A Study On The Inversion And Validation Of Crop Leaf Area Index Based On Remotely-sensed Data

Leaf area index (LAI) of vegetation plays a key role in the research and application in the fields of agriculture, ecology, meteorology, environment science and so on. The accurate retrieval and systemic validation of regional LAI data based on remotely-sensed data becomes very important and necessary.Taking Shijiazhuang, Hengshui and Xingtai as the case study area, the author tested some key techniques in crop LAI inversion and validation. Base on MODIS, TM, the ground LAI measurement data which combined with Cubic Spline Interpolation and GPS data, the author set up a statistical regression models to retrieve crop LAI from Landsat TM data, then verified the accuracy of NASA MODIS LAI data product in the study area. Meanwhile the author also tested the feasibility of 2 methodologies in LAI retrieval and validation based on remotely-sensed data.The main findings and conclusions of this study are as follows:(1) The application of Cubic Spline Interpolation of in situ crop LAI measurement data interpolation is successful, as is more accurate than using neighboring date's data substitution or linear interpolation method.(2) By testing the respective relation between different vegetation indices and leaf area index, the author found that NDVI is better in estimating crop LAI than other vegetation indices, such as the Simple Ratio (SR), Soil-Adjusted Vegetation Index (SAVI) and Modified Soil-Adjusted Vegetation Index (MSAVI).(3) Through scaling up the LAI imagery retrieved from TM image, the author found that scale can cause very large change in resultant LAI. The parameters retrievaled from in situ measurement or higher-resolution satellite images can't be used directly to replace the pixel value of the coarse resolution satellite image, and scaling is necessary in multi-scale inversion and validation of remotely sensed land surface parameters.(4) Different scaling method can result very different outcome as scaling before or after retrieval. We must choose the proper way to acquired more curate LAI data from remotely-sensed imagery.(5) The author found that there are large differences between TM LAI and NASA MODIS LAI land product in the study area. There are some unexpected values in NASA MODIS LAI dara product. To improve the accuracy of MODIS LAI product, some input land surface parameter especially land cover should be improved.