Bayesian Networks (BN) are useful tools for understanding complex relationships between multiple driver and response variables. In BNs, conditional probabilities are assigned to define relationships between individual environmental drivers (e.g., sediment characteristics, pollutants, water depth) and individual ecosystem response variables (e.g., biodiversity, productivity, denitrification). As with all models, parameterisation is a critical step. For BNs, it is important to gather evidence to support the conditional probabilities that underpin the models.
Using a systematic review of global intertidal and subtidal estuarine literature, we were able to collate an extensive dataset to define driver-response relationships. By gathering information for intertidal and subtidal estuarine habitats separately, our estuarine BN can be ‘partitioned’ into intertidal and subtidal compartments, enabling us to simulate sea level rise scenarios and model spatial shifts in ecosystem functioning, such as changes in biodiversity and net changes to nutrient and carbon cycling that may occur with increasing inundation (loss) of intertidal estuarine habitats as sea levels rise.