Copyright © The RAS Solution 2017. All rights reserved.
Expanding upon Chris’ discussion of where to place bank
stations, what should you do about high terrain somewhere in the middle of your
cross section? Here is an example:
How should we treat the left overbank? It’s hard to tell if the high area next the
left bank is isolated (i.e., it would be an island if the water surface were to
get to elevation 370 or so) or if it is a continuous feature (such as a levee)
that would prevent flow from accessing the left overbank until it is
overtopped. Looking at this another way,
is the lower ground of the left overbank a continuous flow path or is it
an isolated low spot (for example, a mining pit)? Aerial photography can often help
determine the situation; here is the overhead view for our example:
The area in question is vegetated (the terrain goes up
steeply when it gets to the storage yard on the bottom of the photograph) but
it is hard to tell if the high point in the terrain would be constraining flow or
if the low area is a potential flow path.
Looking at the cross sections upstream and downstream of the one in
question will often provide answers, but does not help in this
particular example. In this case, the
best course of action would be to go out to the river and see for yourself,
then imagine how the water would behave.
Depending on your conclusion, there are several ways that this can be modeled.
1. Isolated high spot. If flow can simply go around the high spot in
this particular cross section then we probably don’t need any further
adjustments. You may get a “divided flow” warning in the output that signifies
that the program detected dry ground with water on either side, but no action
is needed to address the warning in this case. Assuming that the computed water
surface elevation is high enough, this solution will also allow flow in the
left overbank.
2. Isolated low spot in overbank. You could model this as per #1 above but in that case you should check flow distribution between the channel and overbanks up and downstream of this cross section for reasonable transitions (see earlier blog post from May 20, 2009). Or, if you think that the low area should only store but not convey water you could set an ineffective flow limit as shown below.
3. Continuous high ground. If the high ground is really a ridge that
would prevent the water from accessing the lower ground in the left overbank,
it should be modeled as a levee. However, in this case another decision needs
to be made depending on what happens after the levee is overtopped. Will the
water be conveyed on the land side, or will it just pond? If the latter you may need to add an
ineffective flow limit at or just to the left of the levee.
4. Something in between. Regardless of whether the high or low
features of the cross section are continuous, water is able to access the left
overbank. Natural streams often have
“backswamp” areas behind either human-made or natural levees that flood and
store water but do not really convey much flow downstream. If the left overbank
in our example is like this, we could model it by using the ineffective flow
limit as per #2 above. However, ineffective flow means zero
conveyance. If we expect some water to
move in the overbank, albeit very slowly, you may want to allow a small
non-zero conveyance. A few sharp-eyed
readers may have noticed that we are using a Manning’s roughness coefficient of
0.3 in the left overbank. Using this value allows a small amount of conveyance in
that overbank without zeroing it out completely.
if you were to set your ineffective flow elevation to something higher than the ground, how would it differ from #3 (making it a levee)?
ReplyDeleteOnce it overtops the high ground water can be on the other side. If the ineffective flow elevation is higher than the high ground point, then the water on the other side will remain ineffective even after the levee is overtopped. It will remain so until the ineffective flow trigger is exceeded by the water.
DeleteChris, I'm working on a Dam Break study, in which the floodplains are crammed with such "backswamps". They are like the "isolated low spots in overbanks", but these spots create continuous paths that follow the river for something about 500 m or so until they reach some higher ground in the overbank (they are potentially floodable areas with lower elevations than average water surface elevation - about 3 feet or so - and I don't know the technical term for them in english). From your post, I've decided the best way to model it is by using ineffective flow areas (permanent) only for the cross sections that contains these "backswamps". By doing nothing, considering just the condition of divided flow in higher stages, my flood wave attenuation was below expected. But the problem is that by considering these ineffective flow areas, the peak discharges downstream went higher and I've got even less attenuation. I expected the peak discharges would fall dramatically, since these ineffective flow areas are large and they act like a series of storage volumes, not accounting for conveyance during the flood wave. Well it didn't happen. So I've increased Manning n coefficients for theses "pools" and finished the job. Do you have any idea why this happen? Or am I just doing something wrong?
ReplyDeleteThanks!
Is it okay to have a bank sta elevation significantly higher on one side(1 ft difference)?
ReplyDeleteYes. That’s no problem
Delete