 四套降水资料在喀喇昆仑山叶尔羌河上游流域的适用性分析[J]. 冰川冻土, 2013, 35(3):710-722.]
[18] Jiang Zhihong, Lu Yao, Ding Yuguo. Analysis of the high-resolution merged precipitation products over China based on the temporal and spatial structure score indices[J]. Acta Meteorologica Sinica, 2013, 71(5):891-900.[江志红, 卢尧, 丁裕国. 基于时空结构指标的中国融合降水资料质量评估[J]. 气象学报, 2013, 71(5):891-900.]
[19] Shen Yan, Pan Yang, Yu Jingjing, et al. Quality assessment of hourly merged precipitation product over China[J]. Transactions of Atmospheric Sciences, 2013, 36(1):37-46.[沈艳, 潘旸, 宇婧婧, 等. 中国区域小时降水量融合产品的质量评估[J]. 大气科学学报, 2013, 36(1):37-46.]
[20] Jia Shaofeng, Zhu Wenbin, Lu Aifeng, et al. A statistical spatial downscaling algorithm of TRMM precipitation based on NDVI and DEM in the Qaidam Basin of China[J]. Remote Sensing of Environment, 2011, 115(12):3069-3079.
[21] Daly C, Neilson R P, Philips D L. A statistical topographic model for mapping climatological precipitation over mountainous terrain[J]. Journal of Applied Meteorology, 1994, 33(2):140-158.
[22] Liu Yaqin, Fan Guangzhou, Zhou Dingwen, et al. Variability of NDVI in winter and spring on the Tibetan Plateau and their relationship with summer precipitation[J]. Acta Meteorologica Sinica, 2007, 65(6):959-967.[刘雅勤, 范广洲, 周定文, 等. 青藏高原冬、春植被归一化指数变化特征及其与高原夏季降水的联系[J]. 气象学报, 2007, 65(6):959-967.]
[23] Beuchat X, Schaefli B, Soutter M, et al. Toward a robust method for subdaily rainfall downscaling from daily data[J]. Water Resources Research, 2011, 47(9):24-42.
[24] Wang Yudan, Nan Zhuotong, Chen Hao, et al. Correction of CMORPH daily precipitation data over the Qinghai-Tibetan Plateau with k-nearest neighbor mode[J]. Remote Sensing Technology and Application, 2016, 31(3):607-616.[王玉丹, 南卓铜, 陈浩, 等. 基于k最近邻模型的青藏高原CMORPH日降水数据的订正研究[J]. 遥感技术与应用, 2016, 31(3):607-616.]
[25] Hart P. The condensed nearest neighbor rule[J]. IEEE Transactions on Information Theory, 1968, 14(3):515-516.
[26] Friedman J H. Multivariate adaptive regression splines[J]. The Annals of Statistics, 1999, 19(1):121-141.
[27] Joachims T. Making large-scale SVM learning practical:LS-8 report 24[R]. Dortmund, Germany:Computer Science Department of University of Dortmund, 1998.
[28] Christensen R. Log-linear models and logistic regression[M]. 2nd ed. New York:Springer, 1997.
[29] Jain A K, Mao J, Mohiuddin K M. Artificial neural networks:a tutorial[J]. Computer, 1996, 29(3):31-44.
[30] Fielding E, Isacks B, Barazangi M, et al. How flat is Tibet?[J]. Geology, 1994, 22(2):163-167.
[31] Xu Shiguang, Niu Zheng, Kuang Da, et al. Estimating summer precipitation over the Tibetan Plateau with geostatistics and remote sensing[J]. Mountain Research and Development, 2013, 33(4):424-436.
[32] Liston G E, Elder K. A meteorological distribution system for high-resolution terrestrial modeling (MicroMet)[J]. Journal of Hydrometeorology, 2006, 7(2):217-234.
[33] Bishop C M. Pattern recognition and machine learning[M]. New York:Springer, 2006.
[34] Pan Xiaoduo, Li Xin, Yang Kun, et al. Comparison of downscaled precipitation data over a mountainous watershed:a case study in the Heihe River basin[J]. Journal of Hydrometeorology, 2014, 15(4):1560-1574.
[35] Mernild S H, Liston G E, Hasholt B. Snow-distribution and melt modelling for glaciers in Zackenberg River drainage basin, north-eastern Greenland[J]. Hydrological Processes, 2007, 21(24):3249-3263.
[36] Wu Xuejiao, Pan Xiaoduo, Shen Yongping, et al. Validation of WRF model on simulating forcing data for Kayiertesi River basin, Xinjiang[J]. Journal of Glaciology and Geocryology, 2016, 38(2):332-340.[吴雪娇, 潘小多, 沈永平, 等. WRF模式制备的气象驱动数据在新疆喀依尔特斯河流域的验证[J]. 冰川冻土, 2016, 38(2):332-340.]
[37] Gu M, Eisenstat S C. A stable and efficient algorithm for the rank-one modification of the symmetrical eigenproblem[J]. SIAM Journal on Matrix Analysis and Applications, 1994, 15(4):1266-1276.

