Coordinated Ocean-ice Reference Experiments phase II (CORE-II)

Transcription

1 Coordinated Ocean-ice Reference Experiments phase II (CORE-II) Gokhan Danabasoglu, Stephen G. Yeager, William G. Large National Center for Atmospheric Research, Boulder, CO, USA Stephen M. Griffies NOAA Geophysical Fluid Dynamic Laboratory, Princeton, NJ, USA Helge Drange University of Bergen, Bergen, Norway Anna Pirani International CLIVAR Project Office, UK

2 CORE-II An experimental protocol for ocean ice coupled simulations forced with inter-annually varying atmospheric data sets for the period (Large and Yeager 2009). This effort is coordinated by the CLIVAR Working Group on Ocean Model Development (WGOMD) new name Ocean Model Development Panel (OMDP). These hindcast simulations provide a framework for evaluation, understanding, and improvement of ocean models, investigation of mechanisms for seasonal, inter-annual, and decadal variability, evaluation of robustness of mechanisms across models, complementing data assimilation in bridging observations and modeling and in providing ocean initial conditions for climate (decadal) prediction simulations.

3 CORE-II PROTOCOL The models are integrated for a minimum of 300 years, corresponding to 5 cycles of the 60-year forcing period. After an assessment of degree of equilibrium achieved, the solutions from the last cycle are analyzed. Participants are free in their choices of ocean parameterizations, their parameter values, surface freshwater / salt flux treatments, and sea-ice models. The CORE datasets are periodically updated (currently through 2009) and collaboratively supported by NCAR and GFDL. They can be accessed via - WGOMD CORE web pages

4 Participating groups (20+ models): - Australia: CSIRO (ACCESS) - France: CERFACS, CNRM - Germany: AWI, IfM-GEOMAR (KIEL) - Italy: CMCC, ICTP - Japan: MRI (free, DA) - Norway: U. Bergen - Russia: RAS (INMOM) - UK: NOCS - USA: FSU (2), GFDL-GOLD, GFDL-MOM (2), MIT, NASA GISS (2), NCAR Level, isopycnal, hybrid, mass, and sigma coordinates; unstructured finite element ocean model; mostly nominal o horizontal resolutions

5 Manuscripts published, submitted, or in preparation (CORE-II Special Issue of Ocean Modelling) North Atlantic and Atlantic meridional overturning circulation (AMOC) Part I: Mean states (Danabasoglu & Yeager), PUBLISHED Part II: Variability (Danabasoglu & Yeager), Global and regional sea level (Griffies & Yin), PUBLISHED Southern Ocean water masses, ventilation, and sea-ice (Downes & Farneti), Antarctic Circumpolar Current and Southern Ocean overturning circulation (Farneti & Downes), Arctic Ocean and sea-ice (Gerdes, Wang, & Drange), South Atlantic simulations (Farneti, Deshayes, & Treguier), Ocean circulation in temperature and salinity space (Nurser & Zika), Indian Ocean (Ravichandran, Rahaman, Harrison, Swathi, & Griffies), Pacific Ocean circulation and its variability (Tseng).

6 AMOC in depth space ( ) Danabasoglu et al. 204, Ocean Modelling, 73, 76-07, 0.06/j.ocemod

7 AMOC at 26.5 o N ( ) Danabasoglu et al. 204, Ocean Modelling, 73, 76-07, 0.06/ j.ocemod

8 AMOC standard deviation ( ) FSU2 Danabasoglu et al. 204b (in preparation)

9 Time series of thermosteric sea level Time series are relative to 993 Vertical line represents an estimate of the spread in the available observational estimates Griffies et al. 204, Ocean Modelling, 78, -89, 0.06/j.ocemod

10 Mixed layer depth: March-mean for (based on 0.03 kg m -3 density change from surface) debm: de Boyer Montegut et al. (2004) MMM: Mul?- model mean Downes et al. 204 (submitted)

11 Mixed layer depth: September-mean for (based on 0.03 kg m -3 density change from surface) debm: de Boyer Montegut et al. (2004) MMM: Mul?- model mean Downes et al. 204 (submitted)

12 Zonal-Mean Temperature: mean for o C contour is shown Downes et al. 204 (submitted)

13 Zonal-Mean Salinity: mean for contour is shown Downes et al. 204 (submitted)

14

15 THE EVOLUTION OF SOUTHERN OCEAN ACC & MOC IN CORE- II Farne7 et al. (204), An assessment of Antarc-c Circumpolar Current and Southern Ocean Meridional Overturning Circula-on sensi-vity in a suite of interannual CORE- II simula-ons. To be submiqed to Ocean Modelling (CORE Special Issue) CORE- II Southern Ocean rfarne?@ictp.it

16 Satura?on vs. Compensa?on EDDY SATURATION: addi?onal energy imparted from the winds is cascaded to the oceanic mesoscale instead of inducing prolonged accelera?ons in the horizontal mean flow (i.e. the ACC) EDDY COMPENSATION: at the same?me, the eddy- induced meridional circula?on is enhanced and opposes the Ekman driven component (i.e. the MOC) Eddy satura?on can be total whereas Eddy compensa?on can only be par?al. (many recent studies support this idea) CORE- II Southern Ocean

17 Forcing changes during [N/m 2 ] [N/m 2 ] ACCESS AWI BERGEN CMCC GFDL GOLD GFDL MOM GISS ICTP INMOM KIEL MRI NCAR NOCS [ ] [ ] Anomalies N/m Time [years] % (a) Magnitude of peak zonal wind stress Time [years] ( o ) (b) Change in percentage of peak zonal wind stress (c) Change in latitude of peak zonal wind stress Time [years] Time [years] [N/m 2 ] Wind stress increases up to 70% and shi^s polewards by 5-6 degrees INMOM shows significant differences due to different winds CORE- II Southern Ocean rfarne?@ictp.it

