Adding fictitious channels for vector-based setup for the outlet nodes

The current version of MESH r1860 does not route streamflow in outlet sub-basins/channels. Therefore, as a quick fix, a fictitious channel could be added to each outlet sub-basin/channel, to facilitate the routing. Since these fictitious channels are considered as "outlet" instead, no routing will be done on them either and modellers could trust the result of vector-based routing.

Copying @mee067

Mohamed's comments on Slack:

for gridded setups, we used to have one more grid to account for that. GK does it automatically. It has a Rank with Next = 0 Earlier versions of the code used not to even run CLASS on that grid but newer code (1860) does run CLASS for that grid. Moving to vector-based setups, CLASS would thus run on the last subbasin which probably is not an extra basin. runoff will be routed to the instream channel but I do not think there will additional channel routing for the flow - I am not 100% sure though. I do not think that's critical. It does produce output at the outlet.

The thing is, MESH needs a NEXT value to route to, and assumes the channel to go from a grid/sub to its NEXT one. It's not from the most u/s point within a grid/sub to its most d/s point. I don't think there is channel routing for the last sub and thus we need to add another fictitious one d/s to get the effect. How much difference would that make for the outlet flow, I can't say, but I don't think it's that large.

The vector workflow does not add that extra sub and therefore, I found that Qo = Qi for the last sub for the vector-based Yukon setup. Qo is the channel outflow, Qi is the inflow from u/s subs + local RFF. I made a quick comparison for the Yukon outlet and I do not feel the local runoff is even included. I see all loops go only to NAA and then this is a special treatment for the outlet on lines 958-960 of rte_module.f90:

https://github.com/MESH-Model/MESH-Dev/blob/ff1de25e67ef552844f1e789c0faffb4175a8dac/Routing_Model/RPN_watroute/sa_mesh_process/rte_module.f90#L957-L959

For large basins such as the MRB, YRB, this may not be important - runoff from the last sub/grid would not be that large compared to the huge flow collected from the basin

but for small basins - like those little basins draining to the sea or a lake, ignoring the runoff and routing within the last sub could be significant - depends on the size and relative contribution of that last sub to the flow of the basin. It could be raining a lot on the shore line of a marine basin or a lake shore to make significant contribution.

CH-Earth / meshflow

Adding fictitious channels for vector-based setup for the outlet nodes #6