Multiple Shape Factors for one model 1

[JANeal]

I have been perplexed by this one. We are working on a project where there is varied topography. Much of the overland flow is on relatively flat slopes, but other portions of the same basin are on steeper terrain. We had a discussion about SCS shape factors, and whether it would be appropriate to assign 484 to the hilly portions, and 323 to the flat portions. Drainage area is approximately 50 square miles.

My immediate hesitation to use multiple shape factors is that while I have found nothing that says you can't, it sounds like a single shape factor was used to develop the equations. The shifts in the distribution of runoff significantly changes the end result. Does anyone have experience or thoughts on this?

 

[Crystalct1]

In my experience, the SCS methodology calls for a shape factor of 484 and that is the one to use if you don't have any good compelling arguments for another factor. I do work in Delaware and they have accepted the Delmarva Unit Hydrograph as the one that best fits certain counties and its shape factor is 284. Other than being specifically told by an agency that another factor is applicable, I'd go with the most conservative, the tried and true, the SCS 484.

 

[cvg]

if you are not required to use SCS methods then you could use another procedure that does not have this particular limitation. For instance, Clark Unit hydrograph with Green and Ampt loss coefficients; or a dimensionless unit hydrograph (S-graph) as commonly used by the USACE and the BUREC. Either way, you should probably divide the watershed into sub-watersheds that are smaller and more uniform in slope, cover, development etc. and route the flows from each one to your concentration point.

 

[JANeal]

Thank you guys for the suggestions!

Crystal - I actually just moved to FL from North GA, and I always had stuck to the tried and true 484. Down here, however, it seems 323 is the number expected to be used based on the flat terrain (originally formally used/required by the St. Johns water management district or FDOT, not sure exactly which.)

Cvg - that's a good idea to look into other methods that do not have that limitation. It is divided into sub-watersheds, and Times of Concentration are developed for each, which partially accounts for different slopes/etc. It was during the development of the model that someone asked whether or not shape factors should vary with the different subwatersheds. As an extreme example, consider a valley next to a mountain - the valley is flat, possibly with several depressions, and the mountain is extremely steep terrain. Both have very different basin geometries and flow patterns. SCS doesn't really call this out as a limitation, but I think it's a great idea to look into other methods.

 

[RWF7437]

The following questions are intended to augment your thinking about this problem.

1. Why are you doing this ? Is it to size a culvert or similar conveyance device ? If so, then you probably care most about the peak flow, and less about the volume of runoff.

If it is to size a detention pond, then you probably care most about volume and less about peak flow.

If it is neither of these, then you are calculating in order to be calculating; which is probably not what your employer is paying you for, unless you are a mathematician.

2. Do you have any choice about whether or not to use an uncalibrated hydrology model ? If you can use a calibrated model you can probably reduce your error of estimate to about +/- 30%. Otherwise, your error will be on the order of +/- 50% to +/- 100%, or even more.

3. In this 50 sq mi basin are there any rain or stream gages ? If there are, you have an opportunity to calibrate your chosen model. Indeed, someone may already have done this. Have you checked to see if they have ? Are there regional regression equations developed for this basin ?

4. How good is the rainfall data you are using ?

5. Have you tried calculating your hydrographs in several ways to see how sensitive they are to "shape factor"?

see FAQ 194-1462

good luck

 

[RWF7437]

http://www.forester.net/sw_0607_voodoo.html

 

[RWF7437]

http://www.stormwaterauthority.org/assets/3alternateapproach.pdf

 

[Surcharged]

I have always heard that SCS method's upper limit is 2000 acres. 50 square miles is greater than 30,000 acres. I would suggest looking at regression equations developed for your area.