The answers are embedded in the text below.
________________________________
From: [email protected] [[email protected]] on
behalf of Pauline Martinet [[email protected]]
Sent: Tuesday, July 21, 2015 2:28 PM
To: ARTS Users; [email protected]
Subject: [arts-users] pencil beam calculation/band width averaging
Dear everybody,
I already sent a message a while ago for a proper pencil beam simulation with
ARTS but I am still confused thus I need some help.
We are trying to compare ARTS with a new fast radiative transfer model for
ground-based microwave radiometers and we would like to start the comparison
with pencil beam calculation but I would be interested in also knowing how the
band-averaged brightness temperatures (BT) are computed with ARTS and if I am
doing that properly.
I am using Qpack for simulations of band-averaged BT but for pencil beam
computation I use arts_y since your advice last time.
I) Questions about band-averaged BT
In our fast radiative transfer model, the band-averaged BTs are computed from
three frequencies only: central frequency (cf), cf-df/2 and cf+df/2 with df
the spectral bandwidth of each channel.
To do that in Qpack, I normally declare Q.SENSOR_RESPONSE.F_BACKEND = central
frequencies of the 14 channels of the MWR
and for each of the 14 channel I declare the
Q.SENSOR_RESPONSE.BACKEND_CHANNEL_RESPONSE as:
{df(i)/2*[-1 1]}
where df corresponds to the spectral band of channel i. Thus for me it seems
that I am doing an averaging of the BT over 2 frequencies: the central
frequency minus the spectral band
divided by 2 and the central frequency plus the spectral band divided by 2.
Am I right ?
=> Jupp, just make sure that Q.SENSOR_RESPONSE.SENSOR_NORM = true so that your
backend response is normalized to have a unit area.
If I want to mimic the fast radiative transfer model by averaging over 3
frequencies, should I just add one value to the vector defining the
BACKEND_CHANNEL_RESPONSE ?
like this:
{-df(i)/2*1 0 df(i)/2*1]}
Here I put a zero to consider the central frequency TB also in the averaging.
=> Jupp, just make sure that when you define your backend channel response two
values are needed a vector containing frequencies and another the response at
these frequencies. The example given in qpack2_demo2.m in the demo folder shows
you how to do this.
Could you explain to me what is the role of Q.F_GRID in qpack ? I am confused
because we already define the frequency grid we want to simulate and the
frequency averaging in Q.SENSOR_RESPONSE.F_BACKEND and
Q.SENSOR_RESPONSE.BACKEND_CHANNEL_RESPONSE. Thus why do we need another vector
defining again a frequency grid ?
=> The radiative transfer model calculations are done on the grid specified by
Q.F_GRID, and thereafter it is interpolated onto the grid of each channel
(F_BACKEND / BACKEND_CHANNEL_RESPONSE).
II) Pencil beam calculations
To do the pencil beam calculations I do not use Qpack but arts_y, this time I
define my 14 frequency grid in Q_F.GRID and I switched Q.SENSOR_DO to false.
Could you just confirm that with this configuration, pencil beam calculations
means no modelling of the antenna beam-width ?
=> Jupp
Thank you very much for your kind help
Best regards,
Pauline
----- Météo-France -----
Dr. Pauline Martinet
Chercheur CNRM/GMEI/LISA
[email protected]
Fixe : +33 561079031
Site web: www.sites.google.com/site/martinetpauline31
<https://www.sites.google.com/site/martinetpauline31/>
_______________________________________________
qpack mailing list
[email protected]
https://www.sat.ltu.se/mailman/listinfo/qpack
