Add static river/catchment attribute data

[lekoenig]

This PR adds new river/catchment attributes to the file 1_fetch/in/target_sciencebase_attributes.csv so that these features get downloaded, processed, and added to a formatted table that contains the static input feature datasets we're interested in (p2_seg_attr_data).

The changes in this PR include the following steps:

• changes to 1_fetch/in/target_sciencebase_attributes.csv to specify new attributes to collect. I gathered these based on my own hypotheses for features that might be important, but I expect we'll ultimately end up using only some portion of them. Please also feel free to suggest any other features that you think should be included ⭐
• processing the downloaded attributes. Most of the datasets are minimally processed since they are already referenced to the NHDPlusv2 network. However, we do further process the land cover data by aggregating the NLCD index groups (e.g. 90 and 91) into reclassified groups (e.g. "wetland") based on the lookup table in 1_fetch/in/nlcd_landcover_reclassification.csv. This file is new and is committed here so that others will have it when cloning the repo.
• Combining all of the individual attribute data frames into a single table in p2_seg_attr_data using the function combine_nhdv2_attr().
• Summarizing the attribute values. As a quick method of QA/QC the pipeline now includes a target in 3_visualize that outputs a csv file containing summary statistics (min/mean/max) for all of the processed attributes. I created a new function, summarize_static_attributes() to create and save that table.

This PR is just focused on downloading and formatting the attribute data. We'll use some method of feature selection to finalize the input variables to the LSTM models in #164.

Closes #51

[galengorski]

Hey @lekoenig , this looks great to me. The only other attribute that I had thought of is a calculation of the amount of reservoir storage/mean flow at each site. I have seen this as a way to capture the degree of hydrologic alteration. I'm not sure if it makes sense for our sites as so many are on the same rivers and they will have very similar values. Otherwise I think it looks great.

[lekoenig]

The only other attribute that I had thought of is a calculation of the amount of reservoir storage/mean flow at each site. I have seen this as a way to capture the degree of hydrologic alteration.

Thanks, @galengorski - great suggestion! After our conversation last week I added these two variables from the larger Wieczorek dataset (NID_STORAGE and NORM_STORAGE). Are these the variables you were referring to? If not, do you have a link to alternative datasets that address reservoir storage?

[galengorski]

Yeah I think that sounds great!