Southern California Bight Ocean Modeling Projects at CESR, UCLA


last modified: June 18, 2010



APPROACH


The regional oceanic responses in the Southern California Bight to the large-scale current system such as California Current, global remote forcing through the coastal wave guide alongshore, local and remote forcing (atmospheric forcing, tides and waves), local complexity in the topography and coastline, and intrinsic variabilities associated with mesoscale to submesoscale transition, have been investigated with comprehensive modeling frameworks. The key component to pursue this is the Regional Oceanic Modelling System (ROMS, Schepetkin and McWilliams, 2005; 2008) that is a horizontal curvilinear and vertical terrain-following coordinate, hydrostatic, incompressible, Boussinesq approximation, free-surface oceanic circulation model with non-conservative forcing, vertical and lateral diffusion, and bottom drag. It makes a baroclinic-barotropic mode split, with explicit fast time-stepping and subsequent conservative averaging of barotropic variables. Tremendous effort has been made to expand ROMS' capabilities to include non-hydrostatic dynamics (kanarska et al., 2007), surface gravity waves (Uchiyama et al., 2010), fundamental sediment transport dynamics (Blaas et al., 2007) and so on. The large-scale influences are taken into account through a nesting technique where multiple model domains are configured to realistically determine significant oceanic signals in many different scales. The actual nested model designs depends on what dynamics you would like to detect, and thus varies from project to project. One example being used for “Submesoscale dynamics” project is as follows.

The ROMS configuration consists of triply nested model domains with an off-line, one-way nesting technique that downscale from 5 km horizontal resolution of the United States west coast (L0), to 1 km resolution of the SCB (L1), to 250 m horizontal resolution of the Santa Monica and San Pedro Shelves (L2). Each domain has 40 bottom topography following levels vertically stretched such that grid cell refinement occurs near the surface and bottom. The model topographies are given by the 30-second global SRTM30 bathymetry (Becker et al., 2008: Marine Geodesy) in general, whereas the 3-second NOAA-NGDC coastal relief data set (http://www.ngdc.noaa.gov/mgg/coastal/crm.html) is used for the near-shore regions depending on data availability. The outer-most L0 is forced by the monthly-averaged SODA version 2.0.4, a POP-based assimilated global oceanic dataset (e.g. Carton et al., 1996: J. Geophys. Res.) as lateral boundary conditions, a monthly-averaged QuikSCAT-ECMWF blended wind data (http://cersat.ifremer.fr/data/discovery/by_product_type/gridded_products/mwf_blended) as a surface momentum stress, a monthly-average AVHRR pathfinder satellite SST (http://podaac.jpl.nasa.gov/DATA_PRODUCT/SST/index.html) and the COADS climatological dataset (http://www.ncdc.noaa.gov/oa/climate/coads/) for the other surface fluxes. On L0, the monthly climatology of runoff from major rivers (Dai and Trenberth, 2002: J. Hydrometeorol.) is taken into account. The intermediate L1 and the inner-most L2 are then driven by the corresponding parent ROMS model solutions with daily (L0 to L1) and 2-hourly (L1 to L2) lateral boundary update. All the surface boundary conditions for L1 and L2 are give by an hourly atmospheric forcing by a double-nested WRF model (e.g., Michalakes et al., 1998: In: Design of a next-generation regional weather research and forecast model : Towards Teracomputing, World Scientific) on 18- and 6-km horizontal grid spacings; the 6-km solution is used to force the ROMS L1 and L2 models with a one-way coupling approach. Tides are included in L1 and L2 with TPXO 7.1 global tidal prediction (e.g., Egbert et al., 1994: J. Geophys. Res.) to force L1 at the lateral boundaries with ten tidal major constituents (M2, S2, N2, K2, K1, O1, P1, Q1, Mf, and Mm) that synthetically provide free-surface elevation and barotropic velocity components at every barotropic time step by superposing on to the daily-averaged boundary conditions from the L0 run. Intrinsic three-dimensional tidal variabilities in L1 are then brought into the L2 run through the high-frequency, 2-hourly boundary update. For more details on the model configurations, one may wish to refer our journal articles, most of which are indicated at the end of each project descriptions.


