2001-2010年青藏高原草地生长状况遥感动态监测  

Remote sensing dynamic monitoring of grass growth in Qinghai-Tibet plateau from 2001 to 2010

作　　者：冯琦胜[1] 高新华[1] 黄晓东[1] 于惠[1] 梁天刚[1] 

机构地区：[1]兰州大学草地农业科技学院农业部草地农业生态系统学重点实验室,兰州730020

出　　处：《兰州大学学报（自然科学版）》2011年第4期75-81,90,共8页Journal of Lanzhou University(Natural Sciences)

基　　金：教育部高等学校科技创新工程重大项目培育资金项目(708089);国家自然科学基金项目(30972135);国家科技支撑计划项目(2009BAC53B01);国家高新技术研究发展专项(2007AA10Z232)

摘　　要：基于1970个地面实测数据,结合MODIS EVI和NDVI数据,利用留一法交叉验证方法确定了适合青藏高原地区草地生长状况的遥感反演模型,估算了2001-2010年草地生物量干重空间分布格局,分析了近10年草地生物量变化动态.结果表明:青藏高原地区MODIS的NDVI较EVI能更好地估算草地生长状况,指数模型反演的草地鲜重和风干重精度最高,而盖度反演适合使用乘幂模型;确立了基于MODIS-NDVI的青藏高原地区草地鲜重、风干重和盖度反演模型;青藏高原地区2001-2010年草地地上生物量总体有增加趋势,但不显著;月际变化较大的草地类型主要有沼泽类、山地草甸类和高寒草甸类.Using leave-one-out cross validation method,a suitable inversion model was established in Qinghai-Tibet plateau based on 1970 ground observed data combined with MODIS EVI and NDVI data.Then the spatialdistribution pattern of grassland dry biomass during 2001-2010 was simulated and grassland biomass dynamicsfor nearly 10 years was analyzed.The results show that NDVI is superior to EVI in Qinghai-Tibet plateau,theexponential model has the highest accuracy for calculating fresh and dry biomass,but the power model is bestfor calculating the coverage.The fresh,dry biomass and coverage inversion models based on MODIS-NDVI datawere established.The grassland aboveground biomass shows an increasing trend but it is not obvious.Marsh,mountain meadow and alpine meadow have a larger monthly change.

关 键 词：草地地上生物量 空间变化 反演模型 留一法 交叉验证 青藏高原 

分 类 号：S127[农业科学—农业基础科学] TP79[自动化与计算机技术—检测技术与自动化装置]