-xg-gsp.fits: -tab dec'04,jun'06,jan'08 Binary FITS table column descriptions for *-xg-gsp.fits files: (Total number of fields: 96.) ggoid : (2J) : Global ID of spectral source Joining these two integers yields a 64 bit integer uniquely identifying this source including the tile number, band, product number, reduction try (path), etc. The bits from most significant to least are: 3 bits: vsn : Version number (of the pipeline) 16 bits: tile : Tile or field or target number 2 bits: type : Observ. type (00=single,01=AIS(multi)) 2 bits: ow : optics wheel(01=drct,10=grsm,11=opaq,00=undef) 4 bits: prod : product number(_visits,_best,etc.) 12 bits: img : visit/sub-visit number {when type=01: 5 bit subvis + 7 bit visit} 3 bits: try : try number (processing number) 2 bits: band : Band (01=nuv,10=fuv,11=both,00=undefined) 20 bits: xid : Extraction ID The extraction ID is the same as the extraction ID in the -xd-mcat.fits file used for this spectral reduction. ggoid_D : (2J) : Global ID of direct image source This is the GGOID for the corresponding source in the direct image catalog (the -xd-mcat.fits file). band : (1J) : 1=nuv only, 2=fuv only, 3=both From the GGOID. id : (1J) : Spectral source ID Extraction ID from the GGOID. alpha_j2000 : (1D) : Right Ascension J2000 (deg) delta_j2000 : (1D) : Declination J2000 (deg) Position from the direct image merged catalog file (-xd-mcat.fits). nuv_nc : (1J) : number of columns in image strip (for NUV) nuv_nr : (1J) : number of rows in image strip This defines the dimensions of spectral image strip (from 'pri' file) from which the spectrum is extracted. The scale is 1 arcsecond per pixel. These numbers are usually the same for all sources, but could vary for extended objects. nuv_mag : (1E) : magnitude from direct source catalog NUV calibrated AB magnitude derived from a NUV source catalog magnitude. See 'mcat' file column descriptions for more details. nuv_detx : (1E) : Detector X position of undeviated wavelength point nuv_dety : (1E) : Detector Y position of undeviated wavelength point Estimate of NUV detector X and Y position of a given source at the 'zero' point of the dispersion. This is approximately the position where the object would appear in the absence of the grism. Since the telescope is dithered and many eclipse observations are combined, this is only an estimate of the X position. The units are in pixels (1.5 arcseconds per pixel) relative to the nominal field center. nuv_Q : (1E) : Q value at undeviated wavelength point NUV Q value taken from the calibrated universal Qmap image file at the position given by 'nuv_detx' and 'nuv_dety'. nuv_mid_detx : (1E) : Detector X position of middle of primary order nuv_mid_dety : (1E) : Detector Y position of middle of primary order Estimate of NUV detector X and Y position of a given source at the middle of the primary order (1st order for NUV). This point is used to find the grism flat field (wavelength dependent) correction factor. The middle of the NUV 1st order is defined as 185 arcseconds from the undeviated wavelength point (about 2386 Angstroms). nuv_mid_Q : (1E) : Q value at middle of primary order NUV Q value taken from the calibrated universal Qmap image file at the position given by 'nuv_mid_detx' and 'nuv_mid_dety'. nuv_mid_ra : (1D) : RA of middle of primary order nuv_mid_dec : (1D) : Declination of middle of primary order The RA and DEC position on the sky where the middle of NUV 1st order appears on the whole sky images (such as the -int.fits image). The undeviated wavelength point would appear at 'alpha_j2000' and 'delta_j2000' described above. fuv_nc : (1J) : number of columns in image strip (for FUV) fuv_nr : (1J) : number of rows in image strip See 'fuv_nc/nr' descriptions above. fuv_mag : (1E) : magnitude from direct source catalog FUV calibrated AB magnitude derived from a FUV source catalog magnitude. See 'mcat' file column descriptions for more details. fuv_detx; : (1E) : Detector X position of undeviated wavelength point fuv_dety; : (1E) : Detector Y position of undeviated wavelength point See 'nuv_detx/y' descriptions above. fuv_Q; : (1E) : Q value at undeviated wavelength point FUV Q value taken from the calibrated universal Qmap image file at the position given by 'fuv_detx' and 'fuv_dety'. fuv_mid_detx : (1E) : Detector X position of middle of primary order fuv_mid_dety : (1E) : Detector Y position of middle of primary order Estimate of FUV detector X and Y position of a given source at the middle of the primary