An example of nested grid configurations used for the SCB projects. The figure shows bathrmetry along the U.S. West Coast and perimeters of triple nested ROMS model domains (a descendant of the ICC configuration used for “submesoscale dynamics” project). Blue box is the outer most L0 domain with a 5-km horizontal grid spacing, laterally forced by the SODA global data, black box: the intermediate L1 domain with a 1-km resolution downscaled from ROMS L0, and red box: the inner-most L2 domain with a 250-m resolution nested in ROMS L1.


Related publication




PROJECTS


1. Multi-year, multi-scale oceanic variabilities in SCB


The oceanic circulation in the Southern California Bight (SCB) is influenced by the large-scale California Current offshore, tropical remote forcing through the coastal wave guide alongshore, and local atmospheric forcing. The region is characterized by local complexity in the topography and coastline. All these factors engender variability in the circulation on interannual, seasonal, and intraseasonal time scales. This study applies the Regional Oceanic Modeling System (ROMS) to the SCB circulation and its multiple-scale variability. The model is configured in three levels of nested grids with the parent grid covering the whole US West Coast. The first child grid covers a large southern domain, and the third grid zooms in on the SCB region. The three horizontal grid resolutions are 20 km, 6.7 km, and 1 km, respectively. The external forcings are momentum, heat, and freshwater flux at the surface and adaptive nudging to gyre-scale SODA reanalysis fields at the boundaries. The momentum flux is from a three-hourly reanalysis mesoscale MM5 wind with a 6 km resolution for the finest grid in the SCB. The oceanic model starts in an equilibrium state from a multiple-year cyclical climatology run, and then it is integrated from years 1996 through 2003. In this paper, the 8-year simulation at the 1 km resolution is analyzed and assessed against extensive observational data: High-Frequency (HF) radar data, current meters, Acoustic Doppler Current Profilers (ADCP) data, hydrographic measurements, tide gauges, drifters, altimeters, and radiometers. The simulation shows that the domain-scale surface circulation in the SCB is characterized by the Southern California Cyclonic Gyre, comprised of the offshore equatorward California Current System and the onshore poleward Southern California Countercurrent. The simulation also exhibits three subdomainscale, persistent (i.e., standing), cyclonic eddies related to the local topography and wind forcing: the Santa Barbara Channel Eddy, the Central-SCB Eddy, and the Catalina-Clemente Eddy. Comparisons with observational data reveal that ROMS reproduces a realistic mean state of the SCB oceanic circulation, as well as its interannual (mainly as a local manifestation of an ENSO event), seasonal, and intraseasonal (eddy-scale) variations. We find high correlations of the wind curl with both the alongshore pressure gradient (APG) and the eddy kinetic energy level in their variations on time scales of seasons and longer. The geostrophic currents are much stronger than the wind-driven Ekman flows at the surface. The model exhibits intrinsic eddy variability with strong topographically related heterogeneity, westward-propagating Rossby waves, and poleward-propagating coastally-trapped waves (albeit with smaller amplitude than observed due to missing high-frequency variations in the southern boundary conditions).



Maps of the annual-mean surface EKE during the years with the largest and smallest eddy energy: 1997 (top) and 2001 (bottom). The largest EKE occurs during in 1997 in association with the ENSO event, and the lowest EKE occurs in 2001, with the 1997 level almost double that in 2001. The largest changes in the EKE pattern are in the offshore central SCB and further offshore in the California Current.


Related publication



2. Submesoscale dynamics in eddying flow regimes


In computational simulations of an idealized subtropical eastern boundary upwelling current system, similar to the California Current, a submesoscale transition occurs in the eddy variability as the horizontal grid scale is reduced to O (1) km. The transition in terms of the emergent flow structure and the associated time-averaged eddy fluxes has been examined with ROMS. In addition to the mesoscale eddies that arise from a primary instability of the alongshore, wind-driven currents, significant energy is transferred into submesoscale fronts and vortices in the upper ocean. The submesoscale arises through surface frontogenesis growing off upwelled cold filaments that are pulled offshore and strained in between the mesoscale eddy centers. In turn, some submesoscale fronts become unstable and develop submesoscale meanders and fragment into roll-up vortices. Associated with this phenomenon are a large vertical vorticity and Rossby number, a large vertical velocity, relatively flat horizontal spectra (contrary to the prevailing view of mesoscale dynamics), a large vertical buoyancy flux acting to restratify the upper ocean, a submesoscale energy conversion from potential to kinetic, a significant spatial and temporal intermittency in the upper ocean, and material exchanges between the surface boundary layer and pycnocline. Comparison with available observations indicates that submesoscale fronts and instabilities occur widely in the upper ocean, with characteristics similar to the simulations.

