Canadian Report GODAE Ocean View Science Team June 2009

CMOS, Kelowna BC, May 2008 C-NOOFS status

Canadian Report GODAE Ocean View Science

Team June 2009Fraser Davidson (1), Hal Ritchie (4,6)Greg Smith(1,4), Andry Ratsimandresy(1), Debbie Anne Power (1) Adam Lundrigan (1), Charles Hannah (2), Frederic Dupont (2), Dan Wright (2), Maud Guaracino (2), Denis Lefaivre (3), Pierre Pellerin(4), Mark Buehner (4),Lt. Darryl Williams (5) Keith Thompson (6)

(1) DFO-Northwest Atlantic Fisheries Centre, St. John's, NL, Canada (2) DFO-Bedford Institute of Oceanography, Dartmouth, NS, Canada(3) DFO-Institute Maurice Lamontagne, Mont-Joli, QU, Canada(4) EC-Meteorological Research Division, Dorval, QU, Canada(5) National Defence, METOC, Halifax, NS, Canada(6) Dalhousie University, Halifax, NS, Canada


Canadian Operational Network of Coupled Environmental PredicTion Systems

CONCEPTS•MOU among 3 Government Departments•No $ •Provides ability to obtain funding

•Short term (projects)•Long term (budgeted development)

Secretariat provided by Marty Taillefer DFO and Pierre Pellerin


CONCEPTS: CORE PROJECTSProject 1: Core CMC Systems InstallationInstall common core ocean model configurations at Canadian Meteorological Centre (CMC) for collaborative projects, evaluations and coupling with GEM atmospheric model, coupled data assimilation, examine potential implementation for operational use.

Project 2: Basin-to-Global Ocean ReanalysesCanada-Mercator Contribution to CONCEPTS Validate and improve the ocean component of the basin-global modelling and assimilation system being developed for use by CONCEPTS and Mercator.

Project 3: Regional Ocean Prediction: C-NOOFS Canada-Newfoundland Operational Ocean Forecasting System

Project 4: Sea Ice Modelling and Data Assimilation


CONCEPTS: PROJECTS OF INTERESTS

1. Ocean Data assimilation, GOAPP (Keith Thompson)

2. Spaceborne Ocean Intelligence Network (Darryl Williams)

3. DRDC (A) Requirements and Initiatives (John Osler)

4. Coupled models for the Gulf of St. Lawrence (Lefaivre & Pellerin)


NEMO Applications in

Canada• Global -- BIO, RPNE, Dalhousie, U Quebec Montreal• North Atlantic -- BIO, Dalhousie • North Pacific -- IOS, Royal Military College• Arctic and CAA -- BIO-CIS-Mercator, U Alberta• North-western Atlantic -- CNOOFS, Dalhousie, Memorial • Gulf of St. Lawrence-Scotian Shelf-Gulf of Maine -- BIO,

Dalhousie• Gulf of St. Lawrence -- BIO, U Quebec Rimouski • Great Lakes -- NWRI/RPNE, BIO


CONCEPTS Project 1 & 2Validation and analysis of the ¼-deg global

NEMO-CONCEPTS* ocean model

Francois Roy(1), Youyu Lu(2), Jean-Marc Belanger(3), Hal Ritchie(3), Greg Smith(4)

1- Canadian Meteorological Centre, Environment Canada 2- Ocean Sciences Division, Bedford Institute of Oceanography, Fisheries and Oceans Canada

3- Meteorological Research Division, Environment Canada4- Biological and Physical Oceanography Section, Northwest Atlantic Fisheries Centre, Fisheries and Oceans Canada

A partnership of EC, DFO, DND, GOAPP and MERCATOR-OCEAN


Two global ocean configurations for core project 1: ¼-deg; 1-deg

Status: ¼-deg tested for 6-yr spin-up; 10-day forecast; 1-yr simulation

1-deg model tested for multi-decade simulations;

