Multi-scale assessment of forest cover in an agricultural landscape of Southeastern Brazil: Implications for management and conservation of stream habitat and water quality

Abstract

Forest cover has important functions for streams. Consequently, deforestation and forest degradation due to agricultural activities tend to have negative impacts on stream ecosystems. We related forest cover to stream variables, expecting to find better habitat and water quality conditions in catchments with better forest cover conditions in order to evaluate forest cover as indicator of stream health in agricultural landscapes. We sampled stream variables and quantified forest cover and physical variables in 60 small agricultural catchments in Southeast Brazil. We used redundancy and regression analysis to relate the landscape predictors to the channel responses. Percent forest cover had low to intermediate values in the spatial scales evaluated. Forest cover was fragmented and mostly located in riparian and steep slope areas. Redundancy analysis showed little influence of forest cover on the response variables, which were more influenced by catchment physical variables. Regression analysis showed that forest cover in the reach and forests located closer to the sampled reach are positively related to wood, habitat diversity, and dissolved oxygen, and negatively related to channel depth, volume, and temperature. We also found that forest cover fragmentation is negatively related to pH, potassium, water acidity, and temperature. Although many of these relationships were fairly weak, it appears that naturally regenerated forest cover is at least moderately effective in protecting streams in agricultural landscapes in the region.