CANDIE

1
Circulation and Drift Pathways in the Northwest
Atlantic Ocean
Jinyu Sheng Oceanography, Dalhousie University
2
Collaborators Richard Greatbatch (DAL) Dan
Wright (BIO/DAL) Peter Jones (BIO) Kumiko
Azetsu-Scott (BIO) Liang Wang (DAL) HuaLin Wong
(DAL)
3

• Financial support provided by
• Canada Foundation for Innovation (CFI)
• Atlantic Canada Opportunities Agency (ACOA)
• Department of Fisheries and Oceans (DFO)
• Meterological Service of Canada (MSC)
• NSERC
• MARTEC

4
Contents

• Introduction
• Semi-Monthly Composite SST Images
• Circulation Determined from Drifters

• Ocean Circulation Model CANDIE
• Diagnostic, Prognostic, and Semi-Prognostic
  Modes

• Model Results
• Seasonal Temperature and Circulation
• Comparison with Observations
• Pathways and Vertical Mixing of Tracers

• Animations
• Surface Temperature and Currents
• Pathways of Passive Tracers

5. Summary and Conclusion
5
Introduction

• NSERC/MSC/MARTEC established two Industrial
  Research Chairs in Regional Ocean Modelling and
  Prediction at Dalhousie University.

• The main objective of the Chairs is to develop
  ocean models that will be used for predicting
  atmosphere-ocean-ice conditions in the Atlantic
  region of Canada.

• Eastern Canadian seaboard is one of the most
  challenging marine environments in the world

Schematic showing major currents in the Northwest
Atlantic Ocean. (By courtesy of Dr. Igor
Yashayaev)
6
(No Transcript)
7
Semi-Monthly Composite SST Images in 1999
(By courtesy of Biological Oceanography Section,
BIO)
8
Semi-Monthly Composite SST Images in 1999
(By courtesy of Biological Oceanography Section,
BIO)
February
May
August
November
9
(No Transcript)
10
Trajectories of Sub -Surface Drifters
(Lavender et al., 2000. Nature, 407)

• More than 200 subsurface floats (PALACE, SOLO)

• Drifting at 400, 700, and 1500m for several
  days, then ascends to the surface to transmit
  data via Argos.

• Measuring T and S upon ascent or descent (7400
  TS profiles).

Red, orange and green arrows are three floats
drifting at 700m. Blue and purple arrows are two
floats drifting at 1500m.
11
Mean circulation at 700m
(Lavender et al., 2000. Nature 407)
Blue arrows are speeds of less than 5 cm/s
(distance traveled over 30 days). Red arrows are
speeds of greater than 5 cm/s (distance traveled
over 8 days).
12
Deep Convection in the Labrador Sea
(Lilly et al., 1999. JPO, 29)
Yearlong temperature record in the Western
Labrador Sea contoured over ten instruments
between 110 and 2510m. White lines show the
pressures at the instruments depth.
13
CANDIE Primitive Equation Ocean Model

• CANDIE stands for CANadian version of Diecast.

• Diecast was developed by Dietrich, C-grid CANDIE
  was developed by Sheng, Wright, Greatbatch, and
  Dietrich (1998), and A-grid CANDIE was primarily
  developed by Wright based on A-grid Diecast.

• Other contributors include Youyu Lu (free
  surface), Liang Wang (passive tracers), Bill
  OConnor (CANDIE Users Guide), David Brickman
  (partial cell), HuaLin Wong (CANDIE website),
  etc.

• A three-dimensional (3D), fully non-linear,
  primitive equation, finite-difference, z-level
  model.

• CANDIE has been subjected to several rigorous
  tests, using test problems with known solutions,
  an important process for building confidence in a
  numerical model.

14
Major Applications

• Wind-driven circulation over an idealized
  coastal Canyon (Sheng, Greatbatch, Wright
  Dietrich, 1998).

• Process studies of the Gaspe Current (Sheng
  2000).

• Seasonal circulation in the Northwest Atlantic
  Ocean (Sheng, Greatbatch Wright, submitted in
  Oct., 2000).

• Internal tide generation over topography (Lu,
  Wright Brickman, 2001 (in press)).

• Tidal Circulation and mixing in the Gulf of St.
  Lawrence (Lu, Thompson Wright, 2001 (in press)).

• Circulation in the Eastern Canadian shelf seas
  (Sheng, Thompson, Dowd Petrie, revised in 2000).

15
Governing Equations
16
Diagnostic vs Prognostic Models
Diagnostic Model Calculates ocean currents
from specified temperature and salinity fields.

• Relatively easy and straightforward to run.
• Robust in multi-year simulations.

• Wrong model for studying the interaction of
  temperature/salinity fields with the flow field.
• Wrong model for studying winter convective
  mixing.

Prognostic Model Calculates ocean currents,
together with temperature and salinity fields.

• Capable of simulating baroclinic instability.
• Capable of estimating winter convective mixing.

• Sensitive to subgrid-scale mixing
  parameterizations
• Deteriorating model skill in longer simulations.

