The Doñana National Park, considered one of Europe’s most valuable wetlands, is expected to lose its marshland in 61 years, according to calculations from a major water-resource monitoring study carried out by the University of Seville. The study has developed an innovative algorithm, based on machine learning, capable of detecting the presence of surface water with high precision using images from the Sentinel-2 satellite.
The study includes a pessimistic scenario of 45 years and an optimistic one of 175 years, depending on future temperature and precipitation trends. These conditions will have a decisive impact on the marshland, an area of extraordinary importance as a stopover, breeding ground and wintering site for thousands of European and African birds.
The research, conducted within the project Application of Digital Image Processing for Water Resource Monitoring in line with the 2030 Agenda, has been led by Emilio Ramírez Juidias, researcher at the Department of Graphic Engineering, together with students Clara Isabel González López and Paula Romero Beltrán, both part of the High Intellectual Abilities programme.
Data collected since 2005 show that between that year and 2024, Doñana has lost around 15% of its average wet surface area (29,824 km²), water volume (11,680 hm³) and water depth (0.023 metres). However, it is striking that the bulk of that 15% (more than 13%) has been lost since 2010, "when temperatures began to rise and, above all, rainfall dropped sharply, compounded by the illegal extraction of water resources in the area," says Ramírez Juidias.
Key solutions
The main aim of the project is to provide an advanced technological tool for monitoring water conditions in vulnerable natural environments, in line with the United Nations Sustainable Development Goals, in particular SDG 6 (clean water and sanitation) and SDG 13 (climate action). In this context, remote sensing and digital processing of satellite images emerge as key tools in the face of growing water scarcity and the ecological deterioration caused by climate change and human activity.
The algorithm developed by the students, under the supervision of Professor Ramírez Juidias, uses machine learning techniques applied to optical data from the Sentinel-2 satellite, specifically the near-infrared and red bands. Using a mathematical formula calibrated specifically for wetland environments such as Doñana, the model reliably distinguishes between bodies of water and vegetation cover, generating updated maps that show both the presence and progressive loss of surface water. The results, validated in the field, have shown a significant correlation, confirming the model’s usefulness as a predictive tool.
Strategies to minimise impact
According to the professor, it is possible to minimise the loss of the marshland by implementing a number of strategies. The first, in his view, must be "drastic" and involves the permanent closure of illegal wells and effective control of water use. This requires intensified inspection efforts, the sealing of illegal wells, and the establishment of a real-time monitoring system to ensure compliance with groundwater extraction regulations.
The second measure involves shifting towards a more sustainable, less water‑intensive agricultural model, promoting crops with low water requirements and efficient irrigation techniques (such as drip irrigation), while gradually replacing intensive crops that are incompatible with the actual availability of water resources. The strategy should also include the recovery and restoration of degraded wetlands, acting on the areas most affected by desiccation through ecological restoration works, reconnecting them hydrologically with the aquifer and reintroducing native vegetation that helps retain water.
Other actions include the reuse of treated wastewater for agricultural and forestry purposes, promoting the use of reclaimed water from treatment plants to reduce pressure on the aquifer and ensure water availability for natural ecosystems; as well as the need to adapt to climate change through integrated water-management planning that incorporates future climate scenarios, including declining rainfall, rising temperatures and their impact on the Park's hydrological cycles.
Technology applicable to other environments
This technology not only identifies areas affected by drought or falling groundwater levels, but also supports decision-making for ecosystem conservation. As a scalable and automated approach, the algorithm can be applied to other natural environments facing similar challenges, thus contributing to more efficient and sustainable water management.
The success of this research highlights the importance of fostering young talent in high-impact scientific projects, especially in key areas such as environmental sustainability, digitalisation and artificial intelligence. The collaboration between the University of Seville's High Capabilities Programme and applied research offers an exemplary model of knowledge transfer and advanced training, geared towards the major global challenges of our time.
Ultimately, this breakthrough places the University of Seville and Professor Ramírez Juidias' research team at the forefront of technological innovation applied to the protection of water resources and the fight against climate change.