[5pt] Erroneous inundation in FIM 4

[CarsonPruitt-NOAA]

We've been seeing some large areas of off-channel inundation in FIM 4 and need to investigate to see what's going on. You should take a look at the catchments layer and maybe the synthetic rating curves.

Below is an example screenshot of the issue just SE of Arkadelphia, AR. HUC 08040102 branch 2200000001 HydroID 18440040. HydroVIS is currently using version 4.0.13.1, so please make sure you're looking at that version. In order to see if our inundating workflow matches what HydroVIS is displaying, please use /tools/gms_tools/inundate_gms.py to inundate the HUC with the flow file that I saved here: /data/inundation_review/archived_flow_files/20230211T1817Z_ana_high_flow_magnitude.csv. It would also be a good idea to check that this catchment has the correct feature_id 22002068.

[DamiEyelade-NOAA]

The main issue here is that the DEM derived channel network is different from the underlying NWM network. The branch 1 stream departs from the true course of the Ouachita river main stem North East of Arkadelphia AR and then follows the flow path of a tributary before reconnecting to the main stem river South East of Arkadelphia AR. It appears the flow that should go through the main stem is being attributed to the branch 1 tributary around Arkadelphia leading to flooding (red flood extent) due to an undersized channel. Note that Branch 0 has the right course for the main stem river and it appears the same flow does not cause excessive flooding around Arkadelphia (Blue flood extent).

A zoomed in view shows that the DEM derived network (Magenta) differs significantly from the true course of the Quachita river (yellow). At the point at which branch 1 departs from the true path of the main stem (green oval) there is flow in the upstream direction before the river then reconnects to a tributary where erroneous flooding occurs.

Possible Root Issue:

1. REM/DEM Artefact

The REM for the erroneously flooded reach around Arkadelphia has an entrenched narrow gorge present. This narrow gorge has a very high elevation difference between the channel and the area that drains into it (elevation difference from gorge floor to bank = 20m). The actual channel of the Ouchita river is visible in the REM and also has a wider channel but the elevation difference between the valley floor and banks is = 3 - 4m on average. Given this DEM/REM artifact the level paths are likely to be ignored and the DEM derived network will still preferentially follow the deeper gorge and route "upstream" (green circle) to cut across the actual channel because it has a lower relative elevation difference than the artefact gorge.

2. Boundary issues: The flooded reach has an originating point outside of the flooded catchment. Extra investigation could show if this is a factor.

[RyanSpies-NOAA]

@BradfordBates-NOAA This ticket can be closed with PR #903. Here is the NWM 100yr inundation for the area of interest in HUC 08040102 for FIM 4.3.11.2 (blue) vs. FIM 4.3.3.4 (orange) -->