Early warning monitoring systems in food-insecure countries aim to detect unfavourable crop and pasture conditions as early as possible during the growing season. This manuscript describes a procedure to estimate the probability of experiencing a biomass production deficit during the on-going season based on the statistical analysis of Earth Observation data. A 15-year time series of the Fraction of Absorbed Photosynthetically Active Radiation from the SPOT-VEGETATION instrument is used to characterize the climatological development of vegetation, its variability and its current status. Forecasts of overall seasonal performances, expressed in terms of the probability of experiencing a critical deficit, are updated regularly whenever a new satellite observation is made available. Results and performances of the method are discussed for croplands and pastures in the Sahel.
REMOTE SENSING OF ENVIRONMENT p. 57-68 vol. 142