Reservoir routing - hydrological problem 7

[SMIAH]

Hi,

I have a project involving the determination of a 100-yr storm peak flow. The watershed has the following characteristics:

Area = 8.7 km2 (3.35 mile2)
Average slope = 13.1%
% woods = 67%
% pasture = 3.4%
% residential areas = 12%
% swmamps = 2%
% lake = 15%

I have a CN number for every subbasin.
I'm thinking of using a SCS Type II storm (6 hour, 12 hour and 24 hour) and compare the results.

My problem is the 15% of lake... Actually the lake is at the downstream end of the watershed (see the file attached).

I have the volume of the lake (1 250 000 m3.)
But regarding the reservoir routing... I'm quite lost.

What should i do to get a good estimate of the peak flow at the downstream end of the lake (where there is a small dam)?

Thanks for any clue regarding the method i should use for the reservoir routing!

 

[beej67]

A couple of comments...

Regulators will let you get away with a CN of 98 for the lake area itself, but the appropriate thing to do is set it at 100 if it doesn't break your software. The evaporation from a lake for design storm is negligible over that time scale, especially during a storm. But as has been mentioned above, the differences aren't really worth arguing about.

Always always always start your reservoir routing at the control elevation of the dam. (That'd be your lowest orifice or your spillway / etc) There are two exceptions I can think of, and they're both driven by the regulatory agency you're doing your work for, so you'll know if they're meaningful.

Some regulatory agencies like to have stormwater treatment ponds that capture and infiltrate water into the ground. (e.g. Florida WMDs) In these cases, you're usually required to prove that this dead storage retention volume is recaptured between storms, by showing infiltration calcs and water table elevations and whatnot. Some regulatory agencies (some munis in Georgia) like you to start higher than the lowest control, if the lowest control is a long term bleeder device for water quality and they don't trust that you'd have all the volume available during the flood routing.

The elevation you start your reservoir routing at is important. Folks get sued over this kind of stuff. I was an expert witness against another engineer who screwed this up badly and was exposed for a lot of damages because of it.

On backwater, there's two issues. One is whether folks upstream of the reservoir get flooded by the reservoir level increasing. For any reasonably sized application just assume the water is flat across the reservoir's surface. The other issue is one engineers also screw up quite a bit in some areas of the country, and that's tailwater on your outlet control structure. This is a big deal in Florida since the discharge from series of connected ponds can be effected by the downstream pond, because the driving head is not just upstream head, it's head differential. You need time series routing tools for these, such as ICPR or XP-SWMM. In regions of the country where you aren't likely to have tailwater on your outlet control system, engineers have grown used to using simpler tools that only consider upstream head. (Hydraflow ... ugh, what a piece of crap) Engineers who don't realize the difference, and use the latter to design the former situations can get into big big big trouble, because their pond doesn't draw down like they expect it to, and stages higher than they expect. Again, fodder for lawsuits.

My comments are largely for small reservoirs in land development, or regional reservoirs that are not directly managed. I'm not talking about municipal drinking reservoirs or other things that are managed by some dude at a control desk. How you do hydrology for those sorts of applications is much more complicated, and I assume the top poster isn't asking about those on a forum such as this.

As always, go with widely accepted practice for your region (if such a thing exists) because hydrology is very, very regionally influenced.

 

[SMIAH]

Very interesting comments.

HEC HMS doesn't support 100... I guess 98 and specifying 0 inches of Initial Infiltration is something suitable.

"The elevation you start your reservoir routing at is important."

I'd tend to start it to the normal pool elevation.
I think that assuming full storage prior to a, let say, 1000-year storm is really conservative.

Indeed this is related to a small reservoir where 2d hydraulic modeling would be... too much.

 

[beej67]

Assuming "normal pool" = the elevation of your lowest control, then yes, start it there unless there's a compelling and documented reason to do otherwise.

 

[orshii]

I have a similar situation with drainage areas and a lake, but I have 3 separate subcatchment areas that outlet to the lake at 3 different points. How can I model the lake runoff through an existing outlet dam with HydroCad? Thanks.

 

[psmart]

orshii - I beleive your question is a double-post. Please see answer(s) in the HydroCAD forum.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[SMIAH]

Back on this subject:

The regulatory agency in charge of reviewing a similar project is asking me to find another way to calculate the outflow from a dam.

I did use HEC HMS and reservoir routing.

They say that they don't get the same outflow using flow rate recorded at a gage station located on a river near the study area and transposing it for the basin related to this study.

I really think that such a comparison is not possible considering the important storage of the lake...

Performing my analysis, the WSE of the lake increased of almost 2' during the simulation. The outflow from the dam was then calculated with standard weir equation = CLH^1.5 = 3.1(8.2)(2)^1.5 = 72 cfs.

Someone have any idea of another method to calculate the outflow from the dam?

Please don’t say “Rational Method”.

 

[psmart]

First: What rainfall are you using in he model? And what rainfall produced the gauge reading you are being asked to match? Unless you are using the exact rainfall distribution that produced their gauge reading, your results will be different.

Second: Since you are modeling a complex 3.5 SM watershed, the margin of error will be significant.

How large is the difference? I would verify all the input data (including the rainfall depth and distribution) to look for errors or reasonable adjustments that would bring the model into closer agreement with the gauge.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[SMIAH]

They are using recorded flow rate at a gauge station located outside of the watershed.
And a statistical analysis (e.g. using log-Pearson II distribution) to determine a 100-year flow rate.
Then they are using the ratio of basin areas to transpose this flow rate to the area of study.

I'm using a 24-hour SCS-II 100-year rainfall.

This particular method of comparaison doesn't make any sense to me (!?)

 

[cvg]

you may also have discrepancies with the outflow for the dam. I assume you used a rating curve for your outlet, not just a single flow rate? Also, where did you get the weir coefficient? It seems high for a broad crested weir and low for a sharp crested one. In fact, most weir coefficients vary with the flow rate depending on the width, depth of flow, contraction, entrance conditions and submergence. This could have a big effect on the outflow from the reservoir.

 

[SMIAH]

3.1 in SA would equal to 1.7 in SI which is critical depth.
Yes the rating curve take in account effective length as state in Design of Small Dams (USBR, 1987). No Submergence effects.

Anyway, my question was more related to a validation method fur a project like this one? Another method than using HEC HMS and reservoir routing.

 

[SMIAH]

1.705 and 3.09 for purist.

 

[cvg]

I think if you were comparing apples to apples, HEC-HMS reservoir routing would be fine. HEC-1 should also work. But it sounds like your hydrograph doesn't compare favorably with the agencies hydrograph. Obviously SCS vs statistical manipulation of a similar (perhaps) watershed could yield entirely different results. You may also want to use an antecedent moisture condition III instead of II. Perhaps starting water surface elevation of the reservoir don't match either and who knows what outlet rating curve they are using.

Perhaps you could ask them to provide their inflow hydrograph and plot the two and compare the differences.

 

[SMIAH]

Thats the point! They're not using any reservoir routing method. They're just using statistical basin transposition and comparing peak flows only (no hydrograph).

 

[cvg]

well duh, I thought they were reconstituting a hydrograph. They are measuring flow on an uncontrolled watershed and you are modeling a controlled watershed. Obviously the flows can never match. I'm surprised they are able to understand the statistical analysis and basin transposition and yet don't understand basic reservoir routing...

 

[SMIAH]

This is such a relief to read someone thinking the same thing than i do.

What to do now (apart from teaching them basic hydrology).
A way to validate the max increase of 0.5 m in the lake.

 

[cvg]

try this

http://www.stormh2o.com/sw_0607_voodoo.html