Too much snow compaction causes errors

[arbennett]

There are two sub-issues that arise from a rare case when too much snow compaction occurs:

1. subsequent calculations of volumetric fractions of ice + liquid end up greater than 1
2. subsequent calculations of layer density can be too small

[bartnijssen]

@arbennett: What do you mean with the second point? I don't quite understand what it means that the density is too small ...

[arbennett]

There is a check in the snow compaction routine which ensures that layer densities are above 40 kg/m^3. This tends to happen when the layers are very small (depths below ~0.1mm) after compaction and after corrections ensuring that volumetric fractions add up to 1.

Of course, at such small depths SUMMA should perform a layer merge, but the layer merge code happens before the compaction code, so this error can still happen in the code's current form.

[arbennett]

This was solved via #364