Environmental variables driving habitat suitability for Popillia japonica in Lombardy

The Japanese beetle Popillia japonica Newman (Coleoptera: Scarabaeidae) is an invasive species native to Japan and the far east of Russia. The adults of P. japonica are polyphagous feeding on over 300 plant species, skeletonizing the leaves and eroding flowers and fruits. The larvae feed on roots, causing yellowing and drying of lawns. Since 1916, the species has been considered established in North America, in the 1970s it was introduced in the Azores and more recently in Northern Italy (2014). From the original area of establishment in few Municipalities in the Northern Ticino valley in 2014, P. japonica is spreading in the Lombardy and the Piedmont regions covering an area that is increasing year by year. The Italian and the regional Phytosanitary Services are implementing the appropriate measures to manage the spread of the species. Understanding the characteristics of the habitat that allows the establishment and the growth of P. japonica populations is fundamental for supporting the assessment and the management of the risks linked to this species. In this work, we investigate the habitat suitability for larvae of P. japonica populations recently established in western Lombardy (Italy). Data used for the analysis refer to occurrence of larval populations between 2016 and 2019 in 8908 sites sampled in the infested area of the Lombardy Region. The samples sites were characterized on the basis of a set of explanatory variables related to land use, soil geo-pedological characteristics, and meteorology. We performed a stepwise logistic regression analysis for testing the statistical significance of explanatory variables on the occurrence of local larval populations. Then, we implemented a Generalized Additive Model to investigate the possible nonlinear relationships between explanatory and dependent variables. Results show the significant role of soil variables (e.g., soil texture, level of organic substance and soil pH) and weather variables (e.g., cumulated rain, air and soil temperature, soil humidity) on the probability of larval occurrence. The distance from the area of first infestation and number of years passed since the first establishment of the species are also related to the probability of larval occurrence. The proposed modelling framework can be used for evaluating how habitat suitability for P. japonica varies across the landscape. The model can be considered as an important tool for generating risk maps helping in prioritize survey and control measures.