The derivation of vegetation indices from AVHRR data

The reliability of a 1-weck composited normalized difference vegetation index has been evaluated by using a cloud-screening algorithm applied to the visible and near-infrared data from the Advanced Very High Resolution Radiometer on NOAA-9. It is found that in some areas ofthe U.S. Great Plains this satellite sensor product may not be reliable due to the high frequency of cloud occurrence. Using the example of day-to-day variation in (he observed clear-sky radiances for one target, the vegetation index is shown to have maxima at high off-nadir and low solar zenith angles: this behaviour has been examined in detail. Some recommendations to improve the compositing technique are given.