Does HydroCAD have Weighted-Q method capability for computing SCS Runoff? 1

[LincolnPE]

Question: Does HydroCAD have Weighted-Q method capability for computing SCS Runoff?

 

[psmart]

On the Advanced tab of the Settings|Calculation screen, there is an option to "Calculate separate pervious/impervious runoff". Rather than combining all the CN values into a single weighted CN, this option calculates separate weighted CN values for the pervious and impervious portions of the subcatchment, calculates the runoff from each portion, and combines the two Q values to create the final runoff.

This option is most often used with the SBUH runoff procedure, but can also be used with the SCS unit hydrograph method if desired. The effect is most notable for lower rainfall depths, where the separate calculation for impervious surfaces generates more runoff in comparison to using a single weighted CN for the entire area.

HydroCAD also provides a special CN weighting procedure for unconnected impervious surfaces which addresses this issue in a different manner. See the TR-55 manual for details on this approach.

Please note: There is a separate HydroCAD forum ( forum789 ) specifically for these questions.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[LincolnPE]

"calculates separate weighted CN values for the pervious and impervious portions of the subcatchment"

This is really great news. Many kudos for providing this capability in the software. I saw the setting and just wanted definite confirmation on the methodology implemented.

Just another clarification please, you mention "weighted CN" above. Is there really any "pre-weighting" going on in the methodology before the individual area volume calculations?

Or, is it using the single CN for each individual area, computing the individual area runoff volumes, and then summing and weighting the Q's for a final weighted-Q?

I'll leave it to the forum moderator, however my hope is we can keep this thread here for further discussion on the use of CN in SCS runoff calculations.

 

[psmart]

As an example, if you enter CN values of 98, 86, and 80 in a single subcatchment, HydroCAD will normally calculate a single area-weighted CN of all three values, which is then used to evaluate the SCS runoff equation.

But if you select the option for "separate pervious/impervious" runoff, it will calculate TWO weighted CN values: One for the impervious area (98) and one for the pervious areas (86 and 80 in this example). The runoff is then calculated for these TWO parts of the subcatchment, and added together to produce the total runoff. Regardless of the number of subareas you enter, this option will always place them into these two categories.



Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[LincolnPE]

Thank you for the clarification. I am looking for a volume calculation for each CN number, for the Weighted-Volume method.

The HydroCAD methodology you describe would be Area-Weighting CN's from 39 to 96 then? This unfortunately sets me back to the problems associated with the Weighted-Area method.

The Weighted-Volume is always accurate per the given data. It has more calculations, but the thought is we are using computers now.

 

[psmart]

Correct, this option causes all pervious CN values (96 and lower) are weighted together. Again, this approach was implemented primarily for the SBUH runoff method. It is rarely used with the SCS method - although it could be.

As for a separate runoff evaluation for each CN, we've never had a request for this capability. Frankly, weighting of CN values is not just for expedience, but also for better accuracy with interspersed ground covers. If the ground covers are not interspersed, and they drain separately to the point of analysis, a better solution is to model each CN using a separate subcatchment. This produces the "weighted-Q" you originally asked about. Using separate subcatchments also lets you use a different Tc for each ground cover.

I believe HydroCAD provides a full range of options in this regard. Use whatever approach you wish.

For further details on the use of weighted CN values please see HydroCAD help and

http://www.hydrocad.net/curvenumber.htm

Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[LincolnPE]

Yes, actually it really should be used now in computer methods.

"Frankly, weighting of CN values is not just for expedience, but also for better accuracy with interspersed ground covers."

If you mean Area-Weighting that is not correct. TR 55 talks about this explicitly, how the combination of weighted pervious and impervious areas can give inaccurate runoff estimates. It is primarily because of the presence of the initial abstraction term Ia. The Ia term essentially gets (incorrectly) weighted across the impervious areas also, implying significant losses that may not occur.

Is there anyway I can supply my hand calculated Weighted-Q into HEC RAS to generate the flood hydrograph?

Any thoughts on offering the weighted-Q method? It is always accurate per the given data.

 

[psmart]

I agree that weighting of pervious and impervious areas should generally be avoided, partly because of the very different Ia. But I'm not convinced that weighting of different pervious values is inappropriate when the sub-areas are intermingled. Nevertheless, your goal of a "weighted-Q" is readily achieved by using a separate subcatchment for each CN value. This avoids CN averaging and produces a weighted runoff, exactly as you are seeking. And it allows you to assign a separate Tc value to each ground cover, which is probably needed if the areas are not intermingled.

Of course, this is ultimately a matter of what approach your stormwater agency will accept, and I'm not aware of any that are using a weighted-Q approach. The closest I've seen is the separate pervious/impervious procedure used with the SBUH runoff method, and that option has been available in HydroCAD for some time.

If others see a need for a "weighted-Q" option, let us know and we'll bump it up on the wish list.




Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[LincolnPE]

Hi Peter,

Along with the incorrect Ia parameter distribution that occurs with area-weighting of Curve Numbers, inaccurate estimates occur due to the non-linear behavior of the SCS runoff equation. Area-weighting of the runoff coefficient is not a problem with the linear characteristics of the Rational Method. It can be a significant problem when the CN numbers are not close in value, as is the case many times in land development.

