Spitzer IRAC/MUSYC Public Legacy Survey in the Extended-CDFS (SIMPLE) Reduction procedures and combined images Document version: 1 June 2007 Data Release 1 (DR1) - 2007/06/01 Data reduction version: v1.1 =============================================================================== This document describes the reduction procedures of the combined SIMPLE mosaics. A more detailed account of the data reduction and products will appear in Damen et al. (in prep), please reference this paper when using these data products in published research. TABLE OF CONTENTS: 1. OBSERVATIONS 1.1 Field 1.2 Program ID, AORs, and IRAC observing strategy 2. DATA REDUCTION 2.1 SSC pipeline processing 2.2 Post-processing of the bcd frames 2.3 Cosmic-ray rejection 2.4 Astrometry 2.5 Image combination and mosaicking 2.6 Flux calibration and zeropoints 2.7 Exposure time and RMS determination 2.8 Flags 3. DATA PRODUCTS 3.1 Images 3.2 Catalog =============================================================================== 1. OBSERVATIONS =============================================================================== 1.1 Field The SIMPLE IRAC Legacy survey consists of deep observations with the Infrared Array Camera (IRAC; Fazio et al. 2004) covering the 0.5 x 0.5 deg area centered on the GOODS IRAC imaging (Dickinson et al., in prep) of the Chandra Deep Field South (Giacconi et al. 2001). The coordinates of the fields are: RA 53.122751, DEC -27.805089 (J2000) 1.2 Program ID, AORs, and IRAC observing strategy The SIMPLE IRAC Legacy program was observed under program number GO 20708 (PI van Dokkum). The observation were divided in series of Astronomical Observation Requests (AORs), each several hours long consisting of 2x3 grids (10' x 15') with 30 minutes integration time per pointing. The AORs are organized in two epochs which cover the entire field, for an exposure time of 54 hours each (~1-1.5 hours per pointing on the sky). Epoch 1 was observed in August/September 2005 and epoch 2 was observed in January/February 2006, hence AORs of the two epochs have somewhat different rotation in the sky. The raw data and the AOR details can be obtained from the Spitzer Archive with the Leopard software package (which can be obtained from http://ssc.spitzer.caltech.edu/propkit/spot/). The observational strategy was to map around the GOODS-S field, which would appears as a hole in the center of the mosaic. For more details on observational strategy, see the proposal (proposal.pdf this directory) =============================================================================== 2. DATA REDUCTION =============================================================================== The reduction was carried out using a custom pipeline (Damen et al., in prep) which post-processes and mosaicks the Basic Calibrated Data (BCD) frames provided by the Spitzer Science center (SSC). The reduction includes the following steps: 1. SSC pipeline processing 2. Post-processing of the bcd frames 3. Cosmic-ray rejection 4. Astrometry 5. Image combination and mosaicking 6. Flux calibration and zeropoint 7. Exposure time and RMS determination 8. Flags 2.1 SSC pipeline processing The starting point for the reduction are the bcd frames produced by SSC pipeline. The epoch 1 AORs were processed by BCD pipeline version S12.4.0. The epoch 2 data was processed pipeline version S13.2.0. The following reduction steps are similar, but not identical, to those applied GOODS (http://data.spitzer.caltech.edu/popular/goods/Documents/goods_dr1.html#3.0) 2.2 Post-processing of the bcd frames We have post-processed the bcd frames to subtract the background and to correct for several artifacts caused by highly exposed pixels (primarily bright stars and cosmic rays). Detailed information on IRAC related artifacts can be found in the IRAC Data Handbook section 4 (http://ssc.spitzer.caltech.edu/irac/dh/). The following frame-level post-processing was applied: 2.2.1 Removal of HDR-mode frames: We discarding the two leading short exposures of each AOR, which were observed in "HDR-mode". 2.2.2 Median image subtraction: At each pointing within an AOR a median sky image was constructed and subtracted from each frame to remove structure of gradients in the background from bias, flatfielding, or persistence effects. 