Snowmelt vs Initial Abstractions 1

[SMIAH]

I'm trying to create a model to be able to calculate the runoff induce by precipitation (of both rain and snow). To be able to access both the volume of runoff and the peak discharges...

So i've created a HEC HMS model with subbasins, reaches, etc. I decided to use the Deficit and Constant method for calculating losses (i usually use SCS and CN number) and the SCS lag time for Transform.

I have in hands, both precipitation and temperature on an hourly basis which i've entered in the model so i can use the Temperature-Index method.

The model is calibrated for "storm-events" based on measured flow rate. And the model is giving good results (i'm almost surprised myself...)

So now ... I'm trying to simulate the snowmelt for a long term event (say on a yearly basis). The model runs and everything seems fine for the months where there's no snow (when compared to measured data).

BUT... When it comes to snowmelt... nothing. It's like if the program takes snowmelt in account but when it creates a precipitation from it to put it back in the model (thats how i see the process) the runoff induced is "taken" by the losses (mainly because the Initial abstractions > runoff).

What to do ? Lower the initial abstraction? Even if it's giving good results vs measured flow rates. Model separately months with and without snow?

Any ideas would be fine right now !

 

[Drew08]

On a yearly basis in regions that can get snow, Infiltration, CN and Ia will not be constant as the ground freezes and thaws. Unfortunately, I don't think HMS allows for a change in infiltration parameters with time. Try breaking the simulation time into discrete seasons. In winter when the ground is frozen, it will act more like an impervious surface (CN = 98). Some texts define this condition as ARC-IV.

 

[SMIAH]

Would you recommend another software do this analysis?

Thanks for your time

 

[Drew08]

I don't know of any program that allows variable infiltration parameters. Perhaps someone else out there does. I think you may need to break the model into seasons or months. HEC-HMS is good for this purpose because it allows for more than one basin model and more than one control specification model (one for each set of parameters). Since you have hourly temperature data, try developing 30-day running averages or equivalent. That may help identify good places to break the model and change parameters (e.g. 30-day running average is > or < 32F, 0C).

Unless you have reservoirs in your model that need initial stages from previous segments, each time segment should be mostly independent from the previous segments.

 

[psmart]

With HydroCAD you can change the Ia/S ratio, the default being Ia= 0.2 S

You can also change the AMC/ARC to III (which raises all CN values for wet conditions) or IV (which increases all CN values to 98 for impervious or frozen ground)


Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[Drew08]

Psmart,
Can HydroCAD automatically change Ia/S and ARC at predetermined times during the simulation? Or do they have to remain constant throughout the simulation?

I've only ever used the free version, and not very much at that. But, I never saw an option for time variable infiltration parameters.

 

[psmart]

Drew08 - Those parameters are constant throughout a given (24-hour) run. Frankly, I don't see how they could be changed during a single event, since it would cause an instantaneous change in the cumulative runoff volume at the time of the change. If you want to change them between different runs, that's easy to do.



Peter Smart
HydroCAD Software

http://www.hydrocad.net

 

[Drew08]

I agree during a single 24-hour event, the infiltration parameters won't (and can't) change. But, I believe the original problem presented by SMIAH was for a 1-year duration simulation, encompassing multiple events with changing seasonal conditions.

 

[SMIAH]

I believe, as you already pointed out, that i should cut the continuous simulation into a multiple event's one.
Or change the Infiltration method.