Forecasting ecosystem services to guide coastal wetland rehabilitation decisions

•Predictive mechanistic models enhance decision-making for coastal wetlands.•Probabilistic analysis is essential to interpret ecosystem service forecasts.•Flood mitigation likely dominates ecosystem service value of wetland restoration.•Non-flood services increase cost-effectiveness of restoration on short timescales. Coastal wetlands provide diverse ecosystem services such as flood protection and recreational value. However, predicting changes in ecosystem service value from restoration or management is challenging because environmental systems are highly complex and uncertain. Furthermore, benefits are diverse and accrue over various timescales. We developed a generalizable mathematical coastal management model to compare restoration expenditures to ecosystem service benefits and apply it to McInnis Marsh, Marin County, California, USA. We find that benefits of restoration outweigh costs for a wide range of assumptions. For instance, costs of restoration range from 8–30% of the increase in ecosystem service value over 50 years depending on discount rate. Flood protection is the dominant monetized service for most payback periods and discount rates, but other services (e.g., recreation) dominate on shorter timescales (>50% of total value for payback periods ≤4 years). We find that the range of total ecosystem service value is narrower than overall variability reported in the literature, supporting the use of mechanistic methods in decision-making around coastal resiliency. However, the magnitude and relative importance of ecosystem services are sensitive to payback period, discount rate and risk tolerance, demonstrating the importance of probabilistic decision analysis. This work provides a modular, transferrable tool to that can also inform coastal resiliency investments elsewhere.