A look at the quality of reservoir waters from 786 km of altitude: Sentinel 2 images

Abstract

Assessing the water quality of lakes and reservoirs is a fundamental step in active environmental management programs. In addition to the classic procedures for in situ water samples collecting, complementary approaches are also interesting. This work aims to study comparatively numerous São Paulo reservoirs, using remote sensing techniques, using images from the Sentinel 2 satellite (A and B) (S2), worked on the SNAP software, both from the European Space Agency (ESA). Specific estimation equations will be tested and some already available as SNAP products, and thematic maps will be generated. As case studies will be used the scene T23KLQ, the reservoirs of the Cantareira System (CS), at present including the Igaratá reservoir. The choice of these reservoirs is due to their importance for water public supply. There is also a public database, with information about chlorophyll a concentration (Chla), turbidity, depth of the Secchi disk (SD) and the cyanobacteria cells number (Cyano), presented by CETESB (Companhia Ambiental do Estado de São Paulo), in the Inland Water Quality Report, for the fiscal years 2015 to 2019. The S2 images with L1C level will be downloaded with dates no more than three days apart from the date of collection and based on the estimates of Chla and DS, the trophic state index (EIT) will be calculated, with the classic Carlson equations. In 2021 and 2022 there will also be in situ sampling in these reservoirs, in at least three zones per reservoir and in three different stations per zone, in profiles (temperature, dissolved oxygen, pH and electrical conductivity - multiparameter probe; the nitrogen series and phosphorus, Chla and metals concentrations; phytoplankton - qualitative and quantitative aspects), used to test and validate the algorithms. It is expected to observe a pattern, with the area of the dam more lentic and less eutrophic, when compared to the upper part, and with the in-situ sampling to obtain a diagnosis about the ecological quality of the studied reservoirs. During this project, courses, lectures and meetings will be offered to SABESP employees and partner companies, aiming to train them in remote sensing techniques, in assessing the quality of the water in the reservoirs. It is also expected to produce a manuscript in the form of a support protocol, describing in detail the steps, from fieldwork to image processing and the production of thematic maps, using SNAP with S2 images. Preliminary tests presented already demonstrate the feasibility in the estimates of Chla, SD and Cyano of the CS reservoirs by means of S2 images. (AU)