Improvements to the Equatorial Pacific Cold Tongue Region in an OGCM

Improvements to the Equatorial Pacific Improvements to the Equatorial Pacific Cold Tongue Region in an OGCMCold Tongue Region in an OGCM

Possible Implications for the NCEP GODAS / CFSPossible Implications for the NCEP GODAS / CFS

Kristopher B. Karnauskas, Kristopher B. Karnauskas, Raghu Murtugudde, Antonio J. BusalacchiRaghu Murtugudde, Antonio J. Busalacchi

NOAA NCEP/EMC 6/20/06

Photo of Isla Isabela NOAA R/V Ka’imimoana 2 May 2005

University of Maryland College Park, Md., USA

In revisions at the Journal of Physical Oceanography

Improvements to the Equatorial Pacific Improvements to the Equatorial Pacific Cold Tongue Region in an OGCMCold Tongue Region in an OGCM

Possible Implications for the NCEP GODAS / CFSPossible Implications for the NCEP GODAS / CFS

““The currents of the sea are rapid and sweep across the [Galapagos] archipelago, The currents of the sea are rapid and sweep across the [Galapagos] archipelago, and the gales of wind are extraordinarily rare…”and the gales of wind are extraordinarily rare…”

““Considering that these islands are placed directly under the equator, the climate is Considering that these islands are placed directly under the equator, the climate is far from being excessively hot; this seems chiefly caused by the singularly low far from being excessively hot; this seems chiefly caused by the singularly low temperatures of the surrounding water…”temperatures of the surrounding water…”

- British Naturalist Charles Darwin in 1859 and 1839

In revisions at the Journal of Physical Oceanography

TMI SST Annual Mean (1998-2005)TMI SST Annual Mean (1998-2005)

Introduction: Introduction: The Pacific cold tongue is…The Pacific cold tongue is…

• The result of coastal upwelling and Ekman divergence, a true coupled air- The result of coastal upwelling and Ekman divergence, a true coupled air- sea interaction process.sea interaction process.

• A zonal band of cold SST along the equator in the east-central Pacific.A zonal band of cold SST along the equator in the east-central Pacific.

• Linked to tropical cloud and precipitation patterns, ocean biological Linked to tropical cloud and precipitation patterns, ocean biological productivity and carbon cycling.productivity and carbon cycling.


• Linked to Linked to tropical cloud and precipitation patternstropical cloud and precipitation patterns, ocean biological , ocean biological productivity and carbon cycling.productivity and carbon cycling.

Mean CMAP PrecipitationMean CMAP Precipitation


Mean SeaWiFS Surface Chlorophyll-aMean SeaWiFS Surface Chlorophyll-a

• Linked to tropical cloud and precipitation patterns, Linked to tropical cloud and precipitation patterns, ocean biological ocean biological productivityproductivity and carbon cycling. and carbon cycling.

• Linked to tropical cloud and precipitation patterns, ocean biological Linked to tropical cloud and precipitation patterns, ocean biological productivity and productivity and carbon cyclingcarbon cycling..• Linked to tropical cloud and precipitation patterns, ocean biological Linked to tropical cloud and precipitation patterns, ocean biological productivity and carbon cycling.productivity and carbon cycling.

• Where ENSO events produce large SST anomalies.Where ENSO events produce large SST anomalies.


Mean Ocean-Atmosphere Carbon Flux (Takahashi et al., 1999)Mean Ocean-Atmosphere Carbon Flux (Takahashi et al., 1999)

• Linked to tropical cloud and precipitation patterns, ocean biological Linked to tropical cloud and precipitation patterns, ocean biological productivity and productivity and carbon cyclingcarbon cycling..


• Where ENSO events produce Where ENSO events produce large SST anomalieslarge SST anomalies..

• Linked to tropical cloud and precipitation patterns, ocean biological Linked to tropical cloud and precipitation patterns, ocean biological productivity and carbon cycling.productivity and carbon cycling.

