Why is hydrograph recession limb stairstepping?

[KatieTX]

Does anyone have any ideas on why the recession limb of this hydrograph would be stairstepping like this?

I have run several different storms on this watershed (6 hour 100 year, 24 hour 100 year, 6 hour Probable Max. Precip. adjusted for our location, and 72 Probable Max Precip. adjusted for our location.) The 72 hour PMP is the only one that does this. Each of the 4 storms used different distributions, so I wondered if the distribution was the culprit. Could there be anything else causing this?

 

[SMIAH]

Maybe instability in the calculation or the hydrograph output interval time step.

 

[KatieTX]

I'm using HEC-1. I did a hydrograph time step of 15 minutes. Is there another place to specify the OUTPUT time step? I ran it again with a time step of 3 minutes (the shortest I could do, since HEC-1 limits you to 2000 ordinates), and got the same result.

 

[psmart]

It looks like your rainfall mass curve includes several straight-line segments. Since a straight-line segment in the mass curve indicates a constant intenisity during that time period, this will cause a step in the runoff hydrograph. Using a smaller hydrograph time step will not solve the problem. You need to use a smaller time step for the mass curve in order to eliminate the steps (or at least reduce their duration and visibility.)

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[KatieTX]

Peter,
If I'm understanding you right, I think I tried that. I was using a rainfall distribution that only had a data point every 6 hours. I went back and interpolated between these points to create data points every 15 minutes. Is this what you're suggesting? I am wondering if the program I was using didn't 'save' it and is still using the old rainfall distribution.

 

[psmart]

Adding intermediate points on a straight line will not change the results. You're still describing the same straight line within each 6-hour period. So the rainfall intensity (the slope of the mass curve) is still constant for each 6-hour period, and you will get steps in the runoff, as the rainfall intensity changes abruptly from one line segment to the next. The only way to avoid the steps is to utilize more frequent rainfall depths, which would presumably produce a "smoother" curve. To avoid visbile steps, the points should be spaced somewhat less than the subcatchment's Tc. This causes the steps in the rainfall intensity to be hidden by the Tc-related runoff effects.

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[KatieTX]

Peter,
Thanks for the explanation. I will have to go back into the project and see what exactly I have. You are talking about the mass curve, and I am talking about a distribution. I will go back into the project later and check.

 

[psmart]

A rainfall distribution can be defined as a mass curve or an intensity curve. A mass curve is more common.

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[KatieTX]

OK Peter -- I've looked at our curves again and familiarized myself with the definition of 'mass curve' and you've nailed it. Thank you! I was given the set of points from our State, so I will have to think about how to 'smooth' them.

 

[cvg]

there is no reason to smooth anything and nothing wrong with a stair stepped hydrograph. The results simply reflect the limited / simplifed input data. If anybody asks, then just explain it to them. How does this affect your engineering analysis of the spillway anyway? It's on the receding limb and I presume you are most interested in the peak flow and the volume.