Wednesday, April 15, 2009

Using HEC-HMS for a dam breach simulation

Written by Chris Goodell, P.E., D. WRE | WEST Consultants
Copyright © 2009. All rights reserved.

Using HMS to perform a dam breach has its advantages, namely that it is easy to set up, the data requirements are minimal, and it is numerically stable. However, routing the breach outflow downstream using HMS is very limited. HMS uses hydrologic streamflow routing which is a simplification of full dynamic routing. Plus, backwater will not be fully accounted for using any of the HMS routing techniques-meaning you can not simulate flow attenuation properly. In short, HMS is okay for routing water from A to B, but if you are interested at all in what is happening between A and B, HMS is not appropriate. Furthermore, there is no direct method for mapping flood inundation from an HMS model.

What I usually suggest is that the breach can be modeled using HMS, but downstream of the dam, RAS should be used. You could go with full unsteady RAS downstream of the dam and use the techniques I described in the dam breach class to get the final inundation mapping. If you are having stability problems due to an overly steep reach, HMS could be used for routing, but the peak flows should be finally run through steady flow RAS to get the flood inundation extents.


  1. Even though I entered in the Dam Breach Data that the bottom breach elevation goes down to the bottom of the cross section, when I run the model, the breach only goes about halfway down the dam. Is there a reason why it isnt performing a full breach? Any tips to correct this?

    1. Hmmm...Things to check:
      1. Are you allowing the simulation to go long enough for the breach to fully form?
      2. Do you have small enough detailed output interval to see the entire breach form?

  2. Please, is there any step by step example for dam beak analysis in hecras?
    I am facing many difficulties in filling different menus and running the program with many errors and warnings.

    1. Not really, however you can look through the dam break example problems. That may help you.

  3. struggling with the hydrology in HEC HMS...can someone DM me please?

  4. Someone can help me with this problem in HEC-HMS?

    ERROR 40441: Subbasin "1" could not be initialized because the grid cells were missing or parameter data was invalid.


  5. Hi Chris !

    I have ran a HMS model to route a outflow breach hydrograph to use it in RAS to simulate an overtopping dam breach scenario at downstream.

    I am getting 1.7ft difference of peak elevation and about 1000cfs of outflow between HMS (155.5ft and 2185cfs) and RAS (156.7ft and 3239cfs) at the dam. I was hoping to close this gap little more if it is possible or suitable.

    My question is, what factors could make the RAS change the results from HMS's like this? I thought my settings are fairly stable so that Courant number is 0.96 ((4sec*12ft/s)/50 cell size). I am using same values/conditions to describe the reservoir (Elev. v. Storage) and dam structure in both HMS and RAS. I have tried changing the Theta values around, but no success. I am running out of things to check to stabilize the models and was wondering if you can catch something that I have not tried.

    The RAS model consists of Storage Area as reservoir and 2D Area as downstream with a outflow hydrograph, boundary condition, and theta as 0.8 to run an unsteady simulation. I hope I addressed everything for you to answer my inquiry! Please let me know if you need any more info otherwise.

    Hope all is well !

    Daniel Kang

    1. Daniel. There are some differences in how HMS and RAS compute weir flow and develop the breach. Tailwater submergence is handled differently as well (I think). I think that is likely the issue. You should be able to see the differences in the manual.

    2. Hey Chris,

      Thank your for your guidance :). After looking at the manuals and some lit, it is the negligence of taking account of tailwater in HMS that often overestimates the peak flow for a dam breach hydrograph compared to RAS.

      So, for the best/most accurate assessment, I have used HMS to route an inflow hydrograph for the upstream and used the inflow hydrograph as the reservoir's boundary condition in RAS. Then I used RAS to trigger the breach, as the inflow from upstream brings the lake to the dam crest, to create the outflow breach hydrograph. Would you say that this is a fair way of 2d modeling a dam breach? I am getting about 0.15ft difference in peak WSEL between HMS and RAS results, fyi.

      I always appreciate your guidance! Thank you

      Also Happy new year!


    3. I agree with your approach. One other advantage that RAS 2D has over HMS is its ability to dynamically simulate the reservoir routing during a breach, as opposed to HMS which does level pool routing.

  6. I have daily rainfall data for my study. i computed lag time and got 600 minutes. i set time inetervals (control specs and time series) as 1 day. here are my questions:
    (1) when i run the simulation, the results gave warning 41784 Simulation time interval is greater than 0.29 *lag for sub-basin; reduce simulation time interval. how do i adjust each time interval? do they need to be the same always?
    (2) if i don't have observed discharge data, how do i calibrate the results?
    I will use the results of hec hms in hec ras.

  7. my work is regarding creating a hydrologic model using HEC HMS, I have created the model and now I am working in the calibration of the model,I have read the manual for calibration from your website, so for calibrating the model, I have the measured depth instead of measured discharge , so my doubt is, is it possible to calibrate with depth in HEC HMS ?

    1. Yes, but better to calibrate to water surface elevation instead of depth.


Note: Only a member of this blog may post a comment.