Watershed Mode Calibration 2

[ALTIME]

Hi All,

I used Eaglepoint Watershed modeling for our stormwater management design. Somehow the local regulartory agency thought our predevelopment flow rate is higher compared to the established value in the vicinity of our project land, and asked us to do model calibration with the nearby stream gauge with a gross drainage area of 1889 square km; while our land is only 64 ha. I'm wondering if I can do a calibration with this gauge flow information.

Any Thoughts?

Altime

 

[SMIAH]

I wouldn't do any transposition of flow rate with a basin of 1889 km2 to a 0.64 km2 one.

I usually use a ratio limit of area between 0.5 to 1.5.

 

[cvg]

with an area that large, I would think you might see very significant aerial reduction. You might end up with less runoff. However, I agree the watersheds should be similar in size, vegetation, development, slope, soils and any other characteristics that would affect runoff rates and volumes.

 

[RWF7437]

What method or methods does Eagle Point use for a 64 ha (158 Ac)drainage area ?

 

[cvg]

you might be better off trying to calibrate (or maybe validate is a better word) using a regional regression equation than by using this particular stream gauge.

 

[ALTIME]

I selected Rational method. The predevelopment flow rate I got from the model is 1.25 cubic meter per second.

 

[cvg]

1.25 cms (44 cfs) from 64 ha (158 acres), sounds a bit low to me.

 

[RWF7437]

Rational Method is probably NOT the best choice for a watershed of this size.

 

[jgailla]

RWF is right, the Rational Method is not appropriate for a basin of 158 acres.
I would only use the Rational Method to size pipes for small catchment areas under 1 or 2 acres.
One method, which is accepted by most authorities in my area (SE US), is the SCS method described in TR-55.
I know there are a lot of arguments about the different methods, but the SCS is widely accepted.
The only problem is that it is much more complicated to do by hand, although there are many computer programs such as ICPR which will do the routing for you.

 

[ALTIME]

If I use SCS method, the predevelopment peak discharge rate would be over 3 cm/s. It sounds a big number to me.

 

[jgailla]

3 cubic meters per second is about 100 cfs.

For a 158 acre drainage area, that sounds pretty low. Of course, it could be anything, since we don't know the soil or slope conditions or the design storm you are using.

For example, I recently had a 75 acre site that was undeveloped and very flat, although the soils were pretty wet. The pre-development runoff for a 25 year/24 hour storm was 180 cfs.

Unless you have a very flat area with sandy soils 100 cfs is low for a 158 acre site, so if I was your reviewer I might be suspicious also.
Maybe you could convince the reviewer to let you use the SCS instead of model calibration with a stream gauge, which would take a lot more time.

I wouldn't use the regression equations either. I'm partial to SCS, but that's probably because it's the most accepted method to reviewers around here.

Usually we have to limit our post-development flow to pre-development levels, since we're in a wet area, so if we come up low on the pre-development rate it's not a big deal.
Are you in a dry area and doing drought calculations?

 

[ALTIME]

Depending on the model selected, the results could be significantly different. If we use the same method to design the detention ponds to restrict the flow to predevelopment rate, the sizes of the ponds required resulting from different methods would be slightly different. What would you think Jgailla?

 

[RWF7437]

"Depending on the model selected, the results could be significantly different."

Very true. Just try it and you'll see differences of plus or minus 100% in many cases.

" If we use the same method to design the detention ponds to restrict the flow to predevelopment rate, the sizes of the ponds required resulting from different methods would be slightly different."

More than likely they will be significantly different. That is why most local reviewing agencies adopt "standard" procedures for design. That is also why you should not mix methods in your design because it leads to much confusion. If you want your project approved you will have no choice but to use the local procedures. But, that doesn't mean you can't test your design for "reasonableness" by any other method you have confidence in.

Comparing hydrology methods is NOT calibration. Calibration compares calculated vs measured parameters. If you have calibration data use it. But if you don't be aware that your answers will be subject to a large range of error.

good luck

 

[jgailla]

altime,
RWF (I believe) has posted a pdf entitled "Voodoo Hydrology" or similar in one of these fora.
If I can find it I will repost it, but you may want to use the google search button at the top of the page to search for it.

 

[ALTIME]

I found the article "Voodoo Hydrology" and will spend some time on it.

Now I'm using EPA SWMM5 to compare the results. One question is how I can convert rain fall data from IDF curve to temporal 24 hour synthetic rainfall distribution and input it in the model? Anyone can help?

Thanks

 

[RWF7437]

ALtime,

You have jumped from the frying pan into the fire.

To get back on track, please tell us what it is you are trying to do. Are you trying to size a culvert, design a storm drainage collection system, design a local detention basin or write the definitive opus on the nature of rain?

SWMM 5 is probably overkill for most of these.

Comparisons between methods will probably only confuse things farther.

You cannot get a "temporal synthetic rainfall distribution" from an IDF curve.

Where are you? What are your intentions? What are you likely to be required to do to accomplish your task and gain approval of whatever agency will review your work ?

 

[SMIAH]

Agree with RWF7437!

I think he's interested in a synthetic probabilistic storm.

So he could use a 24-hour precipitation depth obtained from and IDF curve (e.g. 5 inches) and then put it in an Excel sheet or in a software (e.g. SMADA 6.0) and it would gives a distributed rainfall (e.g. SCS-I).

 

[SMIAH]

Reading this post (entirely this time!) - I'm wondering why rational method can't be the best choice for a sub-basin of 0.64 km2.

If it's for the design of a culvert and the designer's interested in a peak discharge, Rational Method could give an estimate value of the peak discharge (conservative).
The the focus should be made on the calculation of Tc.

I'll probably check for a range of flow rate when sizing the culvert and be sure to model the boundary conditions in a correct manner.

Like RWF7437 pointed out, it really depends on what you have to do.

 

[RWF7437]

The Rational Method was introduced in the U.S. by Emil Kuichling in 1873. It is based a very small number of urban drainage basins and has seldom, if ever, been calibrated. Many reviewing agencies restrict its use to small areas. For example, my local City limits it to 25 acres. Others allow its use for larger basins, up to about 1 square mile (640 acres)but few engineers I know would use it for such a large basin.

Also, the Rational Method is NOT the most conservative estimate one can make. Many Flood studies are based on regional regression equations ( which are based on stream flow records) which typically yield higher peak flows than one would calculate using Q=CIA.

All uncalibrated models are subject to large errors and none are any more than statistical probabilities. That's about as sweeping a generalization as I have ever made so I'll stop now.

 

[RWF7437]

http://www.weather.gov/oh/hdsc/