Quantifying estuarine carbon and nutrients retention at the regional scale using a generic process-based model and Monte Carlo simulations

Most downstream compartments of the continental hydrological network, estuaries are the last biogeochemical filter of the Land-Ocean Aquatic Continuum before the oceanic realm. As such, they receive substantial amounts of carbon and nutrients from rivers and their intense biogeochemical processing allows the removal of part of those inputs, hence potentially contributing to the prevention of coastal eutrophication. Indeed, eutrophication resulting from enhanced nutrients loads from rivers is a pressing global issue, affecting numerous coastal areas and regional seas worldwide. However, simulating ecosystems as intricate as estuaries, characterized by numerous biogeochemical gradients, an intense benthic-pelagic coupling and controlled by complex hydrodynamics is a challenge often associated with intensive computation and data requirements. As a result, the development of numerical models suitable to quantify the filtering function of estuaries is often limited to scarce well studied systems. This highlights the still unresolved challenge of designing and applying a generic modeling strategy able to capture the complexity and intensity of biogeochemical processes for a diversity of often data-limited estuaries along a continuous coastal stretch.

In this study, we present the first spatially explicit, regional, fully transient simulation of the estuarine biogeochemical filter over a multi annual period. This application to 40 estuaries of the Atlantic coast of France from its southern border with Spain to Belgium was performed in the context of the nuts-STeauRY project which aims at illustrating the interest of integrated land-sea modelling approaches to better design spatialized scenarios of agriculture and land-use practice to limit coastal eutrophication in France.

                The simulations were performed using the proven generic estuarine model C-GEM coupled with the OMEN_SED sediment module and constrained, upstream by the pyNuts-Riverstrahler model, which describes the transfer of nutrients and carbon from the headwaters streams to the outlets of river hydrosystems. In its current version, C-GEM resolves tidally induced transport within the estuary along its longitudinal axis and its biogeochemical module includes all the main processes involving carbon and nutrients (i.e. production, remineralization, nitrification, denitrification…). The addition of a new explicit benthic module allows simulating sediments processes and burial which are essential to properly quantify carbon and nutrient retention. The strategy to simulate estuaries devoid of measurements relies on Monte Carlo simulations performed by varying the model’s parameterization constrained by an extensive literature survey and thoroughly validated on well monitored reference systems. Our results over the 2014-2019 period provide an insight into the parameters controlling the temporal and spatial variability of carbon and nutrient retention within a large set of estuaries with varying riverine nutrients loads and ranging from very small (<10km) to the Loire, Seine or Gironde estuaries, which lengths exceed 100km. The simulated retention rates vary widely from just a few percent in the smallest systems to over 40%, 30%, 20% in the largest ones for total organic carbon, total phosphorus and total nitrogen, respectively.