Monitoring food security requires near real time information on crop growing conditions for early detection of possible production deficits. Anomaly maps and time profiles of remote sensing derived indicators related to crop and vegetation conditions can be accessed online thanks to a rapidly growing number of web based portals. However, timely and systematic global analysis and coherent interpretation of such information, remains challenging. With the ASAP system (Anomaly hot Spots of Agricultural Production) we propose a two-step analysis to provide timely warning of production deficits in water-limited agricultural systems worldwide every month. The first step is
fully automated and aims at classifying each sub-national administrative unit (Gaul 1 level, i.e. first sub-national level) into a number of possible warning levels, ranging from “none” to level 4 and depending on the nature and number of anomalies taken into consideration. Warnings are triggered only during the crop growing season, as derived from a remote sensing based phenology. The classification takes into consideration the fraction of the agricultural area for each Gaul 1 unit that is affected by a severe anomaly of two rainfall-based indicators and one biophysical indicator, and the timing during the growing cycle at which the anomaly occurs. Maps and summary information are published on a web GIS. The second step involves the verification of the automatic warnings by agricultural analysts to identify the countries (national level) with potentially critical conditions that are marked as “hotspots”. In their evaluation, the analysts are
assisted by graphs and maps automatically generated in the previous step, agriculture and food security-tailored media analysis, and the automatic detection of active crop area using high resolution imagery (e.g. Landsat 8, Sentinel 1 and 2) processed in Google Earth Engine. Maps and statistics, accompanied by short narratives are then made available on the main webpage and can be used directly by food security analysts with no specific expertise in the use of geo-spatial data, or can contribute to global early warning platforms such as the
GEOGLAM Early Warning Crop Monitor, which perform monthly multi-institutional analysis of early warning information.