Using DNA barcode to relate landscape attributes to small vertebrate roadkill
MetadataShow full item record
Large vertebrates are the main focus of roadkill studies because their greater size facilitates taxonomic identification and the collection of statistical data. However, these studies fail to effectively include and identify small vertebrates and correlate roadkill events with the surrounding landscape. Here we showed the effectiveness of molecular data to identify small vertebrate roadkill, and we correlated landscape structure attributes with the location of roadkill for functional groups of varying mobility. The extraction of DNA from roadkilled individuals was followed by the amplification of two mitochondrial genes. We compared each DNA sequence to a database and used the highest similarity values for species identification. The species were classified according to their taxa and degree of mobility: birds, reptilia and amphibia with low and intermediate movement capability. After calculating the landscape attributes for each roadkill point, we used a competing model approach based on Akaike Information Criteria to determine which landscape variable best explained the occurrence of roadkills. Combining molecular and morphological characteristics, we identified 82.93% of the roadkilled animals. DNA barcoding allowed the identification of 310% more specimens than by morphological characteristics alone. Roadkilled birds with intermediate movement capability were strongly influenced by dominated areas by agriculture and sugar cane monocultures. Roadkilled reptiles with low movement capability were positively correlated with the presence of forest remnants, while those with intermediate movement capability seemed to be more frequent in heavily anthropized landscapes. We showed that molecular data is a powerful tool for precisely identifying small-sized roadkilled animals. Our results also highlight that different landscape structure attributes enable the prediction of roadkill occurrence along roads, which in turn allows us to identify roadkill hotspots and plan appropriate mitigation actions.