The emergent upper-ocean submesoscale fronts are analyzed from phenomenological and dynamical perspectives, using a combination of composite averaging and separation of distinctive subregions of the flow. The initiating dynamical process for the transition is near-surface frontogenesis. The frontal behavior is similar to both observed meteorological surface fronts and solutions of the approximate dynamical model called surface dynamics (i.e., uniform interior potential vorticity q and diagnostic force balance) in the intensification of surface density gradients and secondary circulations in response to a mesoscale strain field. However, there are significant behavioral differences compared to the surface-dynamics model. Wind stress acts on fronts through nonlinear Ekman transport and creation and destruction of potential vorticity. The strain-induced frontogenesis is disrupted by vigorous submesoscale frontal instabilities that in turn lead to secondary frontogenesis events, submesoscale vortices, and excitation of even smaller-scale flows. Intermittent, submesoscale breakdown of geostrophic and gradient-wind force balance occurs during the intense frontogenesis and frontal-instability events.

The mesoscale to submesoscale transition is mainly explained by the emergence of ubiquitous submesoscale density fronts and ageostrophic circulations about them in the weakly stratified surface boundary layer. Here the high-resolution simulations are further analyzed from the perspective of the kinetic energy (KE) spectrum shape and spectral energy fluxes in the mesoscale-to-submesoscale range in the upper ocean. For wavenumbers greater than the mesoscale energy peak, there is a submesoscale power-law regime in the spectrum with an exponent close to −2. In the KE balance an important conversion from potential to kinetic energy takes place at all wavenumbers in both mesoscale and submesoscale ranges; this conversion is the energetic counterpart of the vertical restratification flux and frontogenesis discussed in the earlier papers. A significant forward cascade of KE occurs in the submesoscale range en route to dissipation at even smaller scales. This is contrary to the inverse energy cascade of geostrophic turbulence and it is, in fact, fundamentally associated with the horizontally divergent (i.e., ageostrophic) velocity component. The submesoscale dynamical processes of frontogenesis, frontal instability, and breakdown of diagnostic force balance are all essential elements of the energy cycle of potential energy conversion and forward KE cascade.


Instantaneous surface temperature T (x, y) field at time t = 208 days after ICC initialization, as an example of a suite of computational simulations for an idealized subtropical, eastern boundary, upwelling current system [referred to as the idealized California Current (ICC)]. Note the string of meanders and filaments in 17°–19°C water with wavelengths ≈50 km running along the edges of the offshore eddies. Notice that an instability event is located at (x ≈ –250 km, y ≈ 350 km); it separates waters in the range 16°–18°C, and the temperature front is roughly aligned with the x axis.


Related publication



3. Wave-driven currents and their effects on SCB circulations


A vortex-force formalism for the interaction of surface gravity waves and currents is implemented in a three-dimensional (3D), terrain-following, hydrostatic, oceanic circulation model [Regional Oceanic Modeling System: ROMS; Schepetkin and McWilliams, 2005]. Eulerian wave-averaged current equations for mass, momentum, and tracers are included in ROMS based on an asymptotic theory by (McWilliams et al., 2004) plus non-conservative wave effects due to wave breaking, associated surface roller waves, bottom streaming, and wave-enhanced vertical mixing and bottom drag especially for coastal and nearshore applications. The currents are coupled with a spectrum-peak WKB wave-refraction model that includes the effect of currents on waves, or, alternatively, a spectrum-resolving wave model (e.g., SWAN) is used. The coupled system is applied to the nearshore surf zone during the DUCK94 field measurement campaign. Model results are compared to the observations and effects of parameter choices are investigated with emphasis on simulating and interpreting the vertical profiles for alongshore and cross-shore currents. The model is further compared to another ROMS-based 3D coupled model by (Warner et al., 2008) with depth-dependent radiation stresses on a plane beach. In both tests the present model manifests an onshore surface flow and compensating offshore near-bed undertow near the shoreline and around the breaking point. In contrast, the radiation-stress prescription yields significantly weaker vertical shear. The currents' cross-shore and vertical structure is significantly shaped by the wave effects of near-surface breaker acceleration, vertical component of vortex force, and wave-enhanced pressure force and bottom drag.


