2D HEC-RAS modeling results dependent on precipitation BC data time increment

[Davis Murphy]

Does anyone know if Precipitation BC data is rounded or truncated, or whether it is altered from the referenced DSS file in HEC-RAS 2D? (more details below…)

I have a 2D rain-on-mesh model that uses HEC-HMS to produce effective rainfall hyetographs. These hyetographs are used as precipitation boundary conditions and are entered as DSS file references.
I have been doing some testing in the office and have noticed that when I use different time steps in HEC-HMS it will cause large changes to the results in HEC-RAS. For example, if I use one minute time steps in HEC-HMS it will result in lower total volume and peak flow at the outlet of the 2D area in my HEC-RAS model, as compared to using one hour time steps in HEC-HMS.

My theory is that the one-minute precip data has much smaller individual depths at each time step and more records compared to the one-hour precip data. So, if the precip data is rounded then the one-minute precip data will be affected by rounding and truncation to a much larger degree.

An alternate or complementary theory is that the 2D area filter tolerance values may be playing a part in the matter, but I have not done any testing to see.

[cameron]

Just changing the output interval in HMS from 1 min to 1 hour will have a huge impact on results in HMS (before input into HEC-RAS). Generally in HMS, the lower the output interval the higher the peak discharge you get in HMS.

As for the tolerance values, they do make a difference for rainfall values as they are generally really small.

[Davis Murphy]

Thanks for the reply cameron. I recently ran a series of simulations in 5.0.7 and am no longer seeing the issue.

Here’s a summary of results where I have altered the precip data time increment in a single HEC-RAS 2D rain-on-mesh model (precip only) and compare 2D Inflow Volumes from the computation log file:

[Anonymous]

The HEC-HMS time interval should be selected prior to RAS modeling. If you are trying to calibrate your hydrologic calculations via RAS precipitation modeling, then some of your solution parameters need to be kept constant until converging to a desired result.

For example, if you are using 1-min HEC-HMS intervals to generate hyetographs and then going to RAS and using 30×30 mesh and 1 min time intervals and are not satisfied with the results, you shouldnt mess around with all 3 parameters. Remember in a RAS precipitation model the incremental value of depth is utilized per cell. As cameron said, time intervals in HMS have extremely large impacts on output values. The same goes for RAS, ESPECIALLY in precipitation modeling, and with varying cell sizes AND boundary conditions.

What im trying to say is “inflow” shouldnt be the only thing looked at here. Look at the entire picture; inundation extents, necessary breaklines, outflow…% error…

[Davis Murphy]

Luis, I appreciate the reply and the various concerns you shared, but my question was specific in order to narrow the focus of the answers I received.

To be clear, I am not using anything from HEC-HMS except to handle the rainfall abstractions using curve number. HEC-HMS produces an excess precipitation hyetograph and that is all I need it to do. Everything else produced by HEC-HMS is discarded for the purposes of the analysis. The rest of the modeling is done in HEC-RAS, so the time increment used in the HEC-HMS control file is irrelevant to the analysis. At least, it should be, and that is why I asked the question because I was seeing a possible issue with the way that data was handled in HEC-RAS. Now that I have run some more tests, I have been able to show that hyetograph data time increment does NOT affect the inflow volume to the models I am running in 5.0.7.

With all of that said, I am satisfied that the initial question has been answered.

[Anonymous]

Make sure your initial abstraction is set to ZERO (0). That is always looked over

[Author]

Posts