Difference between revisions of "Sky temperature maps"
(→Header keywords) |
(→HFI processing) |
||
| Line 13: | Line 13: | ||
<!-- ==== FIRAS calibration ==== --> | <!-- ==== FIRAS calibration ==== --> | ||
| − | At the higher frequencies, 545 and 857 GHz, calibration | + | At the higher frequencies, 545 and 857 GHz, calibration was initially determined by fitting HFI data onto FIRAS data, but these are then corrected on the basis of results from planet observations, the CMB itself, and more. These zero point are applied for all frequencies to set them at the same scale. |
<!-- ==== Map-making ==== --> | <!-- ==== Map-making ==== --> | ||
| − | A destriping approach is used, where noise is modelled as the sum of a white noise component and a constant, aka offset, per pointing period which represent the low frequency 1/f noise. After subtracting these offsets, calibrated data are projected onto Healpix maps, with the data of each bolometer weighted by a factor of 1/NET of that bolometer. Maps are build after a simple classical dipole removal based on WMAP measurements. | + | A destriping approach is used, where noise is modelled as the sum of a white noise component and a constant, aka offset, per pointing period which represent the low frequency 1/f noise. After subtracting these offsets, calibrated data are projected onto Healpix maps, with the data of each bolometer weighted by a factor of 1/NET of that bolometer. Maps are build after a simple classical dipole removal based on WMAP measurements. |
| − | Together with signal maps, hit count and variance maps are also produced. The hit maps give the (integer) number of valid TOI-level samples that contribute to the signal of each pixel. All valid samples are counted in the same way, i.e., there is no weighting factor applied. The variance maps project the white noise estimate, provided by the NETs, in the sky domain. They give the covariance between the Stokes parameters I,Q andU inside each pixels, whenever polarization is reconstructed. | + | The Zodiacal light and the Far-side-lobes are modeled separately at HPR-level, and subtracted from the signal HPR before projecting them onto the maps. A product showing the difference between the uncorrected and the corrected maps is provided. |
| + | |||
| + | Together with signal maps, hit count and variance maps are also produced. The hit maps give the (integer) number of valid TOI-level samples that contribute to the signal of each pixel. All valid samples are counted in the same way, i.e., there is no weighting factor applied. The variance maps project the white noise estimate, provided by the NETs, in the sky domain. They give the covariance between the Stokes parameters I,Q andU inside each pixels, whenever polarization is reconstructed. | ||
=== LFI processing === | === LFI processing === | ||
Revision as of 13:14, 3 December 2012
Contents
Introduction[edit]
Frequency maps are produced by combining appropriately the data of several detectors over some period of the mission. They may be intensity only or polarized, meaning that they will consist of a set of three maps of I, Q, and U. These will be accompanied by a hit-count map and by a set of variance maps. All maps are in Healpix format, with Nside of 2048 for HFI and of 1024 for LFI, in Galactic coordinates, and Nested ordering. The maps are packaged into a single BINTABLE extension; the structure of the FITS file is given in the FITS file structure section below.
The FITS filenames are of the form {H|L}FI_fff_nnnn_yyyymmdd_{type}_{mission}.fits, where fff are three digits to indicate the Planck frequency band, and nnnn is the Healpix Nside of the map, the optional type indicates the subset of input data used, and mission indicates the coverage period, i.e., full, nominal, or survey_n (TBC; an alternative of the type {H|L}FI_fff_{type}_{mission}_Rn.mm.fits is being considered, where Rn.mm is the release number). A full list of products, by their names, is given in the List of products below.
HFI processing[edit]
The cleaned TOIs of each detector (calibrated in Watts) with their associated flags (flagged samples and anomalous rings) are first used to build Healpix rings, each ring containing the combined data of one pointing period. These are then calibrated in brightness, cleaned of the dipole signals, and projected onto Healpix maps as explained in the following sections.
Gains are initially determined for each pointing period. These estimates are systematically biased by Galactic foregrounds. These gains are averaged over a time interval in the first survey where the Solar dipole amplitude is large enough to reduce these bias. For 100 to 217 GHz, evidence of apparent gain variations of order 1 to 2% over a few tens to thousands of pointing periods led us to adopt a more sophisticated approach, where we determined relative gains for each pointing period, smoothed with a 50 pointing period width.
