Description: Our research focuses on the detection of ocean carbon uptake regimes that are critical in the context of comprehending climate change. One observation among geoscientific data in Earth System Sciences is that the datasets often contain local and distinct statistical distributions posing a major challenge in applying clustering algorithms for data analysis. The use of global parameters in many clustering algorithms is often inadequate to capture such local distributions. In this study, we propose a novel tool to detect and visualize oceanic carbon uptake clusters. We implement a distance-variance selection method (augmented by BIC scores) on agglomerative hierarchical clustering constructed upon a regional multivariate linear regression model set. Instead of relying on a global distance, users can select the local distance and variance thresholds on our tool to detect the connections on the dendrograms that stand as potential clusters by considering both compactness and similarity.

Description: Our research focuses on detecting and tracking ocean carbon regimes, which are crucial indicators for understanding the impacts of climate change on ocean carbon uptake. Geoscientific datasets in Earth System Sciences often contain local and distinct statistical distributions at a regional scale. This poses a significant challenge in applying conventional clustering algorithms for data analysis. Based on the observed limitations of prominent methods, in our study, we propose a framework that enhances well-established unsupervised machine-learning methods tailored to applications on geoscientific datasets. We define a carbon uptake regime as a region characterized by common relationships between the carbon uptake and its drivers, as simulated by a multi-annual hydrodynamic model simulation. As a first step, we compute multivariate linear regressions capturing local spatial relations between carbon dioxide uptake and its drivers to discover such regimes. This is followed by an agglomerative hierarchical clustering constructed upon the collection of regional multivariate linear regression models. To overcome the emerging limitations of a global cut for partitioning, which is inadequate to capture the local statistical distributions, we present a novel, straightforward and adaptive approach to detect and visualize ocean carbon uptake regimes in this work. This method relies on the distance-variance selection technique and detects multiple local cuts on the dendrogram by considering both the compactness and similarity of the clusters. Detecting meaningful and well-defined carbon uptake regimes is vital for their tracking over time. The tracking is performed through a simple yet effective approach where summary structures derived from the clusters are traced over time. Applied over longer time scales, this novel method will enable marine scientists to effectively monitor the impacts of climate change on the ocean carbon cycle more

Description: The Southern Ocean is a major sink of atmospheric CO2, but the nature and magnitude of its variability remains uncertain and debated. Estimates based on observations suggest substantial variability that is not reproduced by process-based ocean models, with increasingly divergent estimates over the past decade. We examine potential constraints on the nature and magnitude of climate-driven variability of the Southern Ocean CO2 sink from observation-based air-sea O-2 fluxes. On interannual time scales, the variability in the air-sea fluxes of CO2 and O-2 estimated from observations is consistent across the two species and positively correlated with the variability simulated by ocean models. Our analysis suggests that variations in ocean ventilation related to the Southern Annular Mode are responsible for this interannual variability. On decadal time scales, the existence of significant variability in the air-sea CO2 flux estimated from observations also tends to be supported by observation-based estimates of O-2 flux variability. However, the large decadal variability in air-sea CO2 flux is absent from ocean models. Our analysis suggests that issues in representing the balance between the thermal and non-thermal components of the CO2 sink and/or insufficient variability in mode water formation might contribute to the lack of decadal variability in the current generation of ocean models.This article is part of a discussion meeting issue 'Heat and carbon uptake in the Southern Ocean: the state of the art and future priorities'.

