ROMS blows up "DIAG speed trouble"

[stevevangorder]

Hello,

I'm running a single grid ROMS-only COAWST model of the GOM. The model blows up after 1 time step with a "DIAG speed trouble"
error. A plot of the model domain is attached and below are what seem to me to be relavent sections of the standard output file.
The entire standard out is attached.

The only problem I see is that the max barotropic Courant number (1.13952908E+00) is too large. My understanding is that it
should be lowered to something in the range .8 to .85. The large Courant # likely occurs in the very deep region near (80W, 19S)
but the model appears to be blowing up in 2-m water at (at least) two different widely separated locations i,j,k = (873,374,1)
and i,j,k = (370,607,1) as indicated by the output from diag.F below. The deep water barotropic signal propagates at something
like 40-50 Km in 1 3D time step, so I don't see how it could have reached these shallow areas. Does anyone see other problems
that I am missing?

I'd like to see the output at time step 1, but the model only writes the initial time step before it terminates. I have set the
write interval to 1 for all the output files but that doesn't help. Is there a way to force ROMS to write out time step 1?

Does anyone have a better idea than increasing ndtfast to reduce the barotropic Courant number? Filling in the deep areas to 5000
m and/or removing those areas from the domain would only reduce it to .9. Reducing it to .8 would require increasing ndtfast from
30 to 43 at the expense of the model run time.

Thanks for your help.

Steve

----------------------------------------------------------------------------------
Resolution, Grid 01: 1079x0767x032, Parallel Nodes: 8, Tiling: 004x002

Physical Parameters, Grid: 01
=============================
180.000 dt Timestep size (s) for 3-D equations.
30 ndtfast Number of timesteps for 2-D equations between
each 3D timestep.
2 Vtransform S-coordinate transformation equation.
4 Vstretching S-coordinate stretching function.
1.0000E+01 theta_s S-coordinate surface control parameter.
3.0000E+00 theta_b S-coordinate bottom control parameter.
2.000 Tcline S-coordinate surface/bottom layer width (m) used
in vertical coordinate stretching.

Vertical S-coordinate System, Grid 01:

level S-coord Cs-curve Z at hmin at hc half way at hmax

32 0.0000000 0.0000000 0.000 0.000 0.000 0.000
31 -0.0312500 -0.0000141 -0.031 -0.031 -0.120 -0.177
30 -0.0625000 -0.0000578 -0.063 -0.063 -0.359 -0.593
29 -0.0937500 -0.0001355 -0.094 -0.094 -0.736 -1.285
28 -0.1250000 -0.0002547 -0.125 -0.125 -1.281 -2.313
27 -0.1562500 -0.0004271 -0.157 -0.157 -2.042 -3.772
26 -0.1875000 -0.0006698 -0.188 -0.188 -3.088 -5.800
25 -0.2187500 -0.0010066 -0.220 -0.220 -4.514 -8.590
24 -0.2500000 -0.0014704 -0.251 -0.251 -6.455 -12.410
23 -0.2812500 -0.0021068 -0.283 -0.283 -9.094 -17.627
22 -0.3125000 -0.0029780 -0.315 -0.315 -12.685 -24.746
21 -0.3437500 -0.0041691 -0.348 -0.348 -17.572 -34.456
20 -0.3750000 -0.0057960 -0.381 -0.381 -24.223 -47.697
19 -0.4062500 -0.0080164 -0.414 -0.414 -33.278 -65.744
18 -0.4375000 -0.0110444 -0.449 -0.449 -45.603 -90.332
17 -0.4687500 -0.0151693 -0.484 -0.484 -62.371 -123.806
16 -0.5000000 -0.0207812 -0.521 -0.521 -85.162 -169.324
15 -0.5312500 -0.0284030 -0.560 -0.560 -116.092 -231.122
14 -0.5625000 -0.0387299 -0.601 -0.601 -157.977 -314.831
13 -0.5937500 -0.0526769 -0.646 -0.646 -214.524 -427.862
12 -0.6250000 -0.0714305 -0.696 -0.696 -290.537 -579.825
11 -0.6562500 -0.0964959 -0.753 -0.753 -392.112 -782.913
10 -0.6875000 -0.1297235 -0.817 -0.817 -526.744 -1052.114
9 -0.7187500 -0.1732801 -0.892 -0.892 -703.207 -1404.978
8 -0.7500000 -0.2295082 -0.980 -0.980 -930.989 -1860.479
7 -0.7812500 -0.3005917 -1.082 -1.082 -1218.934 -2436.306
6 -0.8125000 -0.3879254 -1.200 -1.200 -1572.691 -3143.757
5 -0.8437500 -0.4911237 -1.335 -1.335 -1990.698 -3979.709
4 -0.8750000 -0.6067588 -1.482 -1.482 -2459.073 -4916.397
3 -0.9062500 -0.7272825 -1.634 -1.634 -2947.247 -5892.682
2 -0.9375000 -0.8411194 -1.779 -1.779 -3408.340 -6814.804
1 -0.9687500 -0.9351759 -1.904 -1.904 -3789.323 -7576.709
0 -1.0000000 -1.0000000 -2.000 -2.000 -4051.918 -8101.836

