Optimizing surface water quality parameters in monitoring networks in a developing sub-tropical region with high anthropogenic pressure (São Paulo State Brazil).

Efficient water quality monitoring is a central aspect of water resources management, especially in developing countries, where water quality is under high anthropogenic pressure and resources for monitoring are usually limited. Here, we evaluated an alternative to optimize water quality parameters (WQPs) in the water quality monitoring network (WQMN) of the most populous state in Brazil (São Paulo State). We focused on the monitoring goal of identifying water quality temporal trends, selecting WQPs with high statistical explanatory power and those that were particularly sensitive to natural and anthropogenic perturbations. We considered 12 initial WQPs (dissolved copper, total zinc, total lead, total chromium, total mercury, total nickel, total cadmium, total iron, total manganese, total aluminum, total copper, and surfactant) with data from 2004 to 2018 for 56 monitoring sites distributed across four major watersheds with contrasting land uses in the state. We performed principal component analysis, followed by objective criteria to refine WQPs recommendation for the WQMN. Our results indicated the opportunity of reducing at least one parameter from the initial set of WQPs in all watersheds. Total iron, total manganese, and total aluminum were the most relevant initial WQPs, since their maintenance in monitoring were indicated in all the analyzed cases. Natural watershed conditions (e.g., geomorphology and water geochemistry) potentially governed their concentrations in surface water. On the other hand, total mercury, total chromium, and dissolved copper had the maintenance indicated in only one watershed, especially due to concentrations consistently below the respective limits of quantification (LoQs). Future investigations can complement our recommendations for these parameters, since changes in LoQs could throw another light on water quality spatial and temporal variations and the need for reference areas for assessing baseline conditions can also be relevant. Moreover, we argue that depending on the monitoring goals of the WQMN, additional sampling of biota and sediments could be useful as many of the studied WQPs' bioconcentrate. Our results illustrated an alternative approach towards adaptive monitoring in São Paulo state in accordance with the intended monitoring goal (i.e., water quality temporal trends), converging with the more flexible monitoring adopted in well-structured networks worldwide. While we did not cover other monitoring goals in our study (as the control of illegal discharge of effluents or industrial spills, for example), we expect our methodology can contribute to establishing technical guidelines for reviewing the existing WQMNs in Brazil and other developing countries with similar challenges.