Description: We assess the Southern Ocean CO2 uptake (1985–2018) using data sets gathered in the REgional Carbon Cycle Assessment and Processes Project Phase 2. The Southern Ocean acted as a sink for CO2 with close agreement between simulation results from global ocean biogeochemistry models (GOBMs, 0.75 ± 0.28 PgC yr−1) and pCO2-observation-based products (0.73 ± 0.07 PgC yr−1). This sink is only half that reported by RECCAP1 for the same region and timeframe. The present-day net uptake is to first order a response to rising atmospheric CO2, driving large amounts of anthropogenic CO2 (Cant) into the ocean, thereby overcompensating the loss of natural CO2 to the atmosphere. An apparent knowledge gap is the increase of the sink since 2000, with pCO2-products suggesting a growth that is more than twice as strong and uncertain as that of GOBMs (0.26 ± 0.06 and 0.11 ± 0.03 Pg C yr−1 decade−1, respectively). This is despite nearly identical pCO2 trends in GOBMs and pCO2-products when both products are compared only at the locations where pCO2 was measured. Seasonal analyses revealed agreement in driving processes in winter with uncertainty in the magnitude of outgassing, whereas discrepancies are more fundamental in summer, when GOBMs exhibit difficulties in simulating the effects of the non-thermal processes of biology and mixing/circulation. Ocean interior accumulation of Cant points to an underestimate of Cant uptake and storage in GOBMs. Future work needs to link surface fluxes and interior ocean transport, build long overdue systematic observation networks and push toward better process understanding of drivers of the carbon cycle. Key Points: - Ocean models and machine learning estimates agree on the mean Southern Ocean CO2 sink, but the trend since 2000 differs by a factor of two - REgional Carbon Cycle Assessment and Processes Project Phase 2 estimates a 50% smaller Southern Ocean CO2 sink for the same region and timeframe as RECCAP1 - Large model spread in summer and winter indicates that sustained efforts are required to understand driving processes in all seasons

Description: The consideration of marine biogeochemistry is essential for simulating the carbon cycle in an Earth system model. Here we present the implementation and evaluation of a marine biogeochemical model, Model of Oceanic Pelagic Stoichiometry (MOPS) in the Flexible Ocean and Climate Infrastructure (FOCI) climate model. FOCI-MOPS enables the simulation of marine biological processes, the marine carbon, nitrogen and oxygen cycles, air-sea gas exchange of CO2 and O2, and simulations with prescribed atmospheric CO2 or CO2 emissions. A series of experiments covering the historical period (1850–2014) were performed following the DECK (Diagnostic, Evaluation and Characterization of Klima) and CMIP6 (Coupled Model Intercomparison Project 6) protocols. Overall, modelled biogeochemical tracer distributions and fluxes, as well as transient evolution in surface air temperature, air-sea CO2 fluxes, and changes of ocean carbon and heat, are in good agreement with observations. Modelled inorganic and organic tracer distributions are quantitatively evaluated by statistically-derived metrics. Results of the FOCI-MOPS model, also including sea surface temperature, surface pH, oxygen (100–600 m), nitrate (0–100 m), and primary production, are within the range of other CMIP6 model results. Overall, the evaluation of FOCI-MOPS indicates its suitability for Earth climate system simulations

Description: Enhanced Southern Ocean ventilation in recent decades has been suggested to be a relevant modulator of the observed changes in ocean heat and carbon uptake. This study focuses on the Southern Ocean midlatitude ventilation changes from the 1960s to the 2010s. A global 1/4° configuration of the NEMO–Louvain-la-Neuve sea ice model, version 2 (LIM2), including the inert tracer CFC-12 (a proxy of ocean ventilation) is forced with the CORE, phase II (CORE-II), and JRA-55 driving ocean (JRA55-do) atmospheric reanalyses. Sensitivity experiments, where the variability of wind stress and/or the buoyancy forcing is suppressed on interannual time scales, are used to unravel the mechanisms driving ventilation changes. Ventilation changes are estimated by comparing CFC-12 interior inventories among the different experiments. All simulations suggest a multidecadal fluctuation of Southern Ocean ventilation, with a decrease until the 1980s–90s and a subsequent increase. This evolution is related to the atmospheric forcing and is caused by the (often counteracting) effects of wind stress and buoyancy forcing. Until the 1980s, increased buoyancy gains caused the ventilation decrease, whereas the subsequent ventilation increase was driven by strengthened wind stress causing deeper mixed layers and a stronger meridional overturning circulation. Wind stress emerges as the main driver of ventilation changes, even though buoyancy forcing modulates its trend and decadal variability. The three Southern Ocean basins take up CFC-12 in distinct density intervals but overall respond similarly to the atmospheric forcing. This study suggests that Southern Ocean ventilation is expected to increase as long as the effect of increasing Southern Hemisphere wind stress overwhelms that of increased stratification.

