Biases in Amphibian Sampling in the Amazon: Using Infrastructure and Accessibility Data to Identify Sampling Gaps

Biogeographic knowledge of Amazonian amphibians is limited by spatial and temporal coverage, resulting in biases that affect the understanding of their diversity patterns. This study analyzed a database of 951 species based on 213,072 georeferenced occurrence records distributed across 24,319 locations in the Amazon. The objective was to identify sampling biases related to infrastructure and accessibility predictors. The overall results indicate that rivers are the primary drivers of amphibian sampling, while roads had a limited influence, reflecting the historical reliance of the region on river transport. Regarding infrastructure, both cities and hydroelectric plants had a moderate effect on sampling, whereas transmission lines had a negligible effect. However, with the expansion of hydroelectric projects from the mid-1970s onwards, intensifying after 2008 with the Brazilian government's Growth Acceleration Plan (PAC), the high volume of records obtained from these ventures distorted the sampling pattern, overestimating rivers and hydroelectric plants while underestimating highways as a source of sampling bias. We conclude, therefore, that amphibian sampling in the Amazon exhibits significant geographic and temporal bias due to unevenly distributed research efforts, which are largely constrained by logistical challenges and inadequate infrastructure. To overcome these challenges, it is necessary to promote collaboration between researchers and decision-makers, invest in research infrastructure, and improve data dissemination. Additionally, we emphasize the importance of conducting a rigorous preliminary evaluation of datasets, particularly when a substantial volume of data are rapidly generated by infrastructure projects such as hydroelectric power plants, to prevent analytical biases and ensure accurate results. These measures aim to strengthen amphibian research and support biodiversity conservation, particularly in response to increasing deforestation and climate change in the Amazon.