Circulation East of Australia

Circulation East of Australia


Links to other pages:

     1a. TAO/model comparisons
     1b. Model exploration 1 (E Pacific mixing)
     2.   Model Exploration 2 (Other things)
     3.   Model Exploration 3 (circ E of Australia)    (This page)
     4.   Two-delta-x wave?
     5.   Throughflow XBT lines

     6.   Writeup

     7.   Continuation of work at PMEL


Take careful note because several different model runs have been used here. All work is with the ACOM3 version of MOM. The early work (before 28 March 02) was done with the model file subdomain_new, which had tidal mixing. This in general increased the viscosity/diffusivity considerably, and the circulation and thermal structure are in some cases quite different. For reasons I do not understand, the annual cycles also changed for some quantities. (See the section below "comparing the tides/notides version of MOM").

Later work was done with the notides model (file subdomain_notides), and plots are gradually being replaced. Plots should be labeled correctly, however, early plots may not be, and I have not had time to remake all plots. If the label is not clear, a good way to tell is to look at the plot and "View -> Page info" and check the date. The notides run was completed late on 27 March, and plots made before that are always with tides. Since there was no sample file made with the notides run, all plots showing subsampled fields (everything E of 180°) are always with tides.

Note regarding streamfunction plots here. Unfortunately, I used different sign conventions for the streamfunctions in different situations (leftover old scripts). Most of the time, it is defined: u=-Psi_y, v=Psi_x, as it should be. However, the streamfunction plots based on vertically-integrating the ACOM3 T and S are the reverse: u=Psi_y, v=-Psi_x. This is the case for plots 1.2, 1.3 and 2.1. Sorry!


  1. Mean quantities
    1. Mean sea level around Australia
    2. Mean currents: Maps   v by latitudes   v by depths   Meridional transport
      Streamfunction   Overlay vectors
    3. Mean steric height (a la RG 97 Fig 2: h, P, sigma(h)):   Rel 1000m 
      Sverdrup and total streamfunction   (Another version)
    4. Mean U and T east of Australia

  2. Annual cycles
    1. Annual cycle of the EAC (bimonthly maps): 
      a. V(z), include mean:  Surface   105m   209m   304m   506m   889m   1582m
      b. V(z), anomalies:  Surface   105m   209m   304m   506m   889m   1582m
      cTransport anomalies
      d. Snapshot "streamfunction": Totals   Anomalies
      eDISH anomalies
    2. 1 cpy Amp/phase   With overlay
    3. Annual cycle of shelf currents (a la RG 97 Fig 3)   Topography on the (u,v) grid
    4. A section across open water E of Australia but W of New Caledonia (158°E-163°E): 
      a. Surface u and T: Totals   Anomalies
      b. Vertical section of u and T: Totals   Anomalies
      c. Zonal transport: Totals   Anomalies

  3. A linear Rossby model:
    Driven by average annual cycle winds

    Note: (7 Apr 02) A factor of 2 error was discovered in the code that integrated along the RW characteristics. Ooops! This makes a linear error in the result. Unless specifically noted here, all magnitudes of the RW solutions should be multiplied by 2 (which improves things!).

    1. Model mean = Sverdrup pressure
    2. Annual cycle of S Pacific gyre: No damping   24-month damping timescale   No damping (include mean)
    3. Comparison ACOM3 offshore Australia
    4. Comparison along coast   Remove ACOM3 33°S signal   
    5. Compare offshore and along coast: ACOM3   Linear RW model
    6. Compare DISH: 
      aJan-Jun   Jul-Dec (factor of 2 fixed)
      b. With 36-month damping: Jan-Jun   Jul-Dec (factor of 2 fixed)
      c. With RW EBC: Jan-Jun   Jul-Dec
      d. (old: RW DISH   ACOM3 DISH)
    7. 1 cpy amp/phase of DISH
    8. Eastern BC   Model mean DISH incl eastern BC
    9. Annual cycle of S Pacific winds: Tau-x   Tau-y   Curl   

    This model has not yet been run with interannual winds. That is one of the next tasks.

  4. Island Rule (ERS winds)
    These calculations assume that the ocean is in steady Sverdrup balance on annual and interannual timescales (!?!). The annual cycle is calculated monthly, and interannual variations are done with 12-month average winds.

