【作者】 计淇才；

【导师】 王宗敏；

【作者基本信息】 郑州大学 ， 水利信息技术， 2011， 硕士

【摘要】 干旱因其持续时间长、影响范围广、灾害损失重等特点成为在世界范围内影响最为严重的自然灾害之一。旱灾的频繁发生使我国经济发展、人民生活等遭受了巨大损失,因此对干旱进行监测,掌握其发生、发展规律及影响因子之间的关系,科学合理地预报旱情,对于减少灾害损失、社会的可持续发展等具有重要的实际意义,而遥感干旱监测则以其监测范围广、数据更新快等优点在干旱监测中发挥了重要作用。本文以河南省为例,以MODIS影像为基础,分别计算了归-化植被指数(NDVI)、距平植被指数(AVI)、植被状态指数(VCI)、温度状态指数(TCI),并分析了在不同时空格局下精度较高的遥感干旱指数,以提高遥感干旱监测的精度。本文首先对气象数据进行处理,利用气象数据中的站点坐标、月降水量等信息计算出降水量距平百分率,然后对河南省近十年的干旱状况进行了一系列的分析,同时对获取的MODIS影像数据运用MRT、ENVI进行重投影、剪切等预处理得到河南省的影像图,根据常用遥感干旱指数的计算公式分别计算出常用遥感干旱指数的影像图。为了对比分析不同时域各遥感干旱指数的精度,按照行政区划对河南省分为豫中、豫东、豫南、豫西、豫北五个区域,分别统计各分区的遥感干旱指数值及传统干旱指数降水量距平百分率。然后分别选取月、季度、年为时间尺度,以各个分区为地域条件分析传统干旱指数与常用遥感干旱指数之间的相关性,以确定不同时空格局条件下精度较高的遥感干旱指数。从研究结果来看,除豫东地区是AVI与平均降水量间的相关性最大外,其余均为NDVI与降水量的相关性最强。因此当以多年平均得到的月均遥感干旱指数值为尺度研究各地区干旱时用NDVI来预测干旱情况具有较高的准确性,而以季度为研究时间尺度时可以得到在豫中地区四个季节中遥感干旱指数与降水量相关性最高的依次为：VCI、VCI, TCI、TCI;豫东地区依次为：AVI、VCI、TCI、NDVI;豫南地区依次为：VCI、AVI、AVI、NDVI;豫西地区依次为：TCI、VCI、VCI、NDVI;豫北地区依次为：AVI、VCI、AVI、TCI。更多还原

【Abstract】 Drought is one of the severest nature disasters for its long duration, a wide range affect, disaster losses weight and other characteristics, and for the frequent occurrence of drought in China, the country’s economic development and daily life suffered a great loss. Therefore, drought monitoring and control the relationship between the development and its influencing factors, forecast drought in scientific and reasonable, has important practical significance to the reduction of disaster losses and sustainable development of society. Remote Sensing is playing an increasingly important role in drought monitoring, with its wide range of monitoring and quickly data updates. Henan province was taken as the case, on the basis of MODIS image, Normalized Difference Vegetation Index (NDVI), Anomaly Vegetation Index (AVI), Vegetation Condition Index (VCI) and Temperature Condition Index (TCI) was calculated separately, and compare and analyze precision of the indices under the same spatial and temporal patterns, to select the Remote Sensing of drought index of high precision under different spatial and temporal patterns, and improve the precision of remote sensing of drought monitoring.This paper process the meteorological data of each measure site from China monthly surface climate data set, select the meteorological data site of Henan province, calculate precipitation anomaly percentage of Henan province using site coordinates and monthly precipitation of meteorological data and other information, and analyze the drought condition of Henan province in nearly decade; re-project and cut the MODIS image(monthly vegetation index of lkm resolution, land surface temperature of lkm resolution) and other pretreatment by MRT and ENVI to obtain the image of Henan province, calculate popular Remote Sensing image of drought index base on the formula. According to administrative divisions, Henan Province is divided into five regions (mid Henan, eastern Henan, southern Henan, western Henan and northern Henan) to analyze the precision of popular Remote Sensing of drought index of different spatial and temporal condition, statistic Remote Sensing of drought index and precipitation anomaly percentage of sub area. Take month, quarter and year as time scale, each sub area as spatial condition to analyze the correlation of traditional drought index and Remote Sensing of drought index, and select the Remote Sensing of high precision in different spatial and temporal condition. From the research results, the best correlation with precipitation is AVI in eastern Henan while the other is NDVI, so the use of NDVI to predict drought conditions with high accuracy when research drought condition of each sub area with monthly Remote Sensing of drought index value of annual average, and when research drought condition by quarter, the highest correlation of Remote Sensing of drought index and precipitation in four seasons is VCI, VCI, TCI and TCI in mid Henan; the follow is AVI, VCI, TCI and NDVI in eastern Henan; VCI, AVI, AVI and NDVI in southern Henan; TCI, VCI, VCI and NDVI in western Henan; AVI, VCI, AVI and TCI in northern Henan.更多还原