[8pt] Add lake masking to stage-based CatFIM processing

[EmilyDeardorff]

Stage-based CatFIM does not currently use any lake masking. This results in it inundating areas that we know are water bodies, which could be misleading.

Update stage-based CatFIM so that it masks out and doesn't inundate areas that are known waterbodies.

Example of stage-based CatFIM at site owsn6 inundating a lake in upstate NY:

[EmilyDeardorff]

Developing the method to remove lakes from stage-based CatFIM

The goal of this task is to remove inaccurate CatFIM inundation in and around lakes.
This issue is clearly showed in the example site auon6:

Initially, I tried just masking out the lake polygons after the inundation layers were created.
However, this still left false inundation around the lake shores:

Next, I tried just filtering out the stream segments that have a LakeID associated with them, so those segments just are not inundated.
This improved the result, but there was still a section of the lake that was wrongfully inundated:

The best answer was to combine these methods. By filtering out the HydroIDs that have an associated LakeID, we refrain from inundating most of the stream segments associated with the lake. And then masking out the lakes at the end cleans up any last sections where the lake still might have gotten inundated.
Here is the final result:

[EmilyDeardorff]

Additional testing of lake masking

I ran stage-based CatFIM with and without the lake masking update for a few different sites.

Lake masking worked exactly as we were hoping in site rudc2.

There is extensive inundation in site dibt2 in Texas. Masking the lake out, while it makes sense, does create a weird looking hole in the inundation.

Lake masking worked as expected in site pgmc2. It's a bit tricky to see because the lake is on the far right of the CatFIM polygons, but the CatFIM is trimmed right where the lake starts.

Site stpm5: This is a funny example! In the current live version of CatFIM, this area had much more inundation and thus needed lake masking. But in our dev version of CatFIM, the inundation is much smaller so the lake doesn't even need to be masked out.

[EmilyDeardorff]

Geopackages of newly masked CatFIM sites and masked areas

In order to facilitate communication about the newly masked CatFIM sites, I've made CatFIM sites geopackages to show which sites are impacted by the new lake masking. There is a lake_masked column, and where the column is 'True,' there is at least 0.02 sq km of lake masked from the CatFIM.

I've also included lake intersection geopackages to show what specific areas are being masked out from the CatFIM with this new update. These files are in a Google Drive folder but available upon request.