search

NIEHS / PrestoGP

Penalized Regression on Spatiotemporal Outcomes using Gaussian Processes a.k.a. PrestoGP
https://niehs.github.io/PrestoGP/
0 stars 0 forks source link

Principal Aquifer - Covariate Calculation #8

Open kyle-messier opened 11 months ago

sigmafelix commented 9 months ago

I will use Principal Aquifer from the USGS metadata. Do we keep all aquifer types/codes when multiple aquifers intersect with HUCs?

kyle-messier commented 9 months ago

@sigmafelix the link is broken. Let's decide on 1 or 2 levels for aquifers (similar to HUC scale). And keep all intersecting aquifers for a given level. There is uncertainty in the well depth and the extent of aquifers so if the exact aquifer is derived from is uncertain and multiple values can reflect that

sigmafelix commented 9 months ago

@Spatiotemporal-Exposures-and-Toxicology USGS Aquifers of the United States states that the data is at 1:2.5m scale and I could not see levels in the aquifer data I am using. There are rock types and codes, aquifer names and codes, and OBJECTID. I think alluvial and glacial aquifers data in the manuscript in Overleaf? As it covers around the Great Lakes, many HUCs in other regions will be assigned NULL for alluvial/glacial aquifer columns.

kyle-messier commented 9 months ago

@sigmafelix that works then. If the well location overlaps with multiple, then I'd say to include both.

sigmafelix commented 9 months ago

09/29/2023 update