Abstract
In the present study, we propose a new index called a Vegetation-Soil-Water (VSW) index to monitor land cover conditions. The VSW index is defined as a natural extension of PVI for monitoring not only vegetationconditions but also soil and water conditions as well. VSW index is defined using a triangle on a NIR-Red scatter plot. The PVI measures only vegetation parameters, whereas the VSW index monitors vegetation, soil and water parameters simultaneously by measuring the distances in the scatter plot. Landsat Thematic Mapper (TM) scenes were used for the analyses. Wetland plants in Kushiro Mire were in the beginning of their growing season in June, and in their maximum growth stage in August. All the vegetation was dead in October. These scenes show how the vegetation changes seasonally and year by year for 5 years. We determined the end member spectral points for vegetation, soil and water (VSW) with an algorithm to fit a triangle to the spectral distribution. This automatic end member determination also standardizes the spectral responses of scenes. Scatter plots of the TM scenes on NIR-red axes were overlaid with the determined VSW end member points. The VSW index was used to monitor land cover change in the Kushiro wetland with Landsat TM images to evaluate the effectiveness of our approach for wetland monitoring. Changes in the vegetative components of the Kushiro wetland, both seasonally and over the full 5 year period we examined, could be easily resolved from these color composite images.
