Richard Davy and Stephen Outten (NERSC) published an article in the Journal of Climate, titled “The Arctic Surface Climate in CMIP6: Status and Developments since CMIP5”. Their work shows how phase 6 of the Coupled Model Intercomparison Project (CMIP) has improved from phase 5, with regard to the representation of Arctic climate in the models used in CMIP.

In the Earth’s climate system, the Arctic plays an important role. Factors that influence the climate system externally, such as greenhouse gas build-up, are called forcings, and they have an especially strong influence on the Arctic. Temperatures in the Arctic have been increasing more than the global average. To understand how the Arctic will react to future climate change is of extremely high importance, and Richard Davy and Stephen Outten investigated different scenarios with weak and strong forcings for the Arctic to do exactly that. Low emissions of greenhouse gases are a weak forcing, whereas high emissions define a strong forcing.

The Coupled Model Intercomparison Project (CMIP) is an international project that has been running since 1995, and the aim is to improve the understanding of climate changes in the past, present, and future. The tools to achieve this aim are climate models which can be used to simulate the climate back in time, as well as into the future. The entire project has been advancing climate science significantly in the past decades, and it contributes to national and international climate change assessments such as the Intergovernmental Panel on Climate Change (IPCC).

CMIP6 is the latest phase of the project, and Davy and Outten set out to investigate how 34 of the models used in CMIP6 perform, compared to 49 models used in CMIP5. The models simulate changes in sea ice concentration, sea ice volume, sea level pressure, and surface air temperature. They also used different forcing scenarios, simulations in which emissions are different, to investigate projections for future climate scenarios using the CMIP6 models. Davy and Outten compared these simula

tions and scenarios and their most important findings described in this study are the following:

1. Figure 1: Time series of the annual sea ice volume produced by CMIP6 models from the historical scenario (grey) and future predictions based on low (SSP126, green) to high (SSP585, purple) emission scenarios. It is obvious that with higher emissions, the annual sea ice volume decreases faster over time. The rapid decline in sea ice volume starting around the 1980’s (black line) is really well captured by the historical scenario (dark grey line), which is produced by the CMIP6 models looking back in time.: Graphic: Davy & Outten, 2020Simulating the historical climate conditions in the Arctic has improved, CMIP6 performs better than CMIP5, especially for the Barents Sea area. They attribute this to changes in how the sea ice edge, the thickness, and ice retreat are being represented in the models.
2. With CMIP6, Davy and Outten compared different scenarios for the 21^st century to investigate sea ice extent and sea ice volume, among other parameters, and two scenarios are especially interesting to compare to each other. The lowest emission scenario (SSP126, weak forcing) predicts that the Arctic will have sea ice year-round in the 21^st century, whereas all higher emission scenarios (e.g. SSP585, strong forcing) predict an ice-free Arctic in the summertime in the coming decades.
3. Another finding by Davy and Outten is that all scenarios, independent on how weak the forcing is, indicate that the Arctic will be warming three times faster than the global average. This means that even if the global average “only” increases by 2 °C, the Arctic will warm by 6 °C!

Richard Davy sums up the importance of their work: “The major success story of CMIP6 over previous model generations is that the ensemble mean of the models really well capture the rapid decline in Arctic sea ice volume that has been observed since 1980. However, there is still a lot of uncertainty about sea ice thickness, both in observations and models. And crucially, there is a lot of uncertainty about how the ocean, atmosphere, and sea ice are coupled together in an Arctic that is mostly made up of thin, first-year ice.”

How well the sea ice volume decline is captured by the models in CMIP6 can be seen in Figure 1.

CMIP6 performs well with respect to the Arctic region and we can use CMIP6 to make even better climate predictions than with CMIP5. Using CMIP6 advances climate predictions, especially for the Arctic, one of the most vulnerable parts of the Earth’s climate system, but improvements are always necessary.

In addition to this study, Davy animated the extent of the sea ice in summer for all models used in the low emission scenario and the high emission scenario. A snapshot at 2050 (Figure 2) shows that 50% of the models running the high emission scenario predict the sea ice extent to be minimal, only a few places north of Greenland and Canada have sea ice in 50% of the models. In the low emission scenario, the sea ice extent is much larger. This is a dire reminder of how greenhouse gas emissions affect the Arctic sea ice extent.Figure 2: Low emission scenario (SSP126) on the left vs. high emission scenario (SSP585) on the right for the year 2050. The red line indicates the extent of sea ice that 50 % of the models used predict for 2050.: Graphic: Richard Davy, NERSC

Reference:

Davy, R., and S. Outten, 2020: The Arctic Surface Climate in CMIP6: Status and Developments since CMIP5. Journal of Climate, 33, 8047–8068, https://doi.org/10.1175/JCLI-D-19-0990.1.