Search for: All records

This study utilizes Deep Neural Networks (DNN) to improve the K‐Profile Parameterization (KPP) for the vertical mixing effects in the ocean's surface boundary layer turbulence. The deep neural networks were trained using 11‐year turbulence‐resolving solutions, obtained by running a large eddy simulation model for Ocean Station Papa, to predict the turbulence velocity scale coefficient and unresolved shear coefficient in the KPP. The DNN‐augmented KPP schemes (KPP_DNN) have been implemented in the General Ocean Turbulence Model (GOTM). The KPP_DNN is stable for long‐term integration and more efficient than existing variants of KPP schemes with wave effects. Three different KPP_DNN schemes, each differing in their input and output variables, have been developed and trained. The performance of models utilizing the KPP_DNN schemes is compared to those employing traditional deterministic first‐order and second‐moment closure turbulent mixing parameterizations. Solution comparisons indicate that the simulated mixed layer becomes cooler and deeper when wave effects are included in parameterizations, aligning closer with observations. In the KPP framework, the velocity scale of unresolved shear, which is used to calculate ocean surface boundary layer depth, has a greater impact on the simulated mixed layer than the magnitude of diffusivity does. In the KPP_DNN, unresolved shear depends not only on wave forcing, but also on the mixed layer depth and buoyancy forcing.

 

In contrast to the large volume of studies on the impact of horizontal resolution in oceanic general circulation models (OGCMs), the impact of vertical resolution has been largely overlooked and there is no consensus on how one should construct the vertical grid to represent the vertical structure of the baroclinic modes as well as the distribution of distinct water masses throughout the global ocean. In this paper, we document the importance of vertical resolution in the representations of vertical modes and water masses in the North Atlantic and show i) that vertical resolution is unlikely to undermine the resolution capability of the horizontal grid in representing the vertical modes and a 32-layer isopycnal configuration is adequate to represent the first five baroclinic modes in mid-latitudes and ii) that vertical resolution should focus on representing water masses. A coarse vertical resolution (16-layer) simulation exhibits virtually no transport in the dense overflow water which leads to a weaker and significantly shallower Atlantic meridional overturning circulation (AMOC) despite resolving the first baroclinic mode throughout the domain, whereas there are overall very small differences in the subtropical and subpolar North Atlantic circulation in the simulations with finer vertical resolution (24 to 96 layers). We argue that accurately representing the water masses is more important than representing the baroclinic modes for an OGCM in modeling the low-frequency large-scale circulation. 

Abstract. The ocean mixed layer is the interface between the ocean interior and the atmosphere or sea ice and plays a key role in climate variability. It isthus critical that numerical models used in climate studies are capable of a good representation of the mixed layer, especially its depth. Here weevaluate the mixed-layer depth (MLD) in six pairs of non-eddying (1∘ grid spacing) and eddy-rich (up to 1/16∘) models from theOcean Model Intercomparison Project (OMIP), forced by a common atmospheric state. For model evaluation, we use an updated MLD dataset computed fromobservations using the OMIP protocol (a constant density threshold). In winter, low-resolution models exhibit large biases in the deep-waterformation regions. These biases are reduced in eddy-rich models but not uniformly across models and regions. The improvement is most noticeable inthe mode-water formation regions of the Northern Hemisphere. Results in the Southern Ocean are more contrasted, with biases of either sign remainingat high resolution. In eddy-rich models, mesoscale eddies control the spatial variability in MLD in winter. Contrary to a hypothesis that thedeepening of the mixed layer in anticyclones would make the MLD larger globally, eddy-rich models tend to have a shallower mixed layer at mostlatitudes than coarser models do. In addition, our study highlights the sensitivity of the MLD computation to the choice of a reference level andthe spatio-temporal sampling, which motivates new recommendations for MLD computation in future model intercomparison projects.

 

The potential role of the New England seamount chain (NESC) on the Gulf Stream pathway and variability has been long recognized, and the series of numerical experiments presented in this paper further emphasize the importance of properly resolving the NESC when modeling the Gulf Stream. The NESC has a strong impact on the Gulf Stream pathway and variability, as demonstrated by comparison experiments with and without the NESC. With the NESC removed from the model bathymetry, the Gulf Stream remains a stable coherent jet much farther east than in the experiment with the NESC. The NESC is the leading factor destabilizing the Gulf Stream and, when it is not properly resolved by the model’s grid, its impact on the Gulf Stream’s pathway and variability is surprisingly large. A high-resolution bathymetry, which better resolves the New England seamounts (i.e., narrower and rising higher in the water column), leads to a tighter Gulf Stream mean path that better agrees with the observed path and a sea surface height variability distribution that is in excellent agreement with the observations.

 

Abstract Atlantic bluefin tuna (ABT) (Thunnus thynnus) travel long distances to spawn in oligotrophic regions of the Gulf of Mexico (GoM) which suggests these regions offer some unique benefit to offspring survival. To better understand how larval survival varies within the GoM a spatially explicit, Lagrangian, individual-based model was developed that simulates dispersal and mortality of ABT early life stages within realistic predator and prey fields during the spawning periods from 1993 to 2012. The model estimates that starvation is the largest cumulative source of mortality associated with an early critical period. However, elevated predation on older larvae is identified as the main factor limiting survival to late postflexion. As a result, first-feeding larvae have higher survival on the shelf where food is abundant, whereas older larvae have higher survival in the open ocean with fewer predators, making the shelf break an optimal spawning area. The modeling framework developed in this study explicitly simulates both physical and biological factors that impact larval survival and hence could be used to support ecosystem based management efforts for ABT under current and future climate conditions. 