Conservative wave effects on inner-shelf circulation are investigated in the Channel Island region off Santa Barbara, California (referred to as SBC) using a quadruple nested ROMS configuration bounded by the global SODA field, downscaled from the Pacific basin at 12.5 km and to a half km grid resolution for Southern California Bight. In addition to the synoptic forcing and tides, idealized moderate waves ( uniform in space and time) are imposed. Diagnosis is made for one spring-neap cycle in September 2006. This is a case where radiation stress does not work well as for uniform wave field leading to no radiation stress divergence. The SBC summer thermocline is known to be shallow, around only 10-20 m, whereas waves increase it by about 8 m, and overall the mixed-layer depth is deepened about 15 % by wave-induced upper-current modification and associated enhancement of vertical shear. A clockwise-rotating, anti-cyclone sits in the middle of SBC, while prominent strengthening and modulation of such a circulation occurs mainly due to Stokes-Coriolis effects in upper ocean associated with waves. The upper-ocean averaged velocity magnitude and angle are also modified by waves substantially: the magnitude is increased by about 30 %, and the direction is rotated clockwise by about 6 degree. The Stokes-Ekman layer is much deeper than the mixed layer depth and reaches at about 60 m deep. Vortex force also plays a crucial role in modifying the circulation through eddy modulation. The surface subtidal eddy kinetic energy (EKE) has the maxima in the channel due to poleward advection of submesoscale eddies, accentuated by waves by about 20 %, partially attributed to change in mean advection of eddies associated with Stokes-Coriolis force. In turn, a vortical Rossby number (relative vertical vorticity over the background rotation), giving a ratio of mean vortex force to mean Stokes-Coriolis force, suggests the VF contribution reaches 60 % of the Stoke-Coriolis contribution where EKE is high. Moreover, VF is predominant in the nearshore area that is a central arena for eddy shedding: hance eddy-wave interaction by vortex force is crucial in modulating eddies, that could lead to feedback to the main circulation through Reynolds stress divergence.



Wave impact on upper-ocean turbulent kinetic energy (EKE) field in the SBC (Santa Barbara Channel, CA). Upper panel: subtital EKE in the SBC with wave effects (EKEwec), lower panel: EKE difference, EKEwec - EKEnw, where EKEnw is the one omitting wave effects. The surface subtidal EKE (upper panel) has the maxima in the channel due to poleward advection of submesoscale eddies, accentuated by waves by about 20 % (lower panel), partially attributed to change in mean advection of eddies associated with Stokes-Coriolis force. In turn, a vortical Rossby number (relative vertical vorticity over the background rotation), giving a ratio of mean vortex force to mean Stokes-Coriolis force, suggests the VF contribution reaches 60 % of the Stoke-Coriolis contribution where EKE is high.


Related publication



4. Generation and propagation of internal tides in SCB


To be added...