TMI SST: December 1997

• ProblemProblem: most OGCMs produce a cold tongue that is : most OGCMs produce a cold tongue that is too cold too cold and and extends too far west…extends too far west…

(Vecchi et al., 2005)(Vecchi et al., 2005)

Introduction: Introduction: Tropical cold biasTropical cold bias• ProblemProblem: most OGCMs produce a cold tongue that is : most OGCMs produce a cold tongue that is too coldtoo cold and and

extends too far westextends too far west……

• Presents obstacles for coupled GCMs to produce realistic tropical cloud Presents obstacles for coupled GCMs to produce realistic tropical cloud and precipitation patterns… one potential culprit in the “double ITCZ” and precipitation patterns… one potential culprit in the “double ITCZ” problem.problem.

• A relevant theme in ongoing and upcoming research programs, e.g. A relevant theme in ongoing and upcoming research programs, e.g. Pacific Upwelling and Mixing Physics (PUMP).Pacific Upwelling and Mixing Physics (PUMP).

• Modeling studiesModeling studies aimed at diagnosing the cold bias have focused on the aimed at diagnosing the cold bias have focused on the surface energy budget (Kiehl, 1997), atmospheric feedbacks (Gordon et surface energy budget (Kiehl, 1997), atmospheric feedbacks (Gordon et al., 2000; Sun et al., 2003), biological attenuation of shortwave radiation al., 2000; Sun et al., 2003), biological attenuation of shortwave radiation (Murtugudde et al., 2002, Marzeion et al., 2005) and otherwise coupled (Murtugudde et al., 2002, Marzeion et al., 2005) and otherwise coupled air-sea interactions (Luo et al., 2004).air-sea interactions (Luo et al., 2004).

• These studies represent progress, but the cold bias remains a serious These studies represent progress, but the cold bias remains a serious problem. There also remain problems with the EUC and SEC.problem. There also remain problems with the EUC and SEC.

1982-2004

Image courtesy Dave Behringer and Yan Xue, http://www.nws.noaa.gov/ost/climate/STIP/GODAS.htm

The The GODASGODAS, a part of the , a part of the CFSCFS, is based on the , is based on the MOM3 OGCMMOM3 OGCM..Currently 1° x 1° with 1/3° meridional within 10° of the equator.Currently 1° x 1° with 1/3° meridional within 10° of the equator.

Introduction: Introduction: Current ocean modelingCurrent ocean modeling



1982-2004



• What has What has not been considerednot been considered explicitly are the effects of horizontal explicitly are the effects of horizontal resolution and the inclusion of the Galapagos Islands in the models.resolution and the inclusion of the Galapagos Islands in the models.

• The Galapagos Islands are The Galapagos Islands are directly on the equatordirectly on the equator, in the midst of the , in the midst of the cold tongue… and in a very critical place in terms of surface fluxes and cold tongue… and in a very critical place in terms of surface fluxes and subsurface currents.subsurface currents.

• Current Current operational ocean modelingoperational ocean modeling at NOAA NCEP does not include at NOAA NCEP does not include the Galapagos Islands. GODAS horizontal resolution is arguably the Galapagos Islands. GODAS horizontal resolution is arguably sufficient to represent them.sufficient to represent them.


Galapagos topography in the GODAS / CFSGalapagos topography in the GODAS / CFS

Thanks to Dave Behringer (NCEP/EMC) for kindly providing this figure.


• What has What has not been considerednot been considered explicitly are the effects of horizontal explicitly are the effects of horizontal resolution and the inclusion of the Galapagos Islands in the models.resolution and the inclusion of the Galapagos Islands in the models.

• The Galapagos Islands are The Galapagos Islands are directly on the equatordirectly on the equator, in the midst of the , in the midst of the cold tongue… and in a very critical place in terms of surface fluxes and cold tongue… and in a very critical place in terms of surface fluxes and subsurface currents.subsurface currents.