Abstract. With the increase in computational power, ocean models with kilometer-scale resolution have emerged over the last decade. These models have been used for quantifying the energetic exchanges between spatial scales, informing the design of eddy parametrizations, and preparing observing networks. The increase in resolution, however, has drastically increased the size of model outputs, making it difficult to transfer and analyze the data. It remains, nonetheless, of primary importance to assess more systematically the realism of these models. Here, we showcase a cloud-based analysis framework proposed by the Pangeo project that aims to tackle such distribution and analysis challenges. We analyze the output of eight submesoscale-permitting simulations, all on the cloud, for a crossover region of the upcoming Surface Water and Ocean Topography (SWOT) altimeter mission near the Gulf Stream separation. The cloud-based analysis framework (i) minimizes the cost of duplicating and storing ghost copies of data and (ii) allows for seamless sharing of analysis results amongst collaborators. We describe the framework and provide example analyses (e.g., sea-surface height variability, submesoscale vertical buoyancy fluxes, and comparison to predictions from the mixed-layer instability parametrization). Basin- to global-scale, submesoscale-permitting models are still at their early stage of development; their cost and carbon footprints are also rather large. It would, therefore, benefit the community to document the different model configurations for future best practices. We also argue that an emphasis on data analysis strategies would be crucial for improving the models themselves. 

Accurate representation of air‐sea interaction is crucial to numerical prediction of the ocean, weather, and climate. Sea surface temperature (SST) gradients and surface currents in the oceanic mesoscale regime are known to have significant influence on air‐sea fluxes of momentum. Studies based on high‐resolution numerical models and observations reveal that SST gradients and surface currents in the submesoscale regime are much stronger than those in the mesoscale. However, the feedback between the submesoscale processes and the air‐sea turbulent fluxes is not well understood. To quantitatively assess the responses between air‐sea flux of momentum and submesoscale processes, a non‐hydrostatic ocean model is implemented in this study. The inclusion of SST gradients and surface currents in air‐sea bulk fluxes are argued to be significant for modeling accurate wind stress in the submesoscale regime. Taking both into account, this study shows that the linear relationship between wind stress curl/divergence and crosswind/downwind SST gradients existing in the mesoscale regime is not obvious in the submesoscale. Instead, a linear relationship between wind stress curl/divergence and surface current curl/divergence is revealed in the submesoscale. Furthermore, the magnitude of wind stress curl introduced by submesoscale processes is much greater than that presented by mesoscale processes. Another key finding is that tracer subduction and potential vorticity distribution in the submesoscale is susceptible to submesoscale‐modified air‐sea turbulent momentum flux. This study serves as a starting point in investigating the feedbacks between atmospheric and oceanic submesoscale processes.

 

Abstract Purpose of Review Assessment of the impact of ocean resolution in Earth System models on the mean state, variability, and future projections and discussion of prospects for improved parameterisations to represent the ocean mesoscale. Recent Findings The majority of centres participating in CMIP6 employ ocean components with resolutions of about 1 degree in their full Earth System models (eddy-parameterising models). In contrast, there are also models submitted to CMIP6 (both DECK and HighResMIP) that employ ocean components of approximately 1/4 degree and 1/10 degree (eddy-present and eddy-rich models). Evidence to date suggests that whether the ocean mesoscale is explicitly represented or parameterised affects not only the mean state of the ocean but also the climate variability and the future climate response, particularly in terms of the Atlantic meridional overturning circulation (AMOC) and the Southern Ocean. Recent developments in scale-aware parameterisations of the mesoscale are being developed and will be included in future Earth System models. Summary Although the choice of ocean resolution in Earth System models will always be limited by computational considerations, for the foreseeable future, this choice is likely to affect projections of climate variability and change as well as other aspects of the Earth System. Future Earth System models will be able to choose increased ocean resolution and/or improved parameterisation of processes to capture physical processes with greater fidelity. 

Uncertainty exists in the time‐mean total transport of the Antarctic Circumpolar Current (ACC), the world's strongest ocean current. The two most recent observational programs in Drake Passage, DRAKE and cDrake, yielded transports of 141 and 173.3 Sv, respectively. In this paper, we use a realistic 1/12° global ocean simulation to interpret these observational estimates and reconcile their differences. We first show that the modeled ACC transport in the upper 1,000 m is in excellent agreement with repeat shipboard acoustic Doppler current profiler (SADCP) transects and that the exponentially decaying transport profile in the model is consistent with the profile derived from repeat hydrographic data. By further comparing the model results to the cDrake and DRAKE observations, we argue that the modeled 157.3 Sv transport, that is, approximately the average of the cDrake and DRAKE estimates, is actually representative of the time‐mean ACC transport through the Drake Passage. The cDrake experiment overestimated the barotropic contribution in part because the array undersampled the deep recirculation southwest of the Shackleton Fracture Zone, whereas the surface geostrophic currents used in the DRAKE estimate yielded a weaker near‐surface transport than implied by the SADCP data. We also find that the modeled baroclinic and barotropic transports are not correlated; thus, monitoring either baroclinic or barotropic transport alone may be insufficient to assess the temporal variability of the total ACC transport.