Adaptation Actions for a Changing Arctic: Perspectives from the Barents Area

75

Chapter 4 · Physical and socio-economic environment

Change in sea-ice concentration

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Change in sea-ice concentration

Change in sea-ice thickness, m

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

Figure 4.13 Change in sea ice concentration and thickness in March. Downscaled GISS sea ice concentration (upper left), NCAR sea ice concentration (upper right), NorESM sea ice concentration (lower left), and NorESM sea ice thickness (lower right). The upper plots are from Sandø et al. (2014a) and show change between present (1981–2000, data from the 20C3M control run) and future (2046–2065, A1B scenario). In the lower plots, change is between 2010–2019 and 2060–2069 using the RCP4.5 scenario (Bentsen et al., 2013).

the regional models through initial and boundary conditions. Differences in the representation of the hydrological cycle in the GCM simulations lead to large differences in the ocean salt budget that regional downscaling cannot change much. The ocean is too inert and the impact of the GCM results from the initial and boundary conditions is too large. So, despite improvements due to increased resolution in regional models, unrealistic biases in the global model projections will influence the final regional results.Figure 4.12 shows the projected change in sea-surface temperature from the downscaled NorESM4.5 model.The downscaledRCP4.5 results fromGISS-AOM,NCAR- CCSM, and NorESM show the greatest temperature increase in the Barents Sea will occur in March, which differs from the rest of the Arctic Ocean. Like the two downscaling studies reported by Sandø et al. (2014a), this model also shows a warming of 1–2°C for March in most of the Barents Sea. This warming is reflected in the sea-ice extent data shown in Figure 4.13, where reductions relative to the present ice concentration can be seen

2014; Sandø et al., 2014a,b), but observations show none of the CMIP5 GCMs are able to simulate sufficient inflow of heat through this region (Sandø et al., 2014a). Results from ocean- downscaling of twoCMIP3 control climate simulations (20C3M) were analyzed by Melsom et al. (2009) and Sandø et al. (2014a). They found that sea ice and hydrographic conditions in the Barents Sea are reproducedwell in the downscaling experiments, despite large regional biases in the GCMs used for boundary conditions. This improvement is attributed to a more realistic ocean circulation and inflow of Atlantic Water in the Barents Sea Opening due to higher resolution in the regional models. Can similar improvements be expected if the future scenarios from the GCMs are downscaled? Comparing the downscaled CMIP3 GCMs for the A1B scenario shows relatively good agreement for future temperature rise, but large differences for future salinity (Sandø et al., 2014a). These differences were attributed to deviations in the GCMs that were transferred to