order (2nd order for FUV). This point is used to find the grism flat field (wavelength dependent) correction factor. The middle of the FUV 2nd order is defined as 300 arcseconds from the undeviated wavelength point (about 1601 Angstroms). fuv_mid_Q : (1E) : Q value at middle of primary order FUV Q value taken from the calibrated universal Qmap image file at the position given by 'fuv_detx' and 'fuv_dety'. fuv_mid_ra : (1D) : RA of middle of primary order fuv_mid_dec : (1D) : Declination of middle of primary order The RA and DEC position on the sky where the middle of FUV 2nd order appears on the whole sky images (such as the -int.fits image). The undeviated wavelength point would appear at 'alpha_j2000' and 'delta_j2000' described above. numpt : (1J) : Number of points in spectra The actual number of real data points in the spectral arrays obj,objerr,etc. Note that this may be smaller than the number of allocated values. objwidth : (1E) : 'arcsec ' : Object extraction width (arcsec) The object extraction window width in arcseconds. This is the region of the spatial profile centered on the expected center of the source used to accumulate the flux at each point along the spectrum. Note that this is the total size of extraction aperture (i.e a diameter, not a radius). This number should be an even integer to allow proper centering of source. bckwidth : (1E) : Background extraction width (arcsec) This is the width in the spatial direction used for the background estimate. The background excludes the object window and also regions immediately adjacent to the object (see objbcksep). If the scale is 1.0 arcsecond/pixel (default), then the total number of rows in the image strip is the background width plus one, since the background width is the distance between the center of the first pixel to the center of the last pixel. The image strip has one additional row which is the spectral response deviations (currently not used). So if bckwidth = 78, the number of rows is 80. This number should be an even integer to allow proper centering of source. objbcksep : (1E) : Object-Background separation (arcsec) The separation between the object window and the background windows which bracket each side of the object window. bckcolrad : (1J) : Radius for averaging background columns (pix) Number of columns to average on either side of a given column to compute the background level at each point along the spectrum. A value of 0 means only use the column at each point to determine the background. A value of 1 means include 1 column on either side of the primary column to compute the background at each spectrum point. truthid : (1J) : Truth catalog ID (matched) Not used. objtype : (1J) : Object Type(0=galaxy,1=star) A guess at the object type based on the direct image colors. temperature : (1E) : Estimated effective temper.(K) From a blackbody fit to the spectrum. redshift : (1E) : Estimated redshift Not used. slope : (1E) : Estimated spectral slope A power-law fit to the spectrum, this is alpha in equation: f_nu (proportional to) nu^alpha. Or, in our case, photons per sec. per cm^2 per Ang. proportional to lambda^(-(1+alpha)). qa : (1E) : Quality Analysis value Not currently used. flag : (1J) : Miscellaneous flag (not used) Not currently used. weight : (2E) : Average weight (each band) Average response in effective seconds for the unmasked pixels in the spectral image strip. (NUV is the first element.) This includes flat field response and neighbor masking. background : (2E) : Average background (each band) Average background in photons per (effective) second per pixel for the unmasked pixels within the background limits. The pixel size in the image strips is nominally 1 square arcsecond. (NUV is the first element (value), FUV is the second value.) median_s2n : (2E) : Median Signal/Noise (each band) Median S/N per spectral pixel (wavelength point) in the flux calibrated spectra. Each spectral pixel is nominally 3.5 Angstroms in size. For NUV the range used for this computation is 1755 to 3000 Angstroms. For FUV the range is 1300 to 1745 Angstroms. (NUV is the first element (value), FUV is the second value.) centroid : (2E) : Relative position(spatial)(each) Centroid of the profile relative to the expected object position. The expected object position is the given by the number of rows in the image strip (-pri.fits or -prc.fits): (object center) = ((number of rows in image strip) - 2) / 2 . This assumes the center of a pixel in the first row is 