A14. Evaluating Sea Ice in E3SM with an Altimetric Satellite Emulator
| Poster Title | Evaluating Sea Ice in E3SM with an Altimetric Satellite Emulator |
|---|---|
| First Author | Andrew Roberts |
| Topic | cryosphere, water cycle |
| Affiliation | LANL |
| Link to document |
Title
Evaluating Sea Ice in E3SM with an Altimetric Satellite Emulator
Authors
Andrew Roberts, Jonathan Wolfe, Adrian Turner, Elizabeth Hunke (Los Alamos National Laboratory), Alice DuVivier (National Center for Atmospheric Research), Wieslaw Maslowski (Naval Postgraduate School), Ron Kwok (Jet Propulsion Laboratory), Sinéad Farrell (University of Maryland)
Abstract
We address a key deficiency in the evaluation of sea ice in the Energy Exascale Earth System Model (E3SM) using a freeboard satellite emulator that compares simulated sea ice height above sea level with that of polar ocean retrievals by space-borne altimeters. Late winter, spring and autumnal measurements of surface topography of the Arctic Ocean by the Geoscience Laser Altimeter System (GLAS) instrument aboard the Ice, Cloud, and land Elevation Satellite (ICESat) are used to evaluate E3SM. Sea ice freeboard is retrieved from model grid cells in close spatiotemporal proximity to GLAS samples, and used to generate basin-wide skill and bias statistics of model simulations. We compare E3SM results with the Community Earth System Model (CESM) Version 2 using both the CAM and WACCM models, and also different datasets of ice freeboard and thickness derived from ICESat. A key outcome of this work is that none of the fully coupled CMIP6 DECK simulations from E3SM nor CESM perform particularly well, whereas the observationally-guided Regional Arctic System Model (RASM) does a much better job in creating the spatial pattern commonly seen in sea ice thickness in the central Arctic. In all models, ice freeboard bias is as much as a full standard deviation removed from observations and this is unaffected by the snow cover on the ice. However, what sets E3SM apart from CESM is its large variability across the DECK ensembles, suggesting problems with E3SM’s standard-resolution polar atmosphere, ocean or sea ice physics. Critically, we are able to assign precise skill scores to characterize the E3SM’s and CESM’s limitations, and thus set a benchmark by which the E3SM V2 polar climate may be judged in future publications.