2.2.3 Column pull-down correction: When a bright star or cosmic ray reaches a level of > ~25 MJy/Sr DN in the channel 1 and 2 arrays, there is a change in the intensity of the column in which the signal is found. The intensity is reduced throughout the column (thus the term "column pull-down"). When the effect occurs, it shifts the intensities of the pixels above and below the position of the offending" pixels. While the intensity shift is slightly different below and above the offending pixel and has a small slope, the shift is nearly constant in practice. We therefore correct for the effect by 1) locating the columns of > ~25 MJy/Sr DN pixels 2) masking all bright sources in the frame, and 3) subtract the median of the affected columns excluding any sources. We favor the simplest correction because its implementation is more robust than fitting a general 2 segment slope. 2.2.4 Muxbleed correction: Muxbleed occurs in IRAC channels 1 and 2 (3.6 and 4.5 microns), appearing as a trail of pixels with enhanced (additive) output level, repeating every 4th column, trailing a bright source on the row. The effect can wrap around rows, but not from the first row to the last. We applied a very simple cosmetic correction. For each offending pixel(> 25 MJy/Sr), we generate a list of pixels selecting every 4th pixels next in the row and previous in the row. Then we median filter the pixel list with a filter width of 20 pixels and subtract the result. This procedure removes the bulk of the muxbleed signal, but not in all cases. Note that effect of residual muxbleed signal on each point in the final mosaic is reduced because the AORs were obtained in two epochs with different field rotation. The data products (see section 4) include a map showing which pixels were muxbleed corrected. 2.2.5 Muxstriping correction: In addition to muxbleed, stars, hot pixels, and particle/radiation hits can generate a muxstripe pattern. Where muxbleed only affects pixels on the same row, the muxstripe pattern may extend over a significant part of the image, albeit to lower levels. Muxstriping appears as an extended jailbar pattern preceeding and/or following the bright pixel. It is a fairly subtle effect, usually only visible in individual frames around very bright stars, but it becomes easily visible in deeper combined frames. We remove this effect in each frame by applying an offset in zones surrounding the offending pixels using a program kindly provided by Leonidas Moustakas of the GOODS-team (please see upcoming GOODS IRAC documentation for a further description). 2.2.6 Persistence masking: Bright sources leave positive residuals on subsequent readouts of the array (persistence), although much of the signal subsides after 6-10 frames. We mask for persistence by creating a mask of all highly exposed pixels in a frame and masking those pixels in the 6 subsequent frames. 2.2.7 background subtraction (full array offset): After correction for artifacts, the pipeline subtracts a constant background by 1) iteratively thresholding and masking pixels associated with sources and calculating the mode and rms of the remaining background pixels, 2) subtracting the mode of the image 2.3 Cosmic-ray rejection For each AOR, a first pass registered image is created from the post-processed bcd frames. For the first pass image the bcd "brmasks" were used as a first guess to mask candidate cosmic rays, and the image was median combined, so it should be free of any deviant pixels. The first pass images is then aligned and subtracted from each exposure, and the result is divided by the asociated bcd "bunc" image, which contains estimates of the uncertainties in each pixel based on a noise model. Pixels are flagged as cosmic-rays in this detection image if they deviate more then a certain threshold. Pixels adjacent to deviant pixels are also flagged using a lower threshold. Hot and cold pixels are also identified and masked in this step. 2.4 Astrometry The SIMPLE astrometry has been calibrated to a compact source catalogue detected in a combined deep BVR image from the Multiwavelength Survey by Yale-Chile in the ECDFS (http://www.astro.yale.edu/MUSYC/). The MUSYC data are planned for public release on 1 July 2007. The calibration was done on per-pointing combined frames within an AOR (216 pointings in total, 108 per epoch), and measuring the positions of the reference sources in these combined images. The astrometric differences between the reference catalog and the SIMPLE pointings were small (up to ~1") and could be corrected by applying a simple shift. There is no evidence for rotation, or higher order geometric distortion. We therefore applied a simple offset to the WCS crval1,crval2 of the bcd-frames to refine the pointing. The pointing refinement solutions determined for the 3.6 and 4.5 micron BCDs were applied to the 