nested models: a synthesis tool tom (zack) powell, uc berkeley u. s. globec pan-regional synthesis...

Nested Models: A Synthesis Tool

Tom (Zack) Powell, UC Berkeley

U. S. GLOBEC Pan-Regional Synthesis Meeting

27-30 Nov, Boulder, CO

“...concentrating on a few systems is preferable.”

“HAS GLOBEC EVOLVED A SUCCESSFUL STRATEGY TO ATTAIN ITS

GOALS?”...“...The effect of greenhouse gas-induced global change is expected to be felt in all oceans, and GLOBEC, to be truly

global, could have devised a program that would simultaneously study all, or at least, many ocean ecosystems

simultaneously.”

- J. Knauss et al., Dec, 1997.

“...Understanding how a given ecosystem will respond to global change generated by greenhouse gas warming is much more difficult and unlikely to be achieved within the lifetime of this

program.”

- J. Knauss et al., Dec, 1997.

“SUMMARY”...

“...concentrating on a few systems is preferable.”

“HAS GLOBEC EVOLVED A SUCCESSFUL STRATEGY TO ATTAIN ITS

GOALS?”...“...The effect of greenhouse gas-induced global change is

expected to be felt in all oceans, and GLOBEC, to be truly global, could have devised a program that would

simultaneously study all, or at least, many ocean ecosystems simultaneously.”

- J. Knauss et al., Dec, 1997.

“...Understanding how a given ecosystem will respond to

global change generated by greenhouse gas warming is much more difficult and unlikely to be achieved within the

lifetime of this program.”

“SUMMARY”...

1. Nested Models

2. Present Assessment

3. Future Prospects

4. Proposal

OUTLINE

Nested Models

Nested Model Hierarchy in the North Pacific

NPac (40 km) NEP (10 km) CCS&CGOA (3 km) PWS (1 km)

40 km 10 km 3 km 1 km

Free surface, hydrostatic, primitive equation model

Generalized, terrain-following vertical coordinates

Boundary fitted, orthogonal curvilinear horizontal coordinates on an Arakawa C-grid

Non-homogeneous time-stepping algorithm

High-order advection schemes

Accurate baroclinic pressure gradient (FV/splines)

Continuous monotonic reconstruction of vertical gradients

ROMS: KERNAL ATTRIBUTES

Model Setup. I.

-Ocean Model: ROMS (community model)

-Domain: Pacific basin; 30S to 65N, 100E to 70W. 476x238 Horizontal points, 30 vertical layers

-Spinup: 10 years with climatological fluxes and daily winds

-Hindcast: 1990-2003, NCEP winds and fluxes. Also 2000-2002 Qscat winds

Model Setup. II.

-Air-sea interaction boundary layer from COARE (Fairall et. al. 1996)

-Oceanic surface boundary layer (KPP; Large et. al. 1994)

-Bathymetry: ETOPO5, isobaths do not intersect coastline

Data

(Bond et al.)

Model (NPAC)

SST anomaly 1999-2002

California Current

Aug 022002

SSTSeaWiFS Chl

Jul 312002

Denman and

Abbott 1994

(pigment, CZCS)

[P,T] decorrelation times (Denman and Abbott 1994)

50 - 100 km

25 - 50 km

Domain: 35-48 N, 120-134 W

514 x 510 x 30 gridpoints

NCEP daily forcing (wind, heat flux)

Initial, boundary conditions from NEP model

Start: 1 Jan 2000

M-Y mixing

Forward run (no data assimilation)

CCS Implementation(ROMS)

. _ . _ . Stats

___ Calcs

Spitz et al. 2003 (JGR 108)

Surface

Section

Satellite Model

Present Assessment

Domain: 20 - 73N, 115 – 210E

ROMS: 226 x 642 x 42 gridpoints

Subdaily (6 hr) T42 CORE wind and fluxes (Large and Yeager)

Initial/boundary conditions provided by CCSM-POP hindcast model

Forward run for 1958-2004—includes multiple El Nino’s, Regime Shifts, and 2002 cold intrusion

Daily averaged physical snapshots of velocity, temperature, etc.

