Numerical Modelling Team
- Entcho Demirov, Memorial University
- Brad de Young, Memorial University <
- Morris Flynn, University of Alberta
- Eric Galbraith, McGill University
- Jaime Palter, McGill University
- Andrea Scott, University of Waterloo
- Bruce Sutherland, University of Alberta
- To improve the ability of high-resolution numerical models to represent gas cycles through their controlling processes and their sensitivity to climatic change.
- To provide an understanding of the climate sensitivity of the "breathing", convection, re-stratification, and biogeochemical process studied through the observational program.
- To offer guidance to the observational program in the form of Observing System Simulation Experiments (OSSEs);
- To improve model representations of key physical and biogeochemical processes in a region of deep water formation; and
- To contextualize the results of the observational work in a several-decade simulation of the entire North Atlantic.
Numerical models provide comprehensive output, which can link the spatial and temporal scales relevant to the Labrador Sea. The central mooring will provide high resolution measurements (in depth and time) that will be used to validate the model's ability to represent the seasonal cycle, mixed layer and stratification evolution, and gas exchange with the deep ocean. The additional moorings, floats and gliders of the horizontal exchange theme will test the model's representation (and parameterizations) of key exchange processes linking the physics, biology and air-sea gas exchange, with the models representing the entire Labrador Sea. This will allow for quantitative estimates based on comparisons at the central mooring to the broader basin. Lateral exchange between the interior of the Labrador Sea, the boundary currents, and the larger ocean will be considered, including the key role of freshwater processes. The models will use the observational program measurements, as well as the previous DFO monitoring work (analyzed by the historical analysis team) to place the present in the context of past variability, as well as address questions of future functioning, and climate sensitivity/vulnerability.
The underlying physical model will be the Nucleus for European Modelling of the Ocean (NEMO) numerical framework version 3.4 (Madec, 2008). It is a coupled ocean/sea ice model, including a three-dimensional, free surface, hydrostatic, primitive-equation ocean general circulation model and a dynamic-thermodynamic sea-ice model (Louvain-la-Neuve: LIM2) (Fichefet and Maqueda 1997), with an elastic-viscous-plastic (EVP) ice rheology (Hunke and Dukowicz 1997).
High Resolution Modelling
To address questions of frontal dynamics, fresh water plumes and pathways, the impacts of synoptic weather events, the manner in which stratifying eddies impact primary and export production, as well as to provide fields to drive biogeochemical simulations, high resolution eddy-resolving models (min. 1/12 degree) are needed (e.g. Chanut et al. 2008; Gelderloos et al. 2011; McGeehan and Maslowski 2011). Thus, for the key process and sensitivity studies linking the modelling to the observations, we will use a 1/12 degree configuration (ANHA12) developed by Environment Canada and DFO as part of the inter-departmental CONCEPTS initiative for an integrated marine Arctic prediction system in support of METAREA monitoring and warnings. All knowledge gained during this project will be transferrable back to the operational development ongoing within the Federal Government departments. Additionally, because gas uptake and biological processes depend critically on the mixed layer and stratification, we will use a new high-resolution atmospheric forcing data set provided by Environment Canada and designed specifically to produce better winds for forcing sea-ice and ocean models. Specific questions will include the role and impact of improved air-sea forcing on mixed-layer evolution, stratification, gas uptake and exchange on/off the shelves. We will also categorize the location, variability, key scales, and driving mechanisms of the main exchange between the boundary currents and the interior of the basin. Finally, we will investigate the relationship to mesoscale eddies and their spatial and temporal variability, and the subsequent impact of the eddies on the transport of tracers (especially freshwater), momentum, heat, and dissolved gases.
As computational cost will limit the number of experiments (and integration length) we can perform with ANHA12, we will use a complementary 1/4 degree configuration (ANHA4) to carry out extended hind-casts (1960 to present) and climate change scenarios (2000-2100). The goal of these simulations is to understand the evolution of convection, stratification, and gas uptake in the interior of the Labrador Sea with regard to climatic evolution of the arctic/sub-arctic system (CAA throughflow, Greenland melt, etc.), as well as the sub-tropics (i.e. impact of North Atlantic Current, changes in sub-polar front, etc.). The model for the climate scenarios will be validated by comparison of the hind-cast with the historical analyses carried out by the Historical Analyses Team, especially with regard to variations in velocities, transports, and freshwater.
