P3: limiting behavior of mass/number ice terminal velocity

From question in #583. Should we return `0` bulk (mass/number weighted) terminal velocity if _either_ `L_ice` or `N_ice` are (near) zero, or should mass-weighted only check for `L_ice` and number-weighted only check for `L_ice`?

Start by reading the question in #583 and tracing how mass-weighted and number-weighted ice terminal velocities are currently defined. Resolve whether each weighting should return zero when L_ice or N_ice is near zero, then document the decision and add coverage for the selected limiting behavior.

Written by the indexing model from the issue text.