5.8 and 8.0 micron images, respectively, as there are generally few bright sources at 5.8 and 8.0 micron to derive them independently. The resulting astrometry accuracy relative to the MUSYC ECDFS BVR catalog is typically ~0.02" (averaged per IRAC pointing), with source-to-source rms of ~0.12" in channel 1/2 and ~0.14" in channel 3/4. Large scale shears, systematic variations on scales of a few arcminutes, are 0.05" or less. Note: the quoted astrometric uncertainties that are relative to the MUSYC BVR catalog, but we verified that the astrometry agrees very well (~0.1" level) with the "wfiRdeep" image used as a basis for the ACS GOODS astrometry. The relative astrometry of all SIMPLE mosaics is therefore probably better than 0.1" over the entire field. Absolute astrometry, from comparison to 2MASS sources, should be better than 0.3". 2.5 Image combination and mosaicking After individual processing, the individual bcd frames are mosaicked onto an astrometric reference grid using the refined astrometric solution in the frame headers. 2.5.1 Reference grid For the reference grid we adopt the tangent point, pixel size, and orientation of the GOODS IRAC images (RA--TAN 53.122751, DEC--TAN -27.805089, 0.6"/pixel, pixel axes are aligned with the J2000 celestial axes, see http://data.spitzer.caltech.edu/popular/goods/Documents/goods_dr1.html). Also following GOODS, we put the tangent point (CRVAL) at a half-integer pixel position (CRPIX). This ensures that images with integer pixel scale ratios (e.g., 0.3", 0.6", 1.2") can (in principle) be directly rebinned (block summed or replicated) into pixel alignment with one another. This puts GOODS, SIMPLE, and the upcoming FIDEL survey (a deep 24/70 micron survey in the ECDFS) on the same astrometric grid. However, note that for direct accurate alignment the astrometric solution of the data sets should also agree. For example, SIMPLE, ACS/GOODS, and FIDEL 24 micron are consistent, but we have found astrometric differences with the GOODS-S IRAC data set of up to 0.4" amplitude. These residuals are intrinsic to the GOODS IRAC data set and will be addressed by the GOODS team in their future releases. The final SIMPLE mosaic extends 38' x 48' (3876 x 4868 pixels). 2.5.2 Image combination For each epoch, the individual post-processed bcd frames are transformed to the reference grid using bicubic interpolation, taking into account geometric distortion of the bcd frame. Cosmic rays and bad pixels are masked and the frames are average combined without additional rejection. The undersampling of the IRAC PSF in the original frames and the interpolation step leads to a modest blurring of the images. The result is that the PSFs in the SIMPLE mosaic have slightly (~10%) larger PSFs then could have been achieved with other techniques (e.g. using the "drizzle" method of Fruchter & Hook 2002). However, drizzling onto our reference pixel scale (0.6") would lead to large variation in the pixel-to-pixel exposure time for the substantial areas of the mosaic that have lower overall exposure, which would complicate outlier rejection. Finally, the seperate epoch1 and epoch2 mosaics were weighted with exposure time combined. By design, the SIMPLE ECDFS observational strategy was to map around the GOODS-S, which would leave a 10' x 10' hole in the combined mosaic. To facilitate easy analysis, we have also added the GOODS-S IRAC data (DR3, mosaic version 0.3, http://data.spitzer.caltech.edu/popular/goods/20051229_enhanced_v1/) to the center of the SIMPLE mosaic. We have shifted the GOODS-S IRAC mosaics by ~0.2" to bring its astrometry in better agreement with SIMPLE. Some systematic astrometric residuals of 0.1 - 0.25" amplitude remain in part of the GOODS field, which will likely be resolved in the next GOODS data release. To ensure a seamless combination between the epoch1, epoch2 and GOODS-S images, we subtracted an additional background from the images before combination. The background algorithm masks sources and measures the mode of the background in tiles of 1 x 1 arcminute. The "mode-map" is then smoothed on scales of 3 x 3 arcminutes and subtracted from the image, resulting in extremely flat images and a zero background level on scales > arcminute. 