GRASSEN Operational Group: Platform to support the management of Mediterranean pastures using near and remote sensors

The dehesa constitutes the most characteristic and representative agrosilvopastoral system in Europe, with 4.5 million hectares in the Iberian Peninsula, 60% of its surface area located in Andalusia and Extremadura. In addition to its surface area, it has a significant impact on the socioeconomic development of areas with limited production and a limited industrial structure, generating niches for specialized labor and high-value-added production. Andalusian Law 7/2010 of July 14, on the Dehesa, defines the dehesa as a farm consisting mainly of pastureland, subject to a land use and management system based primarily on livestock grazing, fruits, and branches, as well as other forestry, hunting, or agricultural uses.

For its part, a dehesa formation is considered to be any forest area occupied by a tree layer, with a cover fraction ranging between 5% and 75%, mainly of the genera Quercus (holm oaks, cork oaks, and gall oaks) and Olea (wild olive trees), which allows for the development of a herbaceous layer for livestock or game species. According to the Dehesa Master Plan, Andalusia has 940,622 hectares of dehesa, distributed across more than 14,000 properties. The regions of Los Pedroches and Alto Guadiato contain the largest continuous area of dehesa on the Iberian Peninsula. Managing herbaceous pastures in dehesas and properly planning their use are not simple tasks. In addition to the high spatial variability of pastures, associated with soil diversity, variable terrain geomorphology, the presence of trees, and diverse agricultural uses, there is also a high intra- and interannual variability linked to the specific characteristics of the Mediterranean climate. Rational management of pastures in the dehesa would benefit from the availability of quantitative and qualitative information on this resource and its spatial and temporal variation.

This would facilitate decision-making on crucial aspects of the pasture, such as grazing planning, the provision of supplementary feed to pastures, the organization of forage crops to provide quality feed for livestock at key times, pasture fertilization, and seed bank enrichment. Estimating pasture production and quality using traditional methods (field sampling and laboratory analysis) is expensive, making it unviable for many livestock farms and a service that is requested only occasionally.

Furthermore, information is not provided immediately, as there is often a delay between sampling and obtaining results, which limits its usefulness in short-term decision-making. The variety of existing and currently operational sensors, whether mounted on nearby and/or remote platforms, have the capacity to provide a large volume of data on the territory at low cost and with a spatial and temporal resolution appropriate to the needs of pasture farms. These multispectral, hyperspectral, and/or radar images implicitly contain information on the productive characteristics of pastures, and this information can be extracted using artificial intelligence techniques such as data mining. This project aims to obtain information on the productive characteristics of pastures in the pastureland from images obtained by the Sentinel satellites (1 and 2) of the EU Copernicus program, appropriately combined with images obtained throughout this project by multispectral sensors mounted on unmanned aerial platforms and field data.

Sentinel images were chosen as the basis for their open access and adequate spatial resolution (10 and 20 m pixel bands), spectral resolution (a multispectral sensor with 13 bands), and high temporal coverage (revisit frequency of about 5 days). The dynamic information generated this way on the production and quality of the pastures in the dehesa will be made available to end users, livestock farmers, owners, and dehesa managers, through the development of a web platform and a mobile application.

Finally, for a pilot pasture farm, we will illustrate how to use the dynamic information contained in the web platform to manage pastures and plan their use through grazing or mowing. The development of this tool could therefore contribute to more efficient and sustainable management of pastures in the pasture.