From erik Fri Jan 12 10:05:14 2001 Subject: CUO_39.2 surface brightness Status: OR GAIA surveys of surface brightness ================================== E. Hoeg, C. Fabricius, J. Knude, V.V. Makarov 25 June 1998 SAG_CUO_39 Issue 2 12 Jan. 2001 ABSTRACT: The detection and measurement of surface brightness in the main GAIA telescope sky mappers is discussed, now with realistic assumptions on readnoise. The sensitivity to the parameters readnoise, global background and the uncertainty of the global background is illustrated in tables. It appears, for instance, that all normal galaxies brighter than about V=18 mag may be detected and that four-colour Sloan photometry at a resolution of 0.2 arcsec may be obtained for their central parts, provided the data can be accomodated in the available telemetry. Astrophysicists are invited to discuss the scientific case of the proposed nebula detection and 12-arcsec mapping with GAIA. Introduction to Issue 2 ----------------------- This new issue was produced in connection with CUO_082, also on the surface brightness. The notation has been updated to agree with the final GAIA Study Report (GSR): The previous PSM is now ASM, and FSM is now BBP. The sampling in the BBP has been changed as described in a note at ***. The assumptions on noise and light fluxes are still valid. Some other assumptions are no longer valid: The BBP is no longer in the Sloan system, but that makes no significant difference for the discussion. Astrophysicists were invited to discuss the scientific case of the proposed nebula detection and 12-arcsec mapping with GAIA. The nebula or galaxy detection has been pursued, but the 12-arcsec mapping was not recommended further. Introduction ------------ A GAIA survey of nebulae and sky background was proposed in CUO_32 in which the basic ideas are described. It is now possible to give reliable performance figures because the read-out noise (RON) has been given and a binning strategy has been proposed in CUO_36, based on new information from MMS. We consider the signal-to-noise ratio (SNR=R) of seven quantities obtained from the ASM1 and the BBPs (see Table 1a). For the BBPs we consider the Sloan i' as an example. G16p : the sum of 16 pixels (2*2 samples = 4*4 pixels) from ASM1 Gbm2 : the trimmed median of the 243 samples in a patch of 2*2 arcsec^2 from ASM1 Gbm12 : the trimmed median of the 8760 samples in a patch of 12*12 arcsec^2 from ASM1 i'1*8 : the sample of 1*8 pixels from BBP1 or BBP3 i'3*8 : the sample of 3*8 pixels from BBP2 or BBP4 i'6*8 : the sample of 6*8 pixels from BBP2 or BBP4 i'bm12 : the trimmed median of the 730 samples of 6*8 pixels in a patch of 12*12 arcsec^2 from BBP2 and 4 *** Note: What is here described as sampling in BBP2 and 4 is now done in the BBP of the Astro-2 telescope, and only samples of 6*8 pixels are taken, not any of 3*8. In Astro-1 all samples are of 1*8 pixels. We note that BBP1 and BBP2 in GSR means the BBPs in respectively Astro 1 and Astro 2. Table 1a shows the SNR as function of the excess surface brightness in B or Ic with realistic baseline assumptions on readnoise and global sky background. The standard error of a magnitude is =1/SNR mag. The directly calculated SNR has been multiplied with a margin factor of 0.80, i.e. the SNR was made 20 per cent smaller. The count rates S in Table 1 for Ic=15 mag are 0.90 times the rates given by LL in file ut_calc_flux (16 Feb 1998), the factor 0.9 