At the higher frequencies, 545 and 857 GHz, calibration was initially determined by fitting HFI data onto FIRAS data, but these are then corrected on the basis of results from planet observations, the CMB itself, and more. These zero point are applied for all frequencies to set them at the same scale.
A destriping approach is used, where noise is modelled as the sum of a white noise component and a constant, aka offset, per pointing period which represent the low frequency 1/f noise. After subtracting these offsets, calibrated data are projected onto Healpix maps, with the data of each bolometer weighted by a factor of 1/NET of that bolometer. Maps are build after a simple classical dipole removal based on WMAP measurements.
The Zodiacal light and the Far-side-lobes are modeled separately at HPR-level, and subtracted from the signal HPR before projecting them onto the maps. A product showing the difference between the uncorrected and the corrected maps is provided.
Together with signal maps, hit count and variance maps are also produced. The hit maps give the (integer) number of valid TOI-level samples that contribute to the signal of each pixel. All valid samples are counted in the same way, i.e., there is no weighting factor applied. The variance maps project the white noise estimate, provided by the NETs, in the sky domain. They give the covariance between the Stokes parameters I,Q andU inside each pixels, whenever polarization is reconstructed.
LFI processing[edit]
TBW
Types of maps[edit]
Full channel maps[edit]
Full channel maps are built using all the valid detectors of a frequency channel and cover the full mission (or the nominal mission for the 1st release). For HFI, the 143-8 and 545-3 bolometers are rejected entirely as they are seriously affected by RTS noise.
Single survey maps[edit]
Single survey maps are built using all valid detectors of a frequency channel, but cover separately the different sky surveys. The single sky surveys are defined in terms of the direction of the satellite's spin axis: the first survey covers from the beginning of the science observations (the First Light Survey) to the time when the spin axis has rotated by 180 degrees (to the nearest pointing period), the following ones covers from 180 to 360, and so on. In the case of the nominal mission, the process stops at the third survey, which is incomplete. In the case of the full mission the 4th survey was interrupted shortly before completing the 180 degree rotation (see LINK), in order to begin observing with a different scanning law. The HFI mission ended slightly before the natural end of the 5th survey, the LFI mission continued to the XXX survey. The coverage of each of these periods in terms of ring number, pointingID, and OD, is given in the table below. Note that the OD numbers are only to indicate during which OD the period boundary occurs.
| Name | Ini_OD | Ini_Ring | Ini_ptgID | End_OD | End_ring | End_ptgID |
|---|---|---|---|---|---|---|
| Nominal | 91 | 240 | 00004200 | 807 | 14724 | 04243900 |
| HFI-Full | 91 | 240 | 00004200 | 563 | 27008 | 06344800 |
| LFI-Full | 91 | 240 | 00004200 | TBD | TBD | TBD |
| SCAN1 | 91 | 240 | 00004200 | 270 | 5720 | 01059820 |
| SCAN2 | 270 | 5721 | 01059830 | 456 | 11194 | 02114520 |
| SCAN3 | 456 | 11195 | 02114530 | 636 | 16691 | 03193660 |
| SCAN4 | 636 | 16692 | 03193670 | 807 | 21720 | 04243990 |
| SCAN5 | 807 | 21721 | 95000020 | 974 | 27008 | 06344800 |
Detector-set maps[edit]
| There are no detset maps in the first release.
