Polynya test case

[einola]

A simple polynya setup from Bjornsson et al. (2001). It is good in that it tests both dynamics and thermodynamics at the same time.

The setup is a 50 times 100 km box, with one of the 50 km boundaries open and the other closed (coast). A wind (10 m/s) blows at an angle so that the ice drifts both out of the open boundary and onto the coast. Ice piles up at the coast and then moves out of the domain, making room for new ice that is created in the "polynya" that forms.

Figure: The setup from Bjornsson et al. (2001).

Figure: The results of one of the runs from Bjornsson et al. (2001).

[einola]

Everything for this test is now implemented in the branch called issue375_polynya_test. However, the test crashes because of an instability at the outflow boundary's lower corner.

[einola]

Sea ice concentration for the first 20 hours (4 hour output time step).

[einola]

U velocity after four hours. The instability is already visible as an inflow of ice in the lower right corner.

[einola]

U-velocity during the first for hours, at a per-time-step output resolution. Red, white, and blue are on the right-hand boundary, bottom, middle, and top, respectively. Green is in the interior. It looks like initial oscillations at the boundary lead to an inflow, which is undefined and causes the model to crash.