Multiple subsets of size/distribution within the same problem type

We had a situation where wind slab that we expected to be size 1 more widely, and in very specific locations it could be size 2 or even size 3. There is no way to express this level of granularity in the forecast currently.

One question would be how does this flow through to the calculated hazard rating and likelihood?

* For likelihood we can just take the upper bound of likelihood.
* For hazard we can take the upper bound of size.

No file, test, or entry point is identified in the issue. Start by tracing how forecast problem types, size, distribution, likelihood, and hazard ratings are represented and calculated. The work is done when one problem type can express multiple size/distribution subsets and the agreed upper-bound behavior is covered by tests.

Written by the indexing model from the issue text.