Chlorophyll-a Concentration in the Salto Grande Reservoir (Americana, Brazil) Estimated by Satellite Images

Abstract

Abstract Humankind has a deep relationship with water, which is essential for life, since it is used to meet different human needs; however, human interactions and activities have degraded the quality of water bodies. In this context, monitoring water bodies becomes an essential tool to understand their spatial and temporal variability, as well as to help in decision-making and water resource management. Traditional monitoring, based on sample collection and laboratory analysis, can be considered a costly, one-off process that fails to present the characteristics of the entire water body. Remote sensing techniques can be used complementarily to traditional monitoring, allowing to observe the entire water body while presenting lower costs and short execution time. This study used remote sensing techniques to evaluate the trophic state of the Salto Grande reservoir using Chlorophyll-a (Chl-a) concentration as an analysis parameter. Located in the municipality of Americana, São Paulo, Brazil, the Salto Grande reservoir is situated on the Atibaia River, near its mouth, amidst a highly industrialized region with expressive monoculture. This reservoir has a history of degradation over time. Chl-a was estimated based on images from the MSI sensor present on the Sentinel-2 satellite, which underwent atmospheric correction by Sen2Cor software. Four algorithms using spectral band ratios were tested, with both linear and nonlinear adjustments. The values obtained were compared with data from analyses performed by CETESB, with close dates, at the study site and available to the public. The adjusted algorithms presented a correlation above 0.8, while the best adjustments for each algorithm showed a correlation greater than 0.9 and an error rate below 10% and 8 µg/L. Mishra and Mishra’s (2012) NDCI algorithm had the best applicability as it returned only positive values, in contrast to some algorithms, which returned negative Chl-a values. With the resulting values from the selected algorithm, the modified Carlson Trophic State Index was applied, and a high eutrophication index was observed in the reservoir.