SWMM 5 questions

[pbix]

I'm new to SWMM and am in the process of learning the ins and outs of the software. I feel that I have a pretty good understanding of the software but question the output that I am getting from a test project that I ran. Specifically, I compared the calculated peak runoff rates from SWMM with the runoff caclulated from TR-55 for an 80 AC site we are working on. The site is mostly rural with 10% being impervious. SWMM calculates a peak runoff of 215 cfs where TR-55 calcs a runoff of 128 cfs. The percent difference is even large when looking at the developed conditions (66% impervious). 391 cfs from SWMM versus 206 from TR-55. What could be causing such a large difference in the two methods?

One question that I have is what format should the subcatchment slope be entered in ft/ft or % (i.e. one percent slope entered as 0.01 or 1)? There seems to be inconsistency in documentation for EPA SWMM and some of the documentation and examples from commercial products like StormNet.

I can attach data files if that would help. Thanks.

 

[tsgrue]

.

I strongly suggest you read through the earlier thread on this forum; “Volume of Runoff Rational Method Formula”, including the link to “Comparison of Nine Uncalibrated Runoff Models to Observed Flows in Two Small Urban Watersheds” (

http://smig.usgs.gov/SMIG/features_0398/runoff.html).

SWMM5 indicates the units of all input variables in the text region at the bottom of entry boxes. You can also click on the 'help' button and context-sensitive help will pop-up.

In general, SWMM has the ability to produce more accurate results than TR-55. The TR-55 methodology/model tends to overestimate flows and such are almost always larger than those from SWMM for a given watershed with proper inputs.

Some folks starting out with SWMM may have some troubles with the watershed width parameter, which is very important. Try checking that out.

.


tsgrue: site engineering, stormwater
management, landscape design, ecosystem
rehabilitation, mathematical simulation

http://hhwq.blogspot.com

 

[pbix]

Quote: "The TR-55 methodology/model tends to overestimate flows and such are almost always larger than those from SWMM for a given watershed with proper inputs"

This is precisely why I question the results. I read the articles that you referenced and understand that no two models will produce the same runoff but was expecting that the results would be closer than they are.

You're right the subcatchment width parameter was a little hard to understand at first. The watershed in question is a fairly symetrical, primarily rural watershed. The width I am using is 2 times the channelized flow length. Do any of the following paramaters seem way out of whack?

N-Imperv: 0.016
N-Perv: 0.17
Dstore-Imperv: 0.05
Dstore-Perv: 0.10
%Zero-Imperv: 25
Infiltration: Green-Ampt (Silty Clay Loam soil)
Suction Head: 10.63
Conductivity: 0.20
Initial Deficit: 0.132
10 year 24hr SCS type II storm

The reason I questioned the slope input is that when I first started learning the software I demo'd StormNet. The tutorials and examples that were included showed the slopes as ft/ft values (even though the text by the input bux said in should be %). This raised a question in my mind as to what the correct format is. I searched several SWMM forums to resolve the question and eventually came across a response that indicated that there was slope bug in one of the beta releases of SWMM5. I read the software changelogs and didn't find any mention of the bug or an indication if it had been fixed.

 

[tsgrue]

.

What is the subcatchment:

Area (80 ac?)
Width
% Slope (1.0?)
Subarea Routing Type
Percent Routed

Go ahead an attach/upload a file.

Your Dstore-Perv and N-Perv might be low, but I'd need to know more about the landuse/landcover to have much of an opinion.

Also, what TR-55 parameters are you using - especially the CN?

.



tsgrue: site engineering, stormwater
management, landscape design, ecosystem
rehabilitation, mathematical simulation

http://hhwq.blogspot.com

 

[pbix]

SWMM file attached.

TR-55 Info:

TC = .555 Hrs

SCS Type II Storm (4.3 in)

Land use:

Commercial & business B 6.9 92
Residential districts (1/3 acre) B 4.3 72
Residential districts (1/2 acre) B 3.3 70
Row Crop SR + Crop residue good) B 65.1 75
Total Area / Weighted Curve Number 79.6 76

 

[tsgrue]

.

