HydroCAD vs. HEC-HMS

[nstong]

Hello All,

I'm working on a 6 sq. mi. drainage study. I first ran HEC-HMS, then HydroCAD to get a quick sense of the storage volume required. The numbers were significantly different (2800 cfs vs. 2400 cfs) and I've set them up to use the same calculation methods. I'm using SCS unit hydrograph method, SCS Type 1A temporal distribution, NOAA 100-yr 24hr rainfall depth of 6.5", TR-55 for Tc (Hey HEC, how about incorporating the TR-55 equations?!)

The most curious is when I isolate one subshed and HMS and HydroCAD yield the same runoff depth (4.3") but have such different peaks (2200 cfs vs. 1900 cfs). 5% error I would still question, but 15% is very high, wouldn't you agree?

Thanks in advance for your thoughts!

Nate

 

[LHA]

I would say your results are typical for size of error versus size of watershed. I have never seen two different models produce anything closer than what you are finding.

Now for the real dilemma...you need to design and which do you choose? Conservative design or over-design?

I would run it on with the actual TR55 program, and choose the model which yields the closest. Submit the actual TR55 results for your hydrology, then submit the HEC or HydroCAD output for hydraulic design basis.

Remember: The Chinese ideogram for “crisis” is comprised of the characters for “danger” and “opportunity.”
-Steve

 

[psmart]

The difference may be due different rainfall distributions. The tables have evolved over time, and its difficult to be sure you've got the same table in each program. HydroCAD also uses polynomial-based tables that contain more detail for short Tc values, and can produce higher peaks as the Tc is reduced. As you noticed, runoff volumes will remain the same, indicating that the same runoff equation and CN value are being used.

You should also check the unit hydrograph and the calculation time step. A comparison to TR-20 would also be informative. But keep in mind that none of these tools is claimed to accurately predict exact runoff rates. They are best at predicting changes in runoff from an existing to a proposed (developed) condition. As long as you use the same techniques througout a given project you should be OK.

 

[nstong]

Thanks to you both for your replies. I looked at the time step results for both and it looks as though the HydroCAD model absorbs more of the initial rainfall (3 hours of a 24 hour storm without runoff) while HMS starts yielding runoff after 11 minutes. Also of note: runoff for large sheds HydroCAD>HMS; for small sheds HydroCAD<HMS.

I suppose I will run the TR- routines at some point. Right now I'm pretty tired of the uncertainty of hydrology. I'd rather study String Theory!

Nate

 

[psmart]

If the runoff is starting at different times, this suggests that a different runoff equation or CN value is being used. Also check the Antecedent Moisture Condition (AMC), which can have a significant effect on the runoff.

However, in your original post you said the runoff depth (and therefore the volume) were the same. This would indicate that the same runoff equation and CN value are being used.

Different Tc values or rainfall distributions will affect the runoff timing and peak flow, but not the volume.

I suggest you go back and re-check all your input data. I suspect this is a simple discrepancy in the input.

 

[queque]

I am constantly finding that users calculate peak flows
using a variety of methods. I have yet to see a rainfall runoff event mathced using several methods for the same event.
Try using actual storm data (USGS Water Resources Division is a good source) to calibrate yes calibrate you model. Then the method you use and variations using other methods can be scientifically explained.

 

[Jessica2424]

I am pretty new to HEC-HMS and I seem to be having many problems with it. I am trying to model a watershed of about 1700 acres. There are many basins, ditches and culverts in my watershed. I am having a huge problem modeling the culverts. Can I model culverts in HEC-HMS or should I leave them out of HMS and input them into HEC-RAS? Any ideas? If I should include them in HEC-HMS, then how should I model them?