1-deg model being coupled to 100km GEM;

Validation and process studies


CMC global ¼ deg 1 year model run NEMO3-LIM2 - ¼ deg ORCA025 - 50 vertical levels (1 m surface layer)

Atmospheric forcing from global GEM at 35 km resolution (Ua,Va,Ta,Ha,SW,LW) CORE bulk formulas from Large and Yeager (2004)

Initial conditions: MERCATOR-OCEAN PSY 3V2 (T,S,U,V, April 18, 2007) SSS restoring to yearly climatology + No SST restoring

• 1 Year simulation 1 (S3): 3-hourly forcing

• 1 Year simulation 2 (S24): 24h-av. forcing


S3

S3 – S24

1 Year simulation: sensitivity to atmospheric forcingSSH yearly standard deviation (using 20 outputs per day, close to hourly)

3 hour wind forcing- 24 hour wind forcing


Yearly simulations: sensitivity to atmospheric forcingYearly averaged surface circulation in ms-1 (first 5 m)

S3

S3 – S24


Wind energy input to oceanic near inertial motions: S3Initial estimates based for April 18-30, 2007High energy flux associated with synoptic storms

•Further analyses to reveal seasonal cycle•Comparison to be made with semi-analytic solution


Theme 1 and 2: Conclusion

NEMO3-LIM2 with GEM forcing produces a realistic seasonal cycle without major SST drift

A non-persistent warm bias occurs during the summers of the north and south hemispheres (slightly reduced with 3-hourly forcing)

Significant variability in SST is added with 3-hourly forcing (diurnal cycle)

SSH variability is increased in coastal areas with 3-hourly forcing and the distribution of meso-scale eddies is changed


Future work

Verification with in situ and satellite derived data

Work to explain the summer warm bias

More global diagnostics

Sensitivity tests to flux parameterization (CORE, GEM physics, …) and TKE parameters

Adding MERCATOR data assimilation system

Pseudo-operational 10 day forecasts using GEM forcing

Two-way coupling of NEMO with GEM referenced to PSY3V2


1º Global Ocean Model & Decadal SimulationModel:

Horizontal: Global tri-polar grids; Nominal resolution 1° in lat/long; Meridional refinement in tropics;

Vertical: 46 vertical NEMO 2.3 GOAPP

Simulations:10-yr spinup with CORE Normal Year forcing

CONTROL: CORE forcing 1958-2004

HEAT: wind stress set to Normal YearWIND: buoyancy forcing set to Normal Year

MJO: wind stress = Normal Year + MJO

Sea-Level Trend 1993-2004 (m/yr)

Model¼ o

Obs

1/4º model

Forcing sensitivity

Wind

Heat

1/4º model

Inter-Annual Sea-Level RMS 1993-2004 (m)

Model

Obs

1/4º model

Forcing sensitivity

Wind

Heat

Summary

• Global 1º model able to reproduce large-scale SSH changes during altimeter era

• Wind stress changes are primary cause of SSH trend and RMS at low and mid latitudes;

• impacts of buoyancy forcing mainly show at high latitudes

• 1º NEMO is being coupled to 100 km GEM; expected to be a good tool for study/prediction of intra-seasonal/seasonal variations


C-NOOFS

C-NOOFS system V0C-NOOFS system V0MERCATOR

Global Output

Initial StateC-NOOFS

Fixed BC’s

Weekly

www.c-noofs.gc.ca

NEMO v2.3

T,S,U,V,W

Data products / plotting

Eve

ry 2

4 h

r

WebpageWebpage

CMC WIND

C-NOOFS system V1C-NOOFS system V1MERCATOR

Global Output

RESTARTFILE

Time varying BC’s

Weekly

www.c-noofs.gc.ca

NEMO LIM2* BULK

T,S,U,V,W, ice, SSH

Data products / plotting

OceanOceanViewViewOf Of

The dayThe day

CMC WINDCMC T,Q,

QSW,QLW,precip

Validation Monitoring

NEMO LIM2* BULK

1/12 th 1/4

Observations

ValidationValidationM vs MM vs MM vs OM vs O

MonitoringMonitoringDiagnosticsDiagnostics

T,S,U,V,W, ice, SSH


Effects of resolution on Model and ObservationC

TD

Sea Island lineSea Island line2008 12 19

1/4 1/12

ADCP (hr)

