Sensitivity of sea ice to changes in air temperature

CMIP6 Hackathon CMIP6 Project Proposals
1

Sensitivity of sea ice to changes in air temperature

Topic proposed by myself, Katie Brennan, and Robin Clancy at UW

Scientific Motivation

The state of global sea ice is strongly related to changes in global mean temperature. This is in part a result of the sea ice-albedo feedback, which amplifies ice loss with increases in Arctic air temperature. The sensitivity of atmospheric temperature response to changes in sea ice mean state (i.e. Rind et al., 1995) highlights the importance of constraining the sea ice response to climate perturbations. However, the magnitude of this response and feedback in global climate models is strongly dependent on parameterizations and tuning of the mean sea ice state. The observed decline in sea ice extent has been more rapid than predicted by nearly all CMIP3 models, as well as most CMIP5 models (Stroeve et al., 2012).

There is a large amount of internal variability inherent in the sea ice system. This variability is drastically different across the models (Swart et al., 2015), and is an important consideration for interpretation of changes in sea ice. The divergence of prior CMIP model predictions from observations is suspected to be in part explained by the role of internal variability, and possibly by an under-representation of the sensitivity to changes in temperature (e.g. Stroeve et al., 2012).

Proposed Hacking

During the CMIP6 hackathon, we hope to evaluate the degree to which changes in air temperature (and in particular, Arctic air temperature) result in changes in the sea ice state.

  • Evaluate changes in sea ice mean state with warming
    • Essential to consider thickness as well as extent, as models with initially thicker ice generally retain more ice throughout the 21st century (Holland et al., 2010)
    • Are there any “tipping points” at which summer sea ice area rapidly declines?
    • Evaluate spatial patterns in sensitivity across models
  • Evaluate the relationship of air temperature with the internal variability in sea ice.
    • Are patterns consistent across models?
    • Are patterns consistent with those found in reanalysis?
    • Are there any lagged effects of heating in Spring that create larger sea ice anomalies in summer?
  • Evaluate differences between models and observations
    • Do any terms in sea ice energy budget help explain model variation from observations?
    • Examine asymmetries between the hemispheres in models and observations

Anticipated Data Needs

Sea ice variables: area (siarean, siareas), concentration (siconc or siconca), thickness (sithick), volume (sivol, sivols, sivoln)
Atmospheric variables: near-surface air temperature (tas), downwelling longwave flux (rld)
Ocean variables: sea surface temperature (tos)
Global variables: grid cell area (areacello)

Experiments:
Pre-industrial control, historical, 4xCO2/some other warming scenario

Other data: Compare results to observations/reanalysis
Historical sea ice concentration (https://nsidc.org/data/g02202#)
Atmospheric temperature (ERA5 or some other reanalysis product)

Anticipated Software Tools

Python/xarray

Desired Collaborators

Anyone with interest in sea ice or polar climate.

1 Like
2

Maddie this is a great proposal! As someone interested in polar regions myself, I’m very glad to see the cryosphere getting some attention at the hackathon.

Have you already made a data request for the variables you need? Time is short at this point, so I want to make sure they are available.

3

Nope, not yet. I’m obviously a bit behind on orienting myself to everything that’s going on, but plan to submit a data request today. (Hopefully not too late!)