| 992 | 8 | 339 |
| 下载次数 | 被引频次 | 阅读次数 |
干旱是由降水不足引起的，受温度、蒸散发等各种环境因素的影响，导致缺水和作物歉收。传统的干旱监测方法主要侧重于气象、水文等单一因子，而对多因子综合干旱监测的研究相对有限。本文利用2001—2015年的温度状态指数(TCI)、植被状态指数(VCI)、植被供水状况指数(VSWI)、降水状态指数(PCI)、土壤湿度状态指数(SMCI)、高程(DEM)及田间持水量(AWC)等7个干旱因子为自变量，以综合气象干旱指数(CI)为因变量，利用随机森林(RF)、增强回归树(BRT)和人工神经网络(ANN)构建干旱监测模型，并以西南五省为研究区进行了评价和分析。结果表明，基于人工神经网络的干旱监测指标(ANN-CI)预测效果最好，在测试集中预测值和观测值间的可决系数(R2)为0.94,均方根误差(RMSE)为0.23。三种基于机器学习的综合干旱指标均在草地区表现最好，林区精度最差。在2001—2015年间，ANN-CI和植被生长状况具有显著的时空相关性(R2=0.70,p<0.01)。最后选用ANN-CI对西南地区2009/2010年干旱事件的发展过程进行监测，并且与帕默尔干旱指数(PDSI)的监测结果进行对比，结果表明ANN-CI能够较好地应用于区域的旱情监测。
Abstract：Drought is caused by a lack of precipitation and is affected by various environmental factors such as temperature and evapotranspiration, leading to water shortages and crop failures. Traditional drought monitoring methods mainly focus on single factors such as meteorology and hydrology, while the research on multi-factor integrated drought monitoring is relatively limited. Seven drought factors, including Temperature Condition Index(TCI), Vegetation Condition Index(VCI), Vegetation Supply Water Index(VSWI), Precipitation Condition Index(PCI), Soil Moisture Condition Index(SMCI), Digital Elevation Model(DEM) and soil Available Water Content(AWC), are used as independent variables in this paper. Taking the Composite index(CI) as the dependent variable, Random Forest(RF), Boosted Regression Trees(BRT) and Artificial Neural Network(ANN) are used to construct the drought monitoring model. The five provinces in Southwest China are taken as the study area for drought monitoring and comparison. The results show that the drought monitoring index based on ANN(ANN-CI) has the best prediction effect, and the determination coefficient(R2) and root mean square error(RMSE) between the predicted and observed values in the test set are 0.94 and 0.23 respectively. The three integrated drought indexes based on machine learning all perform best in the grassy area, while the accuracy in the forest area is disappointing. From 2001 to 2015, there is a significant spatial and temporal correlation between ANN-CI and vegetation growth(R2=0.70, p<0.01). Finally, ANN-CI is selected to monitor the development process of drought events in Southwest China in 2009/2010, and compared with the Palmer Drought Severity Index(PDSI) monitoring results, which show that ANN-CI can be better applied to regional drought monitoring.
[1] SHEFFIELD J,ANDREADIS K M,WOOD E F,et al.Global and continental drought in the second half of the twentieth century:severity-area-duration analysis and temporal variability of large-scale events[J].Journal of Climate,2009,22(8):1962-1981.
[2] WILHITE D A.Drought as a natural hazard:concepts and definitions[J].Drought,A Global Assessment,2000,1:3-18.
[3] ZARGAR A,SADIQ R,NASER B,et al.A review of drought indices[J].Environmental Reviews,2011,19(1):333-349.
[4] KEYANTASH J,DRACUP J A.The quantification of drought:an evaluation of drought indices[J].Bulletin of the American Meteorological Society,2002,83(8):1167-1180.
[5] MCKEE T B,DOESKEN N J,KLEIST J.The relationship of drought frequency and duration to time scales[C]//Eighth conference on Applied climatology.California:American Meteordtgical Society,1993:179-183.
[6] VICENTE-SERRANO S M,BEGUERíA S,LóPEZ-MORENO J I.A multiscalar drought index sensitive to global warming:the standardized precipitation evapotranspiration index[J].Journal of Climate,2010,23(7):1696-1718.
