Fractional vegetation cover estimation in arid and semi-arid environments using HJ-1 satellite hyperspectral data

Abstract

This paper evaluates the usefulness of the hyperspectral imager (HSI) onboard Chinese HJ-1-A small satellite in vegetation mapping. Fractional vegetation cover (FVC) is an important surface microclimate parameter for characterizing land surface vegetation cover as well as the most effective indicator for assessing desertification and crop growth condition. The HJ-1/HSI data were used to calculate the narrow band vegetation index by using the in situ plot FVC data, which was then applied in sub-pixel de-composition model for the FVC estimation, namely the dimidiate pixel model. The FVC information in the Shihezi Area, Xinjiang, China was retrieved based on the dimidiate pixel model. Cross-checked with the in situ measured FVC data, a correlation coefficient square of 0.86, and the root mean square error of 10.9% is statistically achieved. The verification indicates that the FVC result retrieved from the HJ-1/HSI data is well correlated with the in situ measurements, demonstrating that the HJ-1/HSI data are promising for studying the potential impacts of global climate change on the arid and semi-arid landscapes.