Baseflow Diagrams from Model Runs

[glenncampagna]

Baseflow Analysis

We are interested in creating 4 charts:

1. Active groundwater storage - specifically for every hour or every day
2. A stacked graph with active groundwater (AGWS), upper zone (UZS), and lower zone storage(LZS)
3. A plot of baseflow OUT of AGWO, that somehow indicates summer. We can do it as July-September only, or do average by month. Be creative: the goal is to see how the model thinks baseflow influences the water in the stream during dry periods, and the most important dry periods in Virginia are between July-October or November.
4. A plot of summer base-flow, and total flow into river reaches.
   • For this we need to refer to the river h5 file. Does the model store values for baseflow (AGWO), interflow (IFWO) and runoff(SURO) into the river segment, or does it just have a total inflow volume from all sources?
   • Maybe there is something other than the HYDR section of outputs?

These will be created using 'land' feature .h5 files, and they are stored in the directory: /media/model/p6/out/land/hsp2_2022/eos/

Naming Convention: Graph 1: [h5prefix]_AGWSgraph Graph 2: [h5prefix]_GWSgraph

[glenncampagna]

• 4 new graphs created using R for AGWS were uploaded to directory /media/model/p6/out/land/hsp2_2022/eos/ , also accessed at http://deq1.bse.vt.edu:81/p6/out/land/hsp2_2022/eos/
• Each graph represents chart type 1 from above but uses a different time interval for the AGWS values (hourly, daily, weekly, and monthly)
• R was used for easier data and time series manipulation compared to excel while we are determining the timestep we will use in the future

[rburghol]

Great stuff! These examples look like there may be a trend towards lower GW storage over time... interesting.

• Note: we absolutely want to continue using R - not excel :).
• Need: the land use specific prefix ("for" in this example) must be included in the image file name
• One request: paste the sample images in the issue body or an issue comment, with descriptions. That way anyone viewing the issue doesn't need to jump around to see the examples.

[juliabruneau]

AGWS Graphs

( by @megpritch and @glenncampagna )

Active groundwater storage represented in various timeframes: hourly, daily, weekly, and monthly. The average storage in inches was calculated in each timeframes from 19484 until 2020.

Example R Code:

[rburghol]

Thanks @juliabruneau ! These are thought provoking.

[juliabruneau]

GWS Graphs

The yearly active groundwater, lower zone, and upper zone storage is represented in the graph below from 1984 until 2020.

It was easy to create a stacked graph using Excel, but we'll create one using R as well and post the code in this comment.

[rburghol]

This is also very thought provoking. The tiny amount of AGWS is in stark contrast to both of the higher zones. I suppose it could be reasonable that since uzs and lzs represent water in soil pores and AGWS is in rock, that AGWS should be lesser, or at least more transient. But still... big difference (in this one single case study).

At minimum this tells me we need to take a quantitative measure, like mean or median storage during the simulation so we can do quantitative analysis.

[durelles]

This is also very thought provoking. The tiny amount of AGWS is in stark contrast to both of the higher zones. I suppose it could be reasonable that since uzs and lzs represent water in soil pores and AGWS is in rock, that AGWS should be lesser, or at least more transient. But still... big difference (in this one single case study).

At minimum this tells me we need to take a quantitative measure, like mean or median storage during the simulation so we can do quantitative analysis.

Remind me what the definitions are for UZS, LZS, and AGWS? What does the adj. active mean here?

• These three variables are measures of groundwater storage: UZS = upper zone storage [in] LZS = initial lower zone storage [in] AGWS = active groundwater storage at the start of the interval [in]
  • We believe active groundwater storage is "storage that directly controls streamflow dynamics during rainless periods"

[rburghol]

Hey team - I just added another wish to the list (in the top of the issue body) -- I am wondering if the h5 posesses enough info to allow us to determine the base flow (AGWO) into a river reach? That would be more efficient that getting the baseflow out of all 35 land uses, summing them up, then analyzing it for a reach.

[juliabruneau]

The pwater dataset includes AGWO (active groundwater outflow [in/interval]), where the max value seems to be only 0.005 in/hr. That being said there is data, but we are unsure if it is usable, or how we might go about analyzing groundwater flow further.

[rburghol]

@juliabruneau I think for starters, we visualize what IS there. Inches per interval tells us something that is summable (inches per day, month, year), and plottable.

I agree that 0.005 seems small, but, if this is an hourly value, then 240.005 = 0.12 inches per day, and that is real water, since 365 0.1 = 36.5 inches/year, which we now know is far more than flows out of our streams, since > 75% of the 40 inches/year of rainfall is evaporated away.