representing another margin. A global background (as defined in CUO_31) derived from several degrees of scan for each CCD in the ASM, could be used as reference against which comparison is made. In this context, the global background may be derived from the Gbm2 values (cf. CUO_32). The global background is assumed to be zero for the Gbm12 and i'bm12 values because they contain the total estimated surface brightness in 12*12 arcsec^2 patches. We assume for the other quantities that the global background to be subtracted from the estimation has a standard error of 10 per cent. For simplicity, the nebulae and the global background are assumed to have the colour index as a star of G2V, A_V=0. The four Tables 1a-1d show the sensitivity of the SNR depending on various parameters, the readnoise (Table 1b), the relative error of the global background (Table 1c), and the global background itself (Table 1d). An improvement of the readnoise by about 30 per cent assumed in Table 1b seems realistic in view of the margin on RON mentioned in CUO_36. Nebula detection ---------------- Nebula detection is done by detection of excess light in the Gbm2 values, an excess over the global background. Such detection will then trigger a special sampling in the BBP2 and 4 in an area. The star detection is based on the same samples from ASM1 and runs all the time, in parallel with the nebula detection. A bright star could in principle trigger a nebula detection. But this should happen very seldom since a trimmed median (or perhaps an even better method) is used to eliminate the effect of the stars on the surface brightness estimate. It appears that a surface brightness in Ic of about uIc=20 mag/arcsec^2 may be safely detected with SNR=3.0 if areas of 2*2 arcsec^2 from the ASM1 are analyzed (Table 1a). Four-colour Sloan photometry of such patches may be obtained with 0.2 arcsec resolution if the data is transmitted from the BBPs with samples of 6*8 pixels. A superpositioning of e.g. 25 such patches would give a SNR=0.6*sqrt(25)=3.0 in Sloan i' for a surface having uIc=21 mag/arcsec^2. Although a 4 year observing period gives an average of 67 transits we here only assume a very conservative number of 25 useful scans for a given object. It appears from Tables 1b that an improved readnoise would improve these SNRs by about 30 per cent. The dependence on global background is only minor (see Tables 1a and 1d). The SNR=3 is achieved at about 1.0 mag brighter uIc than in Table 1 of CUO_32. This is due to the higher readnoise and the margin of 20 per cent applied in the present study. Nevertheless, the conclusion from CUO_32 about the detection can be essentially maintained. Tentatively, elliptical and spiral galaxies can be detected if their effective radii are larger than about 1.5 arcsec, and four-colour Sloan photometry can be obtained. These numbers imply that the galaxies must be brighter than about Ic=17 mag or V=18 mag. A significant detection may be defined by an isolated Gbm2 value with SNR>6.0 or at least two adjacent Gbm2 values with SNR>3.0 (cf. Table 1a). Samples should be transmitted for a larger area, e.g. 2 arcsec wider. Detection and measurement of galactic nebulosities remain to be studied. The total sky area that would be measured on galaxies and nebulosities and the resulting data rate should be estimated. Simulations should be done. DENIS has detected about 36 000 galaxies in 2/3 of the southern sky, sg_I=0.13 