Observing System Simulation Experiment(s) and Data Assimilation
Observing system simulation experiments (OSSEs) can simulate the impact of a proposed data type on a forecasting system (Arnold and Dey 1986) and consist of a high-resolution model run sampled to create simulated observations. Typically, these simulated observations are examined for information they contain and assimilated into a model to investigate their impact on the forecast. An OSSE can also be used to examine the ability of an observation to constrain a coarse resolution model, an incorrect model parameter (Melet et al. 2012), or a model that has been inaccurately perturbed (eg. given an incorrect freshwater input). We will construct an OSSE for float and glider deployment and data assimilation. The workplan is centered on the following questions: What is the optimal float and glider deployment strategy to measure thermohaline properties, with the objective of estimating lateral fluxes? How well can float and glider data constrain model predictions? How will float and glider data impact model simulation of thermohaline properties and lateral fluxes in a real assimilation experiment? How can they best complement the other observation platforms? Can observational data be used to constrain unknown parameters that influence long term/climate simulations?
We will couple the physical configuration with a biogeochemical model for simulation of the exchange of gases with the deep ocean. The Biogeochemistry with Light Iron Nutrient and Gas (BLING - Galbraith et al. 2010) module uses a minimal number of explicit tracers, but behaves according to a relatively complex set of nonlinear functions that approximate well-established ecosystem functionality. This results in a streamlined codebase that is relatively easy to understand and manipulate, while reproducing many of the best-constrained aspects of biogeochemical cycling, such as variable export ratios and photo-adaptation. Because of the small number of tracers required, it is easy to introduce novel tracers to the model, such as stable isotopes, unusual dissolved gases, and tracers for saturation and preformed dissolved inorganic carbon (useful for assessing physical and biological contributions to carbon storage). Recent development has focused on extending the model to upper trophic levels, using the simple alometric scalings of metabolic ecology theory, and including the surprisingly large effect of vertically migrating animals on respiration within the water column. Key activities will include embedding BLING in NEMO; examining respiration rates within the water column and their spatial variations in the water column; improving gas-exchange with the atmosphere through the use of novel gas tracers (e.g. Ar, N2) and the parameterization of bubble injection.
Filling Box Models
Even coarse-grid ocean/sea-ice algorithms require substantial resources; thus it is helpful to develop a low-order model capable of quickly scanning the parameter space. The model in question will be based on the "filling-box" equations (Manins 1979; Kaye et al. 2010), which describe convection in an open or closed basin. One-dimensional equations specifying the vertical variation of the volume, momentum and buoyancy flux of a descending turbulent plume will be solved and the evolution of the ambient stratification will be predicted. Filling-box models have been corroborated by comparison against similitude laboratory experiments. Hughes and Griffiths (2006) showed that a filling-box model can realistically capture the essential dynamics of an overturning circulation. We thereby plan to identify circumstances wherein atypical vertical transport may arise. These regions of parameter space may then be explored at greater resolution using more detailed (i.e. 1/4 or 1/12 degree) numerical models.
Testing the NS-alpha Model for a Realistic Ocean Problem
This work is complementary to the proposed high resolution and eddy-permitting modeling studies. The objective is to evaluate the performance of the NS-alpha model using different specifications of the model parameter. The NS-alpha model produces strongly eddying solutions at coarse grid resolution (e.g. Hecht et al. 2008; Scott and Lien, 2010a). It is promising for oceanographic applications because the equations allow eddy-generating (baroclinic) instabilities to occur at low wave-numbers (Hecht 2008). Implementation has been hindered by problems (Matthew Hecht, pers. comm.) similar to those found by Scott and Lien (2010b), where they were resolved by allowing the model parameter to change with the flow. The current project provides an excellent opportunity to test the model on a realistic test case in which eddy activity is a critical component and coarse resolution a severe constraint, and carry out validation through comparison of results from ANHA4 with both i) high resolution eddy resolving simulations (ANHA12), and ii) oceanographic observations. Summary: We will test models against long-term data from the DFO monitoring program, and against the high-frequency data revealed by the central mooring array coupled with gliders. We will couple the physical models to biophysical and biogeochemical ones and explore the spatial and process dynamics of the system. We will work to get consistent rates of CO2 and oxygen uptake between eddy-resolving numerical models and those using resolutions typical of climate models and thus improve short-term (sub-decadal), as well as longer-term, climate predictability.
Important aspects of model development and validation will be carried out, including producing a high resolution configuration of the NEMO model able to simulate accurately key aspects of the Labrador Sea, and that incorporates a new and streamlined biogeochemical model, BLING. These models will be used to help understand the stratification, mixed-layer, frontal and eddy processes focused on by the observational campaign. The modelling will place these results in a context related to the past using hind-cast simulations and consider future evolution with climate change scenarios. We will examine parameterizing the key results so that they can be represented in coarser resolution coupled climate models, through the use of filling box models and the NS alpha model.