5. Island-induced wakes and their impact on circulation


With the existence of eight substantial islands in the Southern California Bight, the oceanic circulation is significantly affected by island wakes. In this paper a high-resolution numerical model (on a 1km grid), forced by a high-resolution wind (2 km), is used to study the wakes. Island wakes arise due both to currents moving past islands and to wind wakes that force lee currents in response. A comparison between simulations with and without islands shows the surface enstrophy (i.e., area-integrated square of the vertical component of vorticity at the surface) decreases substantially when the islands in the oceanic model are removed, and the enstrophy decrease mainly takes place in the areas around the islands. Three cases of wake formation and evolution are analyzed for the Channel Islands, San Nicolas Island, and Santa Catalina Island. When flows squeeze through gaps between the Channel Islands, current shears arise, and the bottom drag makes a significant contribution to the vorticity generation. Downstream the vorticity rolls up into submesoscale eddies. When the California Current passes San Nicolas Island from the northwest, a relatively strong flow forms over the shelf break on the northeastern coast and gives rise to a locally large bottom stress that generates anticyclonic vorticity, while on the southwestern side, with an adverse flow pushing the main wake current away from the island, positive vorticity has been generated and a cyclonic eddy detaches into the wake. When the northward Southern California Countercurrent passes the irregular shape of Santa Catalina Island, cyclonic eddies form on the southeastern coast of the island, due primarily to lateral stress rather than bottom stress; they remain coherent as they detach and propagate downstream, and thus they are plausible candidates for the submesoscale ‘‘spirals on the sea’’ seen in many satellite images. Finally, the oceanic response to wind wakes is analyzed in a spin-up experiment with a time-invariant wind that exhibits strips of both positive and negative curl in the island lee. Corresponding vorticity strips in the ocean develop through the mechanism of Ekman pumping.




San Nicolas Island wake: sequence of normalized surface vorticity maps from March 21–28, 2002. March 21=day 0. When current passes San Nicolas Island from the northwest, an island wake forms. A time series of 8 days for the surface vorticity is plotted, showing the formation and detachment of a cyclonic eddy and the decaying progression of an anticyclonic eddy in the wake. While the cyclonic eddy remains coherent, the anticyclonic eddy becomes weaker and weaker as it is advected downstream. Asymmetry in the robustness of cyclonic and anticyclonic wake eddies can be due to the weakening effect of centrifugal instability on the latter when ζ is smaller than -f , as occasionally occurs in the figure.



Related publication



6. Sediments and material dispersal in SCB


Suspended sediment-transport processes in Santa Monica and San Pedro Bay are analyzed using the sediment-transport capabilities of the Regional Oceanic Modeling System (ROMS). A one-month simulation for December 2001 has been carried out with a set of nested domains. The model inputs include tides, winds, surface waves, and idealized initial sediment conditions for sand and non-cohesive silt. Apart from the control run, the sensitivity of the results to surface waves, ripple roughness and bed armoring has been analyzed. From the control experiment, the horizontal transport of sand turns out to be limited to within a few km of the nearshore erosion zones. During high wave events, silt is transported over further distances and also partly offshelf in distinct plumes. The effectiveness of horizontal silt transport depends strongly on vertical mixing due to both surface wind stress and wave-enhanced bottom stress. High wave events coincident with strong winds (hence strong vertical mixing) are the most optimal conditions for sediment-transport. Excluding wave effects in the simulation shows that surface waves are the dominant factor in resuspending bed material on the Southern Californian shelves. The sensitivity experiments also show that the direct influence of additional ripple roughness on erosion and deposition is relatively weak. Switching off bed armoring locally results in increases of near-bottom concentrations by a factor of 20 for silt and a factor of 5 for sand as well as stronger spatial gradients in grain size.




(a) Depth-averaged, tide-averaged velocity and vertically integrated silt concentration, half a day after maximum of December 15 wave event. Concentrations locally up to 500 g/m2. (b) Fraction of time that wave–current bottom stress exceeds the threshold for silt suspension, i.e., τcw>τcr,silt. Dotted lines: 20–65 m isobaths (15 m interval). (c) Ripple height (η) at the peak of 10 December wave event. (d) Bed roughness (including ripple roughness and bed-load roughness) at the peak of 10 December wave event. (e) Net change in bed thickness over December 2001 control experiment. Red: net deposition; blue: net erosion. Color range -4 to 4 mm, data range -28 to 11 mm (extremes just off Redondo Beach). Also shown are 15, 30, 60, 120, 240, 480 m isobaths. (f) As (e) but for net change in silt fraction of the active layer. Color range: -0.12 to 0.12, data range: -0.12 to 0.39 (extremes in nearshore areas).


Related publication




RESOURCES


National Center for Supercomputing Applications

University of California Shared Research Computing Services



PEOPLE


James McWilliams

Yusuke Uchiyama

Maarten Buijsman

Jeroen Molemaker

Changming Dong

Alexander Shchepetkin



Copyright © University of California, Los Angeles, Dept. Atmospheric and Ocean Sciences, Center for Earth System Research. All rights reserved.