I looked over the SWMM file you uploaded. Everything looked okay from the inputs and general information I have. I think the SWMM to TR-55 difference is mainly due to the CN values selection, with the watershed width, slope, and pervious storage values having some affect as well. I would also think that in a watershed with that small amount of impervious area that at least half would route from impervious to pervious, but I don't know that without more information. (Is the impervious cover directed to inlets and subsurface conveyances/conduits?)

If you have a silty clay loam soil as described, the Hydrologic Soil Group (HSG) would almost certainly not be B. It would be a C or D Hydrologic Soil Group, and probably D if it is cultivated (row crops). (That must be some sticky wet mess after a rain!) Using your information, at an HSG of C, the CN would be 83 and at an HSG of D, the CN would be 86. Try checking out those and variations of the other values.

Is average depressional storage in pervious areas really only 0.10 inch? With the SCS methodology, you are probably assuming an initial abstraction of 0.20, which would be about 0.63 inch at a CN of 76 (about 6x as much as you estimate for pervious cover). Check out "

http://frontpage.et.byu.edu/ce531/Papers/CNInitialAbstraction.pdf"

for a discussion of initial abstraction at 0.20 versus 0.05. Note the statement; "Altering [Ia/S] requires a change of handbook CNs. That is, if [Ia/S] is changed, a different CN must be used." (page 8). I think your SWMM input value is low and the SCS Ia of 0.2(S) is high.

As this is a fairly extreme event (10 year) in a dominantly agricultural setting (and rather flat), the SCS methodology should perform at it's best - that is the type of scenario for which it was largely developed. SWMM still has the ability to perform better and should do so.

.


tsgrue: site engineering, stormwater
management, landscape design, ecosystem
rehabilitation, mathematical simulation

http://hhwq.blogspot.com

 

[onyeuko]

I did not mean to sing someone's praises but it is hard for me to hide my feeling. After reading the reference in the lead-

http://smig.usgs.gov/SMIG/features_0398/runoff.html

posted by tsgrue, I felt that this article is whort everyone's reading as to be fully informed and equiped to the fact that it is not always what you see that you get (or that is right). I am saying this in the sense that we have to apply caution in the interpretation and the application of results from models stimulations, using the soft wares. And this information is available to some of us only through the meticulous contribution of the likes of tsgrue through the forum. Please keep the flag flying
Teddy

 

[onyeuko]

tsgrue writes:-
"Some folks starting out with SWMM may have some troubles with the watershed width parameter, which is very important. Try checking that out".
Could you please throw more light on this?
Teddy

 

[tsgrue]

.

The EPA SWMM5 documentation does not cover SWMM computations in the depth of previous EPA SWMM documentation (at least from what I have seen).

The EPA SWMM5 documentation presents the subcatchment 'width' parameter as: "Characteristic width of the overland flow path for sheet flow runoff (feet or meters). An initial estimate of the characteristic width is given by the subcatchment area divided by the average maximum overland flow length. The maximum overland flow length is the length of the flow path from the inlet to the furthest drainage point of the subcatchment. Maximum lengths from several different possible flow paths should be averaged. These paths should reflect slow flow, such as over pervious surfaces, more than rapid flow over pavement, for example. Adjustments should be made to the width parameter to produce good fits to measured runoff hydrographs."

For a better treatment, I suggest you read pages 87 to 98 of "Section 4: Runoff Block" from the EPA SWMM4 manual (EPA/600/3-88/001a). This can be downloaded from "

http://eng.odu.edu/cee/resources/model/swmm.shtml".

The subcatchment 'width' parameter in SWMM5 is incorporated in the same conceptual manner as in SWMM4.

.


tsgrue: site engineering, stormwater
management, landscape design, ecosystem
rehabilitation, mathematical simulation

http://hhwq.blogspot.com