Hello again, On 19 July 2015 at 21:55, Patrick Eriksson <[email protected]> wrote: > I am on vacation and the answer will be very short.
Ok, enjoy your vacation! This matter is not urgent, please do wait
until after August 3 :)
> This differs if the conversion is done in yCalc or in yApplyUnit.
>
> If done in yCalc, the conversion is performed for each monochromatic
> frequency and the spectral response is applied on the Tb values.
>
> If done in yApplyUnit, the conversion will be done for y_f, that should
> match f_backend.
>
> I can not answer if any of this will match exactly what is done for
> ATSR. Internally we have had similar discussions around AMSU, ICI ... It
> is surprising that the mimicking the calibration is quite problematic.
> We can discuss further when I am back at work Aug 3.
I have done some more digging. The NOAA KLM User's Guide reports a
"band correction" in Equation 7.3.3-4, which actually refers to
microwave channels, but the same mathematics applies to infrared.
Coefficients for infrared channels are listed in tables D.1-4, D.2-4,
D.3-10, D.4-10, D.5-10, D.6-10, and D.7-10. The KLM User's Guide does
not report where those are from, but according to James Hocking[1]
they are based on Weinreb et al. (1981). Weinreb et al. consider a
black-body at a temperature T, integrate this with the SRF, then use
the resulting channel radiance with the central frequency to calculate
a new brightness temperature T_e, i.e.
T_e = B^{-1} [v, \int_{v_1}^{v_2} B(v, T) SRF(v) dv]
In general, T_e ≠ T (T_e = T only for monochromatic channels or
channels of specific shapes).
I have attached a figure of T_e - T as a function of T for HIRS
channels on NOAA-12 (I have similar figures for other satellites).
For most channels, differences are smaller than 0.05 K; but for
channels 10 (8.17 µm), 12 (6.76 µm) and the shortwave channel 19 (3.77
µm), differences at low temperatures can exceed 0.25 K. Thes is
consistent with the coefficients in the aforementioned NOAA KLM
Tables, where channels 12 and 19 have larger coefficients than other
channels. As this effect is systematic, it should be worth thinking
more about how to handle this. The effect is quite linear with
temperature, which justifies that NOAA approximates it linearly in its
band corrections.
I'm not sure how exact the method by Weinreb et al. (1981) is.
Weinreb et al. (1981) refer to Fleming (1981) for a "more accurate
approach" but I have not gotten my hand on that paper yet, nor on a
more recent study on the topic.
[1] http://www.nwpsaf.eu/forum/viewtopic.php?f=8&t=184
[2] Weinreb, M.P., Fleming, H.E., McMillin, L.M., Neuendorffer, A.C,
Transmittances for the TIROS operational vertical sounder,1981, NOAA
Technical Report NESS 85.
regards,
Gerrit.
BT_corrections_NOAA12.pdf
Description: Adobe PDF document
_______________________________________________ arts-users mailing list [email protected] https://www.sat.ltu.se/mailman/listinfo/arts-users