17
The Semi-Prognostic Method
Sheng, Greatbatch, and Wright recently proposed a
semi-prognostic method to improve the utility of
the ocean model. The main idea is to replace the
hydrostatic equation
by
(1)
with temperature and salinity equations unchanged.
The Semi-Prognostic method is
Better than the robust diagnostic approach
proposed by Sarmiento and Bryan 1982, since the
new method does not constrain the temperature and
salinity equations. Different from the
assimilative approach examined by Woodgate and
Killworth 1997, since the new method does not
add any relaxation terms directly in the momentum
equations.
18
Application of the Semi-Prognostic Method to the
Northwest Atlantic Ocean
19
Model Parameters

• Model resolution 1/3 degree by 1/3 degree in
  horizontal and 31 unevenly spaced z-levels in
  vertical.

• Fourth-order accurate numerical scheme for
  space differencing and Thuburns flux limiter
  scheme for T/S advection terms.

• Smagorinsky 1963 scheme for the horizontal
  eddy viscosity co-efficient.

• Csanady 1982 scheme for vertical mixing
  co-efficients in the surface Ekman layer.

• Ri-dependent scheme Large et al., 1994 for
  vertical mixing co-efficients below the Ekman
  layer.

• Static instability was removed by an
  instantaneous convective adjustment scheme.

• Quadratic bottom stress with CD set to 0.0015.

20
Climatology
o
o

• (1/6) X (1/6) monthly mean temperature and
  salinity climatologies constructed recently by
  Geshelin, Sheng, and Greatbatch (1999) for the
  Northwest Atlantic.

o
o

• (1/2) X (1/2) monthly mean COADS wind
  stress (COADS Comprehensive Ocean-Atmosphere
  Data Set).

o
o

• (1/2) X (1/2) monthly mean COADS net heat
  flux.

o
o

• 1 X 1 annual mean transport
  streamfunction diagnosed by Greatbatch, Fanning,
  Goulding, and Levitus (1991).

21
Sea Surface and Lateral Boundary Conditions

• CANDIE was driven by monthly COADS wind and
  annual depth-mean boundary flow calculated by
  Greatbatch et al. (1991).

• Sea surface salinity was restored to the monthly
  mean climatology generated by Geshelin et al.
  (1991).

• The net heat flux through the Sea surface
  is approximated by

(2)
Where and
are respectively monthly mean COADS net
heat flux and Geshelin et al.s SST.

• Sommerfeld radiation conditions were applied to
  T, S and normal velocity at open boundaries

22
Model Results

• Semi-Prognostic model results
• Sea surface temperature and currents.
• Subsurface temperature and currents.
• Comparison with SST images.
• Comparison with drift velocity data.

• Passive tracer experiments
• Major pathways of tracers released in the
  Labrador Sea.
• Vertical fluxes of tracers in the Labrador Sea.

23
Sea Surface Temperature and Circulation Predicted
by the Semi-Prognostic Model
24
Sea Surface Temperature and Circulation Predicted
by the Semi-Prognostic Model
February
May
August
November
25
Sub -Surface Temperature and Circulation
Predicted by the Semi-Prognostic Model
February
May
August
November
26
Time-Depth Distribution of Temperature
27
Predicted
Observed
Sea Surface Fields in February
Sea Surface Fields in May
28
Predicted
Observed
Sea Surface Fields in August
Sea Surface Fields in November
29
Comparison of Predicted and Observed Sub -Surface
Circulation
Predicted
Observed
Blue arrows are speeds of less than 5 cm/s. Red
arrows are speeds of greater than 5 cm/s.
30
Major Pathways of Passive Tracers

• Major pathways of passive tracers released in
  the Labrador Sea over
• (a) Upper water columns of the slope-water
  region (Tracer2)
• (b) Upper water columns of the deep-water
  region (Tracer 4).

• Comparison of results produced by
• (a) Semi-Prognostic Model
• (b) Pure-Diagnostic Model
• (c) Pure-Prognostic Model

• Vertical fluxes of tracers in the Labrador Sea

31
Predicted by the Semi-Prognostic Model Tracer 2
(a) Concentration in the Upper-Ocean (0-383m)
(b) Concentration in the Lower-Ocean (383-5000m)
32
Comparison of Model Results Tracer 2
(a) Semi-Prognostic
(b) Pure-Diagnostic
(c ) Pure-Prognostic
33
Comparison of Circulation at 61m Produced by
Semi- and Pure-Prognostic Models
34
Comparison of Model Results Tracer 4
(a) Semi-Prognostic
(b) Pure-Diagnostic
(c ) Pure-Prognostic
35
Winter Convective Mixing Tracer 4
36
Animations of Semi-Prognostic Results
Movie 1 Near-surface temperature and currents.
Movie 2 Normalized concentration of Tracer 2.
Movie 3 Concentration of Tracer 2 at 967 m.
37
Summary and Conclusion

• We developed a semi-prognostic method to improve
  the utility of the primitive equation ocean model.

• The newly-developed semi-prognostic method
  adjusts the flow field towards climatology, while
  allowing the temperature and salinity fields to
  evolve freely with the flow.

• Multi-year model results using the
  semi-prognostic method show a significant
  improvement over those produced by
  pure-diagnostic and pure-prognostic models.

• The results show here represent our early
  progress of developing the next generation of
  DALCOAST.

38
Thank you