MONTE CARLO–BASED ESTIMATION OFREGIONAL ECOSYSTEM PARAMETERS IN THE SETO INLAND SEA

Abstract

This study developed and applied a Monte Carlo–based food-chain box model to identify region-specific ecosystem parameter sets for 13 sub-regions of the Seto Inland Sea. For each region, large ensembles of Monte Carlo simulations (approximately 10,000–700,000 realizations) were explored to identify 1,000 feasible parameter sets for 28 ecosystem parameters that simultaneously reproduced the observed ranges of dissolved inorganic nitrogen (DIN) concentrations and phytoplankton biomass in both the 1990s and the 2010s. This approach allowed parameter uncertainty to be systematically constrained using long-term monitoring data rather than subjective tuning.  The estimated parameters exhibited clear inter-regional differences, particularly in the maximum growth rates of phytoplankton and zooplankton and in phytoplankton sinking velocities. These differences were consistent with observed regional variations in phytoplankton community composition, nutrient conditions, and physical environments. Overall, the results highlight the importance of region-specific ecosystem parameter settings for accurately simulating coastal ecosystem dynamics and for improving the reliability of high-resolution numerical models.