18 The ACC over the 5 cycles Sv (0 9 kg/s) Sv (0 9 kg/s) INMOM ACCESS AWI BERGEN CMCC FSU GFDL GOLD GFDL MOM GFDL MOM Time (Years) GISS ICTP INMOM KIEL KIEL025 MRI NCAR NOCS Time (Years) Black dot: Cunningham et al. (2003); White dot: Chidichimo et al. (204); Black square: CMIP5; White square: CMIP3 CORE- II Southern Ocean rfarne?@ictp.it

19 The ACC and the density field ACC (Sv) ACCESS AWI BERGEN CMCC FSU GFDL GOLD GFDL MOM GFDL MOM025 GISS ICTP INMOM KIEL KIEL025 MRI NCAR NOCS MMM (kg/m 3 ) ACC trend (Sv/decade) ACCESS AWI BERGEN 2.75 CMCC FSU 2.5 GFDL GOLD GFDL MOM GFDL MOM025 GISS 2.25 ICTP INMOM KIEL 2 KIEL025 MRI.75 NCAR NOCS MMM trend (kg/m 3 /decade) Changes in wind stress and/or buoyancy forcing lead to changes in the stra?fica?on The ACC transport (variability) is in thermal balance with the density field. CORE- II Southern Ocean rfarne?@ictp.it

20 The ACC: 5 th cycle (w.r.t ) eddy satura?on Sv (0 9 kg/s) ACCESS AWI BERGEN CMCC FSU GFDL GOLD GFDL MOM GFDL MOM025 GISS Sv (0 9 kg/s) ICTP INMOM KIEL KIEL Time (Years) Time (Years) Most models show varia?ons <0% (eddy satura?on) MRI NCAR NOCS CORE- II Southern Ocean rfarne?@ictp.it

21 The MOC: par?al compensa?on ψ * or κ (relative) /2 power /2 power full comp. No comp. ACCESS AWI BERGEN CMCC ψ * or κ (relative) /2 power /2 power full comp. No comp. GFDLMOM GFDLMOM025 GISS ICTP ψ * or κ (relative) /2 power /2 power full comp. No comp. KIEL KIEL025 MRI NCAR NOCS Wind stress (relative) Wind stress (relative) Wind stress (relative) See Meredith et al. (203) - Good degree of compensa7on: BERGEN, GFDL- MOM025, KIEL025, NCAR, (AWI, GISS) - Not so good: ACCESS, CMCC, GFDL- MOM, ICTP, KIEL, MRI, NOCS CORE- II Southern Ocean rfarne?@ictp.it

22 SUMMARY Entrained participation of many (ocean) modeling groups. CORE-II framework is being used for: o Evaluation, understanding, and improvement of ocean components of coupled climate models, o Investigation of variability mechanisms, o Evaluation of robustness of mechanisms across models, o Initialization of decadal prediction simulations. CORE-II simulations show differences in solutions, both in the mean and variability. No grouping of model solutions based on model family or vertical coordinate representation is obvious. Solution differences are primarily due to differences in ocean model parameterizations and their parameter choices. Use of a wide variety of sea-ice models also contributes to the solution differences. Initiated preliminary inter-comparison projects (e.g., related to AMOC and MLD) with the data assimilation community.

23 The residual MOC (?me- mean) 2 (kg/m 3 2 (kg/m 3 2 (kg/m 3 2 (kg/m 3 ) ) ) ) ACCESS FSU GISS 9 2 KIEL AWI GFDL GOLD ICTP 2 2 MRI 5 24 BERGEN GFDL MOM NCAR CMCC 5 2 GFDL MOM025 5 Sv KIEL NOCS 5 2 CORE- II Southern Ocean rfarne?@ictp.it

24 The eddy- induced MOC (?me- mean).5 ACCESS AWI BERGEN CMCC 2 (kg/m 3 ) GFDL MOM GFDL MOM Sv (kg/m 3 ) GISS 8 ICTP 0 KIEL (kg/m 3 ) KIEL MRI 3 NCAR 4 8 NOCS 6 6 CORE- II Southern Ocean rfarne?@ictp.it

25 The residual MOC (decadal trend).5 ACCESS AWI BERGEN CMCC 2 (kg/m 3 ) FSU 2 2 GFDL GOLD GFDL MOM GFDL MOM (kg/m 3 ) GISS ICTP KIEL Sv/decade 2 (kg/m 3 ) KIEL025 MRI NCAR NOCS 2 (kg/m 3 ) CORE- II Southern Ocean rfarne?@ictp.it

26 The eddy- induced MOC (decadal trend).5 ACCESS AWI BERGEN CMCC 2 (kg/m 3 ) GFDL MOM GFDL MOM GISS ICTP 0.2 KIEL Sv/decade 2 (kg/m 3 ) KIEL025 MRI NCAR NOCS 2 (kg/m 3 ) CORE- II Southern Ocean rfarne?@ictp.it

27 Conclusions Eddy satura7on: most models show a good level of satura?on (given the wind- stress changes and concomitant buoyancy changes) Eddy compensa7on: compensa?on is much weaker and achieved only by a few models GM- like coefficient: reinforced the idea of a 3D?me- varying Kappa. CORE- II Southern Ocean rfarne?@ictp.it