基于机器学习模型的青藏高原日降水数据的订正研究

Correction of the daily precipitation data over the Tibetan Plateau with machine learning models

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 0

[1]	徐田利, 邬光剑, 张学磊, 燕妮, 杨松. 基于MODIS数据的青藏高原冰川反照率时空分布及变化研究[J]. 冰川冻土, 2018, 40(5): 875-883.
[2]	李玲萍, 刘维成, 杨梅, 李岩瑛. 1971-2015年青藏高原东北边坡降水特征及主要影响因子分析[J]. 冰川冻土, 2018, 40(5): 916-924.
[3]	常思静, 杨蕊琪, 章高森, 刘光琇, 陈拓. 青藏高原土壤中一株原油降解菌的作用机制探究[J]. 冰川冻土, 2018, 40(5): 1037-1046.
[4]	陈志恒, 张杰, 徐玮平. 青藏高原初春积雪的多尺度变化与北大西洋海温的关系[J]. 冰川冻土, 2018, 40(4): 655-665.
[5]	王陆阳, 吴青柏, 蒋观利. 风沙堆积对下伏多年冻土影响的数值模拟[J]. 冰川冻土, 2018, 40(4): 738-747.
[6]	刘鑫, 王一博, 吕明侠, 孙岩, 杨文静, 赵金鹏. 基于主成分分析的青藏高原多年冻土区高寒草地土壤质量评价[J]. 冰川冻土, 2018, 40(3): 469-479.
[7]	吴小波, 南卓铜, 王维真, 赵林. 基于Noah陆面过程模型模拟青藏高原植被和土壤特征对多年冻土的影响[J]. 冰川冻土, 2018, 40(2): 279-287.
[8]	宋炫颖, 刘勇勤. 高寒地区湖泊浮游病毒时空分布特征[J]. 冰川冻土, 2018, 40(2): 395-403.
[9]	蒋观利, 吴青柏, 张中琼. 青藏高原不同高寒生态系统类型下多年冻土活动层水热过程差异研究[J]. 冰川冻土, 2018, 40(1): 7-17.
[10]	杜玉娥, 刘宝康, 贺卫国, 段水强, 侯扶江, 王宗礼. 1976-2017年青藏高原可可西里盐湖面积动态变化及成因分析[J]. 冰川冻土, 2018, 40(1): 47-54.
[11]	张宝贵, 刘晓娇, 刘敏, 张威, 章高森, 伍修琨, 陈拓, 刘光琇. 青藏高原疏勒河上游不同类型冻土可培养细菌多样性特征研究[J]. 冰川冻土, 2018, 40(1): 156-165.
[12]	刘亚军, 张玉兰, 康世昌, 李小飞, 陈鹏飞, 郭军明. 青藏高原东南部冰川雪冰重金属元素特征[J]. 冰川冻土, 2017, 39(6): 1200-1211.
[13]	朱美壮, 王根绪, 肖瑶, 胡兆永, 宋春林, 黄克威. 青藏高原多年冻土区高寒草甸土壤水分入渗变化研究[J]. 冰川冻土, 2017, 39(6): 1316-1325.
[14]	张志刚, 王建, 何元庆, 何则, 齐翠姗, 李盼盼. MIS 3时期青藏高原东南部稻城古冰帽冰进事件研究[J]. 冰川冻土, 2017, 39(5): 957-966.
[15]	张艳阁, 徐建中, 余光明. 祁连山老虎沟地区夏季大气颗粒物中水溶性离子的变化特征[J]. 冰川冻土, 2017, 39(5): 1022-1028.