Plateau Wetland Vegetation Biomass Estimation Method And The Change Of Time And Space Remote Sensing Research On The Ground

Vegetation biomass is an important indicator to reflect the eco-system environment. Traditional biomass estimation methods are difficult to achieve large regional and multi-temporal observations because of its time-consuming and destructive. Thanks to the development of remote sensing technology, spatial and temporal scales of the regional biomass estimation become possible. At present, biomass estimation used remote sensing data focused primarily on research in forest and shrub species. Herbaceous species plateau wetlands, as a typical fragile ecosystems, has important scientific significance and value for monitoring and early warning of dynamic change of ecological environment in ecologically fragile areas.The model of biomass in the study area using remote sensing data is built. It is based on statistical regression model. Different resolution remote sensing data are selected as inputs of the biomass model derived at different resolutions from oulpuls of the model is analyzed to study scale converse. Scaling method from TM to MODIS is constructed using statistical mean value method, and biomass results derived from MODIS with low resolution are corrected by TM result conversed.TM data which is got on Aug.20th2011and MODIS product which is synthesized on the225th2011are thoroughly used to analysis the main parameters. The spatial-temporal variations of these parameters in the biomass model are analyzed. Simulated biomass results in the paper are respectively compared with total field measured data in September2010, partial field measured data on Aug.20th2011and biomass results estimated by ASAR data. The main research results and conclusions are as follows:(1) The correlation among the parameters obtained by remote sensing data in the study area is analyzed deeply. Principal component regression analysis is used to replacing for the multiple linear regression model to solve the effective of the multicollinearity. Normalized Difference Vegetation Index(NDVI)、Ratio Vegetation Index(RVI)、Difference Vegetation Index(DVI)、Enhanced Vegetation Index(EVI) and LAI are defined as explanatory variable of multiple linear regression model. (2) The coefficient of determination (R2) of principal component regression model constructed by TM data (30meters spatial resolution) reached0.766,P equals0.000, each regression parameter variables are statistically significant. Reduction to the original five parameters, the coefficient of determination reaches0.887, the value of F is23.584, the test value of sig is0.000. Comparing the calculation biomass and the reservation biomass, we can get the conclusion that the results are fit with the requirement, and the coefficient of determination reaches0.966. Taking EVI as explanatory variables in linear regression model is applied to estimate biomass in the study area using MODIS product (250meter spatial resolution). The coefficient of determination is0.826, the value of F is15.734and two tailed test value is0.000. By contrast the calculation biomass with the field measured biomass, the model is corrected.(3) The model is aimed at calculating biomass in specific area. Therefore, different remote sensing methods are focused on different remote sensing data. Using different spatial resolution biomass results to analysis the characteristics of temporal and spatial variation.(4) Conducted a sample survey of biomass in September2010and June&August2011in the study area.