Modelling the spatial distribution of alien invasive species from citizen science data: limits and advantages

Decision-makers and governmental organizations are increasing the use of citizen volunteers to enhance the capacity to monitor species occurrence and distribution around the world and to investigate major aspects of global change impacts on biodiversity. For citizens the motivation is to contribute to “real” science, public information and conservation. For scientists, citizen science offers a way to collect information that would otherwise be not affordable. CSMON-LIFE (Citizen Science MONitoring) is one of the first Italian projects which focuses on using a citizen science approach on biodiversity. It is funded by the European Commission under the LIFE+ programme. CSMON has come to its latest year with more than 4000 registered users and over 20000 validated reports focused on several environmental issues, such as: the loss of biodiversity because of the presence of alien species; the effects of climate change; the conservation of rare species; the impact of human activities on the environment. The data validated feed the National Biodiversity Network established by the Italian Ministry for the Environment in order to make freely available biodiversity data. We used data collected by citizen volunteers for four alien plant species: Ailanthus altissima, Robinia pseudoacacia, Carpobrotus edulis and Opuntia ficus-indica to model their ecological niche using MaxEnt (Maximum Entropy) software. As environmental factors we used the WorldClim bioclimatic variables with a spatial resolution of 1 km2, as well as soil, land use and topography parameters. We set aside 25% of the samples for testing and the rest used for niche modelling training. The analyses used a threshold to make a binary prediction, with conditions above the threshold as suitable and those below as unsuitable. To identify variables that are most important for the species being modelled the jackknife test we used, and a set of response curves elaborated to interpret if there are strong correlations between variables. Finally, it was necessary to include in MaxEnt’s run a bias file to represent sampling effort to reduce the sampling bias performed during volunteer monitoring due to the different accessibility of the sites explored and the different amount of time dedicated to these operations. Despite these limitations data coming from the activity of citizen volunteers significantly increased the knowledge on the spatial distribution of target species. Based on these results, the establishment of permanent biodiversity observatories formed by citizen volunteers is envisaged.