Contemporary land cover and land use patterns in India estimated by different regional and global data sets

Discrepancies in various land cover and land use (LCLU) data sets could induce big uncertainties in assessing interactions among human activities, climate, and ecosystem. In this study, we analyzed inventory LCLU records from Department of Economics and Statistics (DES) and remote-sensing-based data sets obtained from Resourcesat-1, MODIS, Globcover, and HYDE 3.1 for 2005 in India. Based on the DES and Resourcesat-1 data sets, the best estimate of agricultural area was 143 million ha which was lower than MODIS and Globcover data sets (158–167 million ha). Global data sets have underestimated forest area by 20–30 million ha, comparing to DES and Resourcesat-1 data sets (63–66 million ha). All the remote sensing data sets showed a wide range for urban area (1.5–8.4 million ha) which was lower than inventory data sets. Our results caution scientific community and policymakers to carefully use global LCLU data sets that are significantly different from DES records in India.