We built this dataset to assess the impacts of hydropower plants on the distribution of an endemic and imperiled freshwater turtle with very unique ecological requirements, the Williams' side-necked turtle (Phrynops williamsi). To prevent and mitigate impacts, we prioritized sites for species conservation by classifying planned HPP locations according to their predicted adverse effects on species distribution. The dataset has two files: i) species occurrence records and ii) hydropower plant data. The first dataset was fully built by the authors and the second was modified from the Brazilian Electricity Regulatory Agency (ANEEL) georeferenced data system.

Species occurrence records

We built a presence-only database based on our own fieldwork records, data from six scientific collections (i.e., Museu de Ciências Naturais da Fundação Zoobotânica do Rio Grande do Sul, Museu de Ciências da Pontifícia Universidade Católica do Rio Grande do Sul, Museu de História Natural Capão da Imbuia, Colección de la Universidad de la República (Uruguay), Museo Nacional de Historia Natural del Paraguay, University of Michigan Muzeum of Zoology, and Museo Miguel Lillo de Ciencias Naturales (Argentina)), and literature records with pictures or voucher specimens that allowed validation of species taxonomic identification. The identification of the specimens was checked because P. williamsi occurs in sympatry with P. hilarii in Rio Grande do Sul and P. geoffroanus in Paraná, which refutes the observation that the species does not occur in sympatry with other congeners, as mentioned by Rhodin and Mittermeier (1983). Part of the species-occurrence dataset was published by Kunz et al. (2018). We obtained 96 species records that were cleaned, removing (i) 44 records with imprecise coordinates and (ii) 17 locations where HPPs were built after the species record, retaining 35 geographical coordinates for the ecological niche modeling (Table S1). To avoid sampling bias, we filtered the species occurrence data set using the environmental space (following Varela et al., 2014), which revealed that all species records were environmentally distinct. Once our study was based on a species records dataset and did not involve capture of animals, permissions for field work or ethical approval from an animal ethics committee were not necessary.

Hydroelectric power plant data

We obtained the data of hydroelectric projects from the SIGEL (Georeferenced Information System for the Electric Sector) database of ANEEL (the Brazilian National Electric Energy Agency). ANEEL regulates and supervises the production, transmission, distribution, and sale of electric energy in Brazil (https://sigel.aneel.gov.br/). By the SIGEL system, ANEEL makes available georeferenced information (i.e., the shapefile of the dam point) about the operation and expansion of the Brazilian electric system. In Brazil, hydroelectric power plants are classified according to the capacity of electricity production or the extent of the reservoirs (ANA, 2005): micro hydropower (MH) produces up to 1kW, small hydropower (SH) produces from 1 kW to 30 kW or have reservoirs up to 3 km² (but exceptionally some up to 13 km² are acceptable; Damasceno et al. 2014), and large hydropower (LH) generates above 30 kW or more than it when the project does not match the SH specifications. Using the SIGEL, we obtained 687 HPP records within our modeling background delimitation (see Study Area section) and reclassified the 16 ANEEL licensing classes into 3 categories: i) in operation (252 MHs,  121 SHs, and 33 LHs); ii) under construction (5 MHs and 43 SHs); and iii) planned (66 MHs, 154 SHs, and 13 LHs).