Experiences and expectations of NEMO

1
Experiences and expectations of NEMO
Andrew C. Coward, Steve Alderson and Beverly de
Cuevas
2
(No Transcript)
3
Benefits of NEMO Curvilinear grids - tripolar
4
(No Transcript)
5

• NEMO Users meeting Thursday 22nd (afternoon)
  and Wednesday 23rd (morning) May 2007?
• Coupled models
• CICLE project at the CNRM - M. Lucas ?
• MERSEA project and ECMWF system 4 prototype - A.
  Troccoli ?
• First analyses of the new version of the
  sintex-F coupled model - C. De Boyer Montegut ?
• Progress on incorporating NEMO into the next Met
  Office climate model - C. Harris
• Grid computing with ARPEGE-NEMO coupled model -
  E. Maisonnave
• Operational Oceanography ?
• UK operational ocean forecasting with NEMO - A.
  Sellar ?
• 1/12? ocean simulations with Mercator
  configurations - Y. Drillet ?
• The MERSEA/Mercator Ocean global 1/4? analysis
  and forecasting system V2 first results - G.
  Garric
• Implementing NEMO in the new version of the
  Mediterranean Forecasting Sytem - S. dobricic for
  P. Oddo
• Biogeochemistry of the ocean
• Iron and C13 in NEMO-PISCES - A. Tagliabue ?
• Biowaves project first results from coupled
  physical/biogeochemical simulations in the N.
  Atlantic - G. Charria?
• Modeling the neodymium isotopic composition with
  the NEMO global ocean circulation model - T.
  Arsouze

6
OCEANS 2025 Themes and selected scientific
objectives
Theme 9 Next Generation Ocean Prediction Systems

• How sensitive are climate models to the manner
  in which sea ice is coupled?
• Can nested models be trusted to give accurate
  results?
• Can an ocean model be made energetically
  self-consistent?
• What is the most appropriate level of complexity
  of biogeochemical models in climate studies?
• Approaches and methodologies
• Develop NEMO as the core OGCM for use by the
  scientific community in the UK,
• at resolutions of 1, ¼and 1/12, and with
  nested grids (WP 9.10).
• Develop an ocean model testbed permitting
  objective intercomparison and validation of a
  range
• of ecosystem models, with a view to embedding the
  most promising in OGCMs (WP 9.11).

7
OCEANS 2025 Themes and selected scientific
objectives
Theme 2 Marine Biogeochemical Cycles

• To determine the sensitivity to future climate
  change of the mechanisms sustaining total
  nutrient
• supply to the photic zone over the three major
  biomes of the North Atlantic.
• Approaches and methodologies
• Quantify the magnitude and sensitivity of
  nutrient fluxes associated with winter
  overturning
• and Ekman pumping. For overturning, this will be
  achieved using time-series stations,
• Argo floats and mooring data together with
  previous studies and basin-scale simulations
• (NEMO both at ¼º and with a smaller scale nested
  component at 1/12º in the North Atlantic).

8
OCEANS 2025 Themes and selected scientific
objectives
Theme 1 Climate, Ocean Circulation, and Sea Level

• Model simulations of climate change in the ocean
• Identifying the causes of recent climate change
  in the ocean
• Physical-biogeochemical budgets and mixing in
  the Southern Ocean (DIMES)
• Research plan and deliverables
• 2008 Completed simulation of changes in the
  ocean over the period 1950 - 2006 obtained by
• running NEMO globally at 1/4 resolution (and
  with a nested 1/12 North Atlantic grid)
• using NCEP/NCAR (and possibly ECMWF) derived
  surface flux fields (WP 1.1b)

9

• The UK is moving towards a more coordinated
  approach based on the NEMO model
• This forms the core of OCEANS 2025 with plans
  for (amongst others)
• Global 1/4o models with biogeochemistry and
  1/12o nested regions
• Global 1/12o physics only model by 2010

10

• Theme 9 NOCS NEMO activities
• Produced global 1 degree NEMO configuration
  (ORCA1) shared with UKMO
• Tested ORCA1 with LOBSTER NPZD model, about to
  test new NOCS biogeochemistry model MEDUSA
• Developed interpolation on the fly option for
  handling surface forcing fields
• Constructed ORCA1 forcing datasets which are
  compatible with those used with the current
  DRAKKAR (global 1/4o) model
• Completed multi-decadal integrations of ORCA025
• Developed capability to run AGRIF nests in
  sea-ice regions

11
Performance
ORCA1 15 processors 3.59s per timestep
(2.75yrs/day) ORCA1 30 processors 1.97s per
timestep (5yrs/day)
ORCA1 32 processors with LOBSTER 3.57s per
timestep
ORCA025 56 processors Avg ts 14.2634 ( 9
entries ) Min 12.016 Max 28.412

(0.28yrs/day)
On HECToR ORCA025 221 processors Avg ts 2.44
( 14400 entries ) Min 2.13 Max 7.533

(1.6yrs/day)
16 processors 1.5s per timestep ( 7 years/day)
64 processors 4.5s per timestep ( 0.6 years/day)
512 processors 7.2s per timestep (0.12 years/day)
12

• Current status of Drakkar configurations
• at NOCS
• Steven Alderson, Andrew Coward, Beverly de
  Cuevas, Adrian New, George Nurser, Yevgeny
  Aksenov, Jeff Blundell, Jeremy Grist

13
N025_100 run (NOCS)

• Drakkar code with DFS3 forcing
• On-line interpolation of atmospheric fields
• cdmp 0.0333 (5 times weaker than G70)
• icedmp 25 (same as G70)
• 1958 - 2004 (currently1997)

14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
Drake Passage and Bering Strait transport
comparison
N025_100 1958 - 1975
ORCA025_G70 1958 - 2004
19
(No Transcript)
20
NOCS plans for 2008

• Complete and analyse N025_100
• Investigate CICE vs LIM3
• Extend tests with AGRIF in North-west Atlantic
• Run ORCA025 with biogeochemistry

21
Experiments with AGRIF at NOCS

• Steven Alderson

Oceans 2025
Work Package 2.5 NEMO will be run globally at
1/4 with biogeochemistry and with a higher
resolution 1/12 nested model .
22
AGRIF nesting
Bathymetry data
Nesting tools
T/S data
23
NEMO conversion to AGRIF
NEMO
Conversion routine
AGRIF/NEMO code
AGRIF/ NEMO
AGRIF interface routines
AGRIF library
24
Order-1 nest in 1 global NEMO
25
Order-1 nest in 1 model surface differences
(30d)
26
Schematic for calculation of fluxes at water
surface
Outer model
Nest (modified)
27
Order-1 nest in 1 model surface differences
(30d)
28
1 model vs order-3 nest in 1 model SST (4mnth)
latitude
longitude
longitude
29
No nest model vs order-3 nest in 1 model
30
Further work

• Continuity across restart files
• Higher resolution run (standard mpi_send)
• Blanking of areas (e.g. Pacific)
• AGRIF with biogeochemical models
• Southern Ocean region