In the Food Security thematic area, BETTER investigates the utilization of large EO and non-EO biophysical and socio-economic heterogeneous datasets to support and optimize the different operational needs in the management of a humanitarian crisis. The datasets are mostly related to crop monitoring and meteorology and can be derived from different sources, such as e.g. satellite data. By processing this data, BETTER helps the UN’s World Food Program - which is the sponsor of the challenge - to improve its preparedness in case of humanitarian crises.
In the second cycle of the project, 3 new challenges have been identified.
The 4th Food Security challenge (1st one in the 2nd cycle) monitors Land Surface Temperature (LST) for Natural Hazards and Food Security Early Warnings. Among other things, this involves the investigation of four data pipelines in this challenge: Time series of Smoothed Gap-Filled LST, Aggregated LST, Long Term Averages of the LST, and anomalies of it. In the first phase, the LST data will be used in preliminary studies that will analyze the information potential of this variable. One question could be: “can LST identify better (or earlier) evolving drought events compared with vegetation indices?”
The 5th Food Security challenge monitors snow cover patterns and its seasonality in Central Asia. Detecting major deficits in snow extent can provide timely indications of potential shortfalls of crop production and thereby it can help to find indicators that lead to food security early warnings in WFP’s operations. As a technical result of this challenge, the outputs of this pipeline articulate with those that produce vegetation index parameters, so that impacts of snow cover variations can be validated against the behavior of vegetation related information over crop-producing regions.
The 6th Food Security challenge helps to overcome a significant problem in the usage of high-resolution optical data, which is that, for many areas of the globe, cloudiness, haze, and gaps arising from revisit time limitations reduce the usefulness and information content of these satellite- data streams. A smoothing filter developed by WFP (Whittaker filter) is able to minimize most if not all of these problems. This significantly improves the usability of Sentinel-2 data and provides much more accurate land cover classifications, for example for crop type mapping applications that WFP is engaged in.
As the pipelines developed to address the above three challenges become more mature, results will also be added to the above three linked individual challenge pages.