Assessing the effect of landscape features on pond colonisation by an elusive amphibian invader using environmental DNA

Environmental DNA (eDNA) is becoming an essential tool for detecting aquatic invasive species and investigating their spread. Surprisingly, this technique has been very rarely used to investigate habitat selection, site occupancy, and colonisation despite its higher capacity to detect many species. The African clawed frog (Xenopus laevis) is a principally aquatic amphibian introduced in several continents from South Africa. In western France, no recent systematic survey of the invasion range has been attempted, mainly because of the elusive nature of the species. Furthermore, the influence of landscape features on invasion has never been investigated, even if adults and juveniles are known to disperse overland and along river networks. Using presence–absence data generated by an eDNA survey conducted across the known invasion front of X. laevis in western France, we aimed to determine whether and how the landscape features surrounding a pond influence the probability that a pond is colonised. Xenopus laevis was detected well beyond the formerly known invasive distribution and at the outward end of some transects, suggesting that we did not reach the actual invasion front in these parts of the range. The landscape variables that best predicted the presence of X. laevis in a pond were topographic wetness index and grass cover within a buffer of 250 m. Higher values of both topographic wetness index and grass cover were negatively related to the occurrence probability. The effects of these two variables more likely to reflect dispersal behaviour than habitat preferences at the pond scale. By combining the high detection probability of eDNA survey techniques and a landscape ecology approach, we may gain valuable insight into the colonisation process of water bodies by elusive invasive species. Such information is crucial to prevent access to specific sites and locate invasion front areas where connectivity can be disrupted, thus increasing the effectiveness of management countermeasures.