Description: This contribution to the RECCAP2 (REgional Carbon Cycle Assessment and Processes) assessment analyzes the processes that determine the global ocean carbon sink, and its trends and variability over the period 1985-2018, using a combination of models and observation-based products. The mean sea-air CO2 flux from 1985 to 2018 is -1.6 +/- 0.2 PgC yr(-1) based on an ensemble of reconstructions of the history of sea surface pCO(2) (pCO(2) products). Models indicate that the dominant component of this flux is the net oceanic uptake of anthropogenic CO2, which is estimated at -2.1 +/- 0.3 PgC yr(-1) by an ensemble of ocean biogeochemical models, and -2.4 +/- 0.1 PgC yr(-1) by two ocean circulation inverse models. The ocean also degasses about 0.65 +/- 0.3 PgC yr(-1) of terrestrially derived CO2, but this process is not fully resolved by any of the models used here. From 2001 to 2018, the pCO2 products reconstruct a trend in the ocean carbon sink of -0.61 +/- 0.12 PgC yr(-1) decade(-1), while biogeochemical models and inverse models diagnose an anthropogenic CO2-driven trend of -0.34 +/- 0.06 and -0.41 +/- 0.03 PgC yr(-1) decade(-1), respectively. This implies a climate-forced acceleration of the ocean carbon sink in recent decades, but there are still large uncertainties on the magnitude and cause of this trend. The interannual to decadal variability of the global carbon sink is mainly driven by climate variability, with the climate-driven variability exceeding the CO2-forced variability by 2-3 times. These results suggest that anthropogenic CO2 dominates the ocean CO2 sink, while climate-driven variability is potentially large but highly uncertain and not consistently captured across different methods.

Description: In the framework of a changing climate, it is useful to devise methods capable of effectively assessing and monitoring the changing landscape of air-sea CO2 fluxes. In this study, we developed an integrated machine learning tool to objectively classify and track marine carbon biomes under seasonally and interannually changing environmental conditions. The tool was applied to the monthly output of a global ocean biogeochemistry model at 0.25° resolution run under atmospheric forcing for the period 1958–2018. Carbon biomes are defined as regions having consistent relations between surface CO2 fugacity (fCO2) and its main drivers (temperature, dissolved inorganic carbon, alkalinity). We detected carbon biomes by using an agglomerative hierarchical clustering (HC) methodology applied to spatial target-driver relationships, whereby a novel adaptive approach to cut the HC dendrogram based on the compactness and similarity of the clusters was employed. Based only on the spatial variability of the target-driver relationships and with no prior knowledge on the cluster location, we were able to detect well-defined and geographically meaningful carbon biomes. A deep learning model was constructed to track the seasonal and interannual evolution of the carbon biomes, wherein a feed-forward neural network was trained to assign labels to detected biomes. We find that the area covered by the carbon biomes responds robustly to seasonal variations in environmental conditions. A seasonal alternation between different biomes is observed over the North Atlantic and Southern Ocean. Long-term trends in biome coverage over the 1958–2018 period, namely a 10 % expansion of the subtropical biome in the North Atlantic and a 10 % expansion of the subpolar biome in the Southern Ocean, are suggestive of long-term climate shifts. Our approach thus provides a framework that can facilitate the monitoring of the impacts of climate change on the ocean carbon cycle and the evaluation of carbon cycle projections across Earth System Models.