The role of natural versus forced change in future rapid summer Arctic ice loss

Climate model simulations from the Community Climate System Model, version 3 (CCSM3) suggest that Arctic sea ice could undergo rapid September ice retreat in the 21st century. A previous study indicated that this results from a thinning of sea ice to more vulnerable conditions, a "kick" in...

Full description

Bibliographic Details
Other Authors: Holland, Marika (author), DeWeaver, Eric (editor), Bitz, Cecilia (author), Bitz, Cecilia (editor), Tremblay, L. (author), Tremblay, L. (editor), Bailey, David (author)
Format: Article in Journal/Newspaper
Language:English
Published: American Geophysical Union 2008
Subjects:
Online Access:http://nldr.library.ucar.edu/repository/collections/OSGC-000-000-002-728
https://doi.org/10.1029/180GM10
Description
Summary:Climate model simulations from the Community Climate System Model, version 3 (CCSM3) suggest that Arctic sea ice could undergo rapid September ice retreat in the 21st century. A previous study indicated that this results from a thinning of sea ice to more vulnerable conditions, a "kick" in the form of pulse-like increases in ocean heat transport and positive feedbacks that accelerate the retreat. Here we further examine the factors affecting these events, including the role of natural versus forced change and the possibility of threshold-like behavior in the simulated sea ice cover. We find little indication that a critical sea ice state is reached that then leads to rapid ice loss. Instead, our results suggest that the rapid ice loss events result from anthropogenic change reinforced by growing intrinsic variability. The natural variability in summer ice extent increases in the 21st century because of the thinning ice cover. As the ice thins, large regions can easily melt out, resulting in considerable ice extent variations. The important role of natural variability in the simulated rapid ice loss is such that we find little capability for predicting these events based on a knowledge of prior ice and ocean conditions. This is supported by results from sensitivity simulations initialized several years prior to an event, which exhibit little predictive skill.