Seagrass–Sediment Feedback: An Exploration Using a Non-recursive Structural Equation Model

The reciprocal effects between sediment texture and seagrass density are assumed to play an important role in the dynamics and stability of intertidal–coastal ecosystems. However, this feedback relationship has been difficult to study empirically on an ecosystem scale, so that knowledge is mainly based on theoretical models and small-scale (experimental) studies. In this paper we apply a non-recursive structural equation model (SEM) to empirically investigate, at large spatial scale, the mutual dependence between seagrass (Zostera noltii) density and sediment texture, on the pristine, seagrass-dominated, intertidal mudflats of the Banc d’Arguin, Mauritania. The non-recursive SEM allows consistent estimation and testing of a direct feedback between sediment and seagrass whilst statistically controlling for the effects of nutrients and abiotic stress. The resulting model is consistent with the hypothesized negative feedback: grain size decreases with seagrass density, whereas fine grain size has a negative impact on seagrass density because it decreases pore water exchange which leads to hypoxic sediment conditions. Another finding is that seagrass density increases with sediment organic material content up to a threshold level beyond which it levels off. In combination with decreasing grain size, accumulation of organic matter creates hypoxic sediment conditions which lead to the production of toxic hydrogen sulfide which slows down seagrass growth. The negative feedback loop implies that intertidal Z. noltii modifies its own environment, thus controlling its growing conditions. To the best of our knowledge, this study is the first to demonstrate a direct negative feedback relationship in ecosystems by means of a non-recursive SEM.