An Earth Observations based Environmental Performance Tool

The tool measures the environmental performance of the agricultural sector of countries or regions in a ‘multilayer’ environmental scoreboard. It is not a tool designed to evaluate the effects of policy on specific individual environmental indicators, and the value of the tool lies in the unique underlying database. The DIONE environmental metrics document makes a preliminary presentation of all possible data sources, and the evaluator can examine their availability. All information will be adaptable in its spatial and temporal dimensions, stretching from the local/regional to the national and from contemporary to the latest available historical data. 

Thus, the evaluator can make use of the database underlying the tool in three ways. First, by extracting data that integrate multiple monitoring sources, the evaluator may approximate the value of indicators. For example, the soil organic matter indicator will be compiled by Copernicus Sentinel-1 and Sentinel-2 data, LUCAS and data from spectral handheld sensors that will provide space-borne and in situ spectral and laboratory measurements. Second, if the layers are combined with the compliance monitoring tool at the farm or parcel levels, the evaluator may pursue a quantitative assessment of the policy’s effects. Third, by utilising the ‘environmental scoreboard’, the melting pot for nine environmental indicators. The environmental scoreboard and its nine layers at the national level can support various geospatial analyses or data triangulation. Finally, the tool provides an excellent visualisation device that assists those carrying out monitoring and evaluation operations to communicate their findings to decision-makers and other stakeholders. 

The tool is not automatically transferable to other regions and Member States for two reasons. First, data availability varies from country to country, depending on ancillary data. For example, certain Member States have already detailed soil maps from their extensive soil surveys, and it is easy to connect EO and in situ data. Second, all the machine learning and AI algorithms require training with the new data to recognise certain environmental variables. So, depending on data availability, the tool may require adaptation of algorithms and re-training.