[MITgcm-support] A query about EXF

[Martin Losch]

Hi,

your forcing data set (as you describe it) is imcomplete.

The model requires a net heat flux at the ocean surface. The net heat flux (Qnet) consistes of different components, but the ocean doesn't care about them (except for the short wave heat Qsw, but that's not so important here). Typically you have

Qnet = Qsw(shortwave) + Qlw(longwave) + Qsh(sensible) + Qlh(latent)
Qsw = QswDown (from the sun refracted by the atmosphere) + QswUp (reflected from the surface) = Qswdown*(1-albedo)
Qlw = QlwDown (also called counter radiation by the atmosphere) + QlwUp (=BoltzmannConstant * emissivity * surfaceTemp^4)

The turbulent fluxes Qsh and Qlh are usually computed from bulk formulae. To first order Qsh ~ windSpeed*(Tair-SST), and Qlh ~ windSpeed*(qair - qSurface), where qSurface is the saturation  specific humidity at the surface temperature SST (a nonlinear function, either a polynomial or an exponential function or what the meteorologist come up with). One example of these bulk formulae for the ocean is implemented in pkg/exf/exf_bulkformulae.F, a somewhat simpler set is in pkg/seaice/seaice_ocean_budget.F. As input they need the wind speed (derived from uwind/vwind or read in separately), the surface temperature provided by the ocean (or sea ice) model, the air temperature Tair and specific humidity qair (typically at 2m height above the surface, but that's a detail). A data set that provides both Tair/qair and Qsh/Qlh has somehow computed Qsh/Qlh based on surface temperatures from an ice-ocean different from what you have (and unknown to you).

The extremely important contribution QswDown can be computed from astronomical parameters (see e.g. Parkinson and Washington, 1979), but the refraction by the atmosphere needs to be estimated, which is difficult if you don't have an estimate of a clould cover. 

A very large contribution to the net heat flux is the long wave contribution Qlw. QlwUp can be computed from the surface temperature, but QlwDown needs to be specified. You can approximate it based on the surface temperature (someCorrectionFactor*BoltzmannConstant*Tair^4) but that's really not very good, because you actually need the temperature at the bottom of the clouds, and there is no way to reconstruct that reliably from surface Tair.

Currently the exf does not have the option of reading in turbulent heat fluxes. Even if you modify the code to make it do so, your plan of running the model just with wind, Tair, qair, and turbulent heat fluxes will fail (the model will cool down very quickly and explode), if you don't provide incoming radiation. (Just setting them to zero will not do it!!) 
I repeat Dimitris suggestion that, if you really want to use these precomputed fluxes, you construct Qsw(net!!!) and Qlw(net!!!) somehow yourself, then add up the four components and read in the total heat and fresh water flux (so uwind, vwind, hflux, swflux, sflux, stress will be computed). This will only work if you don't have an ice-model, which requires the "default" uwind, vwind, tair, qair, swdonw, lwdown.

Martin


On Oct 10, 2013, at 10:03 AM, Abhisek Chakraborty <abhisek.sac at gmail.com> wrote:

> Dear Dimitris Menemenlis,
> 
> Actually I am presently running a global ocean (excluding poles) configuration using forcings from NCEP reanalysis. However, I am planning to use forcings from some other data sets (which have 10 m winds, 2m air temperature, 2m sp. humidity, sensible and latent heat fluxes only). For such a data set, I wish to use the EXF package. The present configuration of EXF needs other parameters ,.e.g., shortwave, longwave radiations, precipitations etc. So my query is how to use EXF for the mentioned data sets as far as the forcing of the global ocean model set up is considered.
> 
> regards,
> Abhisek
> 
> 
> On Thu, Oct 10, 2013 at 11:28 AM, Menemenlis, Dimitris (3248) <Dimitris.Menemenlis at jpl.nasa.gov> wrote:
> what's so special about this hs/hl? what are you trying to study/achieve?
> 
> Dimitris Menemenlis
> 
> On Oct 9, 2013, at 10:35 PM, Abhisek Chakraborty wrote:
> 
>> Dear Dimitris Menemenlis,
>> 
>> Thanks for your prompt reply.
>> 
>> I understand that for the default configuration of EXF I need to use ustress, vstress (from 10m winds) and hflux, swflux and sflux.  However, I don't have any fluxes (e.g. shortwave or long wave, precipitaion etc.) other than HS and HL.  In such a scenario, I can compute hflux as the sum of HS and HL. But what I should do for the other fluxes (swflux, sflux etc.). Should I use some climatology or simply zero?
>> 
>> regards,
>> Abhisek
>> 
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support