Tropical Forest cover type characterisation in central highlands of India, using multi-temporal IRS-1C WiFS data

Abstract

In this study, we explored the potential of multi-temporal IRS-1D WiFS (Wide Field Sensor) data for characterization of tropical forest in Central India. As the WiFS has a red (R) and near infrared (NIR) band that is sensitive to vegetation, soil moisture and leaf water content, the Normalised Difference Vegetation Index (NDVI) and other matrices of NDVI were calculated. A forest cover map of Central India was generated from a hybrid classification approach with 8 month WiFS composite NDVI data. Eight different forest categories were distinguished from the 188 m spatial resolution WiFS data. The WiFS forest map was compared to estimates of forest area derived from IRS-1C LISS III product of Forest Survey of India (FSI). There was a good agreement on spatial distribution and area of forest between WiFS product and the LISS III images. In the present study the forest cover of the central highlands is accounted as 34.68% whereas the FSI reports 34.84% forest cover. However, the WiFS product provided additional information on forest types, viz., tropical moist deciduous, dry deciduous and mixed deciduous. Analysis of matrices of NDVI over different seasons allowed for the identification of distinct phenological forest cover types. It is evident that WiFS data can be used to provide timely and detailed forest maps with limited ancillary data needed. The WiFS derived forest maps could be very useful as input to biogeochemical models (particularly carbon cycle models) that require timely estimates of forest area and type.