Stream water level forecasting in a built-up basin in Campinas, SP, using machine learning and aiming at structuring a flash flood early warning system

• Grant number: 20/00058-2
• Support type: Scholarships in Brazil - Master
• Effective date (Start): October 01, 2020
• Effective date (End): March 31, 2022
• Field of knowledge: Physical Sciences and Mathematics - Geosciences - Geology
• Principal researcher: Ana Elisa Silva de Abreu
• Grantee: Vinicius Araujo
• Home Institution: Instituto de Geociências (IG). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Associated research grant: 17/50343-2 - Institutional development plan in the area of digital transformation: advanced manufacturing and smart and sustainable cities (PDIp), AP.PDIP

Abstract

One of the most important tasks for flood management is to be able to accurately model the hydrologic behavior of a watershed. This involves understanding the stream level variation in time and incorporating this knowledge in flood early warning systems. One of the approaches that are growing in watershed hydrological modeling is the use of data driven models. Unlike conventional hydrological models, data driven models do not require the characterization of many physical parameters, because they are able to capture the intrinsic and complex relationships between the various parameters without the need to fully understand their controls and interaction mechanisms. In this context, this project aims at correlating stream water levels and rainfall in a sub-watershed of Ribeirão Anhumas in Campinas (SP), where flash floods involving loss of property and lives have been repeatedly reported. To achieve this goal already available data on stream water level and rainfall will be used, as well as data from sensors to be installed. This data will form the database for characterizing and modeling the stream water levels using Machine Learning algorithms. The results of this research are expected foster the use of ultrasonic sensors for urban streams water level monitoring and to contribute to a better understanding of how Machine Learning algorithms can be incorporated in flash flood early warning systems. (AU)