Coupled GCM The Challenges of linking the atmosphere and ocean circulation.
Post on 24-Dec-2015
216 Views
Preview:
Transcript
Coupled GCM
The Challenges of linking the atmosphere and ocean
circulation
Brief History of LRF
Statistical and Analog - earliest Simple models and
Teleconnections Coupled Models with dynamic and
statistical components Dynamic Coupled Atmosphere-
Ocean Models
GCM Model Matters
Grid Spacing• dependent on
coordinate system for globe
• dependent on computer space
Time Step• dependent on
resolution and length of forecst
Terrain and Ocean Mapping• generally rough
with little detail Parameterize
• solar radiation• convection• heat flux• wind stress
The Interface
Oceans• Sea Surface
Temperature• evaporation
• Mixed Layer• heat flux/transport
• Annual cycles• upwelling• Pacific circulation
Atmosphere• Solar Energy
• sun angle• cloud cover
• Wind Stress• mixing layer
• estimate profile
• Heat & Moisture Transport
• shallow and deep convection
Coupled GCM’s Focus
Tropical Oceans - Pacific• Initial Conditions
• Atmosphere• inferred from spotty observations and detailed
satellite analysis
• Ocean• uses a data set developed by Florida State
University which shows climatology of temperature and wind stress
CGCM’s - Many Models
Center for Ocean-Land-Atmosphere Std
Geophysical Fluid Dynamics Lab (GFDL)
NASA-Lamont Doherty (Columbia Univ)
Scripps Institute UCLA NCEP Max Plank
Institute Bureau of
Meteorology Research Centre
The COLA’s Model
The Ocean Portion• Adapted from GFDL - for Pacific
Domain from 30S-45N &130E-80W• Resolution: x=1.5 y=.5 (20S-20N) 1.5
degrees elsewhere• 20 vertical levels to 4000m - 1-16 are
within the top 40m• non-linear vertical mixing of heat,
salinity and momentum
The COLA’s Model
The Atmosphere Portion• Global Spectral Model with 30 wave limit• 18 layers on a sigma coordinate• Solar radiation is parameterized• Deep convection - modified Kuo• Shallow convection - Tiedtke • Complex scheme for exchange of heat,
moisture and momentum
Coupling Strategy
Several Methods• Interpolated Exchange• Anomaly Coupling • Mixed Methods
Significance of Ocean-Atmosphere Exchange is especially important in the Tropics
Coupling Strategy
Interpolated Exchange• Daily mean values are exchanged
• OGCM produces SST for Atmosphere• AGCM produces surface heat flux,
momentum and freshwater (rainfall) for the Oceans
• These values are parameterized and interpolated for grid points in each model
Coupling Strategy
Anomaly Coupling• Each part of the model predicts and
anomaly component compared with a set model climatology.
• Atmosphere climatology - 45 years (1949-94)
• Ocean climatology - 30 years (1964-94)
Coupling Strategy
Start with Atmosphere (AGCM) predicts solar-radiation to estimate SST for Ocean
SST is used to predict a wind profile in the tropical boundary layer - the anomaly component of this profile is used for adding to the wind stress on the ocean.
Experimental Long Lead Models
Coupled GCM from COLA - now uses anomaly of initial conditions from an in-house ocean data assimilation analysis
Coupled GCM from COLA using interpolated values from AGCM and OCM
Hybrid Coupled Ocean-Atmos Model - Scripps-Max Plank
2004 Model Forecast
CPC/EMC• GFDL Ocean• MRF reduced• Ensemble-16• updated wkly
• http://www.emc.ncep.noaa.gov/cmb/sst-forecasts/
2004 Model Forecasts
Scripps Plank
• Hybrid • 30S-30N• 13 vertical• AGM - Stat• mainly wind
stress
2004 Model Forecasts
Japan Meteo Agency
AGCM (T42/40 levels)
OGCM (T 20 levels)• 2.5 x 2.0• Flux Exchanges
every 24 hrs for mean values
2004 Model Forecast
LDEO Model - wind stress Focus on
initialization Ensemble of
3 wind stresses • FSU,NCEP,QUIKSCAT
2004 Model Forecast
Markov Model of SST - CPC
Linear Statistics trained 1980-95
Verified by 1964-1979
2004 Model Forecast
LIM (Linear Inverse Model) from CIRES/CDC - Boulder
Uses a specific Stat function (Green)
2004 Model Forecast
Constructed Analog (Van Den Dool)
Uses past anomalies as predictors
2004 Model Forecast
IRI Summary All Models;
Statistical & Dynamic
Long Lead Predictions
Summary of 2004 Model Forecasts
Long Lead Predictions
Forecast of SST in Tropical Pacific with a Markov Model - NCEP (linear statistical)
Tropical SST’s using a Linear Inverse Model- CIRES - Boulder
Tropical Pacific SST using and intermediate ocean and statistical atmosphere model - Earth Environmental Studies - Seoul
Further Readings
http://grads.iges.org/ellfb/contents.htm • - (updated every 3 months)
top related