Especially want to thank Enrique Curchitser (Rutgers) and Kate Hedstrom (UAF) for providing these model fields.

NEP Implementation

ATM

OCE

LND CPL ICE

CCSM 3.0

Domain: 20 - 73N, 115 – 210E

ROMS: 226 x 642 x 42 gridpoints

Subdaily (6 hr) T42 CORE wind and fluxes (Large and Yeager)

Initial/boundary conditions provided by CCSM-POP hindcast model

Forward run for 1958-2004—includes multiple El Nino’s, Regime Shifts, and 2002 cold intrusion

Daily averaged physical snapshots of velocity, temperature, etc.

Especially want to thank Enrique Curchitser (Rutgers) and Kate Hedstrom (UAF) for providing these model fields.

NEP Implementation

How well do the ROMS NEP physics match our perception and data from the real NEP ocean?

1) Compare SSH from the model with altimetry

2) Large scale climatology

3) Seasonality

4) Compare SST

5) Compare Subsurface Temperatures

6) Interannual Variability in Strength of the Alaskan Gyre Circulation and Bifurcation of the North Pacific Current (particle tracking)

Climatological Seasonal SSH Anomaly from Strub and

James (2002)

MODEL

Both are contoured at 2cm intervals

Positive SSH AnomaliesAlong Coast

AC – IntensifiedCC - Weakened

January SSH Anomaly

Climatological Seasonal SSH Anomaly from Strub and James (2002)

MODEL

Both are contoured at 2cm intervals

Negative SSH AnomaliesAlong Coast

AC – WeakenedCC - Intensified

July SSH Anomaly

Subsurface Temperature: Line P Seasonal Climatology

Model has warmer subthermocline temperatures, and weaker cross-shore isotherm slope suggesting model is underestimating Alaska Gyre current velocity.

From

Line P

web site

May

May

1987 Trajectories

1989 Trajectories

15 April –14 June

May

May

1987 Trajectories

1989 Trajectories

May 87

May 89

Summary1) The current NEP model hindcast for 1958-2004 is a significant improvement over earlier

simulations.

2) Climatological SSH and SST match observations well at largest scales.

3) Model shows reasonable seasonal variation in coastal velocities (intensified AK gyre in winter; intensified CC in summer).

4) Modeled subsurface seasonal temperatures (at 200-600m) along Line P are ca. 2°C warmer than observations.

5) Alaska Gyre geostrophic transport through Line P is weak relative to observations.

6) 10 m particle trajectories respond to instantaneous velocities including substantial mesoscale velocity (eddies).

7) However, is the NEP model product sufficient to use as the basis for coupled biophysical modeling of salmon, euphausiids, and LTL dynamics?

8) Temperature is important for bioenergetic models, but juvenile salmon are mostly in the upper 20-30 meters (where the modeled T may be OK).

9) Mixed layer depth and stratification intensity are important for controlling nutrient flux into the euphotic zone. Increased nutrient fluxes due to weak stratification may lead to early and possibly sustained phytoplankton blooms, etc.

Future Prospects

“HAS GLOBEC EVOLVED A SUCCESSFUL STRATEGY TO ATTAIN ITS

GOALS?”...

“...lack of a specific program tied to investigate nutrients and food production for the targeted organisms;...”

- J. Knauss et al., Dec, 1997.

NEMURO.FISH

Observations- Biological and Physical

NCEP 6 hourly data1948-2002

(includes interannual variability)

COCO – COCO – TokyoTokyo

3-D Nemuro3-D Nemuro

Zooplankton andZooplankton and temperature temperature

time seriestime seriesNemuro Pacific herring model

Validation

Time series output

PROPOSAL

*

ROMS IMPLEMENTATIONS

**

US GLOBEC

OTHER

**

*

*

*

*

*

** *

1. Simultaneously drive all U. S. GLOBEC regional ROMS + ecosystem models with a

single, global climate model.

2. Simultaneously drive ALL GLOBEC International regional ROMS models with a

single, global climate model.

3. Simultaneously drive all GLOBEC regional ROMS + E-2-E ecosystem models with a single,

global climate model.

Ultimately

END