Integration of Orbital Sensors and Machine Learning for Monitoring Public Water Supply Reservoirs: Chlorophyll-a Estimation in the Taiaçupeba Reservoir (SP).

• Grant number: 25/07449-0
• Support Opportunities: Scholarships abroad - Research Internship - Master's degree
• Start date: September 01, 2025
• End date: February 28, 2026
• Field of knowledge: Engineering - Sanitary Engineering - Water Resources
• Principal Investigator: Marcelo Luiz Martins Pompêo
• Grantee: Bianca Carolina Sant'Ana Silva
• Supervisor: Paolo Ettore Gamba
• Host Institution: Instituto de Biociências (IB). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: Università degli Studi di Pavia, Italy
• Associated to the scholarship: 24/02468-4 - Use of images with multispectral resolution as a tool for monitoring and managing the quality of reservoirs for public supply: with emphasis on chlorophyll a and Secchi disk, BP.MS

Abstract

The quality of inland waters has been significantly impacted by anthropogenic activities, demanding more effective monitoring methods. Traditional chlorophyll-a (Chl-a) monitoring, an important indicator of phytoplankton biomass, eutrophication, and aquatic ecosystem health, has limitations in terms of spatial scale, frequency, and associated costs. In this context, remote sensing combined with machine learning techniques emerges as a promising alternative for broader and more automated water quality assessments. This study proposes the development of a machine learning model to estimate Chl-a concentration in the Taiaçupeba Reservoir (SP), using reflectance data from orbital sensors (OLI/Landsat-8, MSI/Sentinel-2, and OLCI/Sentinel-3), corrected with tools such as ACOLITE, Sen2Cor, C2RCC, and L8PAR. The model will be based on machine learning algorithms, specifically Cubist and Extreme Gradient Boosting (XGB), and validated using in situ data provided by SABESP from 2016 to 2023. Model performance will be evaluated using statistical metrics such as R², MAE, RMSE, and MAPE, aiming to provide accurate and automated estimates of Chl-a concentration to support effective large-scale water quality monitoring. (AU)