Length could be represented better in the reference fabric.
The reference fabric currently uses the length of the burn line event.
In cases where headwater flowlines are trimmed to a little short stub, they are going to be strange and in places where divergences have been trimmed they are also going to be a bit odd.
An initial solution would be to not replace the flowline length attribute at all as is done here: https://code.usgs.gov/wma/nhgf/reference-hydrofabric/-/blob/main/workspace/R/utils.R#L51
A more nuanced approach would be to only recalculate the length for headwater flowlines.
We should check with the SPARROW team and find out which they would prefer.
We could easily overwrite the length attribute in the initial refactor workflow to fix the length attributes in the mean time.
@abock does that make sense? What do you think?