This section will serve for later releases. |
|---|
Detector-set (detset) maps are built for the full (or nominal) mission using a minimal set of detectors. This concept is applicable to polarization maps, which are built using two PSB pairs at the proper orientations. The HFI polarized channels are designed to provide two detsets (or quads) each, namely:
| 100–ds1: 100-1a,100-1b,100-4a,100-4b | 100–ds2: 100-2a,100-2b,100-3a,100-3b |
| 143–ds1: 143-1a,143-1b,143-3a,143-3b | 143–ds2: 143-2a,143-2b,143-4a,143-4b |
| 217–ds1: 217-5a,217-5b,217-7a,217-7b | 217–ds2: 217-6a,217-6b,217-8a,217-8b |
| 353–ds1: 353-5a,353-5b,353-3a,353-3b | 353–ds2: 353-6a,353-6b,353-4a,353-4b |
Some maps built using a set of a single detector are also provided, and are described in Single detector maps
The LFI Detector-set maps are built using pairs of horns in the same scanning row, namely:
| 18_23: 18M,18S,23M,23S |
| 19_22: 19M,19S,22M,22S |
| 20_21: 20M,20S,21M,21S |
| 24: 24M,24S |
| 25_26: 25M,25S,26M,26S |
Half-ring maps[edit]
Half-ring maps are built using only the first or the second half of the stable pointing period data. There are thus two half-ring maps per frequency channel named ringhalf_1 and ringhalf_2 respectively. These maps are built for characterization purposes in order to perform null tests.
List of products[edit]
First release[edit]
For the first release, all maps will be Intensity-only
- Full channel maps
LFI_030_2048_yyyymmdd_nominal.fits LFI_044_2048_yyyymmdd_nominal.fits LFI_070_2048_yyyymmdd_nominal.fits HFI_100_2048_yyyymmdd_nominal.fits HFI_143_2048_yyyymmdd_nominal.fits HFI_217_2048_yyyymmdd_nominal.fits HFI_353_2048_yyyymmdd_nominal.fits HFI_545_2048_yyyymmdd_nominal.fits HFI_857_2048_yyyymmdd_nominal.fits
- Single survey maps
LFI_030_2048_yyyymmdd_survey_1.fits LFI_030_2048_yyyymmdd_survey_2.fits LFI_030_2048_yyyymmdd_survey_3.fits LFI_044_2048_yyyymmdd_survey_1.fits LFI_044_2048_yyyymmdd_survey_2.fits LFI_044_2048_yyyymmdd_survey_3.fits LFI_070_2048_yyyymmdd_survey_1.fits LFI_070_2048_yyyymmdd_survey_2.fits LFI_070_2048_yyyymmdd_survey_3.fits HFI_100_2048_yyyymmdd_survey_1.fits HFI_100_2048_yyyymmdd_survey_2.fits HFI_100_2048_yyyymmdd_survey_3.fits HFI_143_2048_yyyymmdd_survey_1.fits HFI_143_2048_yyyymmdd_survey_2.fits HFI_143_2048_yyyymmdd_survey_3.fits HFI_217_2048_yyyymmdd_survey_1.fits HFI_217_2048_yyyymmdd_survey_2.fits HFI_217_2048_yyyymmdd_survey_3.fits HFI_353_2048_yyyymmdd_survey_1.fits HFI_353_2048_yyyymmdd_survey_2.fits HFI_353_2048_yyyymmdd_survey_3.fits HFI_545_2048_yyyymmdd_survey_1.fits HFI_545_2048_yyyymmdd_survey_2.fits HFI_545_2048_yyyymmdd_survey_3.fits HFI_857_2048_yyyymmdd_survey_1.fits HFI_857_2048_yyyymmdd_survey_2.fits HFI_857_2048_yyyymmdd_survey_3.fits
- Half-ring maps
LFI_030_2048_yyyymmdd_ringhalf_1_nominal.fits LFI_030_2048_yyyymmdd_ringhalf_2_nominal.fits LFI_044_2048_yyyymmdd_ringhalf_1_nominal.fits LFI_044_2048_yyyymmdd_ringhalf_2_nominal.fits LFI_070_2048_yyyymmdd_ringhalf_1_nominal.fits LFI_070_2048_yyyymmdd_ringhalf_2_nominal.fits HFI_100_2048_yyyymmdd_ringhalf_1_nominal.fits HFI_100_2048_yyyymmdd_ringhalf_2_nominal.fits HFI_143_2048_yyyymmdd_ringhalf_1_nominal.fits HFI_143_2048_yyyymmdd_ringhalf_2_nominal.fits HFI_217_2048_yyyymmdd_ringhalf_1_nominal.fits HFI_217_2048_yyyymmdd_ringhalf_2_nominal.fits HFI_353_2048_yyyymmdd_ringhalf_1_nominal.fits HFI_353_2048_yyyymmdd_ringhalf_2_nominal.fits HFI_545_2048_yyyymmdd_ringhalf_1_nominal.fits HFI_545_2048_yyyymmdd_ringhalf_2_nominal.fits HFI_857_2048_yyyymmdd_ringhalf_1_nominal.fits HFI_857_2048_yyyymmdd_ringhalf_2_nominal.fits
Second release[edit]
All the above + TBD, including polarization for the polarized frequency channels.