• Current Current operational ocean modelingoperational ocean modeling at NOAA NCEP does not include at NOAA NCEP does not include the Galapagos Islands. GODAS horizontal resolution is arguably the Galapagos Islands. GODAS horizontal resolution is arguably sufficient to resolve them.sufficient to resolve them.

• Effects of the Galapagos IslandsEffects of the Galapagos Islands in an OGCM (MOM3) were looked at in an OGCM (MOM3) were looked at by Eden & Timmerman (2004) but there are some concerns about how by Eden & Timmerman (2004) but there are some concerns about how that was done. Also, ET-04 focused on very local-scale effects (TIWs that was done. Also, ET-04 focused on very local-scale effects (TIWs and island upwelling). However, ET-04 provides useful comparison with and island upwelling). However, ET-04 provides useful comparison with our study.our study.


Experimental philosophyExperimental philosophy


• Gent & Cane (1989) OGCM coupled to an atmospheric mixed layer Gent & Cane (1989) OGCM coupled to an atmospheric mixed layer (just surface fluxes; Murtugudde et al., 1996) and (just surface fluxes; Murtugudde et al., 1996) and hybrid vertical hybrid vertical mixing scheme (Chen et al., 1994a)mixing scheme (Chen et al., 1994a). This model has been . This model has been extensively tested and used successfully in the tropical Pacific (e.g. extensively tested and used successfully in the tropical Pacific (e.g. Kessler et al., 1998; Chen et al., 1994a,b and many others).Kessler et al., 1998; Chen et al., 1994a,b and many others).

Kraus-Turner (1967) type mixed layer modelKraus-Turner (1967) type mixed layer model Price et al. (1986) dynamical instability modelPrice et al. (1986) dynamical instability model

Chen et al. (1994) HYBRID...Chen et al. (1994) HYBRID...

Nicely captures the 3 major physical processes of vertical entrainment-mixing:

1. Mixed layer entrainment-detrainment (related to atmospheric forcing)

2. Shear flow instability (Richardson-dependent)

3. Free convection in thermocline (instant adjustment)


• Gent & Cane (1989) OGCM coupled to an atmospheric mixed layer Gent & Cane (1989) OGCM coupled to an atmospheric mixed layer (just surface fluxes; Murtugudde et al., 1996) and (just surface fluxes; Murtugudde et al., 1996) and hybrid vertical hybrid vertical mixing scheme (Chen et al., 1994a)mixing scheme (Chen et al., 1994a). This model has been . This model has been extensively tested and used successfully in the tropical Pacific (e.g. extensively tested and used successfully in the tropical Pacific (e.g. Kessler et al., 1998; Chen et al., 1994a,b and many others).Kessler et al., 1998; Chen et al., 1994a,b and many others).

• Some modifications: higher-order Shapiro filter (8Some modifications: higher-order Shapiro filter (8thth), shorter time-step ), shorter time-step (30 min.), and grid-stretching in the zonal direction as opposed to just (30 min.), and grid-stretching in the zonal direction as opposed to just y-stretching…y-stretching…

• A total of 4 one-year climatology simulations… after being spun up onto A total of 4 one-year climatology simulations… after being spun up onto each grid independently (60 years plus “spin-overs”).each grid independently (60 years plus “spin-overs”).

CoarseCoarse Uniform zonal resolution, no GalapagosUniform zonal resolution, no Galapagos

Coarse+GCoarse+G Uniform zonal resolution, with GalapagosUniform zonal resolution, with Galapagos

FineFine Higher (stretched) zonal resolution, no GalapagosHigher (stretched) zonal resolution, no Galapagos

Fine+GFine+G Higher (stretched) zonal resolution, with Higher (stretched) zonal resolution, with GalapagosGalapagos

• Forcing was climatological ECMWF winds, Xie & Arkin precipitation, ISCCP Forcing was climatological ECMWF winds, Xie & Arkin precipitation, ISCCP clouds, and ERBE shortwave.clouds, and ERBE shortwave.

• Validation data used were Reynolds, TMI and TAO.Validation data used were Reynolds, TMI and TAO.


