The Arctic sea ice cover has declined dramatically over the past 20-30 years and the fastest rate
of the decline has occurred in the Pacific sector of the Arctic Ocean over the Bering, Chukchi and
Beaufort Seas. The shrinking of sea ice cover has posed serious threats for climate, environment,
ecosystems and human societies not only within the Arctic but also in the midlatitudes and tropics.
This global impact may become more obvious and severe in the future considering that the Arctic
is projected to be ice free in summer by the middle of the century. However, there is a large
uncertainty associated with the timing of the first sea ice-free summer in the Arctic due to the
chaotic nature of internal variability in climate models and the uncertainty in future CO2 scenarios.
Recent studies suggest that internal climate variability might be as important as anthropogenic
influences on the observed Arctic sea ice decline over the past decades. Our recent studies further
suggest that a summertime atmospheric regional barotropic height increase over Greenland and
the Arctic Ocean, which has been partially driven by tropical SST variability, is an important
contributor to sea ice loss, especially over the Pacific side. This tropical-Arctic teleconnection may
thus serve as a main internal source of uncertainty in future projections of the Arctic summer
climate. However, most CMIP5 historical experiments driven by observed anthropogenic forcing
do not reproduce this observed warming and melting process successfully, which is possibly due
to an inability of the models to accurately replicate the observed tropical-Arctic teleconnection.
These biases cast doubt over CMIP6’s credibility in projecting the future of Arctic sea ice. Better
understanding of the origin of these biases in CMIP6 relies on a process-oriented evaluation. The
large diversity of models and experiments in CMIP6 can shed light on understanding to what extent
tropical teleconnections, and their resulting circulation response over the Arctic, influence sea ice.
We will evaluate the performance of CMIP6 models in reproducing the recent 40-year tropical–
sea-ice teleconnection and diagnose successes and failures. There are two principal goals in this
proposal: 1) develop and evaluate process-based metrics that characterize how well CMIP6 models
reproduce the observed tropical–high-latitude circulation–sea ice connection. Based on how well
CMIP6 models perform on our metrics and other criteria such as their ability to capture climate
features of the Arctic sea ice domain, we will select a subgroup to examine future projections. We
hypothesize that by selecting a subgroup of CMIP6 and subsequent models that can faithfully
simulate these tropical-Arctic connections we may increase confidence in the projection of Arctic
climate change over the next 10–20 years and help reduce uncertainty associated with future
projections of the first ice-free summer in the Arctic. 2) perform a comparison of the observations
with control, historical, pacemaker and AMIP simulations of CMIP6 can shed light on how
different mechanisms have contributed to observed changes in the Arctic over the past 40 years.
This approach offers a way to quantify the relative contribution of each factor in the recent sea ice
melting and may potentially also serve as a constraint on future predictions.
Our proposed research focuses on a process-oriented evaluation of CMIP6 that could improve
our physical understanding of the models’ biases in reproducing the observed tropical-Arctic
interaction and thus characterize CMIP6 models’ confidence and uncertainties in projecting the
first ice-free summer in the Arctic. Therefore, this proposal targets the first and second research
areas of the CPO competition 6’s solicitation: MAPP-21st Century Integrated U.S. Climate
Predictions and Projections. Moreover, all activities are strictly related to NOAA’s long-term
climate goals for “providing the essential and highest quality environmental information vital to
our Nation’s safety, prosperity and resilience”.