干旱区生态环境知识资源中心

Time series vegetation index using observations of the earth from space platform is a valuable source to derive several plant biophysical parameters for ecological, hydrological, climate models and to study land-use land-cover change dynamics. Indian geostationary satellite (INSAT 3A) sensor (CCD) observes the earth surface with continental (Asia) coverage at 1 km x 1 km spatial resolution and high temporal frequency (half-an-hour) at constant view direction. This study was aimed at defining and implementing an algorithm to retrieve normalized difference vegetation index (NDVI) at continental scale from INSAT 3A CCD surface reflectances in red (0.62-0.68 mu m), near infrared (0.77-0.86 mu m) bands and evaluate it with the global product. The methodology includes vicarious calibration, cloud screening, atmospheric correction of at-sensor reflectances and development of protocol. The temporal dynamics of 16-day NDVI composite at 0500 GMT (10 : 30 local mean time) for a growing year (June 2008-April 2009) showed matching profiles with respect to global products (e.g. MODIS TERRA) over known land targets such as agriculture, forest and desert. The root mean square deviation between the two was 0.13 with correlation coefficient (r) 0.83. The differences were attributed to surface anisotropy, view angle difference and differences in spectral bandwidths with their relative positions in the electromagnetic spectrum.