Inversion Model Study Of Suaeda Salsa Biomass Based On Hyperspectral Remote Sensing

Liaohe estuary coastal wetland is located in the interchange between sea and land. It has a major role in regulation of climate, water conservation, distributed flood and conservation of biological diversity. Coastal wetland is not only a sensitive area of climate change, but also a fragile area of the ecological environment due to human activity. In recent years, coastal wetland environment is suffering from various threats with the expansion of human activities. How to rationally and effectively use the coastal tidal resource for achieving sustainable development of the resources has attracted more and more attentions nowadays.Vegetation biomass is an important index for evaluating wetland ecosystem health status. Remote sensing has many advantages, such as repeat detection, real time, non-destructive, etc. It can be applied in the vegetation biomass quantitative inversion field. Vegetation index has been proved more sensitive in detecting vegetation physiological parameters than single band. There is a strong correlation between vegetation index and many vegetation elements(e.g. vegetation biomass). Liaohe estuary coastal wetland has given birth to world wonders red beach. Suaeda is one of dominant vegetations in the coastal wetland of Liaodong Bay. It also is the pioneer plant developing from land to sea. This paper takes Suaeda as the object, and carries out biomass inversion by remote sensing.The main works of this paper are as follows:(1) Based on sensitivity analysis, this paper chooses the bands which are sensitive to biomass and 8 vegetation indices which have higher correlation of Suaeda biomass by field datas(measured spectral data and biomass data). Linear and nonlinear biomass quantitative retrieval models were constructed based on vegetation indices. The study found that NDVI, RVI and TSAVI are fit for inverting Suaeda biomass.(2) Satellite biomass inversion models were constructed based on simulated Landsat OLI data, and the result shows that the linear model of TSAVI and power model of NDVI have better inversion effects, the determination coefficients(r2) of linear model and power model are 0.852 and 0.832, F values are 137.792 and 118.519 respectively.(3) This paper obtained the Suaeda biomass Satellite distribution maps of the study area through the linear model of TSAVI and power model of NDVI. The total Suaeda inversion biomasses were 365.6350t(TSAVI) and 386.7063t(NDVI) in the study area in 2013, and relative errors were 10.95% and 5.83% respectively. Considering the model accuracy and relative error, this paper thinks that the NDVI-power model is the best inversion model for mature Suaeda.(4) The paper obtained Suaeda biomass temporal and spatial distribution based on NDVI-power model during 2013-2015, and carried out the analysis of Suaeda biomass and distribution area.