Thank you so much!
I reprojected all my streams, merged them, and re-imported them, looked good. It updated my previous Nodes table with the new streams, but that still ended up mid-ocean.
So I deleted all your RiverGIS and PostGIS layers, and started over:
Running Create RAS Layers and Tables...
Import data into RAS PostGIS tables...
Importing data from Merged...
Now there are two identical-looking Stream Centerlines layers in QGIS (only one in PostGIS). I'm assuming they are two views of the same file? Both seem to get updated when I add things.
Building stream centerlines topology...
New NodesTable (How did it know the flow directions? From the line segment order? )
Calculating river reach(es) lengths and their end stations...
New Endpoints layer
Copying stream centerlines into flowpaths...
(Nothing obvious happened) Flowpaths file has 2 FpID entries called channel...
I guess I'm ready to create cutlines! I found this - thank you again!
You could try SAGA processing algorithm: Cross Profiles.
In QGIS 3.0.3:
Processing -> Toolbox -> SAGA -> Terrain Analysis - Profiles -> Cross profiles
Looks like exactly what I need. But it errors:
'50.0' is not a correct numeric value for option 'NUM_PROFILE'.
Tried the default:
'10.0' is not a correct numeric value for option 'NUM_PROFILE'.
Looks like NUM_PROFILE parameter is different:
Usage: saga_cmd ta_profiles 3 [-DEM <str>] [-LINES <str>] [-PROFILES <str>] [-DIST_LINE <double>] [-DIST_PROFILE <double>] [-NUM_PROFILE <num>]
-NUM_PROFILE:<num> Profile Samples
I tried to delete the decimal point and zeros from the SAGA dialog, but it puts them right back! I guess that would be a QGIS bug?
the problem appears to be caused by a Real format value being supplied where an Integer value is required
Applied in changeset qgis|c2f224b4462887bcc252a48b2c3c602bf5aee6bc.
But they only fixed the specific one that was reported:
I'd first thought maybe the DEM needed to be explicitly reprojected to my project's EPSG:2226, but that added a new warning:
warning: top-to-bottom and left-to-right cell sizes differ.
using cellsize: 26.677698
Apparently that was not a good thing to do.
I realize this is way beyond your responsibility, but maybe it will help someone who finds this page in the future.