Metadata: 05_predictions

[lindsayplatt]

For this step,

• [x] Update the basics: title, abstract, purpose, start/end date (of data), themekeywords, process-date, indirect-spatial, latitude-res, longitude-rest, data-name, data-description, file-format (see mntoha examples and hyperscales examples)
• [x] Remove the entries that only need to exist in the shared text file (authors, pubdate, doi, contacts (incl metadata contact), accuracy test, funding creds, process description, distribution person, liability statement)
• [x] Add entities for each file
• [x] Add attributes for contents of each file
• [x] Include appropriate citations ("larger citations" and cross citations)
• [x] Add build environment
• [x] Build the XML using scmake("out_xml/[STEP]_sb.xml")
• [x] Validate the XML output using https://mrdata.usgs.gov/validation/

[lindsayplatt]

Cites GLM 3.0. Does it need to be 2.0?

[jordansread]

That paper though describes the model in detail. We should cite that paper but make it clear we used GLM v2

[lindsayplatt]

I am seeing both a date and DateTime column in the prediction outputs. The only difference I can see is that pb0 has a timestamp of 00:00:00 and pball has a timestamp of 07:00:00. Are both of these columns needed?

[jordansread]

whoops. No, DateTime shouldn't exist and it looks like I need to take a look at the pball export to verify it is the right day.