Abstract
Marine ecosystems integrate biotic and abiotic processes increasingly disrupted by human activities. We present a Petri net–based computational framework integrated with process mining to quantify respiration, assimilation, and mortality in Venice Lagoon compartments: phytoplankton, bacterioplankton, microzooplankton, Manila clam (Ruditapes philippinarum), and detritus. Using literature flux ranges, synthetic event logs were analyzed with the α-algorithm, inductive miner, and fuzzy miner; token-based replay exposed bottlenecks. Lower-trophic compartments (detritus, phytoplankton, bacterioplankton) exhibited higher throughput than higher-trophic compartments (clam, microzooplankton). Event-level diagnostics explained low microzooplankton assimilation when predators were absent and elevated phytoplankton mortality from unmodeled grazers; bacterioplankton/phytoplankton deviations reflected structural invariants and grazing assumptions. Fuzzy-miner abstractions clarified dominant pathways. Coupling advances ecosystem diagnostics and pinpoints intervention points for biodiversity conservation and climate resilience.
