Reservoir outflow simulation

[Kelly S]

Could someone talk to me about the choosing which simulation to use for the IRESCO variable? I have 7 reservoirs in my watershed and there are 4-5 different owners (most are private companies - only one is managed by USGS/Army Corps of Engineers). I'm not sure what data I will be able to obtain, so each reservoir might be handled differently.

Part of my problem is not completely understanding the SWAT 2012 I/O Documentation.  The options for IRESCO - outflow simulation code are

0  compute outflow for uncontrolled reservoir with avg annual release rate

1  measured monthly outflow

2  simulated controlled outflow - target release

3  measured daily outflow

I think I understand the measured monthly and daily outflow and would use the resmonthly.dbf or resdaily.dbf files for those simulations.  What are the other two (0,2) all about?  Is there a reason to use one over the other? 

Thanks so much for your help!

Kelly

Masters Student, Natural Resources/GIS 

North Carolina State University

[charlesi @iastate.edu]

Kelly,

Here is my understanding, but hopefully some others will chime in and perhaps add some more specifics.

Options 1 and 3 require measured outflow as an input parameter.  I've never used this, but I assume that the simulated reservoir stage/volume/area would be based on simulated inflow from the watershed and the measured outflow input to the reservoir.  The problem I see with this is that unless you can calibrate flow into the reservoir or reservoir stage, you have no idea how your model is performing upstream of the reservoir.  I think these options would be useful in certain, limited situations.  Like I said, if you have measured data upstream of the reservoir and out of the reservoir, these are valid.

Option 0 is quite simplistic.  The user simply inputs an average annual release rate.  I believe the daily flow out of the reservoir is simply calculated as the volume of water that exceeds normal pool (if volume is greater than normal pool but less than emergency pool) or the average release rate (if volume exceeds emergency pool).  This is fairly well documented in the 2009 Theoretical Documentation.

Option 2 allows the user to specify a target release rate.  This is useful if you know how the reservoir is operated and/or have historical records.  Again, the calculations are well documented.

Unfortunately, there is not currently a stage-storage-discharge type of calculation for reservoirs in SWAT, but I am working with the folks in Temple, Texas to incorporate and test that.  I think that will be appealing to a number of users.

--Charles

--
You received this message because you are subscribed to the Google Groups "ArcSWAT" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arcswat+u...@googlegroups.com.
To post to this group, send email to arc...@googlegroups.com.
Visit this group at https://groups.google.com/group/arcswat.
For more options, visit https://groups.google.com/d/optout.

[Kelly S]

Charles,

Thanks so much.  I guess for me a lot will depend on what kind of data I am able to obtain.  In looking around the user group to find out more about reservoir modeling, I noticed a question about consumptive water removal from reservoirs.  Since many reservoirs provide water for irrigation or public water supply in addition to hydroelectric power production, is it necessary to determine this variable (WURESN) or is the amount not large enough to be important?  Looks like an optional parameter.  Which option have you worked with?

Kelly

[Jim Almendinger]

This remains a confusing topic...

Short answer:

For most natural lakes and minimally controlled reservoirs, choose IRESCO = 2, make your entire year a "flood" season, and set all 12 monthly target volumes (STARG values) to the principal volume (the volume in the reservoir when just at the outlet threshold).  Make the emergency volume substantially larger than this -- I think I've used 1.5 to 2x the principal volume.  You may have an actual emergency volume if you've got good specifications for your reservoir.  This gives the reservoir some "room" to allow gradual drawdown, which occurs only between normal and emergency volumes.  All volume above emergency gets spilled that same day.  And, start with NDTARGR = 3 or so.  I've used 2, and up to 10, but 3 seems to be pretty common in my watersheds.  Drawdown occurs as a geometric series of (1/NDTARGR)^days.  Even though some of my reservoirs have been "controlled," I still have the best results with IRESCO=2. 

Longer answer:

For the same reasons that Charles gives, I've never used options 1 or 3.  Option 0 aims to have a fairly regular outflow (perhaps as might be wished for power generation, or milling?), except that all water above the emergency volume spills that day. 

I've only used option 2, which seems to work best for more-or-less natural lakes and reservoirs.  The mind-set here is apparently for reservoirs that are used for water supply, and so they target to keep as much water in storage as possible (up to emergency volumes), EXCEPT during flood-prone seasons, when the water is lowered to allow for some "bounce" and storage in the reservoir to mitigate runoff peaks.  For only the "flood" months, SWAT assumes a "target volume" (not release rate), and calculates outflow based on how much above the target volume the simulated volume is for that day.  There is a formula that estimates soil moisture, with wet soils being more prone to cause floods and so the reservoir target volume is lowered to provide more storage for safety.  Somewhere I suppose reservoirs may be operated in this way (perhaps in the more arid west), but not in my watersheds.  I just subvert the soil-moisture formula by specifying monthly target volumes, which take priority. 

   It would be an improvement and simplification in SWAT to just allow the user to chose the principal volume as the target volume as one of the options -- I would make that the default, with a default NDTARGR=3.  It might help to allow the user to input a stage-discharge curve instead of specifying NDTARGR, but the NDTARGR method actually seems to work pretty well (although I think it should be a real number, rather than an integer).  And I think Ponds and Wetlands ought to have the same simplification (with target volume = normal volume, not dependent on soil moisture, since these small features are not commonly manipulated for water supply). 

More than you asked for -- and when you learn more, let us know! 

Cheers,

-- Jim

---

From: "charlesi @iastate.edu" <char...@iastate.edu>
To: "Kelly" <kmsu...@gmail.com>
Cc: "ArcSWAT" <arc...@googlegroups.com>
Sent: Thursday, January 21, 2016 12:01:58 PM
Subject: Re: [ArcSWAT:7846] Reservoir outflow simulation

---

NOTE: This message was trained as non-spam. If this is wrong, please correct the training as soon as possible.
Spam
Not spam
Forget previous vote

--

Dr. James E. Almendinger
St. Croix Watershed Research Station
Science Museum of Minnesota
16910 152nd St N
Marine on St. Croix, MN  55047
tel: 651-433-5953 ext 19