rapid surface cooling by an hour after initialization

[tereddy]

I am just getting started with ROMS, setting up an application in the Arctic for the Alaskan and Mackenzie Shelf region. I did a quick and dirty setup, just to see if I could get up and running. I have created the model grid and am able to run the model with some simple initialization and forcing, but I find that the ocean temperatures are behaving strangely. I set initial temperature conditions from Levitus values. Near the Bering Strait, the temperatures are around 6deg initially. After running the model for only an hour and a half, the temperatures drops to around 2 degrees for this area which is quite large. Where could all this heat be going and so quickly? (This run was done without sea ice.) Any thoughts are much appreciated. Thanks!

Initial Temperature - Surface (Level 1)

1.5 Hr. Temperature - Surface (Level 1)

Initial Temperature - Level 3

1.5 Hr. Temperature - Level 3

cppdefs:
#define UV_ADV
#define UV_COR
#define UV_SADVECTION

#define TS_U3HADVECTION
#define TS_SVADVECTION
#define TS_PSOURCE
#define SALINITY
#define NONLIN_EOS

#define SOLVE3D
#define CURVGRID
#define MASKING
#define AVERAGES
#define AVERAGES_QUADRATIC
#define AVERAGES_FLUXES
#define DIAGNOSTICS_UV
#define DIAGNOSTICS_TS

#define SPLINES
#define DJ_GRADPS
#define ANA_PSOURCE
#define STATIONS

#define ICE_BULK_FLUXES
#define BULK_FLUXES
#define SOLAR_SOURCE
#define LONGWAVE_OUT
#define EMINUSP
#define UV_QDRAG
#define ANA_SSFLUX
#define ANA_BSFLUX
#define ANA_BTFLUX

#define MY25_MIXING
#ifdef MY25_MIXING
# define N2S2_HORAVG
# define KANTHA_CLAYSON
#endif
#ifdef LMD_MIXING
# define LMD_RIMIX
# define LMD_CONVEC
# define LMD_SKPP
# define LMD_BKPP
# define ANA_CLOUD
#endif

# define RADIATION_2D

# define NORTH_FSCHAPMAN
# define NORTH_M2FLATHER
# define NORTH_M3RADIATION
# define NORTH_M3NUDGING
# define NORTH_TRADIATION

# define SOUTH_FSCHAPMAN
# define SOUTH_M2FLATHER
# define SOUTH_M3RADIATION
# define SOUTH_M3NUDGING
# define SOUTH_TRADIATION

[tereddy]

After reading this post:
viewtopic.php?f=17&t=1205

I changed my cppdefs to:
#undef SPLINES
#undef UV_SADVECTION
#undef TS_SVADVECTION
#define TS_C4VADVECTION
This had little impact on the results of temperatures cooling quickly.

I also noticed the vertical temperature diffusion, could this be part of the problem? But cooling at 0.0002 degC/sec should only lead to less than a degree temperature change in an hour, not the 4-5 degrees seen in the model results.

[arango]

I don't know what is your problem but it likely a set-up one. You need to check your forcing fields.

One thing that I noticed is that your land/sea masking is very wrong in many places. You have one point bays, islands and lakes. You need to use the editmask script to fine tune your land/sea masking. This will take you from 2 to 4 hours of hand editing work.

The editmask.m script is distributed in the matlab repository.

[enrique]

The bulk flux logic handling of humidity breaks in the high-latitudes. It gets confused between relative and specific humidity which up there will cause what you are seeing. Make sure you are indeed using the units you think you have.

[tereddy]

I played around with the humidity but that didn't have an effect on the temperature problem I was seeing.

It turns out there was something wrong with my initialization file, or so it appears. Using ANA_INITIAL prevented the problem from occurring.