0.0 . Note that the last row of an image strip is for the spectral response variation (currently 1.0). The centroid value is relative to the expected object center, so a perfectly aligned object would have a centroid value of 0.0. (NUV is the first element (value), FUV is the second value.) specfwhm : (2E) : Estimated spatial FWHM (each band) This is an approximate estimate of the FWHM (in arcseconds) of the spatial profile. Assuming a gaussian, it uses the formula: FWHM = 0.93944 x (area of profile) / (peak pixel value) (NUV is the first element (value), FUV is the second value.) effexp : (2E) : Effective exposure time in spectral image strip For each band (NUV and FUV, respectively), the effective exposure time in seconds for the entire spectral image strip (object and background regions) for a given source. This value takes into account the varying response of the flat field and corrections for dead spots. It does not include masking of neighboring sources by spectral extraction programs. effmsk : (2E) : Effective masking fraction in spectral image For each band, the effective masking fraction over the entire spectral image strip. This includes masking of neighboring sources. No masking would be 0., masking of 10% of the image strip would be 0.10. Multiplying "effexp" and "1.0 - effmsk" gives you approximately the "weight" value. nuv_numpro : (1J) : Number of points in NUV profile Number of points in the NUV profile array. This is one less than the number of rows in the spectral image strip, since the last row in the image strip is the spectral response variation. fuv_numpro : (1J) : Number of points in FUV profile Number of points in the FUV profile array. nuv_profile : (198E) : NUV profile (spatial) NUV profile array. Values are in photons/second/pixel (a pixel is nominally 1 square arcsecond). The column range used to compute the average profile is offsets +50 to +275 arcseconds for NUV (offset=0 is the undeviated wavelength position on the image strip). This is roughly the NUV first order portion of the spectrum. Masked pixels are ignored. See the description of the "centroid" column for the definition of the "object center row". fuv_profile : (198E) : FUV profile (spatial) FUV profile array. Values are in photons/second/pixel (a pixel is nominally 1 square arcsecond). The column range used to compute the average profile is offsets +175 to +375 arcseconds for FUV. This is roughly the FUV second order portion of the spectrum. Masked pixels are ignored. nuv : (1E) : Direct image flux rate (NUV) NUV flux in photons per (effective) second from direct image source extraction. This is the "nuv_flux" column in the bandmerged direct image source catalog (-xd-mcat.fits), but converted to photons per second rather than micro Janskys. nuve : (1E) : Error in direct image flux rate One sigma error in "nuv" column value. fuv : (1E) : Direct image flux rate (FUV) FUV flux in photons per (effective) second from direct image source extraction. fuve : (1E) : Error in direct image flux rate One sigma error in "fuv" column value. E_bv : (1E) : E(B-V) Galactic Reddening. Copied from direct image catalog (see 'mcat' file description). nuv_dfwhm : (1E) : NUV FWHM from direct image in arcseconds. fuv_dfwhm : (1E) : FUV FWHM from direct image in arcseconds. These are the direct catalog 'SExtractor' values of 'FWHM_IMAGE' converted to arcseconds (see 'mcat' file description). nuv_artifact : (1J) : NUV Artifact flag from direct image. fuv_artifact : (1J) : FUV Artifact flag from direct image. Copied from direct image catalog (see 'mcat' file description). nuv_nc_r2 : (1I) : NUV neighbor count from direct image within radius R2. fuv_nc_r2 : (1I) : FUV neighbor count from direct image within radius R2. Copied from direct image catalog (see 'mcat' file description). nuv_nf_r2 : (1E) : NUV neighbor flux from direct image within radius R2. fuv_nf_r2 : (1E) : FUV neighbor flux from direct image within radius R2. Copied from direct image catalog (see 'mcat' file description). cn_radius : (1E) : Closest neighbor radius from direct image (arcsec). Copied from direct image catalog (see 'mcat' file description). nuvs : (1E) : Spectral flux scaled to direct NUV flux derived from the extracted spectrum. This is computed by summing the products of the flux in each pixel of the flux calibrated spectrum by the direct image effective area and the dispersion. The wavelength range used in the NUV is 1800 to 2800 Angstroms. Since this does not include the edges