NOAA Operational Basin Scale Ocean Modeling for Research and Operations

1
NOAAOperational Basin Scale Ocean
ModelingforResearch and Operations

• Fred Toepfer
• NOAA, Environmental Modeling Program Manager
• Deputy Director, Environmental Modeling Center
• November 28th, 2006

2
Outline

• Driving Requirements
• Applications

3
Driving Requirements

• Long-term Climate Projections
• Climate Change and Global Warming
• Sea-level Rise
• Abrupt Climate Change
• Seasonal and Inter-annual Predictions
• Seasonal Outlooks
• Ocean Forecasting
• Maritime Safety and Efficiency
• Public Safety

4
Current Status and Plans

• Operational and Research Models
• GFDL Climate Model
• NCEP Operational Climate Forecast System
• Real-time Ocean Forecast System Atlantic
• Wave-Watch 3 Wave Model

5
GFDL Climate Model

• CM2.1 State of the art climate model
• Atmosphere AM2, 20 horizontal resolution, 24
  vertical levels
• Ocean OM3, 10 horizontal resolution, 50
  vertical levels
• Newly developed land, sea ice component models
• State of the art model physics
• Common software infrastructure (ESMF prototype)
• Used in IPCC climate assessment process
• Routinely used for seasonal to interannual
  forecasting
• Used for a wide variety of research projects
• CM2.1 is among the best models in the world, as
  judged by several recent independent assessments
• Widespread public dissemination of model output
• New Earth System model (ESM2.1) adds ocean/land
  biogeochemical cycles to CM2.1
• Carbon
• Nitrogen
• Phosphorous

Observed global temperature
Modeled global temperature
6
NCEP Climate Forecast System
7
Global Ocean Data Assimilation System (GODAS)

• The operational GODAS
• Built on the GFDL Modular Ocean Model, version 3
  (MOMv3)
• Quasi-global extending from 75oS to 65oN, 1o
  resolution (1/3o in the tropics)
• Forced by surface fluxes (momentum, heat, fresh
  water) from NCEP Reanalysis 2
• A 3-dimensional variational (3DVAR) method
  assimilates ocean observations
• Univariate in temperature and salinity
• Assimilates data down to 750 meters
• Fixed moorings (TAO, Triton, Pirata) provide key
  tropical data
• XBTs and Argo autonomous drifters provide global
  coverage
• Provides daily analyses (archived as 5-day
  averages)

• The next version of GODAS
• Built on the GFDL MOMv4
• Fully global, 1/2o resolution (1/4o in the
  tropics), includes Arctic Ocean and ice model
• Forced by surface fluxes (momentum, heat, fresh
  water) from the NCEP CFS
• Uses 3DVAR
• Univariate in temperature and salinity,
  multivariate in velocity
• Will assimilate data down to 2200 meters
• Fixed moorings (TAO, Triton, Pirata, etc.) will
  remain key tropical data
• Argo will provide possibility of unprecedented
  global coverage of T and S
• Satellite altimetry observations will increase
  the global temporal resolution of the observations

8
Planned upgrade of the CFS from MOM3 to MOM4

• Interactive sea ice model instead of
  climatological ice fraction and cover
• Higher horizontal resolution of ocean and ice
  model. .5 degree vs 1 degree for ocean and a
  tripolar grid for northern hemisphere sea ice
• Hourly coupling interval vs daily coupling
• MOM4 going to 2200 m deep vs 750 m for MOM3

9
Planned Climate Forecast System

• One major difference between MOM4 and MOM3 is
  that MOM4 has the option to couple to a sea-ice
  model, so it covers the global ocean with a
  possible sea-ice cover on top in the polar
  regions. Fluxes to and from the ocean must be
  passed through sea-ice model.
• Why add a sea-ice model Sea-ice amplifies any
  change of climate due to its positive feedback.
  It restricts the exchange of heat/water between
  the air and ocean, and modifies air/sea momentum
  transfer, ocean fresh water balance and ocean
  circulation.

