Using weighted expert judgement and nonlinear data analysis to improve Bayesian belief network models for riverine...

Area of Science:

• Environmental Science
• Ecology
• Hydrology

Background:

• Rivers are vital for the hydrological cycle and water resources but face anthropogenic threats.
• Integrating ecological and socio-economic factors in river management is complex due to heterogeneous methodologies.
• Existing models often fail to capture the non-linear dynamics common in ecosystem responses.

Purpose of the Study:

• To advance a Bayesian belief network (BBN) model for integrating ecosystem services into river management.
• To causally link catchment stressors with ecosystem services using diverse evidence.
• To apply the BBN model to Irish catchments to demonstrate stressor impacts on ecosystem services.

Main Methods:

• Developed a Bayesian belief network (BBN) model linking catchment stressors to ecosystem services.
• Utilized expert workshops, legislation, and literature for causal links.
• Calibrated the BBN with national monitoring data, expert judgment, and identified non-linear relationships.
• Quantified model outputs using a novel expected index of desirability.

Main Results:

• Non-linear relationships were common in ecosystem responses to stressors.
• Deficiency in riparian shading was a significant factor affecting multiple ecosystem services.
• Sediment load had a lower impact in flashy rivers but synergistic effects with organic matter and phosphate.
• Integrated management of stressor pairs yielded greater societal benefits than individual management.

Conclusions:

• The developed BBN model effectively integrates ecosystem services into river management.
• Riparian shading is a key factor for improving societal benefits from rivers.
• Simultaneous management of synergistic stressors offers enhanced benefits.
• The BBN approach and desirability index are adaptable for global river management applications.