Hi Ole Martin, I agree with you this is why I am trying to compare my radiative transfer models with simple pencil beam simulations in one direction (90° at the moment) and this is where I saw I have unreasonable values when I put sensor_do to false.
For the gridsize of the atmosphere, I am using my input profile (from an NWP model) to define my Q.P_GRID and I use this grid for the other radiative transfer model. I am sorry but I don't know how to use the arts_y function ? I am using Qpack2 at the moment. Thank you again ! Pauline ----- Météo-France ----- Dr. Pauline Martinet Chercheur CNRM/GMEI/LISA [email protected] Fixe : +33 561079031 Site web: www.sites.google.com/site/martinetpauline31 ----- Mail original ----- De: "Ole Martin Christensen" <[email protected]> À: "Patrick Eriksson" <[email protected]>, "Pauline Martinet" <[email protected]> Cc: [email protected] Envoyé: Mardi 19 Mai 2015 11:30:21 Objet: RE: [Qpack] monochromatic pencil beam calculation To do proper pencil beam simulations I will strongly recommend you to use the function "arts_y", to ensure that you are not overriding some of your Q settings with the measurement structure Y (I assume you are running qpack2 now?). furthermore if you are cross comparing different Radiative Transfer models I would start with very simple pencil beam simulations looking in one direction to ensure that nothing in the instrument simulation is different. As Patrick mentions there are lot of details to check (e.g. ppath_step_length, O2 parametrization and gridsizes of the atmosphere) so try and make them agree on as simple a case as possible before incorporating any advanced features. Ole Martin ________________________________________ From: Patrick Eriksson Sent: Tuesday, May 19, 2015 11:09 AM To: Pauline Martinet; Ole Martin Christensen Cc: [email protected] Subject: Re: [Qpack] monochromatic pencil beam calculation Hi Pauline, Regarding the second part. If you have a very high absorption, try to set Q.PPATH_LMAX to e.g. 50 m. In this way, the optical thickness of each ppath step will be smaller, which should improve the accuracy for opaque situations. Is the same oxygen absorption model used in the different codes? This is another possible cause to the differences. Bye, Patrick On 2015-05-19 10:26, Pauline Martinet wrote: > Dear Ole Martin, > > Thank you for your quick answer. > > I probably made a mistake on the definition of the frequency grid. > > H.F_BACKEND is defined as a vector of 14 values (in the capt structure > that is an input to my Q file see attached file) : > [22.24 , 23.04 , 23.84 , 25.44 , 26.24 , 27.84 , 31.40, 51.26 , 52.28 , > 53.86 , 54.94 , 56.66 , 57.3 , 58] > > Whereas Q.F_GRID is defined as : > Q.F_GRID = linspace (fc(1)-df(1)*0.51, fc(end)+df(end)*0.51,100)'; > > with df the spectral bandwith of the receivers : > df = 1e6 * [230 230 230 230 230 230 230 230 230 230 230 600 1000 2000 ]'; > > I thought Q.F_GRID would define the spectral resolution of the > line-by-line calculation of ARTS but it is not clear for me... > > By the way I have now a second problem. I am comparing ARTS with two > radiative transfer models and we have a 1.5 K difference in the most > opaque channels. > Theses channels should see approximately the same values as the > temperature profile in the lowest layers. In my case the lowest values > of my temperature profile are: > 285.0548 > 284.9542 > 284.7708 > > for the first three atmospheric levels > > but my brightness temperatures are too low (these are the most opaque > channels of the O2 band at 56.66, 57.3 and 58 GHz): > 283.6000 > 283.8423 > 283.9169 > > it seems like my sensor is upper in the atmosphere but my Q.Z_surface is > set to 0. > > I am also working in 1D but as I created the ATMDATA of the temperature > profile by reading the climatological file cira86 I have a 4D field > with 3 grids. I am not sure this is good neither.. > > Sorry for all these questions but thank you very much for all your help. > > 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/> > > ------------------------------------------------------------------------ > *De: *"Ole Martin Christensen" <[email protected]> > *À: *"Pauline Martinet" <[email protected]>, [email protected] > *Envoyé: *Mardi 19 Mai 2015 10:09:32 > *Objet: *RE: monochromatic pencil beam calculation > > What frequency grid are you running on in the two cases? If you have > sensor_do and backend_do your output frequencies will be determined by > you instrument backend, while if you have sensor_do = false it will be > given by f_grid. > > Could you attach your Q file? > > Ole Martin > ------------------------------------------------------------------------ > *From:* [email protected] [[email protected]] on > behalf of Pauline Martinet [[email protected]] > *Sent:* Tuesday, May 19, 2015 8:50 AM > *To:* [email protected] > *Subject:* [Qpack] monochromatic pencil beam calculation > > Hi everybody, > > I am used to simulate ground-based microwave spectra with qpack (22GHz > to 58 GHz). > > I normally define a double side band average but now I would like to > test my simulations with the assumption of monochromatic pencil beam. > > I keep the same qpack code and I only changed the variable Q.SENSOR_DO > to false instead of true. However I obtain unreasonable > brightness temperature in the most opaque channels of the O2 absorption > band. > > Here is the spectrum that I simulate with the double side band average: > > 37.1335 > 36.2098 > 31.8496 > 23.8967 > 21.4893 > 18.8657 > 17.9993 > 109.5801 > 151.6191 > 249.6631 > 278.5141 > 283.6000 > 283.8424 > 283.9169 > > whereas this is what I obtain with a monochromatic pencil beam calculation: > 36.9074 > 37.6239 > 36.9598 > 35.3325 > 33.2136 > 30.9684 > 28.8201 > 26.8801 > 25.1884 > 23.7449 > 22.5305 > 21.5187 > 20.6823 > 19.9959 > > I obtain 19K instead of 283 K for the most opaque channels, this is > weird (the last 7 channels). > > Does anyone see what the problem is, I probably did not define all the > variables needed, I just turned out the SENSOR_DO variable.. > > Thank you very much for your 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 > _______________________________________________ qpack mailing list [email protected] https://www.sat.ltu.se/mailman/listinfo/qpack
