Xujing Jia Davis

1
Numerical and Theoretical Investigations of
North Pacific Subtropical Mode Water with
Implications for Pacific Climate Variability

• Xujing Jia Davis
• Graduate School of Oceanography, University of
  Rhode Island
• Lewis M. Rothstein
• Graduate School of Oceanography, University of
  Rhode Island
• William K. Dewar
• Department of Oceanography, Florida State
  University,
• Dimitris Menemenlis
• Jet Propulsion Laboratory, California Institute
  of Technology

2
North Pacific Subtropical Mode Water (STMW)
Schematic current patterns in western North
Pacific
Location forms and resides south of Kuroshio
Extension (KE)

• Features
• - weakly stratified, low PV
• - upper 500 m of the ocean
• water column
• inhabits thermostads
• between 16 and 19C

STMW formation region
- salinity range of 34.65-34.8psu -
potential density range of 24.8-25.7 kg/m3
(Masuzawa, 1969 Suga et al., 1990 Eitarou et
al., 2004)
3
Questions

• STMW has known seasonal variability, but what is
  the variability of STMW on longer time scales (as
  revealed by models)?
• Do models properly capture the seasonal
  variability of STMW volume?
• What is the relationship (if any) between low
  frequency variability of STMW and known climate
  patterns in the Pacific?
• Supporting dynamics?

4
Top Down Modeling

• MITgcm 3D, z-level, primitive equation OGCM
  (Marshall, 1997)
• ECCO2 global-ocean and sea-ice simulation
• North Pacific data from two ECCO2 MITgcm
  simulations are extracted for analysis
• Cube37 simulation 28-year spin-up prior to its
  initial January 1992 conditions, 1992-2000 NCEP
  forcing converted to fluxes using model SST
  (Large et al, 1981, Menemenlis, 2005)
• Cube76 simulation driven by ERA40 atmospheric
  surface boundary conditions, followed by ECMWF
  analysis after August 2002 when the ERA40
  reanalysis stops weak relaxation to
  climatological seasonal cycle of sea surface
  salinity.
• Common horizontal resolution 1/6o x 1/6o 50
  levels
• Common temporal coverage 1992, Jan 2006 Mar
  (171 months)
• Output has not been constrained by oceanic and
  sea-ice data

5
STMW Region Definition
- 130E 160W, 20N 40N and east of islands of
Japan

6
Compare with Observations
KESS, late May 2004
MITgcm (Cube76), May 2004
MITgcm (Cube37), May 2004
After Rainville,et al., 2007
Meridional cross section at 145 E of potential
density, PV and zonal velocity in MITgcm
experiments and observations. The stippled areas
are the observed and modeled STMW.
7
Temporal Variability STMW Volume
Dominant signal is seasonal Lower frequency
variability apparent
(Cube 37)
8
Seasonal Cycle
9
STMW Seasonal CycleThree Distinct Periods
STMW Volume
Period I STMW is exposed to atmos. forcing
Period II STMW is partially isolated from atmos.
forcing Period III STMW is completely isolated
from atmos. forcing
10
Interannual Variability
11
STMW Interannual Variability
(Cube 37)
12
Pacific Decadal Oscillation (PDO)
SST
Warm phase Cooler SST in STMW region Cool Phase
Warmer SST in STMW region
Warm Phase                                Cool
Phase
4 year Cool Phase
1976/77
1998/99
warm phase
cool phase
www.jisao.washington.edu/pdo/
13
STMW Variability PDO
PDO index (top) and STMW volume in MITgcm
simulations (bottom)
STMW volume variability is correlated with the
PDO index Co0.69 (Cube37) Co0.80 (Cube76)
14
Connection Between STMW PDOLarge-scale
Atmospheric Variability
15
Connection Between STMW PDOLarge-scale
Atmospheric Variations from NCEP
1st EOF (37.7)
1st EOF (37.7)
More heat loss from ocean to atmos.
Weaker Ekman pumping
Year 1996, STMW maximum
16
Connection between STMW PDOLarge-scale
Atmospheric Variations from NCEP
1st EOF (37.7)
1st EOF (37.7)
Less heat loss from ocean to atmos.
Stronger Ekman pumping
Year 1999, STMW minimum
17
Summary (MITgcm)

• The dominant temporal pattern of STMW is
  seasonal the annual cycle can be divided into
  formation, partial isolation and complete
  isolation periods that correspond to distinct
  stages of STMW evolution. Strong air-sea
  interaction is the main feature of the STMW
  formation period.
• An interannual signal is clearly seen in STMW
  variability, and this lower frequency signal
  shows significant correlation with the PDO index
  
• This likely results from variations in large
  scale atmospheric forcing wind stress and/or
  surface heat flux

18
Planetary Geostrophic Theoretical Framework A
Ventilated Thermocline Model of STMW(Dewar et
al. 2005)

• Modified LPS theory is used to describe STMW and
  its connections to large scale ocean/atmosphere
  circulation
• An analytical model of STMW, based on ventilated
  thermocline theory.
• Ventilated Pool Hypothesis Assumes that all of
  the water in the pool (i.e. circulating) region
  is ventilated and, therefore, that all the
  Sverdrup transport is carried in the uppermost,
  outcropped layer.
• A PGOM (Samelson Vallis, 1997) numerically
  approximates the solutions to this theory and is
  used to describe/diagnose STMW characteristics/dyn
  amics.
• PGOM experiments were performed to help interpret
  the role of large-scale wind stress curl and the
  local heat flux in forcing STMW variability, as
  seen in the MITgcm simulations.
• Reasonably simulates the analytical solutions of
  PG framework.

