open TELEMAC-MASCARET The mathematically superior suite of solvers

We (Cerema + SPC + SCHAPI) are currently working on the behaviour of mascaret models in the Loire, Seine, Adour, Canche estuaries. Many efforts have been achieved to get models able to reproduce tidal wave propagation in the estuary without success. One of the common behaviour of these models is the incapacity to reproduce the dynamic tide propagation.

First we thought that the reason was a bad imposition at the downstream boundary condition.
We tried to move seaward the downstream boundary condition (to generate a larger oscillating mass of water)… but the results were not improved.
We also tried to change the Strickler coefficient but never get physical value and good results.



Analyzing and comparing the models showed us that model with only low flow channel (minor river bed) gave the best results. Removing high flow (major river bed) channels improve models behaviour.

We are now convinced that the reason for our problem is linked with the progressive overflow.
As we can’t find information on how the progressive overflow works, we tried to understand it on very simple cases.

We get very strange results and definitely impenetrable...

Let’s start with simple models:

Description of the simple case :

- hydraulic network made of a single reach
Compound channel made of the main river bed (low flow channel) separated from high flow channel (or storage area) by a levee (crest of the levee at 3m).
length : 50km
slope of the low flow channel : (1m in 50km = 0,002 %)

Initial conditions :
water level bellow the floodplain (high flow channel). Z water = 1,0m (bellow the lowest elevation in the floodplain)
→ Its means that the high flow channel is not activated in any case (floodplain lowest point elevation is 1,1m)


Boundary conditions :
- inflow hydrograph upstream (50m3/s)
- stage curve downstream (1,1m < peak water level < 3m)
→ the levee is not submerged.

Use of fluvial kernel. Strickler = 35...


Case 1 :
Just the low flow channel. No floodplain in the geometry.
Case 2 :
The floodplain (high flow channel) is added in the geometry. Flat floodplain separated from the low flow channel by a levee. Crest of the levee at 3,0m.
Progressive overflow in high flow channel inactivated.
→ That is meaning that the high flow channel convey water when water level > 1,1m

As expected, when water level > 1,1m high flow channel start conveying water.

Case 3 :
Same geometry. Activation of progressive overflow in high flow channels.
Progressive overflow in high flow channel activated.
→ That is meaning that the high flow left channel should be activated when the water level overflow the levee... > 3m

/!\ When we analyse in detail... No discharge in the high flow channel.
→ we should have case 3 = case 1
but discharges in low flow channel are not the same???

Case 4 :
Almost same geometry. The downstream cross section is modified to remove the high flow channel.
Progressive overflow in high flow channel activated.
- It is just to verify if the high flow channel is not activated by the downstream cross section
→ Case 4 ~ Case 3

Case 5 :
The high flow channel is replaced by a storage zone.
Progressive overflow in storage zone activated.
→ Case 5 ~ Case 1

Can anyone help me understanding how progressive overflow works?
Why when the high flow channel is not activated (case 3 and 4 : Q floodplain = 0), the low flow channel discharge is modified?

Thanks beforehand.

Hello,
There seems to be a bug in graph's results, if section 1 is not recorded as results. Try to record results in section abs 0, or record results in all sections. Graphs seems to be different, is that better ? (I tried it on your case 3, attached file)
Good luck!

Thank you for your hint.

Fortunetly Fudaa post-processing doesn't affect Mascaret.

Even if results look better (graphicly) it does not change anything.

I agree with you : once z>1.10 m (bottom of the high flow channel) section in high flow is considered as storage area (see col DSS in attached file) despite levee level is 3m!
No discharge on overbank but storage area, so discharge in the main channel is lower! I will seek if there is something on mascaret notice...

OK,

This seems to be an unappropriate activation of the storage in the hight flow channel with use of the option "progressive overflow in hight flow channel".

The case 5 is better. The hight flow channel is replaced by a storage zone and in that case the progressive overflow works as expected.

Anyway, when water level is between 1m and 3m, water should just be in the low flow channel.

Hi,

Just to let you know where we are with this problem.

Thanks to mathieu.alquier, we have reported the problem.

Activation of progressive overflow in hight flow channel doesn't work as expected when the hight flow channel is on the left.

You can find two cases case A and B with the same geometry except that the cross profiles are inverted.

You can see that :
case A is working (no hight flow channel activation under 3m meither for conveyance or storage)
case B activate hight flow channel between Z=1.1m and 3m just for storage (it shouldn't).