HEC-RAS Bank Stations 1

[gopher669]

I'm having to model a roadside ditch to satisfy a reviewer for a proposed culvert. I've been told by the reviewer that "The channel bank stations should be at the Ordinary High Water Mark, not at the top of bank" (which is how I've always done it). In essence, moving the Bank Stations from the red dots, down to the red Xs in the attached sketch. Just to satisfy my curiosity, I took my model, copied it and switched bank stations, and re-ran the model. Just by switching the bank stations, the WSEL went down, by about 0.4' (of an 8' deep ditch). Mannings was already set by station, not LOB/Channel/ROB, so that‘s not a factor.

Anyone able to tell my why the WSEL went down, and whether I should or should not move these bank stations?

 

[francesca]

HEC-RAS is a one-dimensional model and divides the flow into left overbank, channel, and right overbank flow areas. It computes a friction value for each of the three, distributes the water to get equivalent WSE (IIRC) and then outputs a WSE for the cross-section. Something in the balance between the new average friction values for the overbanks/channel changed up the mix of what WSE is the magic equilibrium.

FWIW, I have never seen the bank stations set for OHW, only for top of bank. It certainly isn't how FEMA does it in my experience (albeit limited). Theoretically, flow in the channel is deep, and the velocities are higher. With the section you have, I would put the bank stations at the dots, not the Xs. There is plenty of depth for high velocities and channel flow between the X and the dot. However, if That's How We Do Things Around Here, it's an uphill battle. They will want consistency across their models.

 

[Drew08]

Recently, I completed a model for a temporary stream crossing where small flows were much more critical than a 100-year flood which is highly unlikely to occur given the short duration of the project (reviewer excepted this). Since the low flows never reached the top of the bank, I put the "bank stations” at the top of the channelized portion of the flow area. I wanted to take advantage of the mathematical separation of very different hydraulic conditions between the rocky thalweg and grassy shallows. It didn't matter though; the reviewer has a box to check that says the "bank stations” are at the top of the bank regardless if the design flow even reaches the top of the banks. I put quotes around the word "bank station”, because it's just a name, and because it’s that name, it's ingrained in peoples thinking that it has to be set at the top of bank, regardless of what is mathematically more appropriate. I’m heading the effort to informally rename the “bank station”, the “point of mathematical separation of flow”, and maybe obtain a little more flexibility in its use. OK, rant over.

 

[gopher669]

I could see putting the bank stations at the OHWM if I were modeling a smaller storm event; however, this was for a 100 year storm, which was (probably not coincidentally) just barely contained in the channel. After my initial post, I did a little more research into the RAS calculations and how RAS uses conveyance in the overanks and channel in the calculations. I manually calculated the conveyance for the two bank stations situations and came up with a higher conveyance (and therefore lower WSEL) by moving the bank stations in. All has to do with the hydraulic radius - the original channel with A=83 and WP=26.1 will give a different conveyance than the three separate A=17.2 WP=8.4, A=50.2 WP=9.4, and A=15.6 WP=8.3. Even though the areas and wetted perimiters are the same, the hydraulic radius and therefore conveyance is different.

It is more of a "that's how we do things" and I know it was an unwinnable battle. Said agency has been overly fixated on OHWM and Bankfull for some time. Fortunately, it was for a small section of the roadside ditch just to show the effects of a proposed culvert, and not a larger scale project where I'm establishing BFEs, so whichever bank stations I used in the proposed was also going to be reflected in the existing. I gave the reviewer what they wanted, along with a little explanation of conveyance and how it was lowering WSELs by moving the stations in.

 

[gbam]

gopher669;
As I recall, the default setting in HECRAS computes the average conveyance for each segment of cross section (LOB, MCH, ROB). Unlike HEC2 where it computed the conveyance between each ground point then averaged for the water surface limits. My thought is that if you have segmented Manning's through the main channel it will average the conveyance based on the distributed Manning's and when you squeez the bank stations your roughness/conveyance changed. Is there still the trigger to compute conveyance like HEC2 in the latest version of HECRAS?

My thought is that bank stations should be geometrical and not dependant on the OHWM or any given water surface. It is important to identify these differences to correctly model the hydraulics of any channel.