Observations Observations Bonavista Bay line Bonavista Bay line 2004 07

CTDCTD stations~20 km

CMOS, Kelowna BC, May 2008 C-NOOFS statusC-NOOFS Overview

Configuration name

Operational since

Resolution

Initial Condition

Boundary Condition

Wind Forcing

Other forcing fields

Coupled ice model

NWA025-v0 Sept. 2008 ¼° Daily mean MERCATOR-

PSY3V2 forecast

Fixed to I.C. GEM Climatology None

NWA12-v0 Dec. 2008 – Jan. 2009

1/12° Daily mean MERCATOR-

PSY3V2 forecast

Fixed to I.C. GEM Climatology None

NWA025-v1 May 2009 ¼° MERCATOR-PSY3V2 analysis

Daily mean MERCATOR-

PSY3V2 forecasts

GEM GEM LIM2

NWA12-v1 July. 2009 1/12° MERCATOR-PSY2V1 analysis

Daily mean MERCATOR-

PSY3V2 forecasts

GEM GEM LIM2

NWA025-v2 Apr. 2010 ¼° CNOOFS assimilation

system (based on SAM2)

CMC global ocean forecasts

using SAM2

GEM GEM Not yet decided (LIM2, LIM3,

CICE)


Observations to be assimilatedV2• Sea level anomalies :

– AVISO SSALTO/DUACS• Jason (2cm), ENVISAT and GFO (3.5cm)• Near-coast representivity error

• Mean dynamic topography :– Rio et al., 2005 (~5cm rms error)

• Sea surface temperature :– NCEP SST RTG (~0.7C)

• In situ profile T and S data :– E.g. Argo, XBT, TAO, CTD, …– CORIOLIS (Brest) – Quality Controlled by CLS (Toulouse)


Forecast Viewer• Open access to

archive of static images of SSH and 3D T,S and currents

http://www.c-noofs.gc.ca/viewer/


CNOOFS Dynamic Quick-view Webpagewww.c-noofs.gc.ca:8080/ncWMS/godiva2.html

• Uses Godiva2 viewer (from RESC)

• Uses OpenLayers to quickly and easily visualize data

• User-defined options for colour scale and animations

• Capability to create overlays in kmz format for use in geobrowsers (e.g. GoogleEarth)


C-NOOFS Monitoring/Validation System

• Evaluation against:– AVISO satellite altimetry (SSH)– CMC Sea surface temperature analyses– CMC sea ice concentration analyses– Argo and other in situ data (coming soon…)

• For each data type:– maps of differences– RMS error versus forecast lead time– Mean error versus forecast lead time


Evaluation NWA 025 V1 against AVISO SSH for May 13, 20091-day lead time 3-day lead time

10-day lead time6-day lead time

RMS difference

Mean difference


1-day lead time 3-day lead time

10-day lead time6-day lead time

RMS difference

Mean difference

Evaluation NWA 025 V1 against CMC SST for May 13, 2009

PSY3V1 V0 PSY3


Future developments and plansC-NOOFS 1-2

• Implementation of NWA12-v1 to be done this summer

• Optimization of SAM2v1 assimilation system for Northwest Atlantic:– Ability to deal with tides– High-resolution SST track data– Improvement of error modes for Labrador Sea– Assimilation of seal and other additional in situ data

sources – Sea ice assimilation– Produce analyses daily


Future developments and plansC-NOOFS 2-2

• Model improvements:– Update to NEMOv3.2– Add tides (variable volume)– Upgrade ice model (LIM2-EVP, LIM3 or CICE)– Optimization of ocean and ice physics

• Detailed validation studies, e.g.:– Iceberg tracking– Drifters– Argo, seals, …

• All in situ data we can put our hands on!