XRES: uniform 3/4, YRES: 1/3° stretching to 1°

XRES: 1/4° stretching to 1°, YRES: 1/4° stretching to 1°


1°S1°S

0°0°


CoarseCoarse FineFine

ResultsResults

Results: Results: Comparing annual mean SSTComparing annual mean SST

Coarse

Coarse+G

Fine

Fine+G

TMI

Contour Interval: 1°CShaded: SST < 23°CBold contour: 26°C


Coarse

Coarse+G

Fine

Fine+G

TMI



Coarse

Coarse+G

Fine

Fine+G

TMI



Coarse

Coarse+G

Fine

Fine+G

TMI


Results:Results: SST seasonal cycle SST seasonal cycle

A cold tongue indexA cold tongue index

(1998-2005 Annual Mean TMI SST)


Results:Results: Seasonal SST fields Seasonal SST fields

Coarse

Coarse+G

Fine

Fine+G

TMI

Mar-Apr-May Sep-Oct-Nov




Coarse

Coarse+G

Fine

Fine+G

TMI





Coarse

Coarse+G

Fine

Fine+G

TMI





Coarse

Coarse+G

Fine

Fine+G

TMI





2° x 22° x 2°° box indices box indices

(1998-2005 Annual Mean TMI SST)

Results:Results: SST seasonal cycle SST seasonal cycle Coarse Resolution

Results:Results: SST seasonal cycle SST seasonal cycle High Resolution

ResultsResults

The EUCThe EUC

(Sep-Oct-Nov)TAO Clim.TAO Clim. Adapted from Large, et al. (2001)

Results:Results: Zonal Currents and Temperature Zonal Currents and Temperature

CoarseCoarse

Coarse+GCoarse+G

FineFine

Fine+GFine+G

GODASGODAS

Results:Results: Zonal currents at 100m Zonal currents at 100m

* This is our model “off the shelf.” No tuning.

Results:Results: Zonal currents in the RA6 & GODAS Zonal currents in the RA6 & GODAS

(Adapted from Behringer and Xue, 2004)(Adapted from Behringer and Xue, 2004)

Mean Zonal CurrentsMean Zonal Currents RMSE w.r.t. TAORMSE w.r.t. TAO

FineFine+G

(Annual Mean)

Results:Results: Termination of the EUC Termination of the EUC

FineFine

Fine+GFine+G

Results:Results: Termination of the EUC Termination of the EUC

FineFine

Fine+GFine+G

(Eden & Timmerman, 2004)(Eden & Timmerman, 2004)

Results:Results: (Non-)Termination of the EUC (Non-)Termination of the EUC

Results:Results: The EUC from all angles The EUC from all angles

FineFine

Fine+GFine+G

0m 10m 20m 30m 40m 50m 60m 70m 80m 90m 100m 110m 120m 130m 140m 150m 160m 170m 180m 190m 200m


FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G



FineFine

Fine+GFine+G

Zonal Velocity at 93°WZonal Velocity at 93°W

Dept

h [1

0 m

]De

pth

[10

m]

Dept

h [1

0 m

]De

pth

[10

m]


FineFine

Fine+GFine+G


Dept

h [1

0 m

]De

pth

[10

m]

Dept

h [1

0 m

]De

pth

[10

m]


FineFine

Fine+GFine+G


Dept

h [1

0 m

]De

pth

[10

m]

Dept

h [1

0 m

]De

pth

[10

m]

Results:Results: Response of SST and A Response of SST and AO NHFO NHF

Fine+G – FineFine+G – Fine

NHF = Qsw – [ εσBSST4 + LV ρairCMUA( qs–qA) + CPρairCHUA(SST–θA) ]

Shading: ΔSSTShading: ΔSSTContours: ΔNHFContours: ΔNHF

every 10 w mevery 10 w m-2-2

(Sep-Oct-Nov)

(Vecchi et al., 2005)(Vecchi et al., 2005)

Results:Results: SST response SST response


(Sep-Oct-Nov)



(Sep-Oct-Nov)


Results:Results: LocalLocal SST response SST response

(Eden and Timmerman, 2004)(Eden and Timmerman, 2004)

(Sep-Oct-Nov)

Results:Results: Why results differ from ET-04?Why results differ from ET-04?

