Advancing near-real-time water-quality monitoring in Brazil through remote sensing: the MAPAQUALI and MAPAQUALI-IA platforms

Monitoring water quality in inland and coastal waters is essential for understanding biogeochemical cycles and the impacts of anthropogenic pressures such as land use and land cover change, mining, deforestation, dam construction, and climate change. Traditional field surveys, conducted bimonthly or quarterly at limited sampling stations, are valuable but do not offer the spatial and temporal coverage necessary to fully support public policies for sustainable aquatic system management. In this context, remote sensing plays a key role in enabling large-scale water quality monitoring across extensive and remote regions, such as Brazilian Amazon. Despite recent advances, there remains a lack of accessible platforms that deliver validated remote sensing products and algorithms to researchers, stakeholders, and decision-makers. To address this gap, the Instrumentational Laboratory for Aquatic Ecosystems (LabISA) at the Brazilian National Institute for Space Research (INPE) is developing MAPAQUALI, a semi-automatic cloud-based platform designed to generate and distribute water quality products at high spatial and temporal resolution for aquatic ecosystems in Brazil. MAPAQUALI integrates a set of semi-analytical and machine-learning algorithms developed and validated by INPE’s research team. These algorithms retrieve key water quality parameters, including chlorophyll-a, phycocyanin, Secchi disk depth, and total suspended solids, using observations from ESA and NASA multispectral sensors (Sentinel-2 MSI, Sentinel-3 OLCI, and Landsat-8/9 OLI) with a focus on specific reservoirs and lakes in Brazil. In addition to the MAPAQUALI, a new project named MAPAQUALI-IA is leveraging large-scale mapping of water quality in Brazil using artificial intelligence (i.e., machine learning and deep learning methods) to provide these water quality parameters using a single large-scale algorithm. The project will develop algorithms with the help of newly released open datasets, such as BRAZA and GLORIA. The MAPAQUALI/MAPAQUALI-IA processing pipeline incorporates advanced aquatic atmospheric correction techniques, specifically ACOLITE and 6SV, as well as corrections for glint and adjacency effects. A STAC-compliant data cube environment (Brazil Data Cube platform) allows to generate and store data enabling rapid access, visualization, and analysis. This publication introduces the current MAPAQUALI/MAPAQUALI-IA prototype, a modular and continuous monitoring system implemented for representative Brazilian aquatic environments, including Amazonian lakes, eutrophic cascade reservoir system, and coastal waters. Future developments will expand sensor compatibility, include new water-quality algorithms, and extend coverage to additional inland and coastal environments. Ultimately, MAPAQUALI aims to bridge the gap between scientific data and operational application, supporting more informed decision-making to improve aquatic ecosystem conservation and management in Brazil.