Author(s): Philippe Gourbesville; Hezouwe Amaou Talle; Masoud Ghulami

Linked Author(s): Philippe Gourbesville

Keywords: GPU; Decision Support System; AquaVar; Deterministic modeling

Abstract: Challenges related to water issues are becoming more and more complex for local communities who have to answer to the basic needs of the populations, provide sufficient services for economical activities and preserve the resources in a long term perspective. The growing competition among uses requests to harmonize the strategies and to conciliate the various field operations. Within this perspective, the use of Decision Support Systems comes a priority for many cities and metropolis areas. In order to answer to this challenge, Nice Côte d'Azur Metropolis has decided to develop a DSS based on deterministic modeling tools that can represent the full hydrological behavior of the Var catchment (about 3000 km2). The development of the AquaVar system is based on the coupling of 3 models covering hydrology, free surface flows and underground resources. The integrative solution is based on Mike SHE, Mike 21 FM and Feflow. The 3 models are operated in realtime in order to produce, every hour, a forecast for the next 3 days and for all the variables of the different,nt models. The computing architecture is using the GPU resources in order to be able to deliver the results in the expected time frame. The access to the tool is achieved through a Web service that is allowing to visualize all results and to access to field data in real time. The presentation gives an overview of the selected architecture and of the performances of the selected GPU solution. The approach has demonstrated the efficiency of the high performance solution and the interest to invest within deterministic approaches that were, up to now, not so frequently used within the DSS dedicated to the water management. The AquaVar architecture is generic and allows to address topics like floods forecasting and longterm underground resources exploitation. The modeling system allows also the possibility to run climate change scenarios and to identify trends for various hydrological processes. The added value of the GPU approach is clear and represent a major axis for the development of DSS in the water management field.

DOI: https://doi.org/10.3850/IAHR-39WC2521711920221397

Year: 2022