Exploring Uncertainty and Parameter Sensitivity in Estuarine Carbon Dynamics: A C-GEM Framework Update

Estuarine environments are complex dynamic systems and cornerstone components of the  Land-Ocean Aquatic Continuum (LOAC). As such, they play a crucial role in global biogeochemical cycles, acting both as conduits and processors of carbon and nutrients between the terrestrial and oceanic realms of the Earth system. Despite this significance, considerable uncertainties remain associated with the quantification of biogeochemical fluxes within and through estuarine systems. In particular, the interplay between Organic Carbon (OC) degradation and Dissolved Inorganic Carbon (DIC) oversaturation remains poorly constrained in many estuarine systems. Furthermore, global estimates of estuarine CO2 emissions are still derived from limited and heterogeneous observational datasets, highlighting the need for adequate tools to resolve system-specific dynamics for a wide variety of estuarine set-ups. Reactive Transport Models (RTMs) explicitly simulate both physical and biogeochemical processes controlling these dynamics and can be used to resolve the spatial and temporal gradient of their respective prevalence on net biogeochemical dynamics. Designed to reduce data demands through its one-dimensional structure and generic parameterization, the Carbon-Generic Estuary Model (C-GEM) has emerged as a computationally efficient RTM framework for simulating estuarine biogeochemistry, and has been successfully applied to estuarine systems of diverse morphologies ranging from river to marine dominated. In this work, we present an updated version of C-GEM, designed to answer the growing need for estuarine models capable of addressing the long-term impacts of anthropogenic pressures, climate change, and land-use modifications on carbon cycling, including transient simulations over multi-annual scales, the integration of a module for inorganic carbon dynamics, and improved user-accessibility. We first demonstrate the applicability of the enhanced C-GEM framework to a range of realistic and idealized estuarine systems, exploring the relationships between hydrodynamic characteristics and biogeochemical functioning. In a second step, we explore the propagation of the inherent uncertainty associated with biogeochemical parameters towards integrated carbon budget diagnostics including CO2 exchange with the atmosphere, primary production or net ecosystem metabolism. A global sensitivity analysis (Morris Screening) is first performed for various estuarine set-ups (i.e. morphologies, residence time…) in order to rank the system-specific importance of biogeochemical parameters in driving integrated carbon budget diagnostics, thereby highlighting which processes are the main drivers of the estuarine carbon dynamics and which parameters may most require fine-tuning to better constrain estuarine budgets. Probability distributions are then built for selected key parameters based on published reference value. Finally, Monte Carlo simulations based on such constrained parameter sampling are used to delineate the resulting uncertainty in integrated estuarine GHG budgets. This analysis is specified for estuarine systems of different morphologies and hydrological regimes, as well as different portions of the estuarine systems, over which certain transport and biogeochemical processes prevail. Besides revealing the shadow zone of global estuarine carbon budgets, characterizing the resulting spread in observable states  along estuaries (e.g. nutrients, Org C, ..) also help to identify priority observation efforts that would optimally constrain the overall carbon budget diagnostics.