• Galapagos in ET-04 too small and entirely south of the equator…Galapagos in ET-04 too small and entirely south of the equator…

• The EUC in ET-04 is right on the equator, not ½° or so south…The EUC in ET-04 is right on the equator, not ½° or so south…

• Vertical mixing schemes?Vertical mixing schemes?

• ET-04 imposed surface heat flux with Haney-type formulation?ET-04 imposed surface heat flux with Haney-type formulation?

• ET-04 domain was 10°N to 10°S?ET-04 domain was 10°N to 10°S?

• ET-04 results are actually the average of 10 years of output (the 90’s) ET-04 results are actually the average of 10 years of output (the 90’s) …different from a true “climatology” run?…different from a true “climatology” run?

• Similar:Similar: Strong upwelling/colder SST right up against the Strong upwelling/colder SST right up against the west west side of the Galapagos Islands. Very side of the Galapagos Islands. Very locallocal..

• Different:Different: TerminationTermination of EUC = of EUC = broaderbroader tropical SST tropical SST response.response.

Why?Why?

Why might these improvements result Why might these improvements result from the termination of the EUC?from the termination of the EUC?

Results:Results: Temperature vs. Depth Temperature vs. Depth

(Sep-Oct-Nov)

More diffuse gradientMore diffuse gradient

Diffuse gradientDiffuse gradient

Sharp gradientSharp gradient24°C outcrops ~130°W24°C outcrops ~130°W

24°C outcrops ~115°W24°C outcrops ~115°W

24°C outcrops ~115°W24°C outcrops ~115°W

All averaged 2°N to 2°S


Temperature at 110°W (°C)

(Sep-Oct-Nov)All averaged 2°N to 2°S


dT/dz at 110W (°C per 10m)

(Sep-Oct-Nov)All averaged 2°N to 2°S

~MLD

xx

zz

ThermoclineThermocline

Results:Results: Dynamical adjustment Dynamical adjustment

ττxxWhy might these improvements result from the termination of the EUC?Why might these improvements result from the termination of the EUC?

SvSv

WarmWarm ColdColdSECSEC

For a given For a given ττxx and ocean geometry, you have an equilibrium equatorial dynamical balance. and ocean geometry, you have an equilibrium equatorial dynamical balance.

Change the geometry in such a way that disrupts one component of the balance, and you Change the geometry in such a way that disrupts one component of the balance, and you must have a new, must have a new, adjustedadjusted dynamical balance. dynamical balance.

Zonal Circulation

xx

zz

ηη

ThermoclineThermocline EUCEUC

SECSEC

Results:Results: Dynamical adjustment Dynamical adjustment

ττxxWhy might these improvements result from the termination of the EUC?Why might these improvements result from the termination of the EUC?

SvSv

WarmWarm ColdCold

PP



Zonal Circulation

xx

zz

ηη

ThermoclineThermocline EUCEUC

SECSEC

Results:Results: Dynamical adjustment Dynamical adjustmentWhy might these improvements result from the termination of the EUC?Why might these improvements result from the termination of the EUC?

SvSv

WarmWarm ColdCold

PP



Zonal Circulation

xx

zz

ηη

ThermoclineThermocline

SECSEC


SvSv

WarmWarm CoolCool

PP


Change the geometry in such a way that disrupts one component of the balance, and Change the geometry in such a way that disrupts one component of the balance, and you must have a new, adjusted dynamical balance.you must have a new, adjusted dynamical balance.

EUCEUC

Zonal Circulation

Gal

apag

osG

alap

agos

CoolCool


(Sep-Oct-Nov)

η = h(1) b(1) + Σ h(k) b(k)

g g

nz

k=2

b(k) = α g (T – Tref)

Zonal Circulation


Zonal Circulation

(Sep-Oct-Nov)

Realistic Indonesian throughflow?