of the bandpass, a correction factor (guess) of 1.0254 is applied to the sum. Some discrepancy will remain between the direct and grism results, but this value is useful for quality analysis. nuvse : (1E) : Error in nuvs One sigma error in "nuvs" column value. fuvs : (1E) : Spectral flux scaled to direct FUV flux derived from the extracted spectrum. This is computed by summing the products of the flux in each pixel of the flux calibrated spectrum by the direct image effective area and the dispersion. The wavelength range used in the FUV is 1350 to 1700 Angstroms. Since this does not include the edges of the bandpass, a correction factor (guess) of 1.1071 is applied to the sum. Some discrepancy will remain between the direct and grism results, but this value is useful for quality analysis. fuvse : (1E) : Error in fuvs One sigma error in "fuvs" column value. nuv0s : (1E) : NUV Zeroth order flux rate NUV flux derived from only the 0th order spectrum. The wavelength range is 2000 to 2300 Angstroms and no correction is applied. Although smaller than the value "nuvs", each of the values of nuv0s,nuv1s,nuv2s should be the same and can be used for quality analysis. nuv0se : (1E) : Error in NUV Zeroth order flux rate One sigma error in "nuv0s" column value. nuv1s : (1E) : NUV First order flux rate NUV flux derived from only the 1st order spectrum. The wavelength range is 2000 to 2300 Angstroms and no correction is applied. nuv1se : (1E) : Error in NUV First order flux rate One sigma error in "nuv1s" column value. nuv2s : (1E) : NUV Second order flux rate NUV flux derived from only the 2nd order spectrum. The wavelength range is 2000 to 2300 Angstroms and no correction is applied. nuv2se : (1E) : Error in NUV Second order flux rate One sigma error in "nuv2s" column value. fuv1s : (1E) : FUV First order flux rate FUV flux derived from only the 1st order spectrum. The wavelength range is 1350 to 1600 Angstroms and no correction is applied. Although smaller than the value "fuvs", each of the values of fuv1s,fuv2s,fuv3s should be the same and can be used for quality analysis. fuv1se : (1E) : Error in FUV First order flux rate One sigma error in "fuv1s" column value. fuv2s : (1E) : FUV Second order flux rate FUV flux derived from only the 2nd order spectrum. The wavelength range is 1350 to 1600 Angstroms and no correction is applied. fuv2se : (1E) : Error in FUV Second order flux rate One sigma error in "fuv2s" column value. fuv3s : (1E) : FUV Third order flux rate FUV flux derived from only the 3rd order spectrum. The wavelength range is 1350 to 1600 Angstroms and no correction is applied. fuv3se : (1E) : Error in FUV Third order flux rate One sigma error in "fuv3s" column value. zero : (1E) : Wavelength of first pixel (Ang.) Zero point for the wavelength scale in Angstroms. The true wavelength is given by: wavelength = zero + ((pixel#) * disp), where the first pixel is pixel# 0. disp : (1E) : Angstroms per pixel Dispersion for the wavelength scale in Angstroms per pixel. The true wavelength is given by: wavelength = zero + ((pixel#) * disp), where the first pixel is pixel# 0. obj : (552E): Flux (pho/sec/cm^2/Ang) (standard) Spectrum array in photons per second per cm^2 per Angstrom. This is the standard reduction which means a simple sum of the pixels in the object region (given by "objwidth", see -gsax.fits file) after dividing by the response image and subtracting an average background level at each wavelength point (or column). Only NUV 1st order and FUV 2nd order are used to compute this spectrum. objerr : (552E): Error on obj One sigma error for obj[] array. opx : (552E): Flux (pho/sec/cm^2/Ang) (optimal) Optimally-weighted extracted spectrum using a profile model. The profile model is the same for the entire field, so this computation is best for unresolved sources. opxerr : (552E): Error on opx One sigma error in optimal extraction spectrum. objs : (552E): Flux (pho/sec/cm^2/Ang) (FUV 3rd + NUV 2nd) Secondary spectrum-- this is similar to "obj" but only the NUV 2nd order and FUV 3rd order are used to extract the spectrum (instead of only NUV 1st and FUV 2nd). This is expected to have a much lower S/N but will have a higher spectral resolution than "obj". objserr : (552E): Error on objs One sigma error for objs[] array. objmdn : (552E): Flux spectrum using median image strip. Spectrum extracted from the median image strip files (-prm.fits). Only used for co-added or visit-combined data. This array is not used for individual visits.