Butterflies are widely recognised as sensitive bioindicators of ecosystem health, whose distribution reflects both climatic and urban changes. Understanding how butterflies move across fragmented urban landscapes is therefore an important challenge for biodiversity conservation. This thesis presents an ongoing research project aimed at studying butterfly movement between urban parks in Barcelona through a mathematical modelling framework. Data are provided by the Urban Butterfly Monitoring Scheme (uBMS), which monitored 24 parks between 2018 and 2024. Population abundances are reconstructed using a neural-network-based pipeline, producing continuous daily time series from field observations. Since direct data on butterfly movements are not available, the main objective is to infer inter-park flows from observed population dynamics.