Drivers of water quality trends in tributaries of the Itaipu Reservoir, Brazil

Código

I-EBHE0157

Autores

FELIPE MARTINS PINEROLI, Vinícius Bogo Portal Chagas, Daniel Bartiko, Diego Alberto Tavares, Jussara Elias de Souza, ROSELI FREDERIGI BENASSI, Pedro Luiz Borges Chaffe

Tema

WG 1.11: Water Quality Under Global Changes

Resumo

The concentration of water quality parameters in rivers is subject to spatial and temporal variations, resulting from interactions between rivers and their contributing catchments. Comprehending the key factors affecting riverine water quality change is essential to promote water security for populations depending on these rivers, especially in ungauged basins in developing countries. Here, we analyze the drivers and trends of riverine water quality in the tributaries of the Itaipu Hydroeletric Power Plant (IHPP), the main power plant in Brazil. The dataset provided by IHPP comprises water quality data of seven parameters collected across 21 monitoring sites from 1985 to 2023. The respective catchments are delimited, spanning from 44 km² to 2200 km², most of them covered by intensive agriculture. We compute trends with the Theil-Sen slope and Mann-Kendall test for total phosphorus, total Kjeldahl nitrogen, nitrate, ammonia, biological oxygen demand, dissolved oxygen and pH. We detect the drivers of water quality variations by contrasting it with hydrological signatures (e.g., streamflow and precipitation regimes), climate seasonality, atmospheric water balance (when lacking streamflow measurements), land cover, and other catchment characteristics described in CAMELS-BR dataset. Our results show large spatial variability of water quality trends, higher than those suggested by climatic trends. It means that the drivers are linked with other variables such as hydrological signatures, physiography and land use. These findings elucidate a few questions on nutrient dynamics and water quality degradation in intensely farmed lands in the tropics and subtropics. Combining water quality and hydrological signatures provides valuable insights for water security strategies in data scarce regions, such as water quality prediction models that use advanced methodologies like artificial intelligence or machine learning techniques.