Incorporating implicit information to disentangle the impacts of hydropower dams and climate change on basin-scale fish habitat distribution

The loss of freshwater fish habitats, exacerbated by climate change and dam constructions, poses a critical environmental concern. The upper Yangtze River basin, noted for its abundant fish fauna and concentrated dam development, serves as a crucial locale for investigating the impacts of climate shifts and dam construction. This study aims to disentangle the impacts of hydroelectric dams and climate change on fish habitat distribution by analyzing species presence data across different periods. Species distribution models were constructed using Maxent for Coreius guichenoti (a warm-water endangered fish) and Schizopygopsis malacanthus (a cold-water endangered fish). The model accuracy was assessed using the area under the curve of the receiver operating characteristic. Habitat distribution modeling and prediction for the pre-dam period (1970–2000) and post-dam period (2001–2020), as well as future climate change under two shared socioeconomic pathways scenarios, were conducted. The impacts of climate change and dam construction on the habitat suitability of two fish species were quantified. The results revealed dam construction predominantly diminished habitat suitability and range, with high-suitability habitats in the post-dam period decreasing by 56.3% (720.18 km) and 67.0% (1665.52 km) for the two fishes, respectively. Climate change would enhance the habitat suitability of Coreius guichenoti, while it would decrease the habitat suitability of Schizopygopsis malacanthus. The impact of dam construction is greater that of climate change for them. This study underscores the profound impacts of dam construction on fish habitats, particularly for cold-water species, and highlights the critical need for habitat restoration in sustainable hydropower development. Our method of disentangling these factors also provides a new approach to evaluating environmental impacts in large river basins.

Methods
For the input data, they have been masked to the river network of the Upper Yangtze Basin, with a grid resolution of 30 arcseconds. These operations were performed using simple R code, which does not need to be provided.vided data.