[juliabruneau]

@juliabruneau I think for starters, we visualize what IS there. Inches per interval tells us something that is summable (inches per day, month, year), and plottable.

I agree that 0.005 seems small, but, if this is an hourly value, then 240.005 = 0.12 inches per day, and that is real water, since 365 0.1 = 36.5 inches/year, which we now know is far more than flows out of our streams, since > 75% of the 40 inches/year of rainfall is evaporated away.

You are correct. Instead of using the average, for the in/interval unit I had to sum the daily values together in order to get the right values. Here is a graph I plotted just now, and I'm working on different time periods as well.

[megpritch]

For item 3 ("A plot of baseflow OUT of AGWO, that somehow indicates summer...") Glenn and I came up with this plot:

[rburghol]

@juliabruneau, @megpritch and @glenncampagna - I like these a lot! Couple things:

• How about we convert inches to cfs/sqmi ?? That will make it on equal units with some of our other runoff variables.
• I'd love to see the 2nd plot for the whole time period.
• Oooh!! What if we showed AGWS in Fall/Winter/Spring and AGWO for summer?
• Or both year round?

[glenncampagna]

@rburghol , here is my rough draft for plot type 2 for the whole time period I'm working on adding a legend, but the blue corresponds to LZS, the red to UZS, and the green to AGWS

[durelles]

agree with converting to per area. For the groundwater storage variable, is there a time component or just depth [inches]? We often think of depth per time (e.g., average rainfall for Virginia, 45 inches per year). With streamflow, we often use CFS [volume per time, where cf is cubic feet and s = seconds)/ When we want to compare to rainfall (input of water into the system), we'll normalize streamflow to the watershed area (thus, end up with volume per time divided by watershed area = length/time or depth per time, the same as rainfall).

@durelles -- note these are average storage(AGWS) not outflow (AGWO), so no time component other than that I believe they are showing hourly or daily storage here.

[glenncampagna]

Response to @durelles:

• For groundwater storage variables (UZS, LZS, and AGWS), the measurements are a depth (in) and do not contain a time component, but the active groundwater outflow (AGWO) is measured in inches/hr and can be converted to a flow per time if we have an area (possibly the area of the land segment? .. likely can be found in the .h5)
• We'd multiply by the area to get a flow (volume/time) and then divide by area to obtain a flow per area (cfs/ sq mi)

Questions for @rburghol:

• Regarding converting to flow/area, we think we need an area to accomplish this? Would it be the total area of the land segment?
• Regarding plot 4, would 'total flow into river reaches' be a sum of base flow (AGWO), runoff (SURO), and interflow (IFWO)? We believe this is represented by either PERO (total outflow from PLS (pervious land segment) or IVOL (sum of total inflows into RCHRES)?
• Also regarding plot 4, are you wanting 2 variables plotted on the same time series? The base flow (AGWO) and the 'total flow into river reaches' (maybe PERO or IVOL)

[glenncampagna]

Progress:

• We believe we've found values for l90_Runit and l90_agwo_Runit from A51800 l90_Runit = 0.009517 cfs/sq mi l90_agwo_Runit = 0.004750 cfs/sq mi

@rburghol when you tell us to 'export' these values, we think you mean adding them to VaHydro under the appropriate land segment feature and model scenario which we reviewed yesterday. However we are unsure of how exactly to use the script from yesterday to upload our values where we want them to go. We are referencing the diagram you provided us but can't quite figure out what all needs to be listed in the command (varkey, propname, prop value, etc.)

[glenncampagna]

See code in https://github.com/HARPgroup/HARParchive/blob/master/HARP-2022-2023/GraphingAndAnalysis/A51800_graphing.Rmd

Summary of 4 different plot types for A51800

1:

2:

3:

4:

Note: the average of the total outflow, which is the sum of runoff (SURO), interflow (IFWO), and baseflow (AGWO), was found to be 0.92 cfs/sq mi , which supports our conversion since the usual value of yearly river flow in VA ~ 1.0 cfs/sq mi

[rburghol]

Thanks @glenncampagna - we will review this process today at the meeting! BTW:

• Excellent chart of baseflow and total flow, and great to see that the total is near 1.0!
• The summer baseflow chart for the full period is definitely workable, not as clear as the 5 year that you all presented earlier, but tells the complete story and is clear enough IMO.
• I find it interesting that in this one case study the l90_agwo_Runit is approximately 50% of the total l90_Runit. It will be interesting to see how this looks over a larger number of segments and land uses.