Improvement in 1985-1993 NDVI time-series for world desert targets using post-launch calibration

Use of the Normalized Difference Vegetation Index (NDVI) for climate and agricultural monitoring purposes requires a multi-year NDVI time-series. The current archive of the NOAA/NESDIS Global Vegetation Index (GVI) product, widely used in research and for operational applications, contains nine years of data (1985-1993), which were collected from two polar orbiting operational satellites, NOAA-9 and NOAA-11. During these years the performance of the Advanced Very High Resolution Radiometer (AVHRR) in the visible (CH1) and near-infrared (CH2) channels differed between satellites and most importantly, deteriorated over time for each of them. NOAA/NESDIS has recently developed a technique to improve channel performance (Rao et al., 1993). The degradation rate was used to calculate postlaunch calibration coefficients for the visible and near-infrared channels. The relative degradation in time was determined for the visible and near-infrared channels using the southeastern Libyan desert, which surface was assumed to be time-invariant. The purpose of this paper is to apply the post-launch calibration to other major deserts of the world, and to estimate whether the stability of NDVI improved for each satellite and between the satellites especially in comparison with the pre-launch calibration coefficients