Geoffrey Gebbie

Paleoceanographic reconstructions indicate a large‐scale cooling in the last millennium, but significant disagreement on the spatial pattern and timing still exists. This disagreement is partially due to the inference of surface conditions from sparse, noisy, near‐surface marine fauna that exhibit seasonal and habitat depth biases. Here, we infer spatially‐varying surface changes from benthic sediment core records by back‐propagation through an ocean circulation model, improving upon prior methodologies that inferred regional surface evolution from subsurface records. We present a surface temperature and history that fits 11 high‐resolution benthic records within error and also supports previous independent estimates from Common Era marine surface records. Our methodology accounts for the opposing and covarying behavior of temperature and on , as well as age model uncertainty. Our estimate is characterized by a high‐latitude (N) North Atlantic‐centered Little Ice Age sea surface temperature (SST) cooling of 0.38 0.13 K (1 s.e.) from 1140 to 1880 CE, which is within error of planktonic SST record compilations and exceeds the cooling exhibited in the Last Millennium Reanalysis. We improve the detection of long‐term climate signals by using benthic sediment core records which record the properties of abyssal waters that are not subject to high‐energy near‐surface variability, instead of planktonic sediment core records which tend to have more noise. These results are consistent with the global SST trends inferred from the Ocean2k planktonic SST compilation and corroborate robust Little Ice Age cooling in the North Atlantic, while providing further insights on its spatial pattern.

A comparison of 19th century HMS Challenger hydrography with modern observations indicated the mid‐depth Pacific has cooled at a rate of centikelvin (cK, ) per century and could be explained as a slow ongoing adjustment to the Little Ice Age. The historical hydrographic data was sparse, however, and limited by 19th century technology. Observations from recent decades are more plentiful, with times more observations are used to constrain the dynamically consistent ocean state estimate produced by the Estimating the Circulation and Climate of the Ocean (ECCO) Consortium, for example. We find the ECCO mid‐depth Pacific cooling trend is similar in magnitude (5 cK/century) and spatial structure to HMS Challenger data. Additionally, cooling in the mid‐depth North Pacific only emerges after the model is optimized to observations. Here, sensitivity experiments are used to isolate the effects of data‐constrained model parameters on temperature trends in this region. The optimized wind stress achieves most (3 cK/century) of the cooling in ECCO through anomalous vertical transport and adiabatic isopycnal upwelling. Optimized initial conditions and mixing coefficients play a less important role, causing cooling rates of 2 cK/century and 1 cK/century, respectively. Furthermore, the ECCO heat budget is found to be kinematically similar to that expected from a centuries‐long simulation of pre‐industrial anomalies being transported into the ocean interior. Thus, mid‐depth cooling in ECCO can be physically interpreted and is within the dynamical expectations of an ocean out of equilibrium with the atmosphere.

Changes in the hydrological cycle can affect ocean circulation and ventilation. Freshwater enters the ocean as meteoric water (MW; precipitation, river runoff, and glacial discharge) and sea ice meltwater (SIM). These inputs are traced using seawater salinity and stable oxygen isotopes in seawater, . We apply a self‐organizing map, a machine learning technique, to water mass properties to estimate the global distribution of the isotopic signature of MW by characterizing distinct salinity‐ relationships from two comprehensive data sets. The inferred is then used in a three‐endmember mixing model to provide a globally coherent MW and SIM contributions to the extratropical ocean freshwater budget. Through the use of , our results show the role of MW and SIM in dense water formation and the resulting interhemispheric asymmetry in the freshwater sources that fill the interior ocean freshwater budget. Trends drawn in ‐S space show a significant decrease in sea ice formation driving the freshening of Antarctic bottom water for the 1980–2023 period, whereas SIM is significantly increasing in parts of the Arctic halocline. The different roles of sea ice in dense water formation has implications for future ocean circulation under climate change, where machine learning techniques applied to have been proven to have utility in detecting such changes.

Ocean ventilation, or the transfer of tracers from the surface boundary layer into the ocean interior, is a critical process in biogeochemical cycles and the climate system. Here, we assess steady‐state ventilation patterns and timescales in three models of ocean transport: a 1 global configuration of the Nucleus for European Modeling of the Ocean (NEMO), a recent 2 solution of the Ocean Circulation Inverse Model (OCIM), and a 2 solution of the Total Matrix Intercomparison (TMI). We release artificial dyes in six surface regions of each model and compare equilibrium dye distributions as well as ideal age distributions. We find good qualitative agreement in large‐scale dye distributions across the three models. However, the distributions indicate that TMI and OCIM are more diffusive than NEMO. A shallow bias of North Atlantic ventilation in NEMO contributes to a stronger presence of the North Atlantic dye in the mid‐depth Southern Ocean and Pacific. This isopycnal communication between the North Atlantic surface and the mid‐depth Pacific is very slow, however, and NEMO simulates a maximum age in the North Pacific (NP) about 900 years higher than the data‐constrained models. Overly slow NP ventilation persists across NEMO sensitivity experiments encompassing our current best knowledge of diapycnal and isopycnal mixing, pointing to biases in subarctic Pacific dynamics. This study provides a synoptic picture of deep ocean ventilation and a framework for assessing its representation in general circulation models.

