PERMEABILITÀ URBANA: UN APPROCCIO MULTIDISCIPLINARE PER LO STUDIO DELLA MOBILITÀ DELLE FARFALLE DI TORINO

Urban environments are fast growing in the last centuries in size and population. The city is known to negatively impact biodiversity, causing land loss and fragmentation, reducing connectivity within urban habitats. For insects, habitat loss and fragmentation have been identified as major causes of decline. Among insects, butterflies can be used as indicator species to study which how urban habitat affects mobility and how make cities more permeable To answer this question a multidisciplinary approach that combined biology, ecology, simulation and physics of complex systems was applied. Indeed, to predict butterfly movements within the city we applied a Multi-Agent System, an artificial intelligence model where butterflies are agents. An agent is a computerised system able to perform decisions and choices independently and to interact with one another and with the environment. To model the agent butterfly, data from fieldwork, bibliography and GIS maps were used. In 2021 in 2 sites within Turin, the Mark Release Recapture method was applied to obtain data on Pieris butterflies, including 3 common species: Pieris rapae, P. napi and P. mannii. In the two sites (called Urban area and Parco Piemonte ) in the city of Turin a total of 1020 Pieris individuals were marked and 465 were recaptured at least once (recapture rate 45.588%). Thus, we estimated that 829 Pieris individuals in Urban area (369.228 females, std. dev. = 65.865; 460.111 males, std. dev. = 39.880) and 580 Pieris individuals (females 182.691, std. dev. = 15.823; males 397.36, std. dev. = 21.666) in the area Parco Piemonte were flying in the sampling period. Data on flight distance were modelled to understand the influence of sex and species on mobility. No differences were found between sex and species. The estimated probabilities (in % of individuals) of long-distance movements (1km) in the Urban area 5.70% for males and 9.89% for females while in Parco Piemonte area were 5.33% for males and 1.04% for females. To obtain a model that can correctly reproduce the dynamics of the species it is necessary to obtain data to train the model. Thanks to field-data and bibliographic research, a first training of the butterfly-agent was performed. In this first operation, the virtual energy of the agents was studied, in order to reproduce at the best the activity rate of the real butterflies. Future steps will be scaling up to simulate the entire city to understand possible management suggestions for Turin municipality. Thus, a revised version of the model might be a valid resource for urban planners.