question about sign of surface heat fluxes.

[syuip1224]

Dear ROMS users,

I have some questions about sign of surface heat fluxes. sorry for "double posting" but I couldn't get a clear answer.

I define NCEP_FLUXES in ROMS(version 3.5) with ice model.
And I'm using ECMWF 'surface latent, sensible heat flux' and 'surface thermal radiation(Net thermal radiation at the surface, longwave)' data as the atmosphere forcing data.
The sign of these ECMWF heat flux data are downward positive.

Then, I found one thing in ncep_flux.F code.

lhtfl = lhflx(i,j)
shtfl = shflx(i,j)
ta = tsrf_d-.0098*zt-shtfl/(rhoa_n(i,j)*cpair*ch_d(i,j)*sa)
qa = qsrf_d-lhtfl/(rhoa_n(i,j)*le*ce_d(i,j)*sa)

I think, this is from Fairall et al., 1996, equation (2a,b), (3) and (4).
If so, do I need to multiply '-1' by ECMWF sensible and latent data?
Because, this code seems to be 'upward positive'(shtfl = (rhoa_n(i,j)*cpair*ch_d(i,j)*sa)*(tsrf_d-(ta+.0098*zt))).


I have another question.
Does the longwave radiation flux also be considered as downward positive(heating) in ncep_flux.F code?

In ncep_flux.F code,

! Net long-wave radiation to ocean
nlwrs = lrflx(i,j)
sigmaw = 5.67E-8_r8
emisw = 0.97_r8
qlww = nlwrs + 4._r8*sigmaw*emisw*ta**3*(tml-tsrf_d)

! Net upward oceanic heat flux without ice, in W/m**2
qao_n(i,j) = (-qsh -qlh -qsww + qlww)

also,

! Net heat flux form ice to atmosphere
qai_n(i,j) = -qsh_i -qlh_i -qswi + qlwi

I thougth that the longwave radiation is 'downward positive' in this line (! Net long-wave radiation "TO OCEAN").
However, it doesn't seems to 'downward positive' in the next line, because the sign of longwave(qlww) is different to other terms(qsh, qlh, qsww).
It looks like 'upward positive'. do I miss anything?

I'm wondering that I need to multiply '-1' by surface thermal radiation data(longwave).

Sorry for some basic questions, it's so confusing the sign of heat fluxes for me as a beginner.
Could anyone tell me the meaning of equations and how to treat the data?

Thanks for your helps.

- SY

[kate]

The NCEP_FLUX code comes from Paul Budgell and might not be used by anyone else. In general, the ROMS convention is to have surface heat flux be heat into the model, positive for heating from above.