In land development it is common to reasonably discretize an area consisting of interspersed different cover complexes into one basin, with one reasonable Tc path. It would be too time consuming, not cost effective, and not physically warranted to subdivide such a basin into separate basins each with its individual CN, Tc path, and hydrograph; and then have to route them together after that. This could easily amount to 3-5 times more basins, Tc paths and hydrographs, across a land development project than is needed, than if the always correct, Weighted-Volume SCS method was used.

Please see my example below comparing the Area-Weighting method and the Weighted-Volume method for a typical soil cover complex interspersed over 10 acres. Note the inaccurate estimate the area-weighting procedure delivers here.

Cover Area (ac) P (in) CN S (in) Ia (in) Q (in)
Woods 2.00 3.00 30 23.33 4.67 0.13
Pasture 3.00 3.00 58 7.24 1.45 0.27
Lawn 3.00 3.00 72 3.89 0.78 0.81
Gravel 2.00 3.00 96 0.42 0.08 2.55
Weighted-CN 10.00 3.00 64.2 5.58 1.12 0.48


Q(W-Vol) Total Vol
0.86 31,241
Q(W-Area) Total Vol
0.48 17,283

[highlight #FCE94F]% difference = 45%[/highlight]

I have attached an image file also for easier viewing if needed.

 

[psmart]

I realize the difference can be significant, but I'm not convinced that CN-weighting for intermingled areas is entirely incorrect. This process makes some allowance, however crude, for the runoff from a higher CN area flowing over a lower CN area.

I would agree that averaging pervious and impervious CN values presents the greatest chance of inaccuracy, but HydroCAD can address this with the pervious/impervious option discussed above, as commonly used with the SBUH method.

Beyond that, you might want to contact the WinTR-20 development team at NRCS. They are essentially the gate-keepers for NRCS-related runoff calculations. Otherwise you're going to have a tough time getting the calculations accepted by a local stormwater agency, since they will not be be able to reproduce your weighted-Q results with any of their tools.

As I mentioned before, this is on the HydroCAD wish list. Implementation will depend on feedback and customer requests.




Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[LincolnPE]

Thanks Peter.

No that is just simply incorrect, those crude allowances are definitely significantly under estimating (and in some cases over-estimating). Many times the a good portion interspersed pervious ends being upstream also, as you can imagine.

Please reference NEH 630 Chapter 10 directly for the correct application of the SCS runoff equation with CN variation within a catchment. This requires the Weighted-Q method . The New Jersey Stormwater Best Management Practices Manual is also a great resource on this issue.

I am not sure how, in general, the stormwater design industry/community has gotten so far away from NEH 630? A lot of engineers are seeing the weighted average CN method on the software packages, and are getting funneled into "hey that's easy I'll list the CN's here for my catchment, the software will weight them for me, and away we go". They don't realize the implications of an area-weighted Ia and the non-linear behavior of the SCS runoff equation discussed in the above mentioned references.

I think it's hit and miss on who you get as a hydrologic professional reviewer, depending on their education and expertise. I think Santa Barbara and New Jersey are on board with correct Weighted-Q SCS application. Along with the Colorado Front Range's Urban Drainage and Flood Control District in various other technical areas.

It is not the same as the linear Rational Method, that allows area-weighting. Maybe that's the source of confusion, along with the technical dilution attributed to TR 55?

Now, in the software packages you could still have the same multiple CN selection dialog box, but the weighted calculation would be one of Volume (and conservation of mass), that is always correct.

 

[psmart]

As a practical matter, if a reviewer cannot reproduce your runoff numbers using TR-20 or TR-55 you're going to have problems. So this is really an issue that needs to be addressed by the NRCS software development team. Adding this capability to HydroCAD may be a great feature, but until the practice becomes broadly accepted it's going to be a challenge for any engineer that chooses to use this approach.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[LincolnPE]

Thank you Peter for the good practical suggestion. I am writing them right now.

They are currently taking comments for the new release.

 

[psmart]

I'll make sure we give this a bump on the HydroCAD wish-list. Should be interesting.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[beej67]

Checkboxing this thread.

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

http://www.campbellcivil.com

 

[LincolnPE]

Yes, to clarify more:

In NEH 630 Chapter 9, equations 9-1 and 9-2, and subsequent watershed examples, addressing urban basins with connected and unconnected impervious areas, utilize a single pervious CN value (CNp) for the catchment. The chapter mentions nothing about conducting Weighted-Area averaging of pervious CN values.

NEH Chapter 10 provides caution about the practice of using a Weighted-Area CN value in the non-linear SCS runoff equation. It discusses how this practice may not create unacceptable accuracies only if the CN values to be Area-Weighted are close in value. The chapter discusses the inaccuracies of the Weighted-Area method for CN values that have significant variation.

Chapter 10 states the Weighted-Volume approach as "always correct" per the given data.

It also discusses how the runoff equation as a predictor of runoff does reasonably well when the runoff is a substantial fraction of the rainfall, but poorly when the runoff is just a small fraction of the rainfall. So, caution with large loss results relative to rainfall...

 

[psmart]

You can expect to see a "weighted-Q" option in an upcoming build of HydroCAD-10. The options will look something like this:

1) Use composite CN for each subcatchment (Weighted-CN)
2) Calculate separate runoff for each CN (Weighted-Q)
3) Calculate separate Pervious/Impervious runoff (SBUH weighting)

Option 2 is new. The other options are longstanding and will continue to work exactly as before.

Peter Smart
HydroCAD Software

http://www.hydrocad.net