Suggested conflation logic for NWM reaches spanning multiple source RAS reaches

[sclaw]

Several issues have been identified in model subsetting when the conflation step selected cross-sections on different RAS source model reaches for the upstream and downstream limits of a sub model. Here I outline the cases that should be handled in new logic. The new logic could be placed in the process_as_multiple_ras_reach function in subset_gpkg.py or as part of the original model conflation.

Case 1 (hydrologically connected, separated with junction)
Example: Here, NWM reach 10433968 traverses pink and orange source model reaches with a junction in between.

Handling: As currently implemented, walk junction table to identify junction(s) between the source reaches. Add intermediate reaches if any exist. Join cross-section list from u/s, d/s and intermediate reaches to create sub model.

Case 2 (confluencing, separated by junction)
Example: Here, NWM reach 10434076 picks up pink sections on the reach it confluences with as the d/s limit. A junction is present.

Handling: When junctions are present but no junction is found that links u/s reach and d/s reach, intersect cross-sections from both u/s and d/s reaches with NWM centerline and toss out the reach with fewer intersecting cross-sections. Identify a new d/s XS as the XS closest to the d/s limit of the NWM centerline.

Case 3 (no junction)
Example: see #216
Handling: treat as you would case 2. Additionally, add quality note to conflation metrics stating that the sub model is missing a junction.

[sclaw]

Example of issue supporting Case 2:

Collection: mip_05120114
Model: SA_8_Big_Muddy_East
Sub model: 10434076
Description of issue: FIM is super deep
Debugging notes: reach 10434076 gets tagged with sections on REACH 1. It then puts its actual d/s section (12467) upstream of 11121 (u/s section of REACH 1). 100yr WSE at 11121 is 482.75, which is way above the expected WSE for 12467 of 443.5

[sclaw]

Case 2

Collection: mip_05120114
Model: SA_8_Big_Muddy_East
Sub model: 10433892

[sclaw]

For Case 1, we should also add handling for reaches u/s of a junction w/ zero d/s reach length. When put into one sub model, those reaches make the model invalid. I suggest we read the length from the junction and force that as the d/s reach length for the XS u/s of the junction.

Example:

Collection: mip_05120114
Model: salt_creek_tribA
Sub model: 10432210
Description of issue: No FIM
Debugging notes: d/s reach length at 14500 is 0. This is fine when there’s a junction, but it won’t run when we merge to one model. Suggest adding junction distance handling to restationing and updating d/s reach lengths accordingly.

[sclaw]

another Case 1

Collection: mip_05120114
Model: Final_Model
Sub model: 10432006 and others in that area
Description of issue: No FIM
Debugging notes: same junction 0 d/s length issue