2.6 Flux calibration and zeropoints The epoch 1 and epoch2 science images were scaled to a common zeropoint so that their data units agree. For convenience, we adopt the GOODS-S IRAC zeropoints (see also 00README_photometry). Calibration to the GOODS-S images was done by minimizing the count rate differences of bright, non-saturated stars in kron-like apertures in regions where the images overlap. The relative accuracy of the zeropoint can be estimate by a 'blind' comparison to the GOODS fluxes using the original SIMPLE pipeline zeropoints provided by the SSC. This indicates that the fluxes agree within ~3%. The absolute accuracy of IRAC flux calibration is likely < 5% for channel 1 and 2 and < 10% for channel 3 and 4 (see http://ssc.spitzer.caltech.edu/irac/dh/). The zeropoints of final images is such that a countrate of 1 DN/sec corresponds to flux densities and AB magnitudes of: channel wavelength fluxconv(microJy/(DN/s)) zeropoint(AB) 1 3.6mu 3.922 22.416 2 4.5mu 4.808 22.195 3 5.8mu 20.833 20.603 4 8.0mu 7.042 21.781 2.7 Exposure time and RMS determination The exposure time maps are created by multiplying, at each position, the number of bcd-frames that were used to form the final image by the integration time of each frame. The exposure map thus reflects the exposure time in seconds on that position of the sky, not the average exposure time per final output pixel. The oversampling and image combination have resulted in correlated noise in the pixels. Therefore the pixel-to-pixel variation measured directly from the final mosaic will underestimate the true variance. To enable better error determinations, this release provides RMS maps. The RMS maps were created by 1) multiplying the final mosaic by the sqrt(exptime map/median(exptime)), 2) rejecting sources, 3) binning the image by a factor 4 x 4, and 4) calculating the rms statistic of the binned pixels in a moving window of 15 x 15 bins. The result is approximately the rms background variation at scales of 2.4" arcseconds at the median exposure time, which does not suffer from any correlations. We multiply this value by sqrt(4)/sqrt(exptime map/median(exptime)) to get the per-pixel rms variation at the mosaic pixel scale for other exposure times (see e.g., Labbe et al. 2003). This rms map does not directly reflect the contribution to the uncertainty of source confusion. The 25%, 50% and 75% percentiles of the final exposure maps (excluding GOODS-S) are ~3100, 5500 and 9100s (0.9, 1.5 and 2.5 hours) for all channels. The corresponding area with at least that exposure time are ~1200, 800 and 400 arcmin^2 respectively. In addition, the central GOODS-S mosaic has ~20 hours per pointing over ~160 arcmin^2. 2.8 Flags We provide a Flag map, which currently only identifies pixels corrected for muxbleed in channel 1 and channel 2. Since these correction are not optimal, it may be useful to find pixels which may have been affected. The flag image is a bit map, i.e. an integer map that represent the sum of bit-wise added values. bit 1 (flag = 1) indicates a muxbleed correction in the first epoch, bit = 2 (flag = 2) indicates a corrrection in the second epoch. =============================================================================== 3. DATA PRODUCTS =============================================================================== 3.1 Images The first data release of SIMPLE consists of FITS images of all IRAC observations in the E-CDFS. We provide science images, exposure time maps, rms maps, and a flag map. These images comprise combined mosaics of all data taken (both epochs), including also the 10' x 15' GOODS IRAC mosaics in the center (see http://data.spitzer.caltech.edu/popular/goods/20051229_enhanced_v1/). In addition, we provide combined mosaics and exposure maps of the data of the individual epochs (without GOODS), which may be useful to study the reliability and/or variability of sources. The units of the science and rms images are DN/second, with the (GOODS) zeropoints as in section 2.6. The units of the exposure time maps are seconds. 3.2 Catalogs We provide a simple object catalog of IRAC-detected sources created with SExtractor (Bertin & Arnouts 1996). The catalog comprises a list of sources detected at ~10 sigma in a combined 3.6+4.5 micron image. Positions and simple photometry in various apertures are provided. We caution however that these catalogs are very crude. The IRAC point spread functions (PSFs) are quite large and many sources are blended or confused. In addition, while the aperture measurements have been corrected for differences in PSFs across the bands, the correction is still rather crude. Careful attention is needed before interpreting the fluxes in this catalog. The photometry is discuss in more detail in 00README_photometry. REFERENCES Bertin & Arnouts 1996, A&AS, 117, 393 Damen et al, 2007, in preparation Labbe et al., 2003, AJ, 125, 1107