Ground insolation or short-wave incoming radiation (SWR) at the ground is an important input in models of vegetation productivity, hydrology, ecological functioning, crop growth, etc. but is measured over limited number of stations, which fail to capture its spatial variability. A physical retrieval scheme was used for mapping insolation using METEOSAT VISSR sensor data over the Indian landmass for the winter season. The analysis covered the months of November-December 1998 and January 1999, and used five daytime acquisitions of meteosat visible (0.3-1.1 μm), thermal (10.5-12.5 μm), water vapour (6.5-7.5 μm) channel data. A total of 450 images per band per month was processed to retrieve instantaneous, daily total and five-day average total SWR over India. The estimated SWR was compared with the ground-measured pyranometer data over 18 Indian land stations acquired from India Meteorological Department. This validation for instantaneous SWR at 10:00 and 14:00 h, and daily total SWR for six selected dates (1, 5, 10, 15, 20, 25) of each month indicated root mean squared error (RMSE) of 62.51-91.93 Wm - 2 , i.e. 15-18% of observed mean and 1.14-1.74 MJm - 2 respectively. When five-day averages for the three-month period were compared, METEOSAT-derived data had a RMSE of 1.93 MJm - 2 , which is 12% of the mean over pooled data.