Hi Patrick, Thanks for this clarification.
I think I just mixed up the use of Q.MBLOCk_ZA_GRID and qpack and qarts. When I call arts just for forward calculations and not retrieval like in the qarts_demo.m file, I give all the viewing angles of the sensor in Q.MBLOCk_ZA_GRID if I am not wrong. This is why I was suprised to give only one value to Q.MBLOCk_ZA_GRID but I have reasonable results now when using multiple viewing angle for my retrieval. One last question about the retrieval with qpack. From what I understand, I assume this is not possible to use different viewing angles for the different frequencies of the sensor ? I mean, for example I would like to perform a retrieval with observations only in the zenith direction for the first two frequencies and then I would like to use a set of 10 viewing angles for the other frequencies of the sensor. Is it possible to do that with qpack ? Thank you a lot for your help. Best regards, Pauline ----- Météo-France ----- Dr. Pauline Martinet Chercheur CNRM/GMEI/LISA [email protected] Fixe : +33 561079031 Site web: https://sites.google.com/site/martinetpauline31/ ----- Mail original ----- De: "Patrick Eriksson" <[email protected]> À: "George P. Ford" <[email protected]>, "MARTINET Pauline" <[email protected]> Cc: [email protected] Envoyé: Mardi 9 Septembre 2014 16:54:35 Objet: Re: [Qpack] retrieval of humidity profiles with qpack and multi viewing angles Hi Pauline and all, Sorry if the documentation is not totally clear. On the other hand, nice with some traffic on this mailing list! Starting with MBLOCK_ZA_GRID. As long as you ignore the antenna pattern (that is, assuming a pencil beam) you should keep: Q.MBLOCk_ZA_GRID = 0 There is some information in the Qpack "user guide" about how to set up a measurement including multiple angles. On last page of the document you find this: ZA: Line-of-sight zenith angle for the measurement. A scalar value, or a vector. The fields *Z_PLATFORM* and *ZA*, as well as *AA* for 3D case, must have the same size. If these fields have a length > 1, the measurement will consist of several spectra (having same *LAT* and *LON). IN ARTS, these spectra are simulated as seperate measurement blocks. Further, *TNOISE* must have a matching size, where the numbers of columns shall match length of *ZA*. Using the qpack2_wind3d_demo as example, the measurement generated in this example looks like: Y = AA: [5x1 double] DAY: 25 F: [12799x1 double] HOUR: {} HSE_P: 10000 HSE_Z: 16000 LATITUDE: 45 LONGITUDE: 2.7183 MINUTE: {} MONTH: 2 SECOND: {} TNOISE: [0.0500 0.0500 0.0500 0.0500 0.0500] Y: [63995x1 double] YEAR: 2008 ZA: [5x1 double] Z_PLATFORM: [5x1 double] Note that Z_PLATFORM and TNOISE also are vectors. For 1D AA should be []. Also note that Y.Y is a single vector, with the five spectra appended. In general Qpack demands vectors to be column vectors, or to be tensor1 in the atmlab terminology. So one of the error messages is just about that. (It seems that Y.TNOISE is not checked if it is a tensor1! But it should be!) I hope this clarifies the remaining issues. Bye, Patrick On 2014-09-09 15:17, Ford, George P. wrote: > I haven’t worked with that parameter at all, just Y.ZA and Y.AA defining > the line(s) of sight for the forward model calculation. Maybe someone > else could clarify what Q.MBLOCK_ZA_GRID does, something related to beam > size, perhaps? I only think this as it is part of the retrieval and not > the forward model, so I’m guessing must be related to properties of the > instrument. Then a single value of 0 would suggest a perfect pencil > beam... but I’d also welcome clarification! > > *From:*MARTINET Pauline [mailto:[email protected]] > *Sent:* 09 September 2014 13:37 > *To:* Ford, George P. > *Cc:* [email protected] > *Subject:* Re: retrieval of humidity profiles with qpack and multi > viewing angles > > Hi George, > > Thank you for your answer ! > > I followed the qpack2_wind3d_demo and I managed to run the retrieval > with two angles. > > There is just something that I don't understand. I defined the different > angles in > > Y.ZA. > > But apparently we don't have to put the list of angles in q_demo in the > variable Q.MBLOCK_ZA_GRID. > > In the qpack2_wind3d_demo, there is only one value equal to zero in > Q.MBLOCK_ZA_GRID.. > > When I tried to copy in Q.MBLOCK_ZA_GRID the same list of angles as in > Y.ZA I had an error so I assumed we > > have to let 0 in Q.MBLOCK_ZA_GRID. > > Can anybody confirm ? > > Thank you very much. > > Best regards, > > Pauline > > ----- Météo-France ----- > Dr. Pauline Martinet > Chercheur CNRM/GMEI/LISA > [email protected] <mailto:[email protected]> > Fixe : +33 561079031 > Site web: https://sites.google.com/site/martinetpauline31/ > > ------------------------------------------------------------------------ > > *De: *"George P. Ford" <[email protected] <mailto:[email protected]>> > *À: *"MARTINET Pauline" <[email