HEC-RAS Flows (seem too high) 1

[dlnva350]

I’m working on several HEC-RAS models for a drainage study. Our stream restoration team has conducted geomorphic surveys of the stream channels. The spreadsheet used for this process provides output that includes bankfull flowrates. The 1.5 year event is said to be responsible for bankfull flow. While it’s not a linear relationship, we split the difference between the 1 and 2-yr event to define the 1.5 yr storm depth. Flows calculated using the SCS method (with HydroCAD) are significantly higher. I’m talking at least 5 times higher than the geomorphic survey spreadsheet expects for bankfull flow. Our TR-55 1.5-YR flows are out of the channel (flooding the overbanks). Generally speaking, I doubt that 1.5-YR events flood streams.

 

[psmart]

This sounds similar to the previous thread
thread162-219930

Since you're using HEC-RAS, I assume this is a watershed of significant size. TR-55 is not well suited for this scale, and even TR-20 (HydroCAD) will require calibration to accurately predict flows on this scale. As francesca mentioned in the previous thread, timing between different tributaries is often a significant factor in determining the total flow, and will require a non-trivial hydrology model to give reasonable flow predictions.

What is the approximate watershed area? How many subcatchments were used in the TR-55 and TR-20 models?


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[dlnva350]

Peter,
When I logged on earlier today I noticed the other post. I just thought my post was of a slightly different nature. First of all, the entire project area is approximately 2 sq. miles. Most of the subcatchments are larger than 100 acres and less than 400 acres. Subcatchments were created to coincide with HEC-RAS survey locations. By the way, all of our hydrology models have used HydroCAD. Sometimes my coworkers refer to SCS methods as TR-55 or TR-20 (depending on who you're talking to). So if TR-20 is not well suited for large scale work, which method is? We've been discussing calibration and I will most likely attempt this tomorrow.
Thanks,
Dean

 

[psmart]

TR-20 is certainly usable on that scale (note that TR-20 uses areas in square miles) but how you delineate and route the subcatchments will have a profound effect on the results.


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[RWF7437]

dlnva350,

The following questions and suggestions are offered for your consideration.

"I'm working on several HEC-RAS models for a drainage study."

What kind of drainage study ? Is it a "flood study" concerned mainly with high flows ? Is it a "stream restoration study" concerned with a wide range of flows, including very low flows ? Or is to something else ? What is the purpose of your study ?

"Our stream restoration team has conducted geomorphic surveys of the stream channels."

Is a gemorphic survey the same thing as a topographic survey ?
Is the survey limited to the "streams" or does it include the overbank areas subject to flooding ?

Do you have good topographic maps of the basins of interest ?
Do your maps extend upstream and downstream from the area of interest ?

"The spreadsheet used for this process provides output that includes bankfull flowrates."

Where did this spreadsheet come from and what is it based on? Is it based on gaged flows over a long period of time ( years ) ? Do you have confidence in the numbers in this spreadsheet ?

"The 1.5 year event is said to be responsible for bankfull flow."

Said by who ? I have heard many such assertions but seldom seen any data for them. Some people say ten year (10%AEP), others say 2 Year ( 50% AEP ). Does it matter for the purposes of your study ?

"Flows calculated using the SCS method (with HydroCAD) are significantly higher."

Higher than what?

As Mr. Smart points out, the TR-20 and TR-55 methods may be used to calculate peak flows for drainage basins of the size you are studying. Those flows depend on a multitude of assumptions. They also depend heavily on the storm pattern assumed. All these methods are UNCALIBRATED models and are subject to numerous errors. For example, for my part of the USA, rainfall amounts are typically no more "accurate" than plus or minus 30%.

HydroCad is principally a Hydrolgy model.

HEC-RAS is principally an Hydraulic model. As such, it does NOT calculate flows; it calculates water surface elevations given presumably "known" flows and "known" water surface elevations at one or more points along the stream.

If you are doing a flood study you will probably be required ( by FEMA) to use "conservatively" estimated flows. Unless you have gage data for the streams you are studying, FEMA often uses flows derived from gaged data. These flows are often calculated using regional regression equations. In my experience, they are never calculated using the Rational Method, TR-20, TR-55, or any of the many other methods used by local regulatory agencies. The word "conservatively" in this context means high, rather than low.

If the purpose of your study includes estimates of very low flows, then "conservative" may have the opposite meaning. For stream restoration, for example, the goal of restoring a stream to its "natural" condition may INCREASE the probability of flooding. I have seen several such projects where the "designers" never even calculated flows or considered either drought or flooding conditions. They just threw a bunch of big logs ( "large woody debris" they call them ) into the creek and declared it a good design.

The particular computer models you choose to use are probably of far less importance than the assumptions you make, the effort at calibration you put in, or the art of balancing possibly conflicting goals you execise.

good luck