Remote sensing data has long been used to monitor global ecosystems for floods and droughts and AVHRR data, as one of the first product, has many users interested in receiving the data within hours of acquisition. With the introduction of a new series of sensors in 2000 (the AVHRR/3 series), the quality of the NDVI datasets available for real time environmental monitoring has declined. This paper provides evidence of problems of cloud contamination, calibration and noise in the real time data which are not present in the historical AVHRR NDVIg dataset. These differences introduce significant uncertainty in the use of the real time data, degrading their utility for detecting climate variations in near real time.