    1. Mean Island Rule streamfunction   Color shift shows contour around Australia.

      Annual cycle

    2. aStreamfunction around Australia   Anomalies   Color shift shows bifurcation of SEC
      b. Compare ACOM3 streamfunction near EAC: Island rule   ACOM3 (reverse sign: u=Psi_y, v=-Psi_x)
    3. Bifurcation latitude at Australian coast
    4. ITF transport   Compare ACOM3 model   Compare reality (Wijffels IX1)

      Interannual variations

    5. Streamfunction around Australia   Anomalies   Color shift shows bifurcation of SEC   Detail
    6. Bifurcation latitude at Australian coast
    7. ITF transport   Compare ACOM3 model   (Show 0-443m separately)   Total transport   Compare reality (Wijffels IX1)

      Unwrapping the wind along the Island Rule paths

    8. Paths
      Annual cycle:
    9. a. Winds along paths: NZ   Australia   Madagascar   
      b. Anomalies: NZ   Australia   Madagascar   
      c. Show transport, too: NZ   Australia   Madagascar   
    10. Transports
      Interannual:
    11. a. Winds along paths: NZ   Australia   Madagascar   
      b. Anomalies: NZ   Australia   Madagascar   
      c. Show transport, too: NZ   Australia   Madagascar   
    12. Transports

    13. It would be very desirable to compare the island rule transports with those from the ACOM3 model. Unfortunately this is not possible unless the model is saved at full resolution (which it is in the ITF region, as in the comparison plots above (4.3 and 4.6)). In other regions (i.e. between New Zealand and S America), the only saved output is at coarse spatial resolution, which presumably aliases eddies and poorly samples the WBC along the coast of New Zealand.
            This is illustrated by mean meridional transport between New Zealand and S America (function of y) These transport should be identical at all latitudes, and they are not even close. Also see a map of mean meridional transport, which hints at the aliasing, and coast-to-coast transport as a function of (y,t). (See the discussion below, under the heading "Cross-equatorial transport in ACOM3 sliced different ways", of similar errors in finding transport across the equator.)
            -> Conclude: The subsampled fields are not useful for velocities!
            See a page that discusses this problem in more detail.
            The suggestion from the Island Rule calculations above is that annual cycle transport between New Zealand and S America varies from zero in June to 45 Sv in February. Is this true? Similarly, does the transport around Madagascar change sign during the year? (see plots at 4.8).
            It is, however, possible to find the ACOM3 transport between Australia and New Zealand, and that should correspond to the difference between the Island Rule transports around Australia minus those around New Zealand:
            Island Rule AU-NZ differences: Annual cycle   Interannual
            Compare ACOM3 transports: Annual cycle   Interannual
      Not very encouraging!
            But let's not forget the difference between ERS and ACOM3 winds ....  Means: Tau-x   Tau-y   Vectors   
            Annual cycles: NZ   Australia   

    14. Compare Island Rule and ACOM3 DISH between Tasmania and Chile:
      a. Island Rule: Island Rule transport (geostrophic and Ekman)   Delta-P
            Ekman (Sverdrup) transport across Equator: Equator   Divergence Eq/44°S   ERS 1997 anomalies
            Advective and friction terms do not make a big difference (G/C model)
      b. Tasmania: See these plots for location of point T: Dish(t) along the Tasmanian coast   Anomalies   
            Mean ACOM3 DISH   Std dev ACOM3 DISH   A few likely points for T   
      c. Chile: Mean ACOM3 DISH   A few likely points for S   
            Dish(t) along the Chilean coast   Anomalies
      dDISH at points T and S   (Rel 1000m)   dP/dx between Tasmania and S America
      edP/dx based on absolute pressure   Overlay Island Rule dP/dx

      Storage estimation:

    15. From sea level: Region selected   Volume change   Implied transport   
    16. p and eta along 43.5°S between NZ and Chile
    17. From depth of sigma=27.6:  Mean depth   Zonal average depth   (Demeaned)   Depth difference (Yr 8 - Yr 1)   Upper layer volume   Implied transport   

    18. Time series along the S American coast:
      a. Density time series: Equator   20°S   40°S   
      b. Lower isotherms are rising while shallow isotherms are deepening: Temperature at 40°S, 90°W
             (Consistency check on sigma=27.6 plots above).
      c. Steric Ht: 1-yr RM   Anomalies   Detail during 1997-98   
      d. DISH: 1-yr RM   

    19. Similar plots (Island Rule and ACOM3) from the annual cycle:
      aDish at best-guess points T and S   Delta-P across S Pacific (IR and ACOM)   
      b. 2nd-best guess: DISH at T and S   Delta-P across S Pacific   

    20. Sill depth of Throughflow:
      aSill depth of throughflow
      b. Detail of Omboi Strait: k=23 (679m) (open)   k=24 (889m) (closed)
    21. ITF transport each strait: Mean vertical structure across ITF   Time series   Overlay NoTides model
      Compare similar plots from ACOM2 model on another page (see plots 2.3).