protected] > <mailto:[email protected]>>, [email protected] > <mailto:[email protected]> > *Envoyé: *Mardi 9 Septembre 2014 11:53:47 > *Objet: *RE: retrieval of humidity profiles with qpack and multi > viewing angles > > Re. Your first issue, I’ve been using the qpack2_wind3d_demo which > already has multiple viewing angles – you could probably modify it for > your purposes, or it would at least be a good reference. It appears it > appends each spectrum on to the end of the Y vector, rather than > creating a new column, hence avoiding the tensor1 issue. > > *From:*MARTINET Pauline [mailto:[email protected]] > *Sent:* 09 September 2014 10:30 > *To:* [email protected] <mailto:[email protected]> > *Subject:* [Qpack] retrieval of humidity profiles with qpack and multi > viewing angles > > Dear all, > > I have some issues to run humidity profile retrievals with qpack and > also to run the retrieval with different viewing angle. > > I am using the observations from a ground based microwave radiometer > with 7 channels and 8 viewing angles. > > I*) Problem with different viewing angles* > > Fistly, I managed to retrieve temperature profiles with qpack but only > with one viewing angle. I use a > > matlab program similar to the qpack_demo.m file. > > So I modify the q_demo routine to include a second angle in > Q.MBLOCK_ZA_GRID ( for example Q.MBLOCK_ZA_GRID= [0 48] instead of 0). > > I also create the structure Y with a function similar to the y_demo > function in the qpack_demo.m file. > > After creating the Y structure, I copy the brightness temperature > observed by the instrument in the Y(1).Y field: Y(1).Y=TB_mesure > > but this time the size of the array TB_mesure is : 2 * 7 instead of 1* 7 > (so 2 angles and 7 channels). > > When I add this line, I have the error message: > > Error using rqre_datatype (line 68) > Y(1).Y must be tensor1. > > Error in qpack2>checkY (line 154) > rqre_datatype( Y(1).Y, @istensor1, 'Y(1).Y' ); > > Error in qpack2 (line 52) > empty_y = checkY( Y ); > > I assume this comes from the fact that there are two dimensions instead > of 1 due to the addition of the new viewing angle. > > Does anyone know how we can run a retrieval with multiple viewing angles ? > > By the way, does the variable Y.ZA refers to the wiewing angles like > Q.MBLOCK_ZA_GRID ? (I tried to modifiy this variable to include the > > second angle also.) > > *II) Problem with the humidity retrieval* > > In my ARTS simulations, I use a humidity profile directly read from a > background profile (thus it is not a profile given with the atmlab > package). When I declare the water vapor species in the q_demo file: > > Q.ABS_SPECIES(2).TAG = {'H2O-PWR98'}; > Q.ABS_SPECIES(2).RETRIEVE = false; > > If Q.ABS_SPECIES(2).RETRIEVE is false, I have no problem during the > computation of the temperature retrieval (even if I use my own humidity > profile for the simulation of the brightness temperature). > > Then, when I try to retrieve the humidity profile in addition to the > temperature profile, I changed the Q.ABS_SPECIES(2).RETRIEVE variable to > true, and I have the following error: > > Run-time error in controlfile: > /tmp/atmlab-martinet-tp9a5f653c_5510_4107_adc5_2463807d555b/cfile_y.arts > Run-time error in method: atmfields_checkedCalc > All values in *vmr_field* must be >= 0. > Stopping ARTS execution. > Goodbye. > > However, I checked the vmr_field of the Q.ABS_SPECIES(2).ATMDATA.GRID > and all the values look higher than 0 (see attached file) and I already > used this humidity profile > > even if I did not retrieve the humidity without no problem in ARTS. > Thus, I don't really understand why I have this error. > > Does anyone have on idea ? > > Thank you very much for your help. > > Best regards, > > Pauline > > ----- Météo-France ----- > Dr. Pauline Martinet > Chercheur CNRM/GMEI/LISA > [email protected] <mailto:[email protected]> > Fixe : +33 561079031 > Site web: https://sites.google.com/site/martinetpauline31/ > > ------------------------------------------------------------------------ > > This message (and any attachments) is for the recipient only. NERC is > subject to the Freedom of Information Act 2000 and the contents of this > email and any reply you make may be disclosed by NERC unless it is > exempt from release under the Act. Any material supplied to NERC may be > stored in an electronic records management system. > > > ------------------------------------------------------------------------ > This message (and any attachments) is for the recipient only. NERC is > subject to the Freedom of Information Act 2000 and the contents of this > email and any reply you make may be disclosed by NERC unless it is > exempt from release under the Act. Any material supplied to NERC may be > stored in an electronic records management system. > > > _______________________________________________ > 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
