Atropelamentos de vertebrados silvestres no semiárido brasileiro: dinâmica temporal, hotspots e fatores ambientais

Abstract

The road network plays an important role in the movement of people and the transportation of goods between regions of the country. However, roads are responsible for a series of negative environmental impacts, such as wildlife roadkills. In general, these roadkills occur in clustered patterns along specific road segments, known as hotspots, and may be influenced by adjacent landscape characteristics and seasonal factors, such as precipitation. Therefore, the present study analyzed wild vertebrate roadkills along a 112-km stretch of the SE-230 highway, located in a Caatinga area in the state of Sergipe, northeast of Brazil, characterizing the affected fauna, its temporal dynamics, the spatial distribution of hotspots, and the influence of landscape and precipitation on these events. Monitoring was conducted biweekly between January and December 2025 by two observers traveling in a motorized vehicle at 50 km/h, during which information regarding specimen identification, date, and geographic coordinates was recorded. Circular analyses (Rayleigh and Mardia–Watson–Wheeler tests) were used to evaluate the temporal distribution of roadkills, while Ripley’s K 2D and 2D HotSpot Identification tests were applied to identify hotspots. In addition, logistic regression models were used to investigate the influence of landscape and precipitation. A total of 524 individuals belonging to 63 vertebrate species were recorded. Amphibia showed the highest number of records (37.79%), followed by Aves (24.62%), Mammalia (24.04%), and Reptilia (13.55%). Circular analyses indicated a non-uniform distribution of roadkills for amphibians, reptiles, and mammals, whereas birds showed a uniform distribution throughout the year, with differences in the temporal distribution of roadkills among vertebrate classes. The location of hotspots varied between dry and rainy months, showing low spatial similarity among taxonomic groups. Logistic regression models indicated that landscape variables influenced roadkills as urban areas and use mosaics. Furthermore, precipitation had a positive effect on amphibians and birds. The results demonstrate that vertebrate roadkills in the Caatinga exhibit temporal and spatial variation. Thus, this study contributes to expanding knowledge about the impacts of roads in semiarid regions and provides support for the implementation of mitigation measures aimed at wildlife conservation.