• Observation quality control


Sea-Ice Data Assimilation Project• Project Goal: to develop an automated ice analysis system for:

• Canadian Ice Service (CIS): ice concentration, thickness/type distribution, pressure, strength and edge, deformed ice at ~1-2 km resolution

• NWP: ice concentration, thickness, albedo, surface emissivity at ~5 km resolution

• Benefit from experience with variational and ensemble-based assimilation for NWP:

• use variational approach: incremental 3D-FGAT (first guess at appropriate time)

• Developed initial prototype analysis system using CIS ice-ocean model for Canadian east-coast plan to port system to other models/regions:

• Canadian Arctic archipelago region (IPY project, Polar-GEM)• Gulf of St. Lawrence (coupled ice-ocean-atmosphere model developed at

RPN/IML)

Experiments with east coast coupled ice-Experiments with east coast coupled ice-ocean modelocean model

• Goal: evaluate 3D-Var approach relative to nudging approach as currently used at CIS

• Configuration:

– Assimilate partial concentration of 24 ice thickness categories from CIS daily ice charts

– Tests with additionally assimilating CIS RadarSAT image analyses

– Experimental period: 5 Dec 2006 - 30 Jun 2007

• Forecast verification in terms of total concentration and effective ice thickness (forecast minus observation)

– Only where the 1970-2000 weekly sea ice climatology indicates a non-zero probability of ice

Canadian east coast model (CIOM)Canadian east coast model (CIOM)

• Multicategory sea-ice model coupled to the Princeton ocean model (Yao et al., JGR 2000).

• Viscous plastic sea-ice rheology (Hibler, JGR 1979).

• GEM model atmospheric forcing every 3 hours.

• Grid resolution: 1/5º longitude x 1/6º latitude, 16 sigma levels.

• Originially developed by Charles Tang at BIO

48-h forecast

East coast 3D-Var experiments: ExampleEast coast 3D-Var experiments: ExampleBackground Estimate

(24h forecast)Observations

(CIS Daily Ice Chart)Analysis

(initialize next forecast)

Total ice concentration error StdDevTotal ice concentration error StdDev

0.00

0.05

0.10

0.15

0.20

0.25

15-h forecast 39-h forecast 0-h 24-h forecast 48-h forecast

Image analysis Daily charts

Persistence

Direct insertion

nudging

3D-Var 1

3D-Var 2

3D-Var 3

3D-Var 4

}large improvement vs. persistence

}small

differences among data assimilation experiments

ARTIC MODELMarch Sea-ice velocity for the last year of a 10 year run

1º model with OMIP forcing

The model consistently overestimates ice velocities which influences the sea ice

distribution. The reasons for the overestimations are not yet clear.

Future developments: Usage of “nesting” approach

NEMO’s AGRIF allows 2-way nesting

Potentially useful for CAA, e.g., 1º global + 1/4º CAA 1/2º Arctic + 1/8º CAA (2-3 km)

BIO-RPN-CIS collaboration in planning

Should be coordinated with CICE & I-DA developments


So Where Are We Going?

GEM / NEMOGEM / NEMO

• Global Coupled Ocean Atmosphere Ice Forecasting System

• Regional Eastern Canada Coupled Ocean Atmosphere Ice Forecasting System – 1 way nest with

MERCATOR– 10 day forecasting

• Ice • Service

– Agrif zoom

Research Direction

Downscaling

Ocean Only Vs Coupled

Ice Assimilation vs Altimetry Assimilation


GODAE Future

• International group helps focus national research & development

• Opportunity for synergies• Much needed collaboration

– Through Working Groups

Documents

Canadian Report GODAE Ocean View Science Team June 2009