Regional%20Coastal%20Ocean%20Modeling:%20Forcing

1
Regional Coastal Ocean Modeling Forcing

• Local, Remote, Tidal

2
Introduction

• Equilibrium, Variability

3
Variability in coastal systems
PERIOD

10-70 y
4-7 y
1 y
lt 3 Months
WARMING TREND
DECADAL (PDO, NAO)
INTERANNUAL (ENSO)
SEASONAL
MESOSCALE, SUB-MESOSCALE

• Observation
• Surface warming in spite of
• Increased upwelling winds
• Processes
• surface heating, lateral
• advection
• (Di Lorenzo et al., 2003)

• Observation
• filaments, squirts, eddies
• Dominant Processes
• Baroclinic/barotropic instabilities
• Ageostrophic instabilities
• (Marchesiello et al., 2003)

• Observation
• Quasi-periodic events
• Processes
• Local winds, propagating
• Kelvin/CTW waves

• Observation
• Periodic oscillation
• Processes
• Local forcing

• Observation
• Regime shifts
• Processes
• Local forcing, lateral advection

4
Equilibrium and 3 Kinds of Variability

• locally forced variability
• Remotely forced variability
• Intrinsic variability

5
Local Forcing From Synoptic Events to Decadal
Oscillations and Trends
Local forcing is not necessarily high-frequency
(synoptic) Pacific Decadal Oscillation
6
Remote forcing El Nino, Coastal Waves, Tides
Coastal waves
Kelvin waves
7
Intrinsic Variability Baroclinic, Barotropic,
Frontal Instabilities, Tidal Residual Currents
U.S. West Coast
8
Shelf dynamics overview
Midshelf geostrophy and nonlinear dynamics start
playing a role
Outershelf zone of exchange with the deep ocean
through the shelf break
Innershelf frictional zone driven by wind stress
and surface/bottom friction
9
Exchanges between Coastal and Oceanic Regions
residence time of shelf water

• Eastern Boundary Current systems (upwelling
  systems) coastal waters are mainly influenced by
  local forcing flushing time is only a few days
• Western boundary Current systems coastal waters
  are more isolated with flushing times of up to a
  few years.

Coccolithophorid bloom in Jervis bay
10
Local ForcingThe Wind
11
Wind and Wind-stress

• Wind stress (kg m-1 s-2 or Newton m-2) is an
  important driving force for coastal ocean
  currents.
• T Cd ?a u (u2 v2)1/2
• Cd is the dimensionless "drag coefficient" (about
  0.0013) F(u,v,stability)
• ?a is air density (about 1.2 kg m-2)
• (u,v) is wind vector at 10 m above sea level
  (m/s).

12
Wind Products for Ocean Modeling

• Voluntary Observing Ship Program
• COADS Comprehensive Ocean-Atmosphere Data Set
• A global coverage
• D poor sampling in some areas (the coastal areas
  in particular), accuracy, coverage in time
• Mooring Data
• A very good coverage in time
• D very bad in space
• Satellite Scatterometers
• A good coverage, high accuracy
• D young data set, no coverage within 50km from
  the coast
• Models
• A very good coverage in space and time and at
  the coast
• D subject to model errors

13
Satellite Sensors

• Polar orbit ( 90) Usually these satellites
  have height between 500 and 2,000 km and a period
  of about 1 to 2 hours.
• As the Earth rotates under this orbit the
  satellite effectively scans from north to south
  over one face and south to north across other
  face of the Earth, several times each day,
  achieving much greater surface coverage than if
  it were in a non-polar orbit. 

14
Satellite Sensors on Polar Orbit
15
Satellite Sensors
16
Scatterometers

• Active microwave sensor
• Measures sea surface roughness
• Roughness is indicative of the magnitude of the
  wind stress applied at the ocean surface
  (Beaufort scale)
• ERS1, ERS2, NSCAT, QuickSCAT

Microwave scatterometer is based on the principle
of the resonant Bragg scattering. For a smooth
surface, oblique viewing of the surface with
active radar yields virtually no return. If the
surface is rough, significant backscatter occurs.
17
Scatterometers

• Accuracy 1m/s
• Unaffected by clouds
• Global coverage twice a day at 25km resolution
  (QuickSCAT)
• BUT
• Not reliable during precipitation event
• Temporal-spatial sampling biases
• Temporal Land-sea breeze system difficult to
  sample
• Spatial NO DATA WITHIN 50km FROM THE COAST

