Benchmarking Global Ocean Carbon Cycle models: Uncertainties in anthropogenic CO2 uptake estimation

The ocean carbon cycle plays a crucial role in the uptake and storage of atmospheric carbon dioxide. The Global Carbon Budget (GCB) provides annual estimates of this oceanic carbon sink, with the latest estimate for 2022 indicating a net uptake of 2.8 ± 0.4 Gt C yr^-1 (Friedlingstein et al., 2023). However, the estimates by global ocean biogeochemistry models (GOBMs, 2.5 ± 0.4 Gt C yr^-1 ) are substantially lower than the estimates based on upscaled observations (3.1 Gt C yr^-1 [2.5, 3.3]) and the range of the sink estimates covered by the GOBMs is substantial (1.1 Gt C yr^-1 ). Biases and uncertainties in the GOBM estimates of the ocean carbon sink may be due to imperfections in the representation of physical (e.g., transport, mixing) and biogeochemical processes, as well as in the forcing fields.

To address these uncertainties, we employ the International Ocean Model Benchmarking (IOMB) system to evaluate the performance of GCB models against state-of-the-art observations. IOMB is a Python-based open-source software package that is used to evaluate the performance of Earth System Models and the counterpart to ILAMB that is used to evaluate the dynamic vegetation models for the land sink in the GCB. Using IOMB, we analyze the physical (upper ocean temperature, vertical temperature gradient, mixed layer depth, salinity) and biogeochemical (nutrients, chlorophyll, oxygen, total alkalinity, dissolved inorganic carbon (DIC), anthropogenic DIC) variables of interest. but all the variables in the current version of IOMB evaluate the general performance of the ocean biogeochemistry models, not specifically the ocean carbon uptake.The addition of new targeted metrics such as AMOC, stratification indices, CFCs and Revelle factor will help to reduce the source of uncertainty in these models. IOMB evaluates the model performances using statistical measures such as bias, root mean square error (RMSE), annual cycle phase, spatial distribution, interannual variability and a score will be estimated for each model, providing a benchmark for the current state of GCB models and highlighting the areas for further model development.

The goal is to improve the accuracy and reliability of ocean carbon uptake estimates, which are essential for informing climate policy. Through this process, IOMB will provide valuable feedback to the modeling community, offering guidance on how to improve ocean biogeochemical simulations and better constrain the oceanic carbon sink in the context of global carbon budgets. This is a necessary first step towards weighting GCB models in estimating the ocean carbon sink.