[7] PALMER W C.Meteorological drought[M].Marylond:US Department of Commerce,Weather Bureau,1965.
[8] SWAIN S,WARDLOW B D,NARUMALANI S,et al.Assessment of vegetation response to drought in Nebraska using Terra-MODIS land surface temperature and normalized difference vegetation index[J].GIScience & Remote Sensing,2011,48(3):432-455.
[9] 温庆志，孙鹏，张强，等.基于多源遥感数据的农业干旱监测模型构建及应用[J].生态学报，2019,39(20):7757-7770.WEN Qingzhi,SUN Peng,ZHANG Qiang,et al.An integrated agricultural drought monitoring model based on multi-source remote sensing data:model development and application [J].Acta Ecologica Sinica,2019,39(20) :7757-7770.
[10] 沈润平，郭佳，张婧娴，等.基于随机森林的遥感干旱监测模型的构建[J].地球信息科学学报，2017,19(1):125-133.SHEN Runping,GUO Jia,ZHANG Jingxian,et al.Construction of a drought monitoring model using the Random Forest based remote sensing[J].Journal of Geo-information Science,2017,19(1):125-133.
[11] 张婧娴.综合多源遥感数据的干旱监测模型研究[D].南京：南京信息工程大学，2015.ZHANG Jingxian.Study on drought monitoring modeling integrated multi-source remote sensing data[D].Nanjing:Nanjing University of Information Science & Technology,2015.
[12] 孙嵩松.基于多源遥感数据构建的改进干旱监测模型[D].西安：陕西师范大学，2018.SUN Songsong,An improved drought monitoring model based on multi-source remote sensing data[D].Xi’an:Shaanxi Normal University,2018.
[13] PARK S,IM J,JANG E,et al.Drought assessment and monitoring through blending of multi-sensor indices using machine learning approaches for different climate regions[J].Agricultural and Forest Meteorology,2016,216:157-169.
[14] KOGAN F N.Application of vegetation index and brightness temperature for drought detection[J].Advances in Space Research,1995,15(11):91-100.
[15] RHEE J,IM J,CARBONE G J.Monitoring agricultural drought for arid and humid regions using multi-sensor remote sensing data[J].Remote Sensing of Environment,2010,114(12):2875-2887.
[16] WU J,ZHOU L,LIU M,et al.Establishing and assessing the Integrated Surface Drought Index (ISDI) for agricultural drought monitoring in mid-eastern China[J].International Journal of Applied Earth Observation and Geoinformation,2013,23:397-410.
[17] ZHOU L,ZHANG J,WU J,et al.Comparison of remotely sensed and meteorological data-derived drought indices in mid-eastern China[J].International Journal of Remote Sensing,2012,33(6):1755-1779.
[18] WANG W,WANG W,LI J,et al.The impact of sustained drought on vegetation ecosystem in Southwest China based on remote sensing[J].Procedia Environmental Sciences,2010,2:1679-1691.
[19] PAULIK C,DORIGO W,WAGNER W,et al.Validation of the ASCAT Soil Water Index using in situ data from the International Soil Moisture Network[J].International Journal of Applied Earth Observation and Geoinformation,2014,30:1-8.
[20] 杜灵通，田庆久，王磊，等.基于多源遥感数据的综合干旱监测模型构建[J].农业工程学报，2014,30(9):126-132.DU Lingtong,TIAN Qingjiu,WANG Lei,et al.A synthesized drought monitoring model based on multi-source remote sensing data[J].Transactions of the Chinese Society of Agricultural Engineering,2014,30(9):126-132.
[21] 吴黎.基于MODIS数据温度植被干旱指数干旱监测指标的等级划分[J].水土保持研究，2017,24(3):130-135.WU Li,Classification of drought grades based on temperature vegetation drought index using the MODIS data [J].Research of Soil and Water Conservation,2017,24(3):130-135.
[22] 晏红波，韦晚秋，卢献健，等.基于TRMM数据与SPI指数的广西地区旱涝演变分析[J].国土资源遥感，2021,33(1):158-166.YAN Hongbo,WEI Wanqiu,LU Xianjian,et al.A study of drought and flood evolution in Guangxi based on TRMM data and SPI drought index[J].Remote Sensing for Land & Resources,2021,33(1):158-166.
