Arctic climate simulation using CRCM5, with correction of sea surface temperatures

Context

The Arctic is warming 2 to 3 times faster than the global average. It is vital to be able to make projections on the future climate for impact and adaptation studies. These are carried out using climate models. The model used in this study is the fifth version of the Canadian Regional Climate Model (CRCM5).

Two sets of future projections were prepared, using both the traditional two-step approach (MCGC-RCM) and a new three-step approach (MCGC-CGCM-RCM) with empirical correction of systematic biases in sea surface temperatures. Before calculating future climate projections, the past performance of the model had to be evaluated.

Objective(s)

To evaluate the past performance of the CRCM5 in order to quantify the structural biases of the model when it is driven by reanalyses, as well as the influence of errors in border conditions when the CRCM5 is driven by simulations from an MCGC
To conduct a sensitivity study to evaluate the effect of the large-scale spectral nudging method when the CRCM5 is driven by reanalyses
To evaluate the effect of the three-step approach
To calculate projected climate change with the three-step approach and with the two-step approach

Methodology

Compare CRCM5 simulations driven by reanalyses (hindcast) with available observations
Compare CRCM5 simulations driven by reanalyses (hindcast) with and without large-scale spectral nudging
Compare the historical simulation from the CRCM5 driven by the MCGC to the one driven by reanalyses
Compare the historical simulations from the CRCM5 driven by the MCGC, using both the two-step approach and the three-step approach
Evaluate projected climate change with the RCP 8.5 scenario using the two-step approach and the three-step approach

Results

The CRCM5 demonstrated a good ability to simulate the general characteristics of the historical climate (1981-2010), particularly atmospheric circulation. The large-scale spectral nudging method does not have a significant effect on the Arctic region. The analyses showed that the three-step dynamic downscaling approach with correction for systematic sea temperature biases has little effect on the Arctic.

For all variables, the projected climate change patterns for the near future period (2041-2070) were similar to the projections for the far future (2071-2100), but the changes were greater in the far future. The future projections with the two-step and three-step approaches were quite similar. According to these simulations at the end of the 21st century, conducted using the CRCM5 driven by the MCGC MPI, the concentration of sea ice in the Arctic Ocean in the fall will drop by 60 to 100%, with the minimum change in sea ice concentration expected in spring, with a reduction of 20%. There are two regions near the Kara Sea and the Chukchi Sea that show the first signs of sea ice reduction during spring and winter. The maximum projected increase in sea surface temperature is 16˚C in the zone of summer ice loss. For the T2m (please see figure below), warming is most pronounced in winter (around 22°C) in the same two places where there is the greatest reduction in sea ice (Kara and Chukchi Seas). As expected, due to higher temperatures and more water vapour in the atmosphere over the coming decades, an increase in precipitation of 0.5 to 2 mm/day is projected by the CRCM5.

Benefits for adaptation

The CRCM5 driven by MPI projects almost no sea ice in the Arctic Ocean in 2100. Furthermore, warming of 22°C is projected until the end of this century. These robust changes can affect the lives of Indigenous people, animals and the whole planet.