Habitat quality on the edge of anthropogenic pressures: Predicting the impact of land use changes in the Brazilian Upper Paraguay river Basin

Abstract

Landscape changes driven by anthropogenic activities often have negative impacts on ecosystems, jeopardizing the provision of vital services that sustain biotic elements and wildlife. Predicting the effects of such transformations is crucial for developing strategies to mitigate ecological damage. This study explores the interaction of land use changes and habitat quality estimates in the Upper Paraguay River Basin in Brazil, using the InVEST Habitat Quality model to estimate spatiotemporal landscape trends and evaluate the potential impact of land use policies for the future. Under a business-as-usual scenario by 2050, potential land use changes are simulated for shaping habitat quality, habitat degradation, and habitat rarity estimates as landscape proxies for biodiversity conservation. Results show an extensive expansion of pastures through deforestation, leading to progressive habitat degradation with a significant decay in the predicted and observed habitat quality. The habitat quality index estimates show a spatial decrease from 0.78 to 0.57 for the 1989–2050 period. The same trend is observed in protected areas for the period 2019–2050, with an increase of anthropogenic land use and a decrease of the habitat quality from 0.80 to 0.57 in conservation units, and from 0.75 to 0.53 in indigenous lands. For improved conservation outcomes, this work introduces new insights for shaping environmental actions that can be flagged as sustainable land management practices and ecological perspectives towards spatial pressures at different scales. Therefore, government institutions responsible for the protection of conservation units and indigenous communities can be informed about potential land use impacts on their territories. The findings suggest that incorporating additional scenario trends and climate-related variables could be a valuable direction for future studies to extend further the modelling approaches explored in this work.