A Physically-Based Model For Canopy Water Content Retrieval

Canopy water content (CWC) is one of the factors that influence photosynthesis and evapotranspiration. Vegetation indices have been proposed to destribe the relationship between spectral profiles and canopy water content. In this paper, we propose a retrieval method using the common swir bands of satellites (such as sentinel-2), and develop a split-window algorithm to weaken the effect of other influential factors. PROSAIL model and the SPARC03 dataset are used for validation by resampling the hyperspectral data to board bands using spectral response function of sentinel-2. The result shows that the method can retrieve the canopy water content of different species of vegetation, with a satisfactory R of 0.669. After dividing the vegetation into two classes, both classes have an improvement with R of 0.783 and 0.939.