HEC-RAS Overtopping of Banks 4

[Trackfiend]

I have modeled a channel within HEC-RAS and am currently trying to calibrate said model with given historical data. I have been given both upstream and downstream elevations and various flows over a certain time period. I've inputed the downstream w.s. elevation and am trying to match (within a certain tolerance) the computed upstream w.s. elevation to the given historical data.

The main issue that I am dealing with is for any high flow situation. I am getting unrealistic upstream w.s. elevations due to HEC-RAS vertically extending the LOB and ROB in order to contain all of the flow. I am wondering what would be the best way to model overtopping of my banks? I see that there is a lateral weir option (with split flow optimization). Would adding a lateral weir (where the low points in the channel banks are located) along with a parallel channel or storage area in order to "store" any excess water that tops the banks? Any water that tops the banks merely flows into surrounding marsh and does not reenter the channel at any point.

Any suggestions?

 

[RWF7437]

First, extend your cross sections to at least the full width of the flood plain. If you can't do this from available mapping, then get additional field surveyed data.

You say, "Any water that tops the banks merely flows into surrounding marsh and does not reenter the channel at any point."

Unless and until you have complete and accurate cross sections you cannot know if this is a true statement, or not.

You also have the complication that the water level in the marsh may, or may not, be known for the time period of the simulation.

good luck

 

[Trackfiend]

Extending the cross sections may not be the best alternative. I have nearly 250 sections, each of which are georeferenced through a csv file. Adding any additional elevations gives errors in the way of "non-georeferenced items" warnings. Albeit I am a fairly inexperienced HEC-RAS user and can't answer if this would cause issues with trying to solve the model. Also getting additional survey data at this point would not be possible.

We are making the assumption based on topography and existing surveys that it is unlikely that any water that tops the banks will not reintroduce itself downstream.
You are correct that estimating the water level in the marsh is complicated in that we are also merely assuming that the downstream water surface elevation is equal to the marsh water surface elevation (downstream w.s. is a bay that borders the marsh).

I guess what I would like to accomplish is in determining the lost flows from overtopping so that I can input the new reduced flow changes in the channel. I basically want to get rid of any overtopping flows in order to accurately calculate what the upstream w.s. elevation would be.

I do appreciate the input. Thanks.

 

[RWF7437]

"Also getting additional survey data at this point would not be possible."

Difficult, expensive, time consuming but IMPOSSIBLE ? Hardly likely.

The "lost" flows are not lost at all in HEC-RAS. The program accounts for all the mass. It does so by including the volume of water stored in ineffective flow areas as well as the water that actually flows downstream. Your marshes may be ineffective flow areas or they may not be. You're making many assumptions here and need to be reasonably sure they are justified. You need to model them realistically but that doesn't mean you have to know their shape to within 0.01 feet on all directions. Any approximation you can devise which is within a few feet in any direction will probably give you a reasonable result.

 

[sammyk]

I agree with RWF7437, additional survey is required.
However you may have enough data already available to you.
Take a look at the floodplain maps and USGS 10' contour data and extend out your cross sections farther out.
Your current model cannot handle the given historical flows because your xsections do not adequately model the system.

Now with the additional x-section data from the floodplain (and beyond if necessary), you can "tweak" your cross sections until you feel there is good correlation between the model and historical data (ie there is enough "ineffective" area to account for a proper US WSE)

Since you are using some "Ass"umptions on this model, please make sure that you document what you have done somewhere, so that the next person that uses/looks at your model understands what you did.

 

[francesca]

It sounds like you need to spend some quality time with the Hydraulic Reference Manual.

 

[Trackfiend]

Quality time...both the User, Applications, and Reference manual seem to be permanent fixtures on my desk although maybe I haven't been spending as much time as I would actually like with them. Too bad SWMM doesn't have three nice novel sized manuals.

I agree that "ass"umptions aren't the best way to approach a model with so many variables, but the client and upper management both have stated to accept these "ass"umptions for this particular project. It would be nice if we had both the time and funds available to model out completely whole areas in order to calibrate these models to natural and realistic tolerances. In this case however, I agree that is not impossible to obtain additional data but it is highly unprobable. Floodmaps and the USGS contour data would be good place to look though, thanks.

