Developing Earth Observations Requirements for Global Agricultural Monitoring

Recognizing the dynamic nature of agricultural cultivation both within and between years and across the globe, the Group on Earth Observations (GEO) is developing an agricultural monitoring (GEO-GLAM) system with the goal of enhancing the availability and use of satellite and in situ Earth observations (EO) for the generation of timely and accurate information on national, regional, and global food supply. One of the key components of the GEO-GLAM system is the coordination of satellite observations so as to ensure sufficient and appropriate data volume and quality for agricultural monitoring. Therefore, it is essential that we develop EO requirements which articulate in a spatially explicit way where, when, how frequently, and at what spatial resolution satellite imagery must be acquired to meet the needs of a variety of agricultural monitoring applications. Accordingly, best-available cropland location information ('where?') in conjunction with ten years of MODIS surface reflectance data have been used to characterize the timing and duration of the agricultural growing season ('when?') in the form of agricultural growing season calendars (GSCs) for all major agricultural areas of the Earth. With respect to temporal resolution, we must first identify the frequency with which we require imagery inputs for monitoring applications such as crop condition, crop type, crop yield estimation, and planted and harvested area estimation. Members of the GEO Agriculture Monitoring Community of Practice - a group of international scientists - have combined their knowledge and expertise to articulate these general requirements. Second, we must determine how cloud cover impacts the ability of optical sensing systems to meet these established temporal resolution requirements. To this end, MODIS Terra (morning; 2000-2011) and Aqua (afternoon; 2002-2011) observations have been analyzed to derive probabilities of a cloud free clear view at different times of day throughout the agricultural growing season. In conjunction with information on field size distribution - which helps inform where finer resolution imagery are required - this information is being synthesized to generate a set of spatially explicit Earth observation requirements that are scalable to different satellite mission-specific swath widths, and provide concrete evidence for a multi-sensor imaging constellation approach to global agricultural monitoring.