COAWST sediment problem

[wrh]

Dear all:
Does anyone have similar problem like this : when the sediment modular activated with SED_DENS defined as following:
#define SEDIMENT
#ifdef SEDIMENT
# define SUSPLOAD
# undef BEDLOAD_SOULSBY
# undef BEDLOAD_MPM
# define SED_DENS
# define SED_MORPH
# undef SED_BIODIFF
# define NONCOHESIVE_BED1
# undef NONCOHESIVE_BED2
# undef COHESIVE_BED
# undef MIXED_BED
#endif

then the model got NAN value at the first step of running then blowing up. So what's the problem? while with the SED_DENS deactivated, model can run smoothly
STEP Day HH:MM:SS KINETIC_ENRG POTEN_ENRG TOTAL_ENRG NET_VOLUME
C => (i,j,k) Cu Cv Cw Max Speed

0 1 00:00:00 2.696751E-03 6.796553E+03 6.796555E+03 2.193817E+14
(142,044,30) 2.344050E-02 3.724131E-03 0.000000E+00 3.489128E-01
DEF_HIS - creating history file, Grid 01: myproject/nscs/input/sed/out/ocean_his_01_0001.nc
WRT_HIS - wrote history fields (Index=1,1) into time record = 0000001
DEF_AVG - creating average file, Grid 01: myproject/nscs/input/sed/out/ocean_avg_01_0001.nc
DEF_FLOATS - creating floats file, Grid 01: myproject/nscs/input/sed/out/ocean_flt.nc
1 1 00:02:00 2.710639E-03 NaN NaN 2.179343E+14
(173,291,02) 4.779574E-03 8.333636E-03 1.270614E-01 1.183538E+00

Blowing-up: Saving latest model state into RESTART file

WRT_RST - wrote re-start fields (Index=1,2) into time record = 0000001

[jcwarner]

the problem is most likely something to do with SED_DENS.
So i did a
grep -nir SED_DENS ROMS/*
and the most important location for the forward model is in ROMS/Nonlinear/rho_eos.F:

# if defined SEDIMENT && defined SED_DENS
SedDen=0.0_r8
DO ised=1,NST
itrc=idsed(ised)
cff1=1.0_r8/Srho(ised,ng)
SedDen=SedDen+ &
& t(i,j,k,nrhs,itrc)* &
& (Srho(ised,ng)-den(i,k))*cff1
END DO
den(i,k)=den(i,k)+SedDen
# endif

so looking at that code, I would guess that Srho is zero and that would cause a NaN.
can you check you sediment input and also look at the stdout and see what it says Srho is.

-john