Appropriateness of SWMM for designation of flows for small areas

A state environment agency wrote the following comment regarding SWMM to defend its use of a volumetric criteria (rainfall volume x area) for stormwater quality design:

"Finally, common hydrologic models (e.g. TR-55, TR-20, and SWMM) cannot accurately predict flows from common small storm events or small sites."

Any takers care to discuss ?

Graham Bryant

Graham,

I opine that their claim is baseless. Does the agency offer any supporting documentation or studies for their claim? In the case of SWMM, the modeler may enter any real-world period of rainfall, including common small storms, and can calibrate the output for flow, and possibly other parameters. Obviously, this isn't anything new to you or other SWMM-ers It's been done hundreds-- if not thousands--of times over the past two decades by users the world over.

Of course, any model is potentially subject to the "garbage in, garbage out" adage. And since modeling is more subtle than "cookbook" engineering formulas like the (ir)Rational Method, the potential exists to predict flows sloppily, especially if calibration across the ENTIRE flow regime is not done. However, it would have to be a pretty sorry modeling job to out-sloppy the Rational Method. Remember Dr. James' posting of 11 March this year? He referred SWMM-ers to a graph by Pe(l?)cher, which can be seen at http://www.eos.uoguelph.ca/webfiles/james/runcoef.gif .

Don Waye
NVPDC

Graham, I always find these statements to be intesting because they tend to reflect an agency attitude about modeling and modelers. In this case, as much as we may want to deny the statement, it does have an element of truth in it. Looking at the statement at face value without any of the context, both TR20 and TR55 use SCS curve numbers to define the runoff depth. The CN is notorious for its inability to handle small storms. This is becasue the Ia initital abstraction loss must be satisfied before infiltration and direct runoff can occur. For the very small events, say less than 0.5 inches, the Ia may never be met and as a consequence no runoff can be produced, assuming a mixed urban area with moderate to low impervious fraction.

The state agency comment, in my view, is more directed to the SCS CN method than it is to stormwater models in general. Unfortunately, to the extent that the SCS CN is incorporated in our models (incl. SWMM, HEC-1, TR20 etc.) the statement does contain an element of truth for the small storm. I doubt however that the agency realizes this. this problem.