18
Merging Wind Products from Scatterometer Regional
Atmospheric Models
Scatterometer
Error
Atmospheric Model
19
Merging Wind Products of Scatterometer and Models
20
Regional Atmospheric Models

• WRF (NCAR)
• MM5 (NCAR)
• COAMPS (NRL)
• Méso-NH (CNRS/Météo-France)

21
Weather Research and Forecast (WRF) modeling
system. The WRF system is in the public domain
and is freely available for community use. It is
designed to be a flexible, state-of-the-art
atmospheric simulation system that is portable
and efficient on available parallel computing
platforms. WRF is suitable for use in a broad
range of applications across scales ranging from
meters to thousands of kilometers, including -
Idealized simulations (e.g. LES, convection,
baroclinic waves)- Parameterization research-
Data assimilation research- Forecast research-
Real-time NWP (Numerical Weather Prediction) -
Coupled-model applications- Teaching
http//www.mmm.ucar.edu/wrf/users/
22
TESTING AIR-SEA-LAND INTERACTIONS THE CANARY
ISLANDS
23
OCEAN RESPONSE TO SMALL SCALE WIND FORCING
WRF
NCEP
24
WEATHER FORECAST AND HINDCAST IN SENEGAL
25
Regional Atmospheric Models Errors Upwelling
Event of March 2002 (Capet, Marchesiello
McWilliams, 2004)
During March 2002, strong upwelling-favorable
winds lead to a large bloom of toxic diatoms
(Pseudo-nitzschia ) which affected local
ecosystem and killed many marine mammals (L.A.
Times)
Pseudo-nitzschia produces domoic acid, a
neurotoxin causing gastrointestinal and
neurological illness (sometimes fatal) in humans,
termed ASP (amnesic shellfish poisoning), and
mortalities among a variety of marine vertebrates
26
SST
CHL
27
UCLA Mooring Observations
T,S,U,V, Wind data at offshore (SMB) and
nearshore (UCLA) stations
MUCLA
SMB 46025
28
COAMPS Regional Atmospheric solution for the US
west coast
COAMPS is the US navy operational model with
different resolutions along the US west coast
27, 9, 3 km
Coastal wind drop off is sensitive to resolution
COAMPS profiles, August 2003
29
Validating and Adjusting the Forcing
BUOY DATA FITTING
COAMPS
BUOY DATA
BUOY DATA
Offshore buoy location
Nearshore buoy location
30
Model Simulation
COAMPS
BUOY FITTING
Offshore buoy location
Nearshore buoy location
31
Remote Forcing

• Large-scale oceanic fluxes

32
Methods for Oceanic Forcing

• Large-scale data
• Open Boundary Conditions (OBCs)
• Nesting Conditions downscaling
• Combination

33
Mercator
ROMS Senegalese coastal upwelling
Senegal 6km
C. Blanc
Canary 20km
Levitus
C. Bojador
C. Vert
C. Blanc
Clipper
C. Vert
34
Large-scale Data

• World Ocean Atlas (WOA) or Levitus Data
• Global gridded in-situ data for T,S (1 deg.
  resolution grid)
• Climatology only
• Only geostrophic currents with arbitrary level of
  no motion
• Irregular sampling Very smooth data large
  decorrelation scales used in smoothing (1000 km)
• Global Model data (ORCA2, ORCA05, Mercator,
  OCCAM, POP, MOM)
• All needed data available at required resolution
• Synoptic and Interannual variability
• Subject to model errors and drifts

35
Open Boundary Conditions

• PE equations are ill-posed with respect to OBCs
  we cannot find a set of OBCs which guarantees a
  unique and stable solution (Oliger and Sundstrom,
  1978).
• Deal with the consequences
• Discontinuities from over-specification
• Drift from under-specification (extrapolation)
• Make assumptions (hyperbolicity) and put safety
  guards (sponge, nudging layers, )

36
Hyperbolic Systems and Characteristic Method
Exemple of 1D-wave Equation

• We can derive the characteristic equations a
  system of 2 independant transport equations
  (Blayo and Debreu, 2005)
• A well-posed hyperbolic open boundary problem
  requires a boundary condition for every incoming
  characteristics C-
• The incoming characterstic quantitiy C- is
  conserved along the characteristic lines C-Cext

incoming characteristic
Sommerfeld radiation condition
37
Mixed Active/Passive OBCs