Discrepancies in historical global mean surface temperature (GMST) estimates largely stem from differences in bias corrections applied to sea surface temperature (SST) records. Here, using the amplitude of the diurnal cycle in SST, we provide evidence that wooden‐to‐canvas bucket transitions were mostly complete by the early 1900s, earlier than commonly assumed by two decades, resulting in strong early 20th century cold biases. We then use this diurnal evidence, together with coral O and Sr/Ca proxies and expected GMST responses to external radiative forcing, to evaluate four different SST estimates: HadSST4, ERSST5, COBESST2, and DCSST. Of these, DCSST, developed by our team through adjustment of SSTs to match coastal land surface air temperatures, shows the closest overall agreement with all three lines of evidence. DCSST features a larger and steadier warming since the 1900s relative to the other estimates and indicates somewhat higher transient climate sensitivity and smaller decadal variability.

Understanding advective‐diffusive dispersal of trace substances in environmental fluids like the global ocean is a ubiquitous challenge in geophysics. Since the turn of the millennium, substantial progress has been made in the theory, implementation in models, and application of such tracers in oceanography. For the first time, this progress is reviewed here in a synthetic way. We focus on tracer techniques in ocean models, including real and virtual tracers that diagnose timescale information, and we emphasize the connection to the Green's function that solves the advection‐diffusion equation. Implementation of these techniques in ocean models is explained in an accessible way. We present example applications of these techniques to questions concerning ocean circulation, transport of biogeochemicals, and paleoceanography, including future opportunities.

Over the last decades, back‐calculation (BC) techniques for ocean anthropogenic carbon (C_ant) estimation have improved and evolved into different methodologies that are not exempt from various assumptions and limitations. No single optimal BC method exists to date for computing C_ant; therefore, it is necessary to continue advancing the broad range of approaches. Here, we present a novel method based on the BC fundamentals that combines marine‐carbonate‐system (MCS) data and the Total Matrix Intercomparison (TMI) framework. This MCS‐TMI approach differs from other BC methods by using the TMI to reconstruct deep‐ocean biogeochemical properties and their preformed conditions. It also incorporates a global sea‐air oxygen disequilibrium term, and a dynamic stoichiometric carbon‐to‐oxygen ratio that depends on the water‐mass ideal time. The MCS‐TMI yields a total C_ant inventory of 124 ± 7 Pg C (referred to 1995), in good agreement with previous global C_ant climatologies. The MCS‐TMI method uncertainty (±5.6 μmol kg⁻¹) is controlled by input‐data errors that, nonetheless, have a minimal impact on the total C_ant inventory. In contrast, our total C_ant inventory uncertainty is governed by methodological errors, specifically those related to the TMI's boundary conditions. Our study demonstrates the effectiveness of MCS data‐based climatologies in reconstructing a 3D gridded C_ant climatology, and the validity of ocean circulation transport operators for obtaining BC preformed conditions.

Ocean ventilation, or the transfer of tracers from the surface boundary layer into the ocean interior, is a critical process in biogeochemical cycles and the climate system. Here, we assess steady-state ventilation patterns and timescales in three models of ocean transport: a 1º global configuration of the Nucleus for European Modelling of the Ocean (NEMO), a recent 2º solution of the Ocean Circulation Inverse Model (OCIM), and a 2º solution of the Total Matrix Intercomparison (TMI). We release artificial dyes in six surface regions of each model and compare equilibrium dye distributions as well as ideal age distributions. We find good qualitative agreement in large-scale dye distributions across the three models. However, the distributions indicate that TMI and OCIM are more diffusive than NEMO. A shallow bias of North Atlantic ventilation in NEMO contributes to a stronger presence of the North Atlantic dye in the mid-depth Southern Ocean and Pacific. This isopycnal communication between the North Atlantic surface and the mid-depth Pacific is very slow, however, and NEMO simulates a maximum age in the North Pacific about 900 years higher than the data-constrained models. Overly slow North Pacific ventilation persists across NEMO sensitivity experiments encompassing our current best knowledge of diapycnal and isopycnal mixing, pointing to biases in subarctic Pacific dynamics. This study provides a synoptic picture of global ocean ventilation and a framework for assessing its representation in general circulation models.

Understanding advective-diffusive dispersal of trace substances in environmental fluids like the global ocean is a ubiquitous challenge in geophysics. Since the turn of the millennium, substantial progress has been made in the theory, implementation in models, and application of such tracers in oceanography. For the first time, this progress is reviewed here in a synthetic way. We focus on tracer techniques in ocean models, including naturally-occurring and hypothetical tracers that diagnose timescale information, and we emphasize the connection to the Green's function that solves the advection-diffusion equation. Implementation of these techniques in ocean models is explained in an accessible way. We present example applications of these techniques to questions concerning ocean circulation, transport of biogeochemicals, and paleooceanography, including future opportunities.

Understanding advective-diffusive dispersal of trace substances in environmental fluids like the global ocean is a ubiquitous challenge in geophysics. Since the turn of the millennium, substantial progress has been made in the theory, implementation in models, and application of such tracers in oceanography. For the first time, this progress is reviewed here in a synthetic way. We focus on tracer techniques in ocean models, including naturally-occurring and hypothetical tracers that diagnose timescale information, and we emphasize the connection to the Green's function that solves the advection-diffusion equation. Implementation of these techniques in ocean models is explained in an accessible way. We present example applications of these techniques to questions concerning ocean circulation, transport of biogeochemicals, and paleooceanography, including future opportunities.