Wetland biomass inversion and space differentiation – Using the Yellow River Delta Nature Reserve as an example

As one of the world's three major ecosystems, the study of wetland health has global environmental implications. Vegetative biomass has emerged as one of the most valuable indicators of wetland health. The ongoing development of remote sensing techniques, coupled with improved data processing and modeling, creates new possibilities to monitor and understand wetlands. Multiple regression model types were employed to find the best fit between Landsat-8 images, vegetation indices, and field measured biomass in the Yellow River Delta Nature Reserve. Then, these models were used to estimate the spatial distribution of wetland vegetative biomass. Further, the relationship between wetland vegetative biomass and soil factors (organic matter, nitrogen, phosphorus, potassium, water soluble salt, pH, and moisture) was correlated and modelled. It was discovered that the Landsat-8 images and vegetative indices are better at predicting biomass when dry weight data is used rather than fresh weight data. Using multiple regression model types, we were able to achieve higher correlation and higher fit accuracy with vegetative indices and Bands 1-5 as independent variables and biomass dry weight as the dependent variable. Several soil factors were discovered, such as soil moisture and salt concentrations, which affect spatial distribution patterns of biomass in this wetland. Continued research using and improving upon these techniques will yield new insights into how to better promote wetland health.