Feedback from internal review

[cnell-usgs]

Must haves:

• Fix link to GWSIP
• Respond to comment and add more detail to data notes. Q: Does the page include manual measurement only wells? You're only using time-series DVs? If so, then we may be might br missing most of the data and should revise the title to say "USGS Groundwater conditions at continuous wells" @lindsayplatt pinging you to respond to this
• Make it extremely clear how sites were chosen for display. WSCs will likely reach out and ask how to get more points on the map! [the reviewer's] understanding is that we are showing is: (1) monitored by transducer (or instrumented) so have a DV (or IVs) available, (2) have data available for each day of the entire calendar year (is that right?), (3) not required to be real-time, (4) have a long enough history (3 years of data)
• Consider filtering to only reviewed data. This is a sticky issue. GW watch shows unapproved provisional data while NGWMN was mandated to only run statistics on reviewed and approved data. The gw data review process is different than stream gages, so worth having a conversation about
• Clarify data in response to comment: 3 years of data could be an issue, but probably ok. A site could have three years of discrete measurements, which may only be 5 or 6 total measurements over 3 years and then become instrumented. This would skew statistics to compare discrete with continuous, you may want to consider a minimum number of observations as another threshold. (this is part of the motivation to develop the Best Available Time Series for GW data)
• language about percentiles is specific to stream gages, use more general language
• Comparing the daily value to the entire period of record does not account for seasonality. Many is not most sites show seasonality that can be caused by recharge, nearby pumping, and other conditions (note: NGWMN statistics only compare observations to monthly statistics to account for this GW oddity). Comparing a December measurement to a July measurement is really an apples to oranges comparison, which is why comparing an observation to a monthly statistic may be a better approach for groundwater and will truly show major issues, such as drought, in a more apparent and less alarmist manner.
• horizontal mobile view jumbles the month names on the line chart

Nice-to-haves

• Take-home message is unclear. What is the story about 2021 groundwater that users should walk away with? How have conditions changed during the 2021 year? From watching, the reviewer sees how conditions change in time and space, but wants more to help digest this information. What happened on the east coast in March and April? Likes the narrative provided by river conditions.
• Make time more prominent. Line chart does not show simultaneously with map on all screen sizes. Even when both are in view, it's hard to visually track between the two.
• Ability to toggle line chart to control playback would be nice (yes it would)
• line chart y-axis label looks like a subtitle and may be confusing
• Parameter codes are confusing. Consider linking to help.waterdata in that part of text.
• state explicitly that the data pulled from NWIS was groundwater, i.e. "...to pull NWIS groundwater level data between..."
• Clarify historical data time period

Other feedback:

• The map graphic is very nice. The explanations are worded
• The Data Processing section indicates that historical data were used through 2020. When will that update include 2021?
• Great work on catching some p-codes needing to cause inversion. Thanks for making it easy to check through code and data, this was fantastic. (shoutout @lindsayplatt )
• True groundwater gurus may take issue to the symbol being a peak or triangle which has been the classic stream gage symbol, where wells are circles
• Consider doing for the water year

[cnell-usgs]

@lindsayplatt - please read through the comments about the data processing and consider any changes we should make to the Data Processing text section to clarify

[lindsayplatt]

• We are only getting continuous data (either DV or UV), so I think renaming to USGS Groundwater conditions at continuous wells makes sense, though we could add the word "continuous" in our data description below. I am currently trying to understand the comment about "might be missing most of the data", which implies the groundwater measurement data makes up most of the GW data. These are manual measurements and from what I can tell are completed maybe monthly.
• Assumptions from the reviewer are true, expect that I don't really understand how not required to be real-time and continuous are not the same? And have data available for each day of the entire calendar year is more specific than we checked. We just did a basic count, which needed to be greater than 365*3. Maybe that check needs to be most sophisticated?
• For reviewed data, do we know how much that may remove?
• Concern about discrete measurements - we are not using any discrete measurements, so compare discrete with continuous is not an action that is happening.
• Language about percentiles I assume is referring to the linked page, which I agree we should update
• comparing the daily value to the entire period of record does not account for seasonality - correct. It would show the different seasons throughout the year. That is what we do for river conditions also. I kind of wonder if we want to show these oddities? recharge, nearby pumping, and other conditions? Or do we want to switch to a comparison of each day to a monthly statistic? I think we may want a rolling 30 day statistic or something because if not we may get weird cutoffs when months switch.
• The Data Processing section indicates that historical data were used through 2020. When will that update include 2021? we did not intend to do that, but maybe we should and then we can freeze the historic data at Dec 31, 2021.

[cnell-usgs]

I struggled to format the date ticker nicely. So, I added it, but it's commented out right now. If Joe deems this as necessary to publish, may need to bring in Hayley to refine that. I think it is very busy to see all the numbers churning like that though, and it doesn't add a lot of value, although I acknowledge the view is difficult to see the dates simultaneously with the map, and I am also not satisfied with that.