mag, resolution >1 arcsec due to seeing. They expect to detect 900 000 galaxies to I=16.5 and to measure in I, J, and K. The 12-arcsec mapping --------------------- The 12-arcsec mapping should basically contain the bm12 values derived as trimmed median values from 12*12 arcsec^2 covering the entire sky several times. It appear from Table 1a that the Sloan i' could be measured with SNR=6.0, or sg_m=0.2 mag if uIc=22.0 mag in each coverage of the sky. The data correspond to a total of 66 kbits/s from ASM1 in G and the four Sloan colours together, according to CUO_36. It would seem sufficient to transmit data during the first 6 months of mission giving a ten-fold coverage of the sky. Scientific case for surface brightness measurement -------------------------------------------------- It seems that enough is now known about the performance of GAIA that the scientific case could be written. Astrophycisists are invited to do so. --------------------------------------------------------- Table 1a. Diffuse surface brightness. Case: Baseline assumptions (given at the end of the table). Detection in G and measurement in Sloan i'. SNR (=R) is given as function of the excess surface brightness of the object for various sizes of the patch. The Gbm12 and i'bm12 are absolute values, but other values are relative to a global background of uI=21.0 mag/arcsec^2 which is assumed to have been measured with a relative error of 10. per cent. Detector ASM1 ASM1 ASM1 BBP1 BBP2 BBP2 BBP2 Band G G G i' i' i' i' Name G16p Gbm2 Gbm12 1*8 3*8 6*8 i'bm12 Ssurf uB uI R R R R R R R m/a^2 m/a^2 17.5 16.0 13.7 84.7 514.7 5.7 10.8 15.6 339.0 18.5 17.0 6.8 41.6 254.5 3.1 6.4 9.5 206.6 19.5 18.0 3.0 18.6 114.4 1.5 3.5 5.5 120.6 20.5 19.0 1.3 7.8 48.3 0.7 1.7 2.9 65.2 21.5 20.0 0.5 3.2 19.7 0.3 0.8 1.4 31.9 22.5 21.0 0.2 1.3 7.9 0.1 0.3 0.6 14.2 23.5 22.0 0.1 0.5 3.2 0.0 0.1 0.3 6.0 24.5 23.0 0.0 0.2 1.3 0.0 0.1 0.1 2.4 --------------------------------------------------------- patch= 0.066 3.992 144. 0.033 0.099 0.197 144. [arcsec^2] uI=15.0:S= 21571 21571 21571 5743 5743 5743 5743 [e-/s/arcsec^2] t= 0.9 0.9 0.9 0.9 0.9 0.9 0.9 [s/sample] npix= 4 4 4 8 24 48 48 [pixels/sample] n= 4 243 8760 1 1 1 730 [samples/patch] Backgr: b= 1.27 1.27 0.00 0.68 2.03 4.06 0.00 [e-/sample] sg_b= 0.13 0.13 0.00 0.07 0.20 0.41 0.00 [e-/sample] RON: r= 9.5 9.5 9.5 4.5 4.5 4.5 4.5 [e-/sample] --------------------------------------------------------- --------------------------------------------------------- Table 1b. Diffuse surface brightness. Case: Readnoise (RON) 30 per cent smaller. Detection in G and measurement in Sloan i'. SNR (=R) is given as function of the excess surface brightness of the object for various sizes of the patch. The Gbm12 and i'bm12 are absolute values, but other values are relative to a global background of uI=21.0 mag/arcsec^2 which is assumed to have been measured with a relative error of 10. per cent. Detector ASM1 ASM1 ASM1 BBP1 BBP2 BBP2 BBP2 Band G G G i' i' i' i' Name G16p Gbm2 Gbm12 1*8 3*8 6*8 i'bm12 Ssurf uB uI R R R R R R R m/a^2 m/a^2 17.5 16.0 15.5 95.1 579.8 6.1 11.1 15.8 343.2 18.5 17.0 8.2 50.3 310.2 3.5 6.7 9.7 212.7 19.5 18.0 4.0 24.0 149.9 1.9 3.9 5.8 128.7 20.5 19.0 1.8 10.5 66.2 0.9 2.1 3.2 74.1 21.5 20.0 0.7 4.4 27.7 0.4 1.0 1.7 39.2 22.5 21.0 0.3 1.8 11.2 0.2 0.4 0.8 18.7 23.5 22.0 0.1 0.7 4.5 0.1 0.2 0.3 8.2 24.5 23.0 0.0 0.3 1.8 0.0 0.1 0.1 3.4 --------------------------------------------------------- patch= 0.066 3.992 144. 