Just being formal about this since I made such a big deal about keeping track of everything with Issues. ;)
numeric-codings and binary-codings don't semantically map well onto the
histogram template you use for measurements and rankings
(i.e. divide values into 10 bins of size (MAX-MIN)/10). It would be
very nice if numeric-codings preserved their actual integer values
irregardless of the number of classes (it's never very many) and if
binary-codings presented exactly three values: 1, 0, nodata.
The histogram outputs for marine.models.coastal/coastal-flood-sink
and marine.models.coastal/geomorphic-flood-sink look very strange.
First, there are multiple rows with the same value. We assume this
has to do with differences in precision that we aren't seeing due
to truncating the double values at just a few significant
digits. We request that you either print sufficient significant
digits in the mean values so that each row has a unique value or
that you combine bins at the same level of precision that you are
printing out their means. Hopefully that's clear.
Second, the aggregated values definitely can't be right. This is a
spatial probabilistic distribution, but you report the aggregated
value as 0.021 ± 0 m. Wait...0m? That means the result is
deterministic, but we know that when you aggregate random
variables, you get wider total variance than is present in the
individual variables. This should definitely be investigated.
Just being formal about this since I made such a big deal about keeping track of everything with Issues. ;)
The histogram outputs for marine.models.coastal/coastal-flood-sink and marine.models.coastal/geomorphic-flood-sink look very strange.
First, there are multiple rows with the same value. We assume this has to do with differences in precision that we aren't seeing due to truncating the double values at just a few significant digits. We request that you either print sufficient significant digits in the mean values so that each row has a unique value or that you combine bins at the same level of precision that you are printing out their means. Hopefully that's clear.
Second, the aggregated values definitely can't be right. This is a spatial probabilistic distribution, but you report the aggregated value as 0.021 ± 0 m. Wait...0m? That means the result is deterministic, but we know that when you aggregate random variables, you get wider total variance than is present in the individual variables. This should definitely be investigated.