(with 1σ shown)

1. Basin-wide adjustment of the zonal pressure gradient beginning with the EUC1. Basin-wide adjustment of the zonal pressure gradient beginning with the EUC

2. Slower SEC so reduced cold SST advection along the surface2. Slower SEC so reduced cold SST advection along the surface

3. Reduced SEC-EUC shear so 3. Reduced SEC-EUC shear so MUCHMUCH less vertical entrainment-mixing (EMX) less vertical entrainment-mixing (EMX)

4. All of these promote a deeper thermocline and warmer SST4. All of these promote a deeper thermocline and warmer SST

Results:Results: Dynamical adjustment Dynamical adjustment• Why might these improvements result Why might these improvements result

from the termination of the EUC?from the termination of the EUC?

Zonal Circulation


NorthNorthSouthSouth

Meridional Circulation



(Sep-Oct-Nov)




or, schematically…or, schematically…

(Sep-Oct-Nov)

* reduced speeds in meridional * reduced speeds in meridional overturning cellsoverturning cells


Results:Results: Entrainment Mixing Entrainment MixingWhy might these improvements result from the termination of the EUC?Why might these improvements result from the termination of the EUC?

(Sep-Oct-Nov)

Fine+G – FineFine+G – FineShading: ΔEMX (WmShading: ΔEMX (Wm-2-2))

Summary & implicationsSummary & implications


• The bottom line:The bottom line: Given sufficient zonal resolution (to get the EUC Given sufficient zonal resolution (to get the EUC right), the Galapagos Islands terminate the EUC and lead to improved right), the Galapagos Islands terminate the EUC and lead to improved modeling of SST in the Pacific cold tongue region, and for good modeling of SST in the Pacific cold tongue region, and for good reasons… adjusted equatorial dynamics and less entrainment-mixing.reasons… adjusted equatorial dynamics and less entrainment-mixing.

Isla Isabela, Galapagos Archipelago From NOAA R/V Ka’imimoana 2 May 2005

Galapagos in Fine+G

Summary & implicationsSummary & implications Overall SST Improvement

ReynoldsReynolds

CoarseCoarse(Sep-Oct-Nov)

(Sep-Oct-Nov)

Fine+GFine+G TMI

TMI





• The cold bias problem has for some time been a mainstay of OGCMs.The cold bias problem has for some time been a mainstay of OGCMs.

• These SST improvements could lead to substantial improvements in These SST improvements could lead to substantial improvements in tropical cloud tropical cloud and precipitation patternsand precipitation patterns, ocean biological productivity, and carbon cycling in , ocean biological productivity, and carbon cycling in coupled models.coupled models.

• The GODAS cold bias could be improved by…The GODAS cold bias could be improved by…

(a) increasing the zonal resolution- just enough to get the EUC right(a) increasing the zonal resolution- just enough to get the EUC right

(b) adding the Galapagos Islands. This should be given (b) adding the Galapagos Islands. This should be given consideration in the next iteration of the NCEP GODAS.consideration in the next iteration of the NCEP GODAS.

• Future work should address the impact of the Galapagos Islands on interannual Future work should address the impact of the Galapagos Islands on interannual variability (e.g. ENSO events) and variability (e.g. ENSO events) and predictionprediction of seasonal and interannual variability in of seasonal and interannual variability in the tropical Pacific (e.g. NINO forecasts).the tropical Pacific (e.g. NINO forecasts).

Thank You.Thank You.

University of MarylandMODIS (Terra) true color March 12, 2002 Courtesy NASA GSFC

Improvements to the Equatorial Pacific Cold Tongue Region in an OGCM

Documents

tropical cold biasproblem

precipitation patterns

ocean biological productivity

carbon cycling

realistic tropical cloud

pacific upwelling

zonal band of cold sst

west introduction