0.033 0.099 0.197 144. [arcsec^2] uI=15.0:S= 21571 21571 21571 5743 5743 5743 5743 [e-/s/arcsec^2] t= 0.9 0.9 0.9 0.9 0.9 0.9 0.9 [s/sample] npix= 4 4 4 8 24 48 48 [pixels/sample] n= 4 243 8760 1 1 1 730 [samples/patch] Backgr: b= 1.27 1.27 0.00 0.68 2.03 4.06 0.00 [e-/sample] sg_b= 0.13 0.13 0.00 0.07 0.20 0.41 0.00 [e-/sample] RON: r= 6.7 6.7 6.7 3.1 3.1 3.1 3.1 [e-/sample] --------------------------------------------------------- --------------------------------------------------------- Table 1c. Diffuse surface brightness. Case: Higher sigma_background and again the baseline RON. Detection in G and measurement in Sloan i'. SNR (=R) is given as function of the excess surface brightness of the object for various sizes of the patch. The Gbm12 and i'bm12 are absolute values, but other values are relative to a global background of uI=21.0 mag/arcsec^2 which is assumed to have been measured with a relative error of 20. per cent. Detector ASM1 ASM1 ASM1 BBP1 BBP2 BBP2 BBP2 Band G G G i' i' i' i' Name G16p Gbm2 Gbm12 1*8 3*8 6*8 i'bm12 Ssurf uB uI R R R R R R R m/a^2 m/a^2 17.5 16.0 13.7 82.6 514.7 5.7 10.8 15.6 339.0 18.5 17.0 6.8 40.0 254.5 3.1 6.4 9.5 206.6 19.5 18.0 3.0 17.7 114.4 1.5 3.5 5.4 120.6 20.5 19.0 1.3 7.4 48.3 0.7 1.7 2.9 65.2 21.5 20.0 0.5 3.0 19.7 0.3 0.8 1.4 31.9 22.5 21.0 0.2 1.2 7.9 0.1 0.3 0.6 14.2 23.5 22.0 0.1 0.5 3.2 0.0 0.1 0.3 6.0 24.5 23.0 0.0 0.2 1.3 0.0 0.1 0.1 2.4 --------------------------------------------------------- patch= 0.066 3.992 144. 0.033 0.099 0.197 144. [arcsec^2] uI=15.0:S= 21571 21571 21571 5743 5743 5743 5743 [e-/s/arcsec^2] t= 0.9 0.9 0.9 0.9 0.9 0.9 0.9 [s/sample] npix= 4 4 4 8 24 48 48 [pixels/sample] n= 4 243 8760 1 1 1 730 [samples/patch] Backgr: b= 1.27 1.27 0.00 0.68 2.03 4.06 0.00 [e-/sample] sg_b= 0.25 0.25 0.00 0.14 0.41 0.81 0.00 [e-/sample] RON: r= 9.5 9.5 9.5 4.5 4.5 4.5 4.5 [e-/sample] --------------------------------------------------------- --------------------------------------------------------- Table 1d. Diffuse surface brightness. Case: Global background = 0. Detection in G and measurement in Sloan i'. SNR (=R) is given as function of the excess surface brightness of the object for various sizes of the patch. The Gbm12 and i'bm12 are absolute values, but other values are relative to a global background of uI=29.0 mag/arcsec^2 which is assumed to have been measured with a relative error of 10. per cent. Detector ASM1 ASM1 ASM1 BBP1 BBP2 BBP2 BBP2 Band G G G i' i' i' i' Name G16p Gbm2 Gbm12 1*8 3*8 6*8 i'bm12 Ssurf uB uI R R R R R R R m/a^2 m/a^2 17.5 16.0 13.8 85.7 514.7 5.8 10.9 15.7 339.0 18.5 17.0 6.8 42.4 254.5 3.1 6.4 9.6 206.6 19.5 18.0 3.1 19.1 114.4 1.5 3.6 5.6 120.6 20.5 19.0 1.3 8.0 48.3 0.7 1.8 3.0 65.2 21.5 20.0 0.5 3.3 19.7 0.3 0.8 1.5 31.9 22.5 21.0 0.2 1.3 7.9 0.1 0.3 0.7 14.2 23.5 22.0 0.1 0.5 3.2 0.0 0.1 0.3 6.0 24.5 23.0 0.0 0.2 1.3 0.0 0.1 0.1 2.4 --------------------------------------------------------- patch= 0.066 3.992 144. 0.033 0.099 0.197 144. [arcsec^2] uI=15.0:S= 21571 21571 21571 5743 5743 5743 5743 [e-/s/arcsec^2] t= 0.9 0.9 0.9 0.9 0.9 0.9 0.9 [s/sample] npix= 4 4 4 8 24 48 48 [pixels/sample] n= 4 243 8760 1 1 1 730 [samples/patch] Backgr: b= 0.00 0.00 0.00 0.00 0.00 0.00 0.00 [e-/sample] sg_b= 0.00 0.00 0.00 0.00 0.00 0.00 0.00 [e-/sample] RON: r= 9.5 9.5 9.5 4.5 4.5 4.5 4.5 [e-/sample] ---------------------------------------------------------************