CCSM Simulations w/CORE Forcing Some preliminary results and a discussion of dataset issues

1. CCSM Simulations w/CORE ForcingSome preliminary results and a discussion of dataset issues Marika Holland With much input from Bill Large Steve Yeager

2. Experiments Using CORE Forcing

3. CORE I Results - Ocean-Ice Coupled Runs MOC Sea Ice equilibrates rapidly

4. CORE-II Ocn-Only Experiments • Purpose of experiments • Attribution of upper ocean biases in CCSM3 (Large and Danabasoglu, ‘05) • Ocean variability and process studies (Yeager and Large, 2004; Capotondi et al., 2005) • Protocol • Initialized with Levitus/PHC, no motion • 4x43 year cycles • Salinity Forcing • Precipitation factor used • Weak restoring with piston velocity of 50m/4yrs globally • Specified ice-ocean flux from coupled run • Frazil ice formation salinity flux

5. Attribution of Upper Ocean Biases in CCSM3 OBS • Equatorial Pacific Zonal Velocity • Coupled run biases • upper ocean shear in west • westward surface flow at 140E • Absent in Ocean-only runs • Associated with the lack of westerly wind bursts that are present in the observations Ocean Only Coupled Large and Danabasoglu, 2005

6. Attribution of Upper Ocean Biases in CCSM3 OBS • Pacific Zonal Velocity at 140W • Coupled and Ocean-Only biases • Westward SEC too weak • Eastward NECC too weak • Coupled biases • Too symmetric about equator • Presence of a SECC • Associated with symmetric atmospheric forcing Ocean Only Coupled Large and Danabasoglu, 2005

7. Attribution of Upper Ocean Biases in CCSM3 OBS • Equatorial Pacific (2S-2N mean) Potential Temperature • Coupled and Ocean-only biases • Warm bias indicative of a reduced temperature gradient • Coupled bias • Cold bias at 100-200m depth, 180E • Boundary layer not penetrating deep enough - related to sfc buoyancy forcing Ocean Only Minus OBS Coupled-OBS Large and Danabasoglu, 2005

8. Ice-Only ExperimentsCORE-II Forcing • Purpose of Experiments • Examine issues/difficulties in validating sea ice models • Perform simulations with different forcing datasets, • Perform simulations with variations in parameter values • quantify uncertainty due to model forcing vs model physics • Currently a single cycle of forcing performed with NCEP and with Large-Yeager (43 years) • Protocol • Initial ice conditions from a previous ice-only run • Ocean heat flux convergence specified (held fixed) from a CCSM3 coupled integration

9. CORE-II Forced Ice Only Experiments Average Sept Arctic Ice Concentration

10. Arctic Ice Variability Summer Variability Quite Different. Associated with Mean Differences. Winter Variability Very Similar

11. NCEP Large-Yeager Antarctic Sept Averaged Sea Ice • NCEP forcing results in thicker ice cover • SH variability very similar between runs

12. Discussion of Dataset IssuesCompiled by Bill Large with input from dataset users

13. New/Extended Data Available • IVF data through 2004 available at NCAR. • Should be checked out early in the new year. • Future CCSM IVF runs will run through 2004. • No intention of recomputing NYF (with data to 2004) • Should the GFDL IVF data base be extended?

14. New/Extended Radiation Data • Radiation data has just arrived • Contains "replacements" for the previous data for January ‘97 - June ‘01. • "replacements may have some minor effects in terms of global means (up to a few tenths W/m2) but there are some large flux values changes (> 100 W/m2) for a few grid cells (primarily land areas) for a few flux components.” • Should we use the "replacement" data?

15. Dataset Issues - Tropical Humidity • Comparison of TOA and NCEP humidity (Jiang et al., 2005 ) fundamentally different than earlier comparison (Wang and McPhaden, 2000) on which tropical corrections were based. • In tropical E Pac, new study consistent with SOC (NOT in west). • New comparison and SOC suggest humidity corrections should NOT be zonally uniform (as currently done). • Considering exploring an objective alternative SOC-based correction that depends on both lat and lon. (ERA-40 consistency) • What should be done for CORE? Should we work with relative or specific humidity ?

16. Dataset Issues - Arctic Humidity • Elizabeth Hunke (LANL) believes that the corrected humidity is still too high over Arctic Sea ice. • She is trying to gather some data sets to quantify the possible problem, so that the correction could be improved. • Are there Arctic humidity data available?

17. Dataset Issues - Normal Year Winds • Mean wind stress in NYF based on 1958-2000 • Southern ocean has trend in zonal wind stress to increasing westerlies. • If ACC transport tuned to winds from later years, a weaker ACC will result when forced with NYF. • Should mean NYF be based on 1958-2000, 1983-2000, 1958-2004, 1983-2004 ?

18. Dataset Issues - Wind Direction • NCEP wind direction has been compared to QSCAT. • There are 2 small regions of systematic bias; in the ITCZ regions of central N Pacific, along the Pacific coast of S America. • Possible correction being considered to adjust the mean and standard deviation of the wind direction • N’=Q + (sQ/sN) (N-N) • So N’ (corrected NCEP direction) is a function of the mean QSCAT direction and the ratio of the standard deviations • Should corrections be applied? How? Globally?

19. Dataset Issues - Solar Radiation Corrections • The 5% reduction made to the solar radiation is supported by measurements from TAO and PIRATA bouys in the tropical Pacific and Atlantic. • Has anyone else found a similar or different result?

20. Dataset Issues - Heat Imbalance +10 W/m2 • Normal year forcing, when used with obs SST has a global ocean heat flux imbalance of -5 W/m2, compared to -1 W/m2 for the IVF (1958-2000). • Is this a problem? • Will be looking into in any case. +5 W/m2 NYF

21. Dataset Issues - Others?

22. Antarctic Winter Ice Variability Nearly identical for the different forcing runs