TR-55 Sheet Flow limit 1

[Markie55]

We have been using sheet flow of various lengths up to the maximum length of 300' per TR-55 for various projects. During a recent review, a question came up regarding the use of the 150-200' sheet flow lengths and the reviewer notes that the NRCS now limits the 1ength to 100'. We use HydroCad which allows up to 300 feet for sheet flow computations. We are not aware of any local agency or any other reviewer that uses the 100 foot limit for sheet flow. The state highway drainage manual also uses 300 foot as the limit for sheet flow. Thoughts?

 

[TerryScan]

From Part 630 Hydrology National Engineering Handbook, Chapter 15:
"Typically, sheet flow occurs for no more than 100 feet before transitioning to shallow
concentrated flow (Merkel 2001)."
and
"Kibler and Aron (1982) and others indicated the maximum sheet flow length is less than 100 feet. To support the sheet flow limit of 100 feet, Merkel (2001) reviewed a number of technical papers on sheet flow. McCuen and Spiess (1995) indicated that use of flow length as the limiting variable in the equation 15–8 could lead to less accurate designs, and proposed that the limitation should instead be based on:
L = 100 * S^.5/n (eq. 15–9) (length, slope, mannings n)
Table 15-2 Maximum sheet flow lengths using the McCuen-Spiess limitation criterion
Cover type, n values,Slope(ft/ft),Length(ft)
Range ,0.13, 0.01, 77
Grass ,0.41, 0.01, 24
Woods ,0.80, 0.01, 12.5
Range ,0.13, 0.05, 172
Grass ,0.41, 0.05, 55
Woods ,0.80, 0.05, 28"

Some time ago regulators in my state performed informal field observations during storm events and have stated that the 300 ft. length is only achieved in ideal conditions on smooth surfaces (perfectly paved parking lot.)

 

[psmart]

Although HydroCAD implements the original TR-55 limit of 300 feet, many agencies apply a lower limit, as noted in TerryScan's post.

For further details see "References on time of concentration with respect to sheet flow" published by NRCS. A PDF version is available at

http://www.hydrocad.net/pdf/SheetFlowRefs.pdf

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[beej67]

Since sheet flow length dominates the Tc equation in TR-55, and hydrologists often have no honest idea what the sheet flow length is for large watersheds anyway, I often times use Kirpich or another method and then back my way into a sheet flow length assumption that reasonably matches the Tc developed by another, more consistent method.

I honestly hate the TR-55 method for Tc. If you've got a 10 acre watershed and one corner of it has 100 feet of sheet flow through dense forest, but the rest of it is paved, do you really think that little forested corner is having that big of an impact on your runoff hydrograph? Obviously not. And if that's the case, then you're making a judgment call to abandon the method situationally based purely on your own judgment. And if that's the case, then what good is the method? Why not just use your own judgment for the Tc on every basin?

I much prefer other methods that use mean basin slope and land cover.

Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[psmart]

We're seeing a lot of engineers using the curve number method (aka Lag Method) for estimating the Tc. In addition to eliminating the subjective determination of sheet flow length , the CN method uses the average land slope and CN over the entire subcachment, so it's sensitive to the entire ground area. Many agencies appreciate the advantage of a more objective procedure which eliminates the endless wrangling over sheet flow length.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[Rye1]

Very nice question. Hydrology is a little bit of voodoo anyway. Typically, we allow sheet flow to about 300' in woods and fields. In urban areas we are at 150' at the most. If we are looking at a small drainage problem, if the sheet flow goes between two homes, it becomes shallow concentrated regardless of length.

Robert Billings

http://www.newrivereng.com

 

[TerryScan]

Robert,

It seems odd that one allows a longer sheet flow in rougher land covers. Surface irregularities tend to cause concentration. (As reflected in the NEH cited above).

 

[Crystalct1]

Always remember also, that the concept of sheet flow is violated for flow depths greater than 1" or 0.1'. I tend to look at a square area in sheet flow and use manning's equation on the flow to check that its depth has not gone too deep. For instance, take a square cross-section that is one foot wide flowing downslope. Its width is 1 and its depth is D, so its area is D. Its hydraulic radius is A/1 = A. Use the appropriate slope and manning's n for sheet flow on your surface and solve it for D, using a flow value from either the rational method or NRCS. Keep D under 0.1 foot. It is an iterative procedure, but makes more sense than just randomly choosing a maximum sheet flow length.

