Validation Of Remote Sensing Leaf Area Index Product Of Forest And Cropland Over China

Because of the need for research on global climate change, carbon sources and sinks change driving mechanism, leaf area index (Leaf Area Index, LAI) as a key input to climate models,the carbon cycle models and other dynamic process model, the research and application of LAI should be required on regional and global scales, so the use of remote sensing satellite data LAI inversion, thereby producing a global remote sensing data LAI products to be more widely used. However, due to different factors in different measurement methods and instruments and canopy structure defines leaf area index, LAI products can significantly change, currently not have the accuracy and consistency of the global and regional applications products. Therefore, in the application of remote sensing LAI products, it is particularly important to the precision of the evaluation of LAI products. In this study, forest Jiagedaqi in northeast China and the typical farmland in Nanjing city of Jiangsu Province as the research area.Get the typical vegetation index based on environmental satellite remote sensing data,building a regression analysis model inverse of woodlands, wheat and rice LAI use of environmental satellite vegetation index and measured LAI.Finally, the use of environmental satellite data retrieval LAI to validate MODIS LAI products and GLOBCARBON LAI products through the scale conversion,and analyze the error source of LAI products,and then to correct the farmland area MODIS LAI products and GLOBCARBON LAI products. The main conclusions of this study are as follows:(1)Forest Jiagedaqi in northeast China, three kinds of LAI data in the largest regional vegetation LAI range for the GLOBCARBON LAI data,its values in the 0.93～4.91, HJ～1kmLAI data and MODIS LAI data range is basically the same, GLOBCARBON LAI highest average, its value than HJ-1kmLAI high 0.29, the error is 11%, while the MODIS LAI data mean is lower than the mean HJ-1kmLAI 0.28, the error was 11.8%, two remote sensing data products LAI precision error in the study area were about 20%, but GLOBCARBON LAI overestimate phenomenon, and MODIS LAI data was lower than the measured value inversion.(2) The study area of crops in nanjing, in wheat area GLOBCARBON LAI mean lower than HJ-30mLAI mean 1.18, the error is 44%, MODIS LAI mean than HJ-30mLAI mean low 1.75, the error is 66%; in rice area, GLOBCARBON LAI mean than HJ-30mLAI mean low 0.84, the error is 25%, MODIS LAI mean lower than HJ-30mLAI mean 1.47, the error is 43%. By the results of the analysis can be seen in the study area of wheat and rice MODIS LAI, GLOBCARBON LAI was significantly lower than the mean value of the environmental satellite LAI inversion, there is a serious underestimation. According to the analysis, due to the Nanjing farmland were scattered like distribution, surface heterogeneity serious, leading to a low-resolution GLOBCARBON LAI and MODIS LAI products mixed pixels exist.(3)Then the GLOBCARBON LAI and MODIS LAI product mixed pixels decomposition in Nanjing crops in the study area,come in wheat area, GLOBCARBON LAI overestimate HJ-30mLAI mean of 0.25, the error from 44% down to 8.6%, MODIS LAI than HJ-30mLAI mean low 0.29, the error from 66% down to 10.9%; in rice area, GLOBCARBON LAI mean higher than HJ-30mLAI mean 0.28, the error from 25% down to 7.6%, MODIS LAI mean lower than HJ-30mLAI mean 0.23, the error from 43% dropped 6.7%. The revised data from the point of view, MODIS LAI GLOBCARBON LAI revised and greatly improved the problem of mixed pixels. But GLOBCARBON LAI overrated phenomenon exists, and MODIS LAI data was lower than the actual value inversion.