Time Splitting Weights for Grid 01: ndtfast = 30 nfast = 42
ndtfast, nfast = 30 42 nfast/ndtfast = 1.40000

Metrics information for Grid 01:
===============================

Minimum X-grid spacing, DXmin = 1.88519400E+00 km
Maximum X-grid spacing, DXmax = 2.11257511E+00 km
Minimum Y-grid spacing, DYmin = 1.88620079E+00 km
Maximum Y-grid spacing, DYmax = 2.11234566E+00 km
Minimum Z-grid spacing, DZmin = 3.12641131E-02 m
Maximum Z-grid spacing, DZmax = 9.76284856E+02 m

Minimum barotropic Courant Number = 1.78492731E-02
Maximum barotropic Courant Number = 1.13952908E+00
Maximum Coriolis Courant Number = 1.38957237E-02

Basin information for Grid 01:

Maximum grid stiffness ratios:
rx0 = 1.000002E-01 (Beckmann and Haidvogel)
rx1 = 3.966128E+00 (Haney)

Initial domain volumes: TotVolume = 4.1050194260E+15 m3
MinCellVol = 1.1287869071E+05 m3
MaxCellVol = 4.3023991784E+09 m3
Max/Min = 3.8115247008E+04

TIME-STEP YYYY-MM-DD hh:mm:ss.ss KINETIC_ENRG POTEN_ENRG TOTAL_ENRG NET_VOLUME
C => (i,j,k) Cu Cv Cw Max Speed

0 2010-04-13 00:00:00.00 1.141082E-02 1.568609E+04 1.568610E+04 4.106616E+15
(0920,704,14) 9.835437E-02 1.573447E-01 0.000000E+00 2.127229E+00
1 2010-04-13 00:03:00.00 1.136669E-02 1.568603E+04 1.568604E+04 4.106607E+15
(0873,374,01) 1.350519E+00 1.055360E+00 2.589530E-04 2.500017E+01

DIAG speed trouble 25.0001661151840
DIAG speed ijk 25.2612649921961 370 607 1
Found Error: 01 Line: 351 Source: ROMS/Nonlinear/main3d.F
Found Error: 01 Line: 320 Source: ROMS/Drivers/nl_ocean.h

ROMS/TOMS - Blows up ................ exit_flag: 1

MAIN: Abnormal termination: BLOWUP.
--------------------------------------------------------------------------------------------

[kate]

Blowing up on the first step with this:

Maximum barotropic Courant Number = 1.13952908E+00

points to needing a shorter barotropic timestep.

[susonic]

First of all, check out the barotropic time step as Kate mentioned above.

Secondly, I saw several weird option with your boundary condition.

I saw gradient condition for temp and salt at the northern boundary.
For West, I saw gradient condition for temp and salt(Not closed).
I was wondering whether you set up this condition intentionally.

-JH

[stevevangorder]

Kate & JH

I'm checking out the tine step issue

I used the same boundary conditions as the COAWST Hurricane Sandy example which is also all land on the western boundary and partially open on the other boundaries. I assumed that the gradient condition specifies the temperature and salinity gradients the on the west boundary from the "bdy" forcing file, but I'll admit that I don't know really know. Do you think this mix of BCs is a problem?

Steve

[kate]

If you have a closed boundary, it should be closed to all variables. If you have an open boundary and a source of boundary values, I would use them for all variables. But really, look at your boundaries - do you think you have a problem there?