GLIMPSE Data Flow version 16-feb-2001 Level A - Validation 1) Basic Calibrated Data (BCD) received from SSC 2) Check each supertile a) Is the AOR executed correctly? - does header position agree with AOR request? - number of images expected in each bandpass - header info agrees with AOR specs (every frame) - think of more tests b) Frame quality (automated checks on all frames) - bad downlink data? (checksums?) - dead pixels - check BCD flags for bad pixels, cosmic rays. count up bad pixels. if above a certain threshold, alert; may need to re-observe. - do our own (simple) bad pixel check to verify: look for pixels above and below certain thresholds; more - striping (??) - file integrity (??) - noise level (pixel-to-pixel variations) - make use of SSC algorithms if possible* - other tests c) photometry spot check - do this on central frame, first observed frame, last observed frame? - position vs. MSX/2MASS (e.g. http://tdc-www.harvard.edu/software/wcstools.html). - fluxes vs. MSX/2MASS - image quality (psf check). (focus can change going from MIPS to IRAC due to temperature change) - sensitivity d) AOR completion (close-out) - Accept, or Re-observe *Note that SSC plans to visually inspect every BQD (Browse Quality Data) before releasing data. I think the BQD is a "mosaic" of all the BCDs in an AOR. They will inspect for missing data, striping, noise, patches of bad pixels, ghost images (see pp. 9-10 of SSC Pipeline Overview handout). 3) Compare these parameters between supertiles to watch for variations in PSF, sensitivity, noise. 4) Track survey completion, revisit each position Level B - Basic processing and Mask Propagation 1) Start with Basic Calibrated Data (BCD) "Most users are expected to begin their analyses from the BCD." These are fits files that contain header information followed by the image data. Along with the calibrated image come many calibration files that will be useful for us (Latent image, cosmic ray image, mask file, log files, and more). The following corrections have been applied or flagged to this point (see p.2 of SSC Pipeline Overview handout): - dark/bias subtraction - flat-fielding - linearization of detector response - cosmic ray detection (maskfile) - saturation detection (tags) - latent image detection (mask file) 2) GLIMPSE processing on BCD Most of the operations in Level B will be flags, masks, instructions in the fits header for corrections to be applied later. The only correction to the images performed in Level B is the banding correction (so far). It's possible other operations on the images can be performed in Level B (as opposed to Level C). Please give your input on this. a) banding correction. is that the same thing as striping? blooming? - Did Tony say something about SSC having a model for this? depends on electronics? b) quality flags (similar to Level A): - noise, pixel-to-pixel variation - pointing accuracy? - others that will come up c) cosmic ray flag (unless we are satisfied with SSC pipeline) d) bad pixel flags (already provided but perhaps we want to do more?) e) tune latency flags (if desirable) - currently all that is provided is a mask file, soft or hard saturation. - SSC may provide Latent Strength Estimation, and therefore latent image which implies that you could subtract out latent fluxes. Question: do we subtract latent image at this stage? - instrument teams may have latency model f) geometric correction instructions to be applied later; uncertain what SSC provides on this. They don't do the correction but may provide the information we need (in the header). - pixel area maps g) Zodi background flag (value in header to subtract from frame at a later stage, based on favorite model, observation date, and position in sky) h) resample to galactic coordinates (0.6"/pixel scale). populate a supermosaic image with information on sky coverage and number of observations contributing at each pixel. Level C - Processing for BSC (Bright Source Catalog) 1) Photometry on each frame - apply necessary corrections from Level B for each frame before doing photometry - replace bad pixels, cosmic rays with median of surrounding pixels. - latent image correction. either throw out flagged pixels or subtract latent image - Zodi subtraction? (using values calculated in Level B). - photometry (DAOPHOT?) 2) BSC - Bright Source Catalog (at least 20 sigma in all four wavelength bands) - add photometry results for multiple observations (in quadrature) - Quality control. - catalog entries: ID, lat, lon, RA, Dec, flux in each band, flux error Level D - Processing for GSC (GLIMPSE Source Catalog) and Mosaics 1) Mosaic - Register and resample images to a galactic coordinate frame - look for solar system objects in multiple frames and flag - apply geometric corrections, flux conserving. - in positions where bad pixels or cosmic rays exist in one frame, use good data wherever possible from multiple observations. - combine into large mosaic with several layers to include information from each image that contributes to flux at a give pixel. Due to memory requirements, the large mosaic will have to be divided up into smaller pieces, the size as yet to be determined, but possibly 1x1 degree. - It's possible we can use SSC post-pipeline software that does the same thing (p. 8 of SSC Pipeline Overview handout). 2) Refine Positional accuracy by comparing to other catalogs - SSC has a a tool for this (p. 7 of SSC Pipeline Overview handout) but no plans to make it available to teams. - see http://tdc-www.harvard.edu/software/wcstools.html - Jeff Percival has a tool for this 3) Photometry on mosaicked images - apply necessary corrections from level B on mosaiced frame, including Zodi subtraction - apply high-pass filter (not sure yet if we will do this). - photometry (probably DAOPHOT) 4) GSC - point source catalog (5 sigma) - include information on extended sources, such as values in different apertures, ellipticity, position angle - Quality control. - Catalog entries: ID, lat, lon, RA, Dec, flux in each band, flux error, point source quality, DAOPHOT info on size ellipticity, PA, background, data quality flags (latency, cosmic ray, blended object). Web modeling Tools Modifications for SIRTF are to be a joint effort between our team and Neal Evan's team with Lee Mundy at UMD A) First Order Products: 1) Extend the model parameter range to lower Luminosity and Lower Teff appropriate for brown dwarfs (L = 0.01 L0, Teff = 1500 K), bracket and fill in the "new" spectral classes. 2) Add in Av = 0 spectral energy distributions. 3) Create new SIRTF specific modules: - Convert units to common SIRTF values (microns, Jy). - For each IRAC band plot model intensities I vs. source size (arcsec). Convolve model sources (at user input distances) with the IRAC filters and (cleaned?) PSFs. - Enter IRAC observations in viewer and search for best fit models - Repeat for MIPs filters. 4) Simple fitting tools - Estimate total shell mass from MIPS, single T fits from band ratios, power law dust emissivities from T and SEDs. B) Second Order Products: - 2-D analysis. Enter map and find radial profile from data. Subrtract off model profiles to find goodness of fit. C) Third Order Products: TBD 1) Add disk emission 2) Add PAH emission to spherical code 3) Add additional large grain models 4) Add PAH emission to PDR code and show IRAC bands from face-on and side-on PDRs Deliverables 1) BSC - Bright Source Catalog 2) GSC - GLIMPSE Source Catalog 1) mosaicked images 2) Web modeling Tools Enhanced Science Products not delivered to SSC. only do this if after Deliverables 1) correlations to other databases (SSC may provide this after they receive our catalogs) 2) SEDs 3) background-matched supermosaics?