FITS file structure[edit]
The FITS file will have one of the following structures:
The exact order of the columns is indicative only, and the details will be given in the keywords. Keywords will also indicate the coordinate system (GALACTIC), the Healpix ordering scheme (NESTED), the units (K_cmb), and of course the detector. Details of the FITS file general structure can be found in the EFDD (ref), and the specifics of certain parameters in ICD-031 (ref)
Header keywords[edit]
A typical header for the data extension of an intensity only map is:
;----------------------------------------------------------------------------- ; EXTENSION 1: FREQ-MAP ; - Header ;----------------------------------------------------------------------------- XTENSION= 'BINTABLE' /Written by IDL: Fri Nov 30 15:13:55 2012 BITPIX = 8 / NAXIS = 2 /Binary table NAXIS1 = 603979776 /Number of bytes per row NAXIS2 = 1 /Number of rows PCOUNT = 0 /Random parameter count GCOUNT = 1 /Group count TFIELDS = 3 /Number of columns COMMENT COMMENT *** End of mandatory fields *** COMMENT COMMENT COMMENT *** Column names *** COMMENT TTYPE1 = 'I_STOKES' / TTYPE2 = 'HITS ' / TTYPE3 = 'II_COV ' / COMMENT COMMENT *** Column formats *** COMMENT TFORM1 = '50331648E' / TFORM2 = '50331648E' / TFORM3 = '50331648E' / COMMENT COMMENT *** Column units *** COMMENT TUNIT1 = 'K_CMB ' / TUNIT2 = ' ' / TUNIT3 = 'K_CMB ' / COMMENT COMMENT *** Planck params *** COMMENT UNITCONV= 0.00000 / (MJy/sr)/K_cmb unit conv. factor EXTNAME = 'FREQ-MAP' / COORSYS = 'GALACTIC' / Coordinate system ORDERING= 'NESTED ' / Healpix ordering NSIDE = 2048 / Healpix Nside FIRSTPIX= 0 / LASTPIX = 50331647 / FILENAME= 'HFI_SkyMap_100_2048_R1.00_nominal.fits' / FITS filename FILEDATE= '2012-11-30 15:13' / File creation date BAD_DATA= -1.63750E+30 / bad pixel value PROCVER = 'DX9-delta' / Product version COMMENT COMMENT ------------------------------------------------------------------------ COMMENT Full channel frequency map for nominal mission COMMENT ------------------------------------------------------------------------ COMMENT HFI-DMC objects: COMMENT in-group: MAP_DX9_2048_GALACTIC_0240_27008/ COMMENT Creation date - object name COMMENT 12-05-29 13:38 - 100GHz_W_TauDeconv_nominal_I COMMENT 12-05-29 13:38 - 100GHz_W_TauDeconv_nominal_H COMMENT 12-05-29 13:38 - 100GHz_W_TauDeconv_nominal_II COMMENT ------------------------------------------------------------------------ END
(Planck) High Frequency Instrument
(Planck) Low Frequency Instrument
Flexible Image Transfer Specification
To be confirmed
Cosmic Microwave background
Noise Equivalent Temperature
random telegraphic signal
Operation Day definition is geometric visibility driven as it runs from the start of a DTCP (satellite Acquisition Of Signal) to the start of the next DTCP. Given the different ground stations and spacecraft will takes which station for how long, the OD duration varies but it is basically once a day.
To be defined / determined
Interface Control Document
Data Management Component, the databases used at the HFI and LFI DPCs