I will try to add additional points within the cross-sections containing overtopping but the issue still remains that all of my survey data was imported via a csv file with both Northing, Easting, and elevations. When adding new points, I get errors stating non-georeferenced points and the geometry file doesn't locate them properly (due to the Northing and Easting). Maybe creating new points in the original csv file could help....

Nobody has mentioned the lateral weir option. Couldn't this situation be treated with this option using an arbitrary storage area large enough to contain what does top the banks?

 

[RWF7437]

If you are doing this study for FEMA, or if it will be reviewed by FEMA, you have no choice but to do the study according to FEMA's guidelines. This is true no matter what your Boss says to assume or what the project budget is.

Or, you can simply not do the study.

If you are only doing this study for your own edification, you will learn more by doing it as correctly as you know how.

The side weir option was not considered by me to be worth trying because:
A. I've never had any luck with it and,
B. It ASSUMES the water that overflows will NOT reenter the stream (as I understand this option ).
C. It doesn't sound, from what you've told us, that you know that for certain.

Are you using HEC-GEO RAS? If so, I can't help much because I have no experience with it.

But, once you have entered cross section data in HEC-RAS it is fairly simple, though tedious, to edit and extend it.

The Hydraulics Manual is indispensable.

HEC-RAS does three basic operations:
1. It solves the Conservation of Energy Equations
2. It solves the Conservation of Momentum Equations
3. It solves the Conservation of Mass Equations

It does this iteratively and quickly but it cannot make up the (approximately) correct data for you. That you must supply or the results are worst than worthless, they are meaningless and possibly misleading.

good luck

 

[Trackfiend]

No, this project is not for FEMA but for a local government entity. The guidelines that I am following are in strict accordance to what they have given me. I do agree 100% with the assessment that the cross-sections could be extended, however in this particular case I was wondering if there was another way to go about modeling this case without having to input additional cross-section data due to manner in which the original data was imported (csv file). Without modifying the orginal imported file, I get numerous errors due to the data being two different types. The geometry file also misplaces the "new" cross-section data due to inconsitencies with the Northing and Easting points. Version 4.0 I believe is the first version to offer this import option (csv file).

You are absolutely correct though, Garbage In = Garbage Out.

I'll keep hammering away at it. I guess I just need more quality time with the program. Thanks for all of the replies. This forum is extremely helpful in getting your peers to help you look at a problem in a different view.

 

[francesca]

When I started out using HEC-RAS I quickly found out that the user manual is useless as the software is very intuitive. I read at least half of the Hydraulic Reference Manual, though, and got a good understanding of how the model does its calculations and that let me know where I could make trivial assumptions and where some assumptions would lead to garbage.

I believe HEC-RAS has an option to model your lateral storage in the wetland. You need to include the geometry and set the elevation at which flow will spill over. You then need to set an ineffective flow area equal to the top of the wetland. (i.e. before the wetland starts to spill over and there is meaningful flow parallel to the stream.) I can't remember what this option is called - maybe a levee station? - but as you said, you have the books on your desk.

 

[francesca]

Yes, look at this:

http://www.bossintl.com/online_help/hec-ras/source/leveeovertoppingandbreaching1.htm

 

[Crystalct1]

I modeled the same type of stream in HEC-RAS. I did use the lateral weir with split flow optimization. This takes out the flow that overtops the banks and yes it can just "lose" it. You also may let the overtopped flow get reintroduced into the system downstream or either let it go into a pond.

This seems like the way to go. My stream was a tributary to a much larger body of water and the overtopped flow went into the other body of water, so for our analysis, it actually was "lost".

 

[Trackfiend]

I ended up using the lateral weir option with storage areas. The only reason I included the storage areas was due to having to run an unsteady analysis (for the steady, modeling "out of system" was more than enough). Split flow optimization was also untilized. The results came out fairly comparable (within certian tolerances) to a spreadsheet with similar characteristics.

I was also able to extend certain cross-sections using the cross-section graphic editor. As opposed to my previous thinking, it did not cause errors, and was an easy way to model the extending floodplain. Thanks for advice. It took me a little while to find the right option.

Slowly but surely I'm (hopefully) getting the hang of HEC-RAS. You are right, the software is pretty intuitive. Thanks for all the replies.