• Radiation conditions with relaxation
• (Marchesiello et al., 2001)
• Transition to external data sponge and nudging
  layers
• Progressively matching the scales of external and
  internal (model) data
• Volume constraint

Computed at previous space/time step
38
Flather Condition for the Shallow Water Equations
(barotropic mode of PE)

• The Flather OBC leads to a well-posed problem for
  the linear shallow-water problem. Using the
  characteristic methods
• Ensures near conservation of mass and energy
  through the open boundary
• Ideal for tidal forcing

39
Mesh Refinement Nesting

• Nesting the external data is provided by a
  simulation on a larger domain
• External and internal data are almost consistent
• Mesh refinement the same model runs on the
  parent and child grids simultaneously (online
  nesting)
• One-way / two-way nesting
• Multi-level refinement
• Cost of the model is driven by the last child
  level

Nesting condition
40
AGRIF
The same model (executable) runs on grids with
different space/time resolutions
2 20 45 34 59 3 3 3 30 55 70 89 3 3 2 0 1 10 30
20 40 5 3 5 0

• Each domain has its own input/output files
• Grids locations specified in input file
• Parallelized
• Forcings, initial conditions can be generated
  with interactive tools (ROMS)
• Local conservation enforcement

41
AGRIF in OPA North Atlantic
42
AGRIF in ROMS NEW CALEDONIA
43
Tidal Modeling

• Forcing, tidal currents, residuals, residence time

44
Forcing Tidal Waves in regional domains
Co-oscillating tides

• Tides in regional seas without a narrow opening
  to the ocean are usually driven by tides outside
  the region.
• These tides are called co-oscillating tides.
• When co-oscillating tides dominate, direct
  astronomical forcing of tides can be neglected
• The effect of co-oscillation has to be prescribed
  through open boundary conditions
• In regions with highly restristed access to the
  outside, direct forcing is quite important
  eastern Mediterranean Sea.

45
Forcing Co-oscillating Tides

• Use Flather open boundary condition
• External data is derived from a global tidal
  model the inverse model of Ebert et al. (1994,
  OSU) combines a hydrodynamic model and T/P
  Altimetric analyses (data available on line)
• Global models provide amplitude and phase of the
  primary tidal constituents Semi-diurnal M2, S2,
  N2, K2 Diurnal K1, O1, P1, Q1 Long-term Mf,
  Mm
• The tidal signal is the sum of these primary
  tidal constituents

46
Forcing Co-oscillating Tides

• Tidal elevation at point (x,y) is the sum of
  primary tidal constituents
• A amplitude, T period, f and V nodal factors
  (18.6-year variations in lunar orbit), ? phase
  including astronomical argument
• Tidal currents
• UUMA cos? cosf UMI sin? sinf
• VUMA sin? cosf UMI sin? cosf
• UMA, UMI semimajor and semiminor axis currents,
  ? angle of semimajor axis, f phase at t

47
Tidal Modeling Applications

• High resolution tidal flows
• Residual circulation
• Eulerian and Lagrangian
• Residence time for water quality
• Sediment transports
• Tidal mixing and impact on ecosystem

48
Tidal oscillation in Bay of Brest
ROMS simulation 300m resolution grid Embedded
into a regional model of the Iroise Sea
49
Residual Currents
50
Residual currents

• Eulerian residual currents
• Eulerian residual transport
• Lagrangian flow
• Lagrangian model dX/dtu(X,t)
• the Lagragian flow results from combination of
  Eulerian residual currents and tidal oscillation
  gives the real nature of residual flow and
  dispersion processes

51
Generation of Residual Currents Sydney Harbour

• Residual currents are non-linear phenomenon
• 2 sources of non-linearity
• Quadratic bottom friction
• has a dual effect generation and dissipation
• Nonlinear advective terms
• Topography interacts with bottom friction to
  generate a torque. When advected, produce
  residuals

52
Linear bottom friction
No advective terms
Max vel 5 cm/s
Max vel 2 cm/s
Constant depth
No bottom friction
Max vel 20 cm/s
Max vel 14 cm/s
53
Tidal Flushing
54
Tidal Flushing
55
Residence Time

• Exponential fit
• C(t) C(0) exp (- t / T)
• More exact definition

56
Residence Time
57
(No Transcript)