Unrealistic 1 m/day Sediment Deposition in River Channels During Floods

[SongCZ]

Hello,
I am currently simulating sediment transport during extreme river floods with a suspended sediment concentration (SSC) of approximately 35 kg/m³ using the ROMS model. However, I am encountering an issue where the model predicts an unrealistic amount of sediment deposition in river channels, reaching up to 1 meter per day during high-concentration events. That is to say, the bed_thickness will increase ~1 meter one day.This excessive deposition severely limits the transport of sediments to the coast.
What I Have Tried:
I enabled the UV_DRAG_GRID option and increased the river channel bottom drag coefficient (rdrag2) by a factor of 4. Unfortunately, this did not improve the situation.
I am looking for ways to reduce this unrealistic deposition. I would greatly appreciate any insights or suggestions on how to address this issue. Thank you in advance for your help!
Attached is my .h document.

[jcwarner]

you are not providing a lot of info to help.
The 35 kg/m3 is a lot (!) of suspended sediment.
How is the model getting that much of a concentration? is this from resuspension? or from a lateral boundary?

[SongCZ]

The high concentration of suspended sediment (35 kg/m³) observed in the model is primarily attributed to the riverine inputs. These inputs are implemented through river forcing, which is applied at the land-sea boundary within the river channel. Specifically, the river forcing is imposed via a SSF file (runoff data file). This SSF file contains the necessary data (such as river_transport;river_sand_01;river_sand_02...)to simulate the river transport and sediment concentration at three specific grid points along the river channel.At these designated grid points, the model continuously releases runoff and sediment, with the data derived from hydrological stations.

[SongCZ]

you are not providing a lot of info to help.
The 35 kg/m3 is a lot (!) of suspended sediment.
How is the model getting that much of a concentration? is this from resuspension? or from a lateral boundary?

The high concentration of suspended sediment (35 kg/m³) observed in the model is primarily attributed to the riverine inputs. These inputs are implemented through river forcing, which is applied at the land-sea boundary within the river channel. Specifically, the river forcing is imposed via a SSF file (runoff data file). This SSF file contains the necessary data (such as river_transport;river_sand_01;river_sand_02...)to simulate the river transport and sediment concentration at three specific grid points along the river channel.At these designated grid points, the model continuously releases runoff and sediment, with the data derived from hydrological stations.

[jcwarner]

ok. check the units of the data, make sure it is not mg/l.
Are you setting that 35 kg/m3 as the same concentration in the entire water column?
Values that high would probably only be near the bed.

[wilkin]

Looking at the last few options in your CPP definitions file ...

Code: Select all

/* bottom boundary layer closure,define only one of the following 5 */
#define UV_DRAG_GRID
#undef UV_LOGDRAG /*use to turn ON or OFF logarithmic bottom friction*/
#undef UV_QDRAG /*use to turn ON or OFF quadratic bottom friction*/
#undef MB_BBL /*use if Meinte Blaas BBL closure*/
#undef SG_BBL /*use if Styles and Glenn (2000) BBL closure*/
#undef SSW_BBL /* Sherwood et al. BBL closure */

... you have no BBL model defined, and neither LOG nor QUADRATIC drag.

If you look in Nonlinear/set_vbc.F where the bottom boundary condition for momentum (i.e. stress) is set you will see that for your options you will inside the block #ifndef BBL_MODEL (for BBL not defined) and then in neither of the options for UV_LOGDRAG or UV_QDRAG. Since I don't see that you've defined UV_LDRAG I'm pretty sure you exit that routine with zero bottom stress. With no stress there will be no resuspension of settling sediment.

You do have #define UV_DRAG_GRID to read spatially varying drag coefficients from the grid file. But if you look for UV_DRAG_GRID in the code (e.g. with grep -f UV_DRAG_GRID .) you;'ll see that in Utility/get_grid.F a block of code ...

Code: Select all

#  if defined UV_DRAG_GRID && !defined ANA_DRAG
#   ifdef UV_LOGDRAG
!  Read in spacially varying bottom roughness length (m).
..
#  ifdef UV_QDRAG
...

You still need UV_QDRAG or UV_LOGDRAG etc. to activate reading and using of the chosen drag scheme.

Look in your log file for a report that those coefficients were read. I suspect you will find they were not.