      Combine a Rossby model with instantaneous Island Rule:

    22. Mean Island Rule DISH along S America and Australia   Mean DISH near Australia
      Annual cycle:
    23. Time series of DISH along the coast and at 160°E: Totals   Anomalies

      Cross-equatorial transport in ACOM3 sliced different ways:
            Attempts to find the total cross-equatorial flow by combining the subsampled data east of 180° and the full resolution data to the west produced hugely inflated transports (especially clear in the "running integral" plots below). Problems with the subsampled fields are illustrated by v N of New Guinea and extended to 180°.   Time series of the difference show that it is systematically positive (by chance???). The vertically-integrated transport differs by a factor of 5! Between 130°E and 180° (where direct comparison is possible) to be 89 Sv from subsampled data but 16.4 Sv from the full resolution. It is likely that the plots below are not believable! (See also the discussion of transport between NZ and S America under the heading "Unwrapping the wind along the Island Rule paths" above).
      See a page collecting plots of this phenomenon (2-delta-x-wave).

    24. Some summary plots: Mean v(x,z)   Std dev (30-day sampling)   Std dev (12-month RM)   
    25. Time series (V(x,t): Top-to-bottom   By levels   By levels (anomalies)
    26. Transport integrated across various depth and longitude ranges:
      Split by longitudes. These plots show integrated from the eastern boundary to 180°, 140°E, the western boundary transport (120°E-140°E) and the total. Shallow plots overlay Sverdrup (Ekman) transport.
            0-100m   0-150m   0-200m   0-500m
            100-200m   200-500m   500-1000m   1000-2000m   2000-bottom   
    27. Running integrals in depth (overlay Sverdrup transport): Whole basin   E of 140°E

  5. Coastal sea level around Australia/New Guinea
    Gridpoints were hand-selected for the nearest point to the coast (on the T grid), and for the point just outside the 1000m isobath.

    Unfortunately, these were all done with the tides model, and have not been remade ....

      First set just included the New Guinea coastline.
      See the later set below with the coast extended to 20S along eastern Australia ....
    1. Coastal point selection: Topography   Both (plain)   Both (overlay mean eta)   Coast points   1000m isobath points   
    2. Mean Eta and DH
    3. Timeseries comparing coastal and 1000m sections: Total   Anomalies   Annual cycle   
    4. Timeseries comparing 1000m and equatorial sections: Total   Anomalies   Annual cycle   

      Extending the coast to 20°S along Australia:

    5. Coastal point selection: Topography and coastal points   Distance along coast   Mean sea level   Compare mean SL and DH
    6. Timeseries comparing coastal and 1000m sections: Total   Anomalies   Annual cycle   
    7. Timeseries comparing coastal and equatorial sections: Anomalies   Annual cycle   

      Now extend it all the way to Bass Strait:

    8. Distance along coast   

      1st annual harmonic of sea level:

    9. Maps: Whole region   Coastal region   Detail region   
    10. N coast only: Along coast   Along coast and Eq
    11. Full coast: Along coast   

      Variance of sea level in frequency bands:

    12. Overall RMS: Full region   Detail   
    13. Interannual RMS: Full region   Detail   
    14. High-pass RMS: Full region   Detail   

      Tau along the coast:

    15. Tau gridpoints along the coast
    16. Annual cycle: Vectors   Amp/Phase
    17. Interannual vectors
    18. Annual cycle of S Pacific winds: Tau-x   Tau-y   Curl   
    19. Coastal sea level data:
      1. Hunter SL stations
      2. 1 cpy Annual cycle: Amplitude   Phase

  6. Comparing the tides/notides versions of MOM
    Some of these plots bring together plots that are also in other sections for comparison .... Here see the tides/no-tides differences.
      Direct comparisons and difference fields:
    1. Streamfunction   Zonal transport (map)   Zonal transport at 160°E
    2. Transport between Australia and New Zealand: Annual cycle   Interannual

      Other plots made with the individual runs:

    3. DISH bimonthly anomalies: No Tides   Tides
    4. Streamfunction: No Tides   Tides
    5. Streamfunction anomalies: No Tides   Tides
    6. Mean steric height: No Tides   Tides
    7. Meridional current along the coast (by lats): No Tides   Tides
    8. Meridional current along the coast (by depths): No Tides   Tides