UNDERC depth data and crosswalk, Isle Royale hypso data and crosswalk

[jordansread]

I put both issue #261 and #176 in here. #261 is only partially resolved because the temperature data are not yet in the pipeline and the light attenuation data (additional processing of "PAR" profiles in the same dataset as the temperature) are required to gain more lakes.

Despite the UNDERC dataset having 100+ lakes, we only gained one from that depth dataset. The others were either already in the pipeline or not in but still missing Kw/secchi data so not considered a "GLM" lake at this point, which is our lowest threshold for modeling.

The Isle Royale (ISRO) dataset added more, since the pipeline seemingly already had temperature and secchi data for these sites (likely from WQP) and so adding hypsography resulted in ten new GLM lakes and eight new PGDL lakes. Also, since these lakes seem to be highly observed, it is worth noting that I only added the low hanging fruit contour lakes which could be easily turned into hypso. I didn't try to handle multilines so I'm not sure if issues like #167 will exist or if these could be cast into polygons without issue. We probably want to spin off another lower priority issue for the rest of the ISRO lakes.

[jordansread]

Oh yeah, a zoom in on the island

for #176

[lindsayplatt]

Ran scmake('8_viz/out/lakes_summary_fig.html.ind') and it did what I expected up until it started to build 7_drivers_munge/out/7_drivers_munge_tasks.ind, so I stopped it.

[lindsayplatt]

I wonder why all of these hashes changed?

[jordansread]

I think this is related to this scipiper issue but I haven't dug into it. Agreed that it is annoying, whatever is causing it.