salinity goes negative and large in a river plume testing

[FengZhou]

Dear all,

I am running a coarse realistic case (18km*18km) including a large fresh water inflow (annual mean is about 30000 m^3/s, largest is close to 50000m^3/s in normal years). In this case, I focus on the river plume and its interaction with ambient water. Before, a general simulation has been finished, results seems to be acceptable.

The model keep stable in one year (what a surprise!) while the salinity completely goes wrong and unrealistic: negative and large value (large than 40) - they comes up together:(

I have searched in the forum and found many good experiences concerning the river plume, but there is no such specific solution on my problem. Could you please help me diagnose or share your solution?


Model Input Parameters: ROMS/TOMS version 3.0
Friday - March 21, 2008 - 4:34:39 PM
Operating system : Linux
CPU/hardware : i686
Compiler system : pgi
Compiler command : mpif90
Compiler flags : -u -Bstatic -fastsse -Mipa=fast -Kieee -Mfree

Resolution, Grid 01: 0083x0121x020, Parallel Nodes: 4, Tiling: 002x002

155520 ntimes Number of timesteps for 3-D equations.
400.000 dt Timestep size (s) for 3-D equations.
30 ndtfast Number of timesteps for 2-D equations between
each 3D timestep.


mask_u(row, column):
22 ->
50 0 0 0 0 0 1 1
51 0 0 0 1 1 1 1
52 0 0 1 1 1 1 1
53 0 1 1 1 0 0 0
54 0 1 0 0 0 0 0
55 0 0 0 0 0 1 1

mask_v:
20 ->
50 0 0 0 0 0 1 1
51 0 0 0 1 1 1 1
52 0 0 1 1 1 0 0
53 0 1 1 0 0 0 0
54 0 1 0 0 0 0 0
55 0 0 0 0 0 1 1

(The first column denotes row number)
Input river source at 11 points are defined as mentioned by Kate etal in

viewtopic.php?t=831&highlight=river

with monthly mean discharge. The data prepared into forcing files is like :

River =

Name: [11x19 char]
Xpos: [22 22 23 24 26 27 27 24 25 26 27]
Ypos: [53 54 52 51 50 49 48 54 53 53 52]
dir: [0 0 0 0 0 0 0 1 1 1 1]
num: [1 2 3 4 5 6 7 8 9 10 11]
flag: [3 3 3 3 3 3 3 3 3 3 3]
time: [36x1 double]
trans: [11x36 double]
vshape: [11x20 double]
temp: [11x20x36 double]
salt: [11x20x36 double]

The vertical profile is as, which I suppose the river flows out in 9 top layers, 20%, 20%, 20%, 20%, 10%, 5%, 3%, 1%, 1%.

The second question is that the most of the outflowing fresh water goes northward not southward, which is against the real. I have found some references but seems no discussions here? Ok, one more post about this might be better.

Thank you!

zhou

[FengZhou]

I have just tried the TS_MPDATA option. There are no negative values even in July after 7 month simulation. However I have checked the bottom salinity that the largest salinity exceeds 70 psu. It occurs in both cases :

1st case : The frseh water is only specified on the top 15 layers, almost linearly decreased from the top layer.

2nd case: averagely distributed in all 20 vertical layers.


zhou

[babsona]

Are you using steep bathymetry? I found that using steep bathymetry with the MPDATA option led to a similar increase in salinity (at the intersection with the bathmetry, but it immediately sinks to the bottom). I guess this is a problem for stretched coordinate models (see Allen et al (2003) for a full explanation). My work-around was to switch to vertical sidewalls where I could. Good luck, Amanda

[FengZhou]

Hi, Amanda,

The bathymetry is steep only at other grids far away from the estuary. While the large salinity value occrus only at the inlet of river, especially on lower levels. The top 5-10 layers sound not weird.