Where to search: the use of opportunistic data for the detection of an invasive forest pest

Abstract

Early detection is important for the management of invasive alien species. In the last decade citizen science has become an important source of such data. Here, we used opportunistic records from the “LIFE ARTEMIS” citizen science project, in which people submitted records from places where they observed tree pests, to understand the distribution of a rapidly-spreading forest pest: the oak lace bug (Corythucha arcuata) in Slovenia. These citizen science records were not distributed randomly. We constructed a species distribution model for C. arcuata that accounted for the biased distribution of citizen science by using the records of other tree pests and diseases from the same project as pseudo-absences (so-called constrained pseudo-absences), and compared this to a model with pseudo-absences selected randomly from across Slovenia. We found that the constrained pseudo-absence model showed that C. arcuata was more likely to be found in east, in places with more oak trees and at lower elevations, and also closer to highways and railways, indicating introduction and dispersal by accidental human transport. The outputs from the model with random pseudo-absences were broadly similar, although estimates from this model tended to be higher and less precise, and some factors that were significant (proximity to minor roads and human settlements) were artefacts of recorder bias, showing the importance of taking the distribution of recording into account wherever possible. The finding that C. arcuata is more likely to be found near highways allows us to design advice for where future citizen science should be directed for efficient early detection.