The role of the deposit feeding bivalve Macomona liliana in a simple biophysical model for predicting estuarine ecosystem tipping points

Abstract

Estuaries are highly productive, coastal marine environments that provide vital ecosystem functions supporting human health. Throughout history, many major settlements have been established adjacent to estuaries, so to take advantage of the resources they provide. After centuries of exploitation and heavy pollution, the health of estuaries have experienced a notable decline. The need for effective management strategies has arisen, but their efficacy continues to be questioned due to the dynamic and unpredictable nature of these environments. Non-linear processes, such as regime shifts, make long-term predictions of ecosystem health a point of concern. To combat this issue in New Zealand, ecosystem interaction networks (EINs) have been used to predict the future health of estuaries using long-term statistical modelling. Unfortunately, these models fail to capture the coupling interactions between biotic and abiotic variables that occur in such a volatile environment. Therefore, we have created a simple model design that replicates the dynamic ecosystem interactions found within the benthic environments of New Zealand estuaries, with a specific focus on the behaviour of the keystone bivalve species Macomona liliana. A multitude of biophysical dynamical models currently exist in this field (e.g. Delft3D and DELWAQ); however none with the ability to replicate the exchange of nutrients between the sediment and benthic layer caused by M. liliana pumping behaviour. Consequently, this makes our model the first of its kind. Our model design simulates the variation of three dynamic variables (biomass, nutrients, detritus) corresponding to the interactions of M. liliana and microphytobenthos (MPB) at a representative site of New Zealand sandy bottom estuaries (Tuapiro Point, Tauranga Harbour). Extensive testing was conducted to ensure that, at the baseline level, the model’s output was consistent with field sampling data seen within the literature. Once completed, the model was tested against external forcing factors that are likely to occur within New Zealand estuaries – turbidity, sedimentation, and external nutrient loading. We found that our system is light limited, with M. liliana and MPB biomasses both declining with increasing turbidity. MPB populations fluctuate significantly following a sediment mud content increase, which we believe demonstrates a reaction to an ecological tipping point being surpassed. Furthermore, we found that a significant influx of external nutrient loading is required to initiate a MPB bloom, which the system cannot maintain despite nutrient levels remaining high. Our model’s results are in agreement with estuarine responses seen in the past, with our design exceeding the guidelines surrounding the efficacy in predicting ecosystem tipping points. Therefore, due to its unique ability to demonstrate the interactions within coupled systems and its broad applicability globally, we believe our model can become a central component of future estuarine management strategies. Recommendations are made for future improvements to the model design, increasing both its range of simulated ecological processes and the validity of its predictions.