Shortwave and Longwave radiation.

[jayashree]

Hi,

I have configured a model for the Indian Ocean using ROMS.
For a long run, I want to make use of JRA 55 data.
I have downloaded following atmospheric fields,
a)Air temperature
b)Specific humidity
c)u-component of wind
d)v-component of wind
e)Pressure reduced to mean sea level
f)Total precipitation
g)Downward solar radiation flux
h)Downward longwave radiation flux
i)Upward Solar Radiation flux
j)Upward longwave radiation flux

Can anyone suggest me how to calculate the net shortwave radiation and net longwave radiation to force my model?

Do I need to subtract 'Upward Solar Radiation flux' from 'Downward solar radiation flux' to get net shortwave radiation?, and
'Upward longwave radiation flux' from 'Downward longwave radiation flux' to get net longwave radiation?

Regards,
Jayashree

[kate]

I always let ROMS compute its own upward longwave flux. There's an option for that.

For the shortwave, it might be worth trying what you say. I use an albedo option because the ice model wants to have its own albedo.

[wilkin]

Yes, you need to subtract 'Upward Solar Radiation flux' from 'Downward solar radiation flux' to get net shortwave radiation.

For longwave you have two options and you can prepare for both.

When using #define BULK_FLUXES the default is to read net longwave radiation from variable lwrad (the name that is set in varinfo.dat) and for this you need to subtract 'Upward longwave radiation flux' from 'Downward longwave radiation flux'.

You might also #define LONGWAVE_OUT in which case ROMS computes outgoing longwave on the basis of the model SST. In this case you must read the downward longwave from the atmosphere as variable lwrad_down.

So in processing the JRA-55 data you might create both lwrad and lwrad_down and then choose how to run your application later. There is no harm in having unused variables in your forcing files.

[cae]

Hi.

What John describes is the usual way in which lwrad_down is generated for the LONGWAVE_OUT option. I wonder if there's an error in this approach.

I interpret 'Downward longwave radiation flux' from atmospheric models to mean downward above the ocean surface. Thus, this radiative term includes the fraction of downward flux that is reflected at the ocean surface. I believe the flux below the ocean surface (i.e., what's going into the ocean) should be reduced to 94% of the value above the surface (1-0.06, where 0.06 is reflected fraction). The value of alb_w in ROMS is 0.06, though lwrad_down does not seem to be reduced by 1-alb_w in the code. Accounting for this reflection could be done either in the code using alb_w or in the preparation of the forcing field.

I'll note that the reflected portion in the atmospheric model should contribute to the 'Upward long wave radiation flux' and so their difference is indeed the net going into the ocean if one is not using LONGWAVE_OUT.

It's possible that 'Downward longwave radiation flux' and 'Upward long wave radiation flux' are defined below the ocean surface, but if that were true, then the reflected portion would not exist in either, which does not seem likely to me.

I'm only modestly confident of this point, so I apologize if it's in error.

[Rafael_Santana]

Hi,

I'm running simulations for the northeast region of New Zealand using JRA55-do (https://climate.mri-jma.go.jp/~htsujino/jra55do.html), which has corrected atmospheric variables for ocean forcing.

The dataset does not have upward shortwave and longwave fluxes. Therefore, I'm using the option LONGWAVE_OUT in ROMS. This simulation (red line in Figure 1) had colder temperatures than AVHRR SST in waters deeper than 200m (green line in Figure 1).

I tried running without the option LONGWAVE_OUT and computed net longwave radiation (red and blue shade in Figure 2) by subtracting upward longwave radiation (from JRA55) from downward longwave radiation (from JRA55do). The results showed even colder temperatures (blue line in Figure 1) (the model still running).

One of the corrections applied to downward longwave radiation in JRA55do was the multiplication by a factor of 1.1 around New Zealand (https://doi.org/10.1016/j.ocemod.2018.07.002). I applied a similar correction using a factor of 1.4 and ran an experiment defining LONGWAVE_OUT (pink line in Figure 1). Is the correction factor of 1.4 feasible? Or should I try different options in ROMS (cppdefs used below)?

Thanks in advance,

Rafael


/*
*******************************************************************
** Northeast shelf run nested in HYCOM **
*******************************************************************
*/

#define SOLVE3D
#define CURVGRID
#define MASKING
#define AVERAGES
#define STATIONS
#define SALINITY
#define NONLIN_EOS
#define RADIATION_2D
#define NO_LBC_ATT
#define ANA_BTFLUX
#define ANA_BSFLUX
#define ANA_DQDSST
#define UV_ADV
#define UV_COR
#define UV_QDRAG
#define TS_U3HADVECTION
#define TS_SVADVECTION
#define DJ_GRADPS

#define LMD_MIXING
#if defined GLS_MIXING
# define KANTHA_CLAYSON
# define N2S2_HORAVG
#elif defined LMD_MIXING
# define LMD_RIMIX
# define LMD_CONVEC
# define LMD_DDMIX
# define LMD_SKPP
# define LMD_NONLOCAL
# define LMD_BKPP
#endif

#define BULK_FLUXES
#ifdef BULK_FLUXES
#define SOLAR_SOURCE /* define solar radiation source term */
#define LONGWAVE_OUT /* Compute net longwave radiation internally */
#define EMINUSP /* turn ON internal calculation of E-P */
#define COOL_SKIN
#else
# define QCORRECTION
#endif


PS: I'm not subtracting upward shortwave radiation (figure 3) from downward shortwave radiation because the former values are small and the impact would be minimal.

[hpftcb]

Hi Rafael,

posting the cpp flags I usually use in case it helps. One of the times I had problem with a colder surface temperature was due to humidity (wrong unit).

/* momentum equations */
#define UV_ADV
#define UV_COR
#define UV_QDRAG

#define UV_U3HADVECTION
#define UV_C4VADVECTION

#define TS_U3HADVECTION
#define TS_C4VADVECTION


#define SALINITY
#define NONLIN_EOS

#define UV_VIS2
#define MIX_S_UV

#define TS_DIF2
#define MIX_GEO_TS

#define SPLINES_VDIFF
#define SPLINES_VVISC
#define RI_SPLINES

#undef DIAGNOSTICS
#define AVERAGES
/* model configuration options */
#define SOLVE3D
#define MASKING
#define CURVGRID


#define DIAGNOSTICS_UV
#define DIAGNOSTICS_TS
# define POWER_LAW

# define VISC_GRID /*Viscosity coefficient scaled by grid resolution*/
# define DIFF_GRID /*Diffusivity coefficient scaled by grid resolution*/

# define SPHERICAL

#define PERFECT_RESTART
/* vertical turbulent mixing scheme */

#define MY25_MIXING

#ifdef MY25_MIXING
# define N2S2_HORAVG /* use if Large et al. (1994) interior closure */
# define KANTHA_CLAYSON /* use if Kantha and Clayson stability function */
#endif

#define DJ_GRADPS

#define WIND_MINUS_CURRENT
/* #define ANA_INITIAL */

#define ANA_BTFLUX
#define ANA_BSFLUX


#define BULK_FLUXES

#define EMINUSP
#define ANA_CLOUD
#define LONGWAVE_OUT
#define SOLAR_SOURCE

#define SSH_TIDES
#define UV_TIDES
#undef RAMP_TIDES
#define ADD_FSOBC
#define ADD_M2OBC




Regards

[Rafael_Santana]

Thanks for your message.