Integration of two spectral indices to monitor loss of moist grasslands within the Jaldapara Wildlife Sanctuary, India

Abstract

We conducted a study within flood-managed grasslands to evaluate the utility of remotely sensed imagery to evaluate the influence of an altered flooding regime on grassland distribution. Grasslands found along the Torsa River, which flows north to south through the Jaldapara Wildlife Sanctuary in West Bengal, India, provided an excellent test case due to the protected nature of this landscape from intensive management and cultivation. Further, during the 1968 flood season, the Torsa River experienced a major shift in its course from the west side of the sanctuary to east. We used remote-sensing data to identify an efficient method to spectrally monitor changes in grassland distribution. Spectrally normalized multi-temporal (1978, 1990, 2001, 2005) Landsat (MSS, TM, ETM) and ASTER data were used to compare changes in grassland distribution between the current and historic floodplain. A combination of the normalized difference vegetation index (NDVI) and a normalized difference dry index (NDDI) proved very useful in identifying and monitoring grasslands. Given the absence of historic ground data, spectral indices derived from historical satellite imagery also proved valuable as a means to understand temporal dynamics of the distribution of grasslands.

Acknowledgements

We are thankful to West Bengal Forest Department, India, for giving us the permission and logistic support to work within the sanctuary. We are thankful to the Ford Foundation International Fellowship Program for funding the majority of this work, the National Geographic Research Grant from Ashoka Trust for Research in Ecology and Environment (ATREE, Bangalore) and the Ecology Center, Utah State University for funding the field work, and the Intermountain Region Digital Image Archive Center – NASA for providing satellite imagery and logistical support.