Yes, in NJ, the DEP and DOT mandate a maximum of 100 sheet flow length. This is bunk! This is just a way to make sure the calculations are conservative enough. We should be able to do better than that.

 

[LincolnPE]

"If you've got a 10 acre watershed and one corner of it has 100 feet of sheet flow through dense forest, but the rest of it is paved, do you really think that little forested corner is having that big of an impact on your runoff hydrograph?"

The paved area should be analyzed separately, then analyze the total composite area, then use the highest peak flow of the two. There are two conditions to evaluate, with two separate Tc's.

 

[beej67]

There are practical limits to how many subwatersheds you can divide your study into, Lincoln. It's not uncommon for an urbanized watershed to be 30% impervious parking and roads, 40% open space, and 30% forest, on varying topography. Where do you draw your Tc path? If you draw it to the farthest point, chances are that's a parking lot since historical roads typically follow ridges, and Wal Marts tend to go next to roads. If you draw it to the "most hydraulically remote point" then you're picking some spot in the woods somewhere half way up the watershed, based on an assumption of ground cover and slope, even though you don't have good survey data of that spot. You can't divide 30 acres of urbanized watershed into 20 subwatersheds because nobody will pay you to do so, and by rights you shouldn't have to if you can get your parameters right.

I vastly prefer the CN method Peter references, and if not that, then some other direct methods such as Kirpich, the "USBR Design of Small Dams" (1973) equation, FAA, Izzard, etc. Any method where an assumption dominates a parameter is a bad method, and one cannot possibly know every sheet flow condition for every 100 foot stretch in a large watershed.

But as long as the reviewers demand it, I'll keep using it. I just check it internally vs other methods to make sure I believe what the method is telling me.

Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[beej67]

Also, for the record, you absolutely should not be breaking a watershed into two watersheds and then throwing one out because the other is higher. I don't know if that's what you intended to say, Lincoln, but that's very wrong. (Perhaps I just misread that?)

When characterizing watersheds, you either model the whole area with a composite CN, or you break the area up and model each sub area with CNs characteristic of that area, then lag and sum the hydrographs of both appropriately to create a composite hydrograph that includes the flow from both sub areas.

Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[LincolnPE]

"When characterizing watersheds, you either model the whole area with a composite CN"

This is incorrect. I think this is why you have confusion when thinking about what is the correct Tc for a catchment with a pervious area draining over an impervious area to the common catchment outfall.

If you were designing an inlet or culvert and you did not do this test, and just did a composite calculation for the catchment, there is a chance that the smaller downstream impervious could have a higher peak flow. You would miss that if you didn't do this test, and you would not have the max peak to correctly size your stormwater appurtenance.

 

[LincolnPE]

Yes to confirm your question about the misreading, you do throw out the scenario that generates the lower peak.

 

[beej67]

In your case Lincoln, where a smaller downstream area has a higher peak, you DO NOT NOT NOT throw the hydrographic from the larger area out. You sum the hydrographs to generate a composite outflow hydrograph.

Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com

 

[LincolnPE]

Yes, you are correct considering the hydrograph methods. The test method I mention is valid only for the Rational Method, a linear method.

One thing to be aware of regarding the NRCS method in this directly connected impervious scenario, with a larger pervious area over a smaller impervious area, is to not use a weighted average CN number. This is due to the computation of Ia using the weighted average CN.

This produces an erroneously high average initial abstraction value for the composite basin, that includes the impervious area. The modeler will not see how, at times, the smaller impervious area is dominating. In certain previous/impervious basin cover ratios virtually all of the runoff is coming from the smaller impervious area.

 

[beej67]

Ahh I see the confusion. Yes, I agree with what you're saying for a rational method, peak flow only analysis.

Hydrology, Drainage Analysis, Flood Studies, and Complex Stormwater Litigation for Atlanta and the South East -

http://www.campbellcivil.com