AOFVCOM Development A System Nested with Global Ocean

AO-FVCOM Development: A System Nested with Global Ocean Models Changsheng Chen University of Massachusetts School of Marine Science, USA Contributors: Guoping Gao, Zhigang Lai, Yu Zhang (UMASSD) Robert C. Beardsley and Andrey Proshutinsky (WHOI) http: //fvcom. smast. umassd. edu/

Outline 1. Review of critical needs for a model to resolve multi-scale processes over the Arctic Ocean; 2. Updated development of unstructured grid Finite-Volume Community Ocean Model system in the Arctic Ocean (AOFVCOM); 3. Examples of applications of AO-FVCOM nested with Global. FVCOM; 1. Procedure of the long-term simulation for 1978 -present; 2. Summary

Two Critical Issues: Basin-shelf interaction Basin scale 1. Multi-scale dynamics: Basinshelf interaction, convection via advection, etc. 2. Open boundary connected to the North Atlantic Ocean and Pacific Ocean Coastal process Ne ste db ou nd ary

Ocean Model Dynamics ~1 ~1000 km Hydrostatic Large-scale motion in which the vertical motion is at least one order of magnitude smaller than the horizontal motion. Vertical convection, over -turning, and high frequency internal waves are not resolved. ~ a few meters Non-hydrostatic Small-scale motion in which vertical motion is the same order of the horizontal motion. Vertical convection, over -turning, and high frequency internal waves can be resolved.

In the ocean, Numerical error speed Small Domain Large Domain Because Δxs ≠ Δx. L; Δys ≠ Δy. L; The surface gravity wave speed propagating from the small domain is not equal at the nesting boundary. Energy accumulation at the boundary !

Modules of FVCOM 3. 1 Nudging/OI Assimilation Ensemble/Reduced Kalman Filters North Pole Nested System Model Field Sampling 3 -D Wet/Dry Treatment Forcings: Tides (Equilibrium+ O. B. ) Winds, Heat flux, Precipitation/Evaporation River discharges, Groundwater O. B. fluxes 3 -D Sediment Model FVCOM-Main Code Cartesian/Spherical Coordinates MPI Parallel Surface Wave Model Multiple Nesting Non-hydrostatic Existing Modules Generalized Biological Model Water Quality Models Adjoint Assimilation Key: General Ocean Turbulence Model (GOTM) Multi OB Radiations Vi. Si. T Monitoring Net. CDF Output GUI Post-process Tools Under Development Solver: Mode-split or semi-implicit; 2 -D and 3 -D Lagrangian-IBM Ice model

Unstructured grid Common boundary Non-hydrostatic process Unstructured nesting approach: Mass conservation

One-way Nesting (implemented) 1. The nesting boundary consists of the boundary nodes and triangles connected to the boundary nodes; 2. The model output includes all variables at boundary nodes and velocities in the triangle cells 3. A nesting file with inclusion of boundary node and cell index is pre-defined when the nesting approach is used 4. The nested domain model runs with the nesting file as the boundary forcing Disadvantage: 1. Two meshes are required for master and nested domains; 2. The operation is just like running two models.

Two-way Nesting (under development) Patched grid:

Two-way Nesting Main domain Local subdomain Main domain The main domain uses the interior meshes of the local subdomain as the boundary, while the local subdomain uses the interior meshes of the main domain as the boundary.

Global-FVCOM (5 -50 km) Nested AO-FVCOM (0. 5 -50 km) The AO-FVCOM system nested with Global-FVCOM Canadian Archipelago (0. 5 -1. 0 km)

Examples of Global-FVCOM Simulation Results

Observed (Averaged over 1979 -94) AO-FVCOM/UG-CICE The sea ice concentration March September 9/21/10 Chen, Gao (UMASSD), Beardsley (WHOI)

The NSIDC data show the averaged drifting velocity over 1979 -94 FVCOM model results were obtained with the climatologic forcing condition averaged over 1979 -94 March 9/21/10 September Chen, Gao (UMASSD), Beardsley (WHOI)

Examination of the impact of horizontal resolution on currents Coarse grid (10 -50 km) Finer grid (1. 0 -25 km)

Monthly averaged currents at 400 m (summer) Coarse grid

Monthly averaged currents at 400 m (summer) Finer grid

Annually averaged along-slope currents at 152 o. W around the Alaska coast Coarse Resolution 9/21/10 Finer Resolution Chen, Gao (UMASSD), Beardsley (WHOI)

Bering Strait, Chukchi Sea and Alaska coast Annual mean vertically averaged currents in the depths of 0 -50 m Coarse grid Transports across Bering Strait: 0. 8 Sv for the coarse grid case 1. 1 Sv for the finer grid case Finer grid

Canadian Archipelago Annual mean vertically averaged currents in the depths of 0 -50 m Coarse grid Finer grid

On-going Work 1. Run Global-FVCOM from 1978 -2010 with inclusion of data assimilation of SST, SSH, T/S profiles-compare the results for currents and ices. 2. AO-FVCOM nested with Global-FVCOM experiments, to examine the need of horizontal resolution to the Arctic Ocean 3. Conduct the process studies to study the ice-current interaction.

Gao’s thesis work: Which mechanism causes the variation of slope current and how can we improve the simulation of the slope current in the Arctic Ocean model? Slope currents vary to different external and internal forcing. Ice formation is one of the most significant seasonal signal in the Arctic Ocean. Its associated processes may play essential roles on variation of slope current. Q Ice cover Brine 1. Advection vs convection 2. Tide rectification and mixing 3. Wind forcing 4. Convection vs Baroclinic instability and eddy formation Eddy formation tide

Gao’s thesis work (Three-dimensional experiments) 1. In order to identify and understand the important contribution of convection and baroclinic instability processes to slope current, a 3 -D experiments will be carried out with a specified Arctic Ocean model through nesting approach in FVCOM. The dynamic processes in the intensify zone will be simulated with high resolution NH-FVCOM with nesting to the large scale Arctic Ocean model.

Summary 1. AO-FVCOM nested with Global-FVCOM provides a new model tool to examine multi-scale responses of the Arctic to climate change. 2. Comparisons between the coarse and fine resolution AO-FVCOM suggest a need for the high-resolution model to resolve the basinshelf interaction.