The CLEPSYDRA Decision Support System (DSS) is an advanced, AI-driven digital platform developed to support groundwater monitoring, forecasting, and sustainable management in water-scarce and environmentally sensitive regions of the Mediterranean. It has been designed within the CLEPSYDRA project to address the increasing pressures on aquifer systems caused by climate change, intensive agriculture, and diffuse pollution, by transforming complex environmental data into actionable knowledge for decision-makers.
The DSS integrates heterogeneous, multi-source datasets – including measurements, meteorological variables (e.g., precipitation and temperature), and other environmental indicators – into a unified, interoperable framework. Through standardized data ingestion and preprocessing pipelines, the system ensures data consistency and reliability, enabling robust analysis across different geographical contexts and monitoring infrastructure.
At its core, the DSS leverages advanced Artificial Intelligence (AI) methodologies, with a particular focus on Machine Learning and deep learning approaches. A suite of predictive models – such as Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) networks, Kolmogorov-Arnold Networks (KAN) and hybrid architecture- is embedded within the platform to capture the nonlinear and nonstationary dynamics of groundwater systems. These models are capable of learning complex temporal dependencies and generating accurate forecasts of groundwater levels, supporting early warning and proactive resources management.
At the current stage of development, the DSS is primarily focused on groundwater level forecasting, which represents a key indicator for assessing aquifer status and sustainability. The prediction of groundwater quality parameters, such as nitrate concentrations, is not yet operational due to limited availability of consistent and long-term datasets across the pilot regions. Nevertheless, the system has been explicitly designed with extensibility in mind, allowing for the future integration of water quality forecasting modules once sufficient data become available. This capability represents a strategic direction for enhancing the DSS toward comprehensive groundwater quality management and pollution risk assessment.
From a technical perspective, the DSS is implemented as a modular and scalable multi-tier architecture, combining a web-based user interfacen, a backend API layer, a dedicated machine learning server, and a centralized database. This design enables efficient data management, secure multi-user access, and the execution of computationally intensive forecasting tasks through asynchronous processing. The platform supports multiple data ingestion modes – including manual uploads, batch processing, and API-based real-time data streams – ensuring flexibility in diverse operational contexts.
A key strength of the DSS lies in its user-oriented design. Through interactive dashboards and visualization tools, users can explore monitoring data, configure predictive models, and analyze forecasting results in an intuitive environment. The system also supports scenario analysis functionalities, enabling stakeholders to simulate the potential impacts of different management strategies, such as variations in groundwater abstraction, irrigation practices, or environmental policies. This feature is particularly relevant for supporting evidence-based decision-making under uncertainty.
The DSS adopts a structured governance framework based on role-based access control, ensuring data security, traceability, and accountability across different user levels. Its architecture has been conceived to guarantee interoperability and transferability, allowing the platform to be adapted to different regions, aquifer types, and institutional settings beyond the CLEPSYDRA pilot sites.
Overall, the CLEPSYDRA DSS represents a significant advancement in the digitalization of groundwater management. By integrating advanced AI-based forecasting, real-time data analysis, and user-friendly decision-support tools, the platform facilitates a transition from reactive to proactive water management strategies. In doing so, it contributes to enhancing the resilience, sustainability, and long-term protection of groundwater resources in the Mediterranean region and beyond.
