基于MODIS和AVHRR数据源的东北地区植被NDVI变化及其与气温和降水间的相关分析  

Correlation Analysis between NDVI and Climate in Northeast China based on AVHRR and GIMMS Data Sources

作　　者：毛德华[1,2] 王宗明[1] 罗玲[1] 杨桄[3] 

机构地区：[1]中国科学院东北地理与农业生态研究所,吉林长春130012 [2]中国科学院研究生院,北京100049 [3]空军航空大学特种专业系,吉林长春130022

出　　处：《遥感技术与应用》2012年第1期77-85,共9页Remote Sensing Technology and Application

基　　金：国家自然科学基金项目(40930527;40871187);中国科学院知识创新工程重要方向项目(KZCX2-YW-341);国家973计划项目(2009CB421103)资助

摘　　要：基于逐像元一元线性回归模型,应用MODIS NDVI数据对AVHRR-GIMMS NDVI进行时间序列拓展,拓展序列通过一致性检验,基于所建立的1982～2009年植被年最大NDVI数据集,在GIS平台上进行了植被NDVI变化和NDVI与年平均气温、年降水量之间的相关分析。研究结果表明:过去28a间,植被年最大NDVI呈3个变化阶段:1982～1992年呈小幅上升趋势,1992～2006年呈缓慢下降趋势,2006～2009年呈缓慢回升态势。由空间变异分析得出NDVI变化相对大的区域主要分布在内蒙干旱和半干旱区。21世纪初和20世纪90年代相对于80年代NDVI值升高,3个阶段平均NDVI变化幅度为±0.3。20世纪初,赤峰地区以及松嫩平原西部地区植被NDVI呈轻度增加的面积占全区6.45%。植被年最大NDVI与年平均气温、年降水量相关性空间差异明显。偏相关系数绝对值,气温大于降水的像元数占54%;综合分析,较降水而言,气温是东北全区植被年最大NDVI的主控影响因子。对于不同植被类型年最大NDVI,受气温影响强度由大到小依次为:森林>草地>沼泽湿地>灌丛>耕地;受降水影响按草地>耕地>灌丛>沼泽湿地>森林依次减弱。Based on the AVHRR GIMMS NDVI and MODIS NDVI,this paper constructed yearly maximum NDVI sequence covering Northeast China from 1982 to 2009 using per-pixel unary Linear Regression Model.The constructed NDVI through the consistency check and well be used in the temporal analysis and correlation analysis with climatic factors.Results showed that there were three change stages in the past 28 years：A modest increase trend during 1982-1992,slow decrease trend during 1982-2006 and a recovery during 2006-2009.Result from variation analysis on spatial pattern showed that obvious changes occurred mainly in Inner Mongolia Automomous Region（IMAR） arid and semi-arid zone.Mean NDVI value in 2000s and 1990s were bigger than 1980s.Change value ranged from-0.3 to 0.3 and increase trend were mainly found in Chifeng and west of Songnen plain which have a proportion about 6.45% in total area.A high spatial heterogeneity was observed in the correlation between yearly maximum NDVI and climate（temperature and precipitation）.Pixels which had a bigger coefficient of partial correlation between NDVI and temperature than between NDVI and precipitation account for 54% of all pixels.At whole scale,temperature was the dominate influence factor compared with precipitation.For different vegetation types,the yearly maximum NDVI was affected by annual average temperature as follows：forests〉grasslands〉marshes〉shrubs〉crops;and by precipitation：grasslands〉crops〉shrubs〉marshes〉forests.

关 键 词：逐像元一元线性回归模型 MODISNDVI AVHRR-GIMMSNDVI 气温和降水 相关分析 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]