Researchers from the Valencian Research Institute for artificial intelligence (VRAIN), the Institute of Water and Environmental Engineering (IIAMA) and ValgrAI have developed a hydrological digital twin designed to anticipate flooding in ravines such as the Poyo during extreme rainfall events.
The tool was demonstrated using rainfall data from the October 2024 dana in Valencia, which caused 234 deaths and extensive damage. The simulation focused on the Poyo basin and also examined how the same storm could have affected areas around the Carraixet ravine.
The system, known as PGTEC, brings together hydrometeorological data from multiple official bodies in a shared, interoperable environment. One of its applications is the creation of digital twins that can forecast the rise of waterways and display on a map how floodwater could move across the surrounding territory.
PGTEC combines the TETIS hydrological model with the HIDRIS tool. TETIS uses recorded rainfall and weather forecasts from different sources to estimate how much water could flow through each section of a channel and when flooding might occur. HIDRIS models the potential spread of water across the terrain, including its expected depth and speed, as well as the buildings and critical infrastructure that could be affected.
The platform presents the information visually rather than requiring emergency managers to compare separate maps and forecasts. TETIS highlights the most vulnerable sections, supports forecasts for the following 48 hours and can run a simulation in roughly one or two minutes. It can also modify a storm’s position or intensity and reproduce historical events to examine how their consequences might change.
PGTEC currently includes digital twins for the basins of l’Horta Sud and the Carraixet ravine. According to Sergio Martí, a VRAIN researcher at the UPV, the platform uses a common semantic framework so that data from different sources have the same meaning. Each organisation retains control over its information and decides who can access it, for what purpose and under which conditions through an agreed data-use contract.
The project also covers longer-term climate risks. PGTEC incorporates WATER4CAST, an artificial intelligence tool intended to improve weather forecasts and help anticipate droughts, water needs and conditions conducive to forest fires.
IIAMA researcher Félix Francés said the project addresses both rapid-onset emergencies, such as flooding that requires decisions within hours, and droughts, which depend on long and consistent historical records. Both rely on the same interoperable data environment, while participating organisations retain control of their own information.
The project was funded by Spain’s Ministry for Digital Transformation and the Civil Service through initiatives promoted by the Directorate-General for Data and the Data Economy Reference Centre. Its presentation concluded with a debate on proactive management of climate crises involving representatives from Aemet, the Júcar Hydrographic Confederation, AVAMET, Valencia’s fire service and the Generalitat Valenciana.
Sources: La Vanguardia.