Abstract
This study evaluates the consistency and validity of the data generated by the citizen science program Observando os Rios (OoR), coordinated by the SOS Mata Atlântica Foundation and active in 17 Brazilian states. Water quality index (WQI) results produced by OoR were compared with those of São Paulo’s environmental agency, Companhia Ambiental do Estado de São Paulo (CETESB), across three hydrographic mesobasins—Alto Tietê, Sorocaba/Médio Tietê, and Piracicaba–Capivari–Jundiaí (PCJ)—from 2002 to 2023. Methodological conversions were applied to harmonize OoR classifications with CETESB standards. Among them, the weighted model (OoR_CET_8*), which adjusts both parameter inclusion and weighting, showed the closest quantitative alignment with CETESB according to Welch’s t-tests. Exact agreement between OoR and CETESB classifications varied—90% in Alto Tietê, 10% in Sorocaba/Médio Tietê, and 65% in PCJ—while ±1-category agreement exceeded 95% in all regions, indicating strong qualitative coherence. Kappa coefficients confirmed substantial agreement in Alto Tietê (0.74), moderate agreement in Sorocaba/Médio Tietê (0.51), and moderate to substantial agreement in PCJ (0.58). Along the Tietê River, finer-scale analyses revealed marked spatial variability, with an average Kappa of 0.43 and reduced convergence in highly urbanized stretches such as Guarulhos and São Paulo, where OoR frequently assigned as “Regular” while CETESB classified the same sites as “Poor” or “Terrible.” This discrepancy reflects the broader amplitude of the Regular category in the OoR index (~32% of the scale), which can smooth short-term deterioration. Despite limitations associated with biological indicators and observer-dependent variability, OoR demonstrated robustness, broad spatial coverage, and strong socio-environmental relevance. With continued methodological refinement and deeper institutional integration, the program constitutes a scalable and replicable model for citizen-science-based water quality monitoring, particularly valuable in regions with limited governmental oversight.