Coast / Estuary

Coastal systems are among the most dynamic physical systems on earth and are subject to a large variety of forces. The morphodynamic changes occurring to coastlines worldwide are of great interest and importance. These changes occur as a result of the erosion of sediments, its subsequent transport as bed load or suspended load, and eventual deposition. 
Estuaries are partly enclosed water bodies that have an open connection to the coast. Estuaries generally have one or more branching channels, intertidal mudflats and/or salt marshes. Intertidal areas are of high ecological importance and trap sediments (sands, silts, clays and organic matter).
Within the Delft3D modelling package a large variation of coastal and estuarine physical and chemical processes can be simulated. These include waves, tidal propagation, wind- or wave-induced water level setup, flow induced by salinity or temperature gradients, sand and mud transport, water quality and changing bathymetry (morphology). Delft3D can also be used operationally e.g. storm, surge and algal bloom forecasting. 
On this discussion page you can post questions, research discussions or just share your experience about modelling coastal and/or estuarine systems with Delft3D FM. 




Sub groups
D-Flow Flexible Mesh
Cohesive sediments & muddy systems


« Back to Coastal/Estuary

RE: Morphodynamics of a 1D river: different output using 1D/2D domain

Combination View Flat View Tree View
Threads [ Previous | Next ]
Dear all,

Recently I started using Delft3D to model the morphodynamics of rivers/estuaries.

At the beginning, I tried to simplify the research problem as much as possible, i.e.,
using a rectangular domain to represent a tidal-river, which is affected by the river/sediment discharge and the tides.

Yangtze River was used as a prototype. Simplified parameters are:
a domain size of 3*560 km (rectangular), initially the bottom has a uniform slope,
in the upstream a water discharge of 30000 m^3/s and sediment concentration of 0.5 kg/m^3,
in the downstream a M2 tide with an amplitude of 2.5 m,
and non-cohesive sediment with a grain size of 0.1 mm.
Slip boundary condition was used. Latitude and longitude are 0.

First I used only 1 grid (dy = 3 km) in one horizontal direction (y-direction) and 400 grids (dx = 1.4 km) in the other horizontal direction (x-direction).
Smooth results were obtained in the end of the simulation (5 hydrodynamic years, with a morphological acceleration factor of 100).

To make sure the above configuration (only 1 grid in the y-direction, default case) gives reliable results, next I used more grids in the y-direction,
e.g., 2, 3, and 5. It turns out that the results using more than 1 grid in the y-direction differ a lot from that in the default case! In the cases with
more than 1 grid in the y-direction, oscillations in the results were observed. Two figures (time evolution of the bed level of a slice in the x-direction)
were attached to show the differences between those cases. One is from the default case, the other is from the case with 3 grids in the y-direction
(the middle slice was shown).

Two questions:
1) given the same settings except the number of grids, what could be the causes for the different results in the above cases?
2) in the cases with multiple grids in the y-direction, for the 1D configuration of the model, it is supposed that the results would
be at least symmetrical about the middle line of the river. However, that is not the case (not shown, but comparison was done
between different slices in the y-direction), how could that be?

Any reply would be appreciated. Thank you!

Kind regards,
Attachments: h_ny2_YZ_rec_check1D1p.pdf (29.0k), h_ny3_YZ_rec_check1D3p.pdf (28.9k)
RE: Morphodynamics of a 1D river: different output using 1D/2D domain
bed level instability morfac
Answer (Unmark)
11/15/17 8:53 PMas a reply to Bing Yuan.
Hi Bing,

Interesting question! I'm don't think I have the answer but here are a few thoughts in no particular order:

1) It looks to me that the oscillations could also be present in the 1D configuration, just with much lower temporal and spatial frequency? There appears to be a bump moving downstream as the bed erodes. Or maybe this is sensible?
2) Given that it appears to be a issue with morphological stability you may want to try either reducing your timestep or your MorFac. I suggest doing a sensitivity test with a reduced MorFac - ideally the model should give results independent of MorFac (for the same Morphological time) but I suspect you may get different (smoother?) results? If so this suggests you need to reduce MorFac.
3) Regarding your question 2, this does seem strange but I guess it only needs one tiny asymmetric perturbation to introduce asymmetry if the model is unstable.
4) Is the instability in time as well as space, i.e. for a given cell does the bed level oscillate up and down over time? You could use observation points to monitor this at a higher temporal frequency.
5) Do you get any warning messages relating to the instability?

It would be great to hear how you get on and what solutions you find.

RE: Morphodynamics of a 1D river: different output using 1D/2D domain
Answer (Unmark)
11/16/17 3:12 AMas a reply to Richard Measures.
Hi Richard,

thank you for your suggestions!

Some replies to your thought:

1) in the default case (one grid in the y-direction), an equilibrium state (changes in the bed level is subtle) was achieved in a longer morphological time (2000 years), although I used a larger morphological acceleration factor MF=400.

2) I did try using smaller MF of 50 and 10 for cases using multiple grids in the y-direction. Unfortunately, at similar morphological time, the bed level near the sea boundary (M2 tide) started to oscillate, and the bed level in the y-direction varied. One thing I just found out is that at the beginning of the oscillation, the bed level was symmetrical about the middle line along the river.

ps. I understand that 2D features could appear in a 2D domain due to nonlinear effect even if the forcing is uniform, but I am not sure if this is the case for the present tests, as the amplitude of the oscillation in the bed level at the beginning was small compared to the water depth.

3) I agree that a tiny asymmetry in the y-direction would result in asymmetric profiles in the y-direction. Quickin was used to generate the file for the initial bed level.

4) I have not checked the temporal evolution of points yet, but could have a look later.

5) No messages related to instability...

RE: Morphodynamics of a 1D river: different output using 1D/2D domain
Answer (Unmark)
11/21/17 12:28 PMas a reply to Bing Yuan.
The asymmetry of the bed level with respect to the middle line along the river might be due to the formation of channel-shoal pattern in a 2D domain, see van der Wegen and Roelvink (2008) and refer. therein. However, notice that in their study, initially a random perturbation of the bed level was applied to trigger the formation of the channel-shoal pattern, whereas in the test here initially no explicit perturbation was applied to the bed level.

Is there anyone with similar experiences as I mentioned in the initial post? Or someone might be interested to conduct similar tests?
RE: Morphodynamics of a 1D river: different output using 1D/2D domain
Answer (Unmark)
11/29/17 12:45 PMas a reply to Bing Yuan.
Hi Bing, Richard,

Here I just put some thoughts here, not necessarily correct.

I would be able to imagine if you have more than 1 cell in y-direction (cross river direction), the morphology results would be different. For 2 reasons, the first: for 1 cell, the uniform velocity cross river is forced, but not for more than 1 cells. You mentioned you have slip boundary condition. A small precision difference will lead to an asymmetry flow profile across the river. The second, due to the nature of ADI algorithm, the computation is carried out from one direction to another direction, which makes the symmetric bedform to the middle line of the river not so easy to achieve. Refer to Bert Jagers's PhD thesis for detailed information.

I hope this will be a little bit helpful,


RE: Morphodynamics of a 1D river: different output using 1D/2D domain
Answer (Unmark)
12/7/17 2:35 AMas a reply to Qinghua Ye.
Hi Qinghua,

thank you for your reply.

Yes, the numerical precision and the numerical scheme (ADI) might be the cause for the asymmetry in the results.

Also thanks for the reference.

Kind regards,