Non-linear responses of ecological indicators to urban environmental drivers across Europe

Urban ecosystems are impacted by multiple environmental drivers which can work at different spatial scales and intensities, negatively impacting biodiversity and ecological processes. Ecological indicators such as lichens are used to measure the impacts of these drivers at the ecosystem level and provide valuable information for urban policy guiding. Lichen responses interpretation is nevertheless complex, as these drivers act together in cities. Furthermore, while lichen responses are often nonlinear, studies typically ignore such complexities. This not only hinders our ability to disentangle the individual effect of each driver but also prevents evaluating the effectiveness of policies to mitigate their impact.
Here, we used a continental lichen biodiversity dataset and a machine-learning approach to disentangle the individual effect of the prevailing drivers on urban lichens. Particularly, we aimed at identifying the intensity/direction of the effect of single drivers across cities and determine, for each driver, the presence of threshold-like responses shaping lichen biodiversity.
We sampled epiphytic lichen biodiversity in seven European cities, from which taxonomic and functional metrics were calculated. Environmental drivers related to temperature, water availability and anthropic pressure were modeled to explain lichen metrics with Random Forest, to account for the potential collinearity and non-linear responses.
Overall, temperature and water availability were the most important drivers for urban lichen biodiversity, followed by anthropic pressure, mainly related to urbanization. Partial dependence responses revealed thresholds of lichen biodiversity change along the environmental gradient considered. The quantification of the main drivers of urban lichen diversity is extremely useful to inform city-level sustainability and biodiversity policies, while the identification of thresholds is crucial to better define environmental safety limits in urban planning.