June 18 Meeting

From Interactive System for Ice sheet Simulation
Jump to: navigation, search

Contents

Held in Breckenridge, Colorado as part of the Community Climate System Model (CCSM) meeting[1].

Webcast Instructions and Information: http://www.ccsm.ucar.edu/webcasts/

Summary of Talks

8:30 a.m. Welcome – Robert Bindschadler

  • Thanks for coming; lots to do during the workshop; blah, blah, blah; let’s get started!
  • The ppt SeaRISE summary Bob showed the day before at the Land Ice WG meeting is posted at the LANL Trac site.

8:40 a.m. Jesse Johnson – University of Montana

  • Model Spin-ups: Comparison and Synthesis (see pdf posted on LANL Trac site)
  • Many of the now-posted parameter fields were shown. The discussion quickly identified that there are a number of improved parameter fields available; consistency difficulties noted with some and need to get others; most sources willing to provide;
    • Bamber to supply improved Ant. and Grnld surface elevations;
    • Arthern or RACMO for better Ant. precip;
    • Joughin has a better blended Ant. surface velocity and is a better velocity “goal” than balance velocity;
    • Csatho may have a better Grnld geometry;
    • NSIDC has a grounding line (from MODIS imagery);
    • Geothermal heat flux is contentious, but potentially incorrect values must be weathered until better values are available; Bea will resolve this parameter with Jesse (Grnld only)
    • Bueler expressed a need for uplift data (Jesse will clarify this with Ed)
  • Spin-up climate fields from ice cores shown; spin ups of Maine/CISM/Penn State models shown individually and side-by-side; some differences (Maine is too thick), lots of similarities; no problems raised but a strong recommendation that spin-ups tune to fit velocity, rather than geometry for the sake of SeaRISE objectives.
  • Assimilation schemes shown (Price for Greenland and Larour for Pine Island). This is an alternate approach to the “paleo-to-present” climatic spin up.
  • Temp and Precip fields shown (Ant and Grnld) of AR4 compilations to be used for control runs

9:00 a.m. Model Reports

  • Status
  • What’s gone well
  • What’s not gone well
  • What you need from SeaRISE

(see ppt on LANL Trac site for the compilation of 2-slide summaries from Wang, Ritz; Greve; Bueler; Ayako.)

Oral presentations by Pollard, Fastook, Price, Martin (not posted yet) Good news is that no required data sets seem to be missing however there is a general desire to get the improved fields listed above. Bad news is that there is a strongly expressed desire for uncertainty fields for all input parameters; w/o the uncertainties, sensitivity tests are more difficult to quantify. Jesse will reexamine the sources of the posted data sets to see what can be provided. There are some additional “nice-to-have” data sets.

  • Basal melt regions
  • Sediment distribution
  • Surface melt
  • Ocean temperature
  • Radar isochrones

9:30 a.m. Spin Ups and Control Runs

This section prompted a very long discussion about control runs—what they are; how to do them to quantify variability. In the end, neither the ice sheet models nor the forcing fields are well enough known to conduct the type of pre-industrial spin-ups and pre-industrial-to-present runs that some would have us do.

Following lunch the consensus was that the original strategy of running each model to present day and then completing three control runs—one with present climate held constant into the future and two others with AR4 moderate and extreme scenarios to 2100, followed by a steady climate. Experiments will be compared to these control runs on a model-specific basis.

Because there is a clear distinction between the spin-up/initialization activity and the future runs (both control and experiments), these are separated below. Spin Ups Input

  • Are data sets adequate?
Yes, 
  • Any required new / improved data sets?
As noted above, some improved data sets should be pursued and uncertainties must be quantified to accompany these field.
  • Are climate forcings data sets ok?
No.  Ice-shelf basal melt rates are needed.  3 possibilities: Pollard has a field or they can be parameterized from ocean temps from AWI).
  • When to “freeze” data sets?
August 31, 2009 

Control Runs

  1. modern climate (20th C)
  2. AR4 climates (2m temp.; total precip; LW SW rad.; spec humid.; air temp from Bracegirdle) hold steady for 100-500 yr. 0-200 yr may be availibale
#A1B
#A1F1 

Input

  • Are data sets adequate?
No.  Atmosphere surface data are needed to drive a surface energy balance scheme
  • Any required new / improved data sets?
No.
  • When to “freeze” data sets?
August 31, 2009 

Output

  • Run duration?
200 years required, 500 years encouraged
  • Output format (time steps and parameters)?
Annually: 
 1D:Vol, 
 2D: Zsurf, H, Zbed, Basal water (amt and pressure),  dh/dt
 3D: Velocity, Viscosity, Temp
 ll on polar stereographic, horizontal 5 km grid

Vertical spacing of logarithmically- spaced sigma, 20 levels

Must prescribe output file formats and names (adopt PISM?) PISM to post on wiki

  • Due date for control runs
**Greenland 
**Antarctica 
  October 31, 2009 (2 months after freeze)
  • Who compares outputs from various models?
  • There is a need for a full-time position (post-doc?, where?) who would perform the majority of the model output comparisons and syntheses for group examination and comment

10:30 a.m. Scenario Runs

Greenland
  • Meltwater increase / effect on velocity
    • What models have water & sliding linked?
  • Outlet glacier activation
    • What should this look like?
  • How to specify external forcings
Increase flux by 100% at specified outlet glaciers by some set of local changes 
and have these local change follow a grounding line retreating at a specified rate.  

Specified outlet glaciers could be all of a certain size or character (e.g., all tidewater outlets)

CAN ALL WHOLE ICE SHEET MODELS IMPOSE THIS CONDITION?

1:00 p.m. Scenario Runs (con’t)

Antarctica
  • Ice shelf thinning / removal
    • How many / how quickly / HOW?
    • Ways to handle grounding line
Increase sub-shelf melt rates by 2, 20 and 200 m/a

CAN ALL WHOLE ICE SHEET MODELS IMPOSE THIS CONDITION?

2:00 p.m. Community Issues

Web site
Continue to use U/Montana and LANL sites until a need becomes clear for something new

Future Workshops: Times and Places

Shadow the CCSM Land Ice WG (large overlap in interest and participants).  

Meetings at CCSM in June and approx. Jan-Feb (Boulder?/) Proposal What is money needed for? DoE may be able to absorb marginal workshop costs if CCSM SSC is supportive of SeaRISE shadowing Land Ice WG. NSF and NASA are interested in supporting SeaRISE Bob has already provided NSF with concept proposal to support a full-time post doc for model output comparisons. US institutions with a need for partial student support have already been advised to contact NASA.

Timeline

There seems to be enough time for 3 iterations of experiments and analysis (see my ppt for the timetable)

IPCC input must be accepted for publication by November 2012. Special Issue for results Multiple journals were discussed (Annals of Glaciology; The Cryosphere; JGR). Open access is popular. Model descriptions are highly suitable for a special issue of Geoscientific Model Development. Bob will pursue this. It was suggested there could be a final symposium wrapping up SeaRISE. Perhaps joint with Ice2Sea.

Regional Models

The general feeling was that these wait until the whole ice sheet models have completed their initial set of control runs.

3:00 p.m. Adjourn

Thanks Bob!