Modelagem estatística para avaliação do impacto do sistema de captação em tempo seco (CTS) na qualidade da água da Lagoa Rodrigo de Freitas, Rio de Janeiro, RJ

Abstract

The lack of sanitation in Brazil worsens diseases, social inequalities, and water pollution. In this context, the research evaluates, through statistical modelling applied to real and official water quality data, the performance of the Dry Weather Sewage Intake (DWSI) system at the Rodrigo de Freitas Lagoon (Rio de Janeiro, Brazil) – an environment with low water renewal, tendency to eutrophication, historical fish kill events, and intense urban occupation in its surroundings. The objective is to quantify the contribution of the Galeria de Cintura DWSI, implemented in 2001, to the reduction of Total Phosphorus (TP) concentrations and loads, an indicator of sewage contamination. To support decision-making, simple statistical models, such as linear regression, were identified as effective tools in the face of limited data. Accordingly, historical series of Total Phosphorus concentrations (1997-2001 and 2017-2024) were analysed by applying logarithmic transformation, outlier removal, and sequential decomposition through simple linear regressions of the influence functions of temporal trend, rainfall peaks on the residuals and dry-weather periods on the second residuals, integrated into a global multiple linear regression model – with robust statistical validation using the Huber-White standard error method to address heteroskedasticity. Combined with a dry-weather mass balance based on daily predictions, this model enabled the estimation of the parameter load and rates, and the evaluation of scenarios without and with DWSI. The results showed that the DWSI reduced by approximately 90% the daily input of Total Phosphorus under dry-weather conditions, effectively mitigating eutrophication and improving water quality, which, in practice, fulfils a structural role as a complementary component of the separate sanitary sewer system