The Study Of Hyperspectral Remote Sensing Estimation Models For Above Ground Biomass Of East Inner Mongolia Grassland

At present studies in carbon cycle have increasingly become a focus of global change and ecological research in the world. The estimation of vegetation biomass is critical for carbon cycle modeling and climate change mitigation programs. Field surveys provide the most accurate method for obtaining vegetative data, but are far too time-consuming, costly to cover large expanses and the spatial scale and time scale is limited. Besides, precise monitoring of agricultural crop biomass and yield quantities is critical for crop production management and prediction. The position of the inflexion point in the red edge region (680 to 780 nm) of the spectral reflectance signature, termed the red edge position (REP), is affected by biochemical and biophysical parameters and has been used as a means to estimate foliar chlorophyll or nitrogen content. Nevertheless, traditional multispectral broadband sensor data have known limitations of sensor saturation and absence of specific narrowbands to target and highlight specific biophysical and biochemical parameters. These factors lead to significant uncertainties in spectro-biophysical and/or biochemical modeling of vegetation biomass assessment. Indeed, it has been shown how hyperspectral data can provide significant improvements in extracting the relationship between spectral information and vegetation biomass. The methods of hyperspectral remote sensing has already been used in estimation of vegetation biomass on large areas.Firstly, the aim of this paper were to constructed the spectral estimation model of above-ground biomass of grassland in Inner Mongolia by using the regression analysis, based on the spectral variables calculated from measured spectral data and the biomass data in the same period. The results showed that the relationship between the grassland biomass and the spectral variables was correlated, and could be fitted by linear equation, logarithm equation, polynomial equation, power equation and exponent equation. In these equations fitting results all reached significant related levels, and the fitting effect of linear and non-linear equations based on HVI was the best, but the best precision of one-sample linear and non-linear method was based on Dr. The result of using multiple linear regression shows that the estimation model based on five vegetation indices which are VI2, VI5, HVI, REP, TCI was better than one-sample linear and non-linear models. The results of this paper indicated that the grassland biomass was estimated by using the hyperspectral variables and vegetation indices.Secondly, this paper were to constructed the spectral estimation model of above-ground biomass of grassland in Xilinhot by using the statistical models, based on the spectral variables calculated from measured ASD spectral data, the biomass data and the Hyperion hyperspectral remote sensing data in the same period. The results showed that the relationship between the grassland biomass and the spectral variables was correlated, and could be fitted by linear equation, logarithm equation, polynomial equation, power equation and exponent equation. In these equations fitting results reached significant related levels, and the estimation effect of linear and non-linear equations based on Triangle Chlorophyll Vegetation Index (TCI) was the best. Besides, we investigated to what degree the calibrated biomass models could be scaled to Hyperion data. The relationship between the ASD data and the Hyperion data was highly correlated, and the exponent estimation model based on Hyperion TCI was best. The results of this paper indicated that the grassland biomass was estimated by using the Hyperion hyperspectral vegetation indices in large area.