19
Planetary Geostrophic Theoretical Framework
Results

• The formation of a deep, vertically homogeneous,
  fluid layer in the northwest corner of the
  subtropical gyre that extends from the surface to
  the base of the ventilated thermocline.
• This ventilated pool is the model analog of the
  observed STMW.

20
PGOM Model Description
Designated SV97, adapted from Samelson and
Vallis 1997, Dewar et al, 2005, based on PG
Equations (non-dimensionalized)

Momentum
Hydrostatic
Continuity
Thermal
Salinity
Eq of state
21
PGOM Model Domain

• Domain
• - Non-dimensionalized box the dimensional
  values are 5000 km horizontally, 5km vertically
• - Central latitude 35N, y2500 km,
• zero Ekman pumping line at y24N and y46N
• Resolution
• - 80 km horizontally
• - vertically stretched to give more resolution
  in the thermocline

Zero Ekman Pumping
5000km
5km
5000km
22
Ekman Surface Layer

• Upper B.C. on T and w Simple slab model of
  frictional surface B.L., fixed depth
• The vertically integrated Ekman balance is
  assumed to hold

Ekman Surface Layer

• TE is obtained from the vertically
• integrated thermodynamic
• equation

Adapted from Welander, 1971
23
2 Classes of PGOM Experiments

• Steady forcing To diagnose the characteristics
  of STMW circulation and upper ocean structure
  under constraints of the modified ventilated
  thermocline scheme.
• Time-dependent forcing To investigate dynamics
  associated with STMW variability for reasonably
  realistic
• - basin-scale wind forcing (in the form of
  Ekman pumping)
• - heat flux (isolated to the northwest corner
  of the STG)

24
PGOM Experiment 2 DesignTime Varying Forcing
Heat flux varying experiment
Wind stress curl varying experiment
25
Results Time-dependent Ekman Pumping
Ekman pumping amplitude (x10-4cm/s)
STMW volume (m3)
x1015
O(1013 m3)
O(1015 m3)
.

• Both seasonal and interannual variations are
  simulated

• Stronger Ekman pumping, more mode water

• Interannual change, with 12 year lag

• The STMW volume is O(1015 m3)

• The magnitude of the oscillation annually
  O(1013 m3)

26
Results Time-dependent Heat Flux
Air-sea heat flux coefficient
STMW volume (m3)
x1014
O(1014 m3)
O(1015 m3)

• Seasonal and Interannual variations simulated

• More heat loss, more STMW volume, with 12 year
  offset

• The STMW volume is O(1015 m3)

• The magnitude of the oscillation annually
  O(1014 m3)

27
PGOM Results

• PGOM experiments demonstrate that the
  interannual variability observed in the MITgcm
  simulation can be driven by variations in the
  large scale air-sea heat flux (zero lag) and wind
  stress patterns (2 year lag) seen in the NCEP
  reanalysis
• The variations in air-sea heat flux play the
  more dominant role during the period in he late
  1990s when STMW discharges its volume, with
  influence of one order of magnitude larger than
  the varying wind stress

28
Acknowledgements

• Roger M. Samelson (OSU)
• Geoffrey K. Vallis (GFDL)
• Young-Oh Kwon (WHOI)
• ECCO2 Estimating the Circulation and Climate of
  the Ocean, Phase II, which is sponsored by the
  NASA Modeling Analysis and Prediction (MAP)
  program

29
Numerical and Theoretical Investigations of
North Pacific Subtropical Mode Water with
Implications to Pacific Climate Variability

• Xujing Jia Davis
• Graduate School of Oceanography, University of
  Rhode Island
• Lewis M. Rothstein
• Graduate School of Oceanography, University of
  Rhode Island
• William K. Dewar
• Department of Oceanography, Florida State
  University,
• Dimitris Menemenlis
• Jet Propulsion Laboratory, California Institute
  of Technology

30
STMW Definition in PGOM

• In the region of
  and
• roughly in the subtropical gyre
• and east of the western boundary current.

31
Seasonal CycleMeridional cross sections
32
Seasonal CycleMeridional cross sections
33
Seasonal CycleMeridional cross sections
34
Seasonal CycleMeridional cross sections
35
Seasonal CycleMeridional cross sections
36
Seasonal CycleMeridional cross sections
37
Experiment 1 Constant Atmosphere Forcing
Ekman pumping (WE)
Initialization from motionless, 18000 years
integration diffusive time TH2/kv , H3750 m,
kv2.5x10-5 m2/s
38
Experiment 1 Thermocline Structure
at the center of the domain
I. ventilated regime
STMW
W0
II. diffusive regime
T (C)
Tz (C/m)
Tzz (C/m2)
W (m/s)
39
Experiment 1 Thermostad
Zero Ekman Pumping line

• A thermostad is found between the two Tz
  maximums, or between the two thermocline regimes
• The isopycnal surfaces of the shallow themostad
  are in the ventilated thermocline
• Around the bottom of the themostad, the isopycnal
  surfaces in the internal thermocline