[23] 赵安周，王冬利，范倩倩，等.TRMM数据在京津冀地区干旱监测适用性研究[J].水资源与水工程学报，2020,31(2):235-242.ZHAO Anzhou,WANG Dongli,FAN Qianqian,et al.Evaluating the applicability of TRMM data in monitoring drought events in Beijing-Tianjin-Hebei region[J].Journal of Water Resources & Water Engineering,2020,31(2):235-242.
[24] 胡蝶，沙莎，王丽娟，等.欧空局主被动微波土壤水分融合产品在甘肃省干旱监测中的应用[J].干旱气象，2019,37(4):517-528.HU Die,SHA Sha,WANG Lijuan,et al.Drought monitoring applications of the European Space Agency climate change initiative soil moisture combined product in Gansu Province[J].Journal of Arid Meteorology,2019,37(4):517-528.
[25] 李维娇，王云鹏.基于VCI的2003—2017年广东省干旱时空变化特征分析[J].华南师范大学学报(自然科学版),2020,52(3):85-91.LI Weijiao,WANG Yunpeng.An analysis of the spatial-temporal charateristics of drought in Guangdong based on vegetation condition index from 2003 to 2017[J].Journal of South China Normal University (Natural Science Edition),2020,52(3):85-91.
[26] 李琪，孙晓宇，王连喜，等.基于不同植被指数的VSWI在河南省春季干旱监测中的应用分析[J].作物杂志，2016(1):162-168.LI Qi,SUN Xiaoyu,WANG Lianxi,et al.The use of Vegetaion Supply Water Index (VSWI) based on different vegetation indices in the spring drought monitoring in Henan Province[J].Crops,2016(1):162-168.
[27] SEILER R A,KOGAN F,SULLIVAN J.AVHRR-based vegetation and temperature condition indices for drought detection in Argentina[J].Advances in Space Research,1998,21(3):481-484.
[28] MCVICAR T R,BIERWIRTH P N.Rapidly assessing the 1997 drought in Papua New Guinea using composite AVHRR imagery[J].International Journal of Remote Sensing,2001,22(11):2109-2128.
[29] 蔡新玲，叶殿秀，李茜，等.基于CI指数的陕西干旱时空变化特征分析[J].干旱地区农业研究，2013,31(5):1-8.CAI Xinling,YE Dianxiu,LI Qian,et al.Analysis of temporal-spatial variation characteristics of drought in Shanxi Province based on compound meteorological drought index (CI)[J].Agricultural Research in the Arid Areas,2013,31(5):1-8.
[30] BREIMAN L.Random Forests[J].Machine Learning,2001,45(1):5-32.
[31] ELITH J,LEATHWICK J R,HASTIE T.A working guide to boosted regression trees[J].Journal of Animal Ecology,2008,77(4):802-813.
[32] RIDGEWAY G.Generalized boosted models:a guide to the gbm package[J].Update,2007,1(1):1-15.
[33] MOKHTARZAD M,ESKANDARI F,VANJANI N J,et al.Drought forecasting by ANN,ANFIS,and SVM and comparison of the models[J].Environmental Earth Sciences,2017,76(21):1-10.
[34] WU X,LIU H,LI X,et al.Differentiating drought legacy effects on vegetation growth over the temperate Northern Hemisphere[J].Global Change Biology,2018,24(1):504-516.
[35] GESSNER U,NAEIMI V,KLEIN I,et al.The relationship between precipitation anomalies and satellite-derived vegetation activity in Central Asia[J].Global and Planetary Change,2013,110:74-87.
基本信息:
DOI：10.13928/j.cnki.wrahe.2022.02.004
中图分类号:P426.616
引用信息:
[1]何志远,钟九生,代仁丽.基于机器学习的综合干旱监测建模及在西南地区应用[J],2022,53(02):43-51.DOI:10.13928/j.cnki.wrahe.2022.02.004.
基金信息:
国家自然科学基金项目(41661081);; 贵州省科技计划项目(黔科合平台人才[2017]5726-56)