Artificial Intelligence Technology Solutions Based on Complex Adaptative Systems in Integrated Water Resources Management

• Grant number: 22/14849-7
• Support Opportunities: Scholarships in Brazil - Doctorate
• Start date: March 01, 2023
• End date: January 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Geosciences - Geology
• Principal Investigator: Didier Gastmans
• Grantee: Carolina Stager Quaggio
• Host Institution: Centro de Estudos Ambientais (CEA). Universidade Estadual Paulista (UNESP). Campus de Rio Claro. Rio Claro , SP, Brazil
• Associated scholarship(s): 23/13079-6 - Artificial intelligence technology solutions based on complex adaptative systems in integrated water resources management, BE.EP.DR

Abstract

The multidisciplinary of the components that determine the water flow intensity and paths between the atmospheric, superficial and underground reservoirs through the Hydrological Cycle, such as the climate, the hydrology and the geology, are conditioned by the mutual adaptation to climate change and land use conversion, arising from anthropogenic actions, and has turned into complex the rational water use. Integrated Water Resources Management (IWRM) is founded in the joint assessment of surface and groundwater, in search of sustainable solutions to meet the demands of consumptive uses, and constitutes a Complex Adaptive System (CAS), since the connection between the reservoirs promotes the constant rebalancing of the components. Artificial Intelligence (IA) techniques are an emerging approach in the construction of CAS, due to the analyses of large volume of data and of correlations between non-linear functions and the description of phenomena in temporal and spatial scale. This project aims to simulate the adaptation of a Water Resource Management Unit (WRMU) from the state of São Paulo facing scenarios of increasing water demand, associated with economic development and climate change, through AI techniques. The balance between reservoirs will be estimated by simple water balance equations, in which will be based the projection of scenarios and the rebalancing in face of CAS adaptations. It is intended to investigate supervised machine learning models of classification and regression. We hope to create a tool that will assist IWRM, improving the knowledge of the water resources panorama in the basin and contributing to the optimization of the management of consumptive uses and of water functions and ecosystem services.