10
Realtime Ocean Forecast System(RTOFS)

• Atlantic Basin Operational
• The RTOFS for the Atlantic is run once per day to
  produce a one day nowcast and five day forecasts
  of currents, temperature, salinity and sea
  surface elevation fields.
• The dynamical model is HYCOM.
• The ocean is forced by three hourly GFS
  atmospheric fluxes ( wind stress, radiative heat
  fluxes, turbulent heat fluxes, precipitation,
  evaporation, and atmospheric pressure).
• River outflows are included (observation in real
  time for US rivers (USGS) and climatology
  (RIVDIS)
• Body tides ( eight constituents) and boundary
  tides (TPX.06) are included.
• During the nowcast cycle the following data is
  assimilated. - Sea Surface Temperatures obtained
  from satellite thermal imagery (GOES AVHRR)  -
  Sea Surface Heights obtained from satellite
  altimetry (SLA and Jason-GFO).  - Temperature
  and Salinity obtained from in-situ oceanographic
  measurements (ARGOS, XBT, CTD).
• Global and Pacific Basin
• Under development 2010 Implementation

11
(No Transcript)
12
Operational Wave Modeling

• Presently operational NOAA WAVEWATCH III (NWW3)
  v. 2.22.
• NWW3 Global, 1.25x1?, 180h, 3-hourly GFS winds,
  with assimilation and a 10 member ensemble.
• AKW Alaskan Waters, 0.5x0.25?, 180h forecast
  based on 3h GFS winds.
• WNA and ENP W. North Atlantic and E. North
  Pacific, 0.25x0.25?, 180h forecast based on 3h
  GFS winds.
• NAH and NPH Hurricane version of WNA and ENP,
  0.25x0.25?, 126h forecast, 1h GFSGFDL winds.
• GLW Great Lakes wave model, 4km, 84h, NAM winds.
• All models 24 directions, 25 frequencies, 4
  daily cycles with 6h hindcasts for continuity.

13
In the Pipeline

• The new multi-grid wave model.
• More to follow.
• Full wave field separation in space and time.
• With first version of multi-grid model.
• Improved physics
• Initial growth and surf-zone wave breaking with
  first multi-grid model (FY 2007-Q4)
• Systematic improvements to follow (FY 2007 and
  following).

14
Multi-grid model
Deep ocean model resolution dictated by GFS model
Higher coastal model resolution dictated by model
economy
Highest model resolution in areas of special
interest
Hurricane nests moving with storm(s) like GFDL
and WRF
15
Back-ups
16
Multi-grid model

• This represents a massive software engineering
  effort.
• Basic model modifications necessary to work with
  different resolutions in a single model have been
  made.
• Full two-way nesting is nearing completion for
  static grids, making first implementations
  feasible.
• Moving (relocatable) grid options will be
  developed this winter.
• An example of static nest capabilities are
  presented in the following slides.
• Alaskan application for NWS capabilities.

17
GODAS performance anomaly correlations with
TOPEX/Jason-1 altimetric sea surface height for
the years 1993-2005
Control assimilates no data and has good
correlations only on Pacific equator.
Assimilation of Tao data expands the region of
high correlation throughout the tropical Pacific.
Assimilation of satellite altimetry improves
correlations globally.
18
GODAS performance anomaly RMS differences with
TOPEX/Jason-1 altimetric sea surface height for
the years 1993-2005
Control assimilates no data and has low RMS error
on Pacific equator and where SSH signal is weak.
Assimilation of T and S expands the region of low
RMS error throughout the tropical Pacific.
Assimilation of satellite altimetry RMS error
globally.
19
Results are from the MOM4 coupled to a dynamic
thermodynamic sea-ice model using CDAS 2 6-hourly
forcing reasonable sea-ice distribution has
been simulated.
Satellite Obs
Model
Siberia
Greenland
Alaska