PACS Instrument and Calibration page for Archive Users
Welcome to the PACS instrument and calibration page. You will find here a set of top-level information about the PACS instrument and its characteristics, together with recommendations on how to best exploit the archival data offered for this instrument in the Herschel Science Archive (HSA). Because more detailed and targeted documentation is often available somewhere else, this page is essentially a gateway to all relevant information needed to best use the PACS products.
PACS consisted of two sub-instruments: an imaging photometer with three broad bands centred at 70, 100 and 160 µm, and a two-camera Integral Field Spectrometer covering 50–210 µm. Where necessary, the information provided here is separated by sub-instrument.
A quite extensive set of documentation is compiled in the PACS section of the Herschel Explanatory Legacy Library (HELL). In order to get acquainted with PACS products, and how to best exploit them, the following top-level documents from the HELL should be read first:
- If you want to know whether PACS is the right instrument for your science, please check the PACS Photometer and PACS Spectrometer Quick-Start Guides. These give an overview of the instrument performance, the observing modes, the calibration and uncertainties, the data-reduction pipelines and the products provided in any observation, and include source-specific advice and information on the science-readiness of the PACS products.
- If you want to use PACS data, and need to become familiar with their characteristics and calibration, you should read the relevant sections of the PACS Handbook.
- In order to be aware of the different PACS product flavours available in the HSA, please read the PACS Products Explained. In here is a description of the layout and contents of a PACS observation, with a guide to those products that the archive user should focus on.
- If considering reducing or analysing your PACS data within the Herschel Data Processing Environment (HIPE), please read the PACS Photometry Launch Pad and PACS Spectroscopy Launch Pad. These are the first chapters of the Data Reduction Guides, and they give an overview of the state of the data obtained from the HSA, cases where re-reducing the data could be useful, which post-processing tasks are necessary or just recommended, and a summary of analysis tasks available in HIPE. If running the pipelines in HIPE, you need to consult the PACS Photometer Data Reduction Guide and the PACS Spectrometer Data Reduction Guide. Information about how the crucial pipeline tasks work is included in these documents, and this can be useful even for those who do not want to re-reduce their data: how the mappers work for photometry and how the cubes are created for spectroscopy.
In addition to these documents, the following web-sites contain useful information:
- Data Products Known Issues: for known issues concerning all Herschel products, for example artefacts that were not corrected or issues to be aware of or just to consider when working with PACS products.
- What’s New in HIPE 15: for a listing of what is new in the software, the data-reduction pipelines, the calibration, and the data products in the latest release of HIPE. Since the final version of HIPE and the data calibration have been release, there is no longer a real need to consult these pages unless you are comparing newly-obtained PACS products to those which were reduce many years ago.
- NHSC PACS pages: there is some information provided on these pages which is complimentary to that given here, for example the tutorials.
PACS Data Products
- Products generated by the standard instrument pipelines: these have been processed essentially by HIPE 14.2 (see also Data Processing section below), and cover observations performed during the Routine Science phase and during the early Performance Validation and Routine Calibration campaigns. Most of the data is public, the main exception being engineering observations and non-standard calibration observations. The range of products, and the levels they can be found in, produced by the pipeline for the different observing modes is summarised in the Quick-Start Guides (photometry or spectroscopy).
- User-Provided Data Products (UPDPs): these are products delivered by science consortia to the Herschel Science Centre. All UPDPs stored in the HSA can be checked and downloaded from https://www.cosmos.esa.int/web/herschel/user-provided-data-products.
Users should be warned that a fraction of the UPDPs were generated with earlier versions of the data processing pipelines and calibration files than the products gotten from the HSA. These older products may have caveats which have since been fixed. However, they may nontheless be useful products. A careful reading of the UPDP release note or the science paper, provided by the Key Project team, of the PACS part of the What’s New and the DP known issues pages for the relevant HIPE tracks should be undertaken before making a decision.
- Highly-Processed Data Products (HPDPs): these are products delivered by experts from the Instrument Control Centres and the (NASA) Herschel Science Centre. They offer essentially quality-enhanced (improved calibration) and/or complementary products to those processed with the standard pipelines.
All HPDPs stored in the HSA can be checked and downloaded from https://www.cosmos.esa.int/web/herschel/highly-processed-data-products#pacs_hpdp
- Ancillary Data Products: these are products not necessarily linked to any observations, sometimes of lesser relevance to the scientific end user. They do, however, provide important information concerning the instrument spatial responses, or more engineer-oriented data taken prior to launch and during the mission. All Ancillary Data Products stored in the HSA can be checked and downloaded from https://www.cosmos.esa.int/web/herschel/ancillary-data-products#pacs_adp
The majority of the PACS products can be considered science-ready in the sense that no further data processing or calibration is necessary. However — and particularly for spectroscopy — the science user of PACS data is strongly encouraged to read the relevant information in the Quick-Start Guides (photometry or spectroscopy) and the PACS Handbook before doing any science: the depth to which an observation can be analysed has a dependence on the observing mode and on the type of source observed, and various post-processing tasks can or should be applied to the different observing-mode/source-type combinations.
For photometry there are no particular issues to mention that are not covered in the Quick-Start Guide.
Several different versions of the same science-grade data products are provided in a spectroscopy observation, and this can be rather confusing for the first-time user. A set of spectroscopy product use-case flowcharts can be found on the HELL pages, and these should be consulted as a guide to the products to work with in an observation depending on the science case and the observing mode, and what pre-analysis needs to or can be done on the data.
Most spectroscopy observations are affected by the non-uniform illumination of the PACS integral field unit (due to the non-contiguous nature of the spaxels), which results in incomplete collection of the flux from a source (over every part of the FoV). Advice for dealing with this can be found in the Quick-Start Guide (spectroscopy). See the Data Processing section for more information.
All standard observations in the HSA have a Quality Control Report (QCR) entry in the HSA search results page. For PACS observations, the most useful part of this report is in the comment section, where relevant warnings about the quality or state of the data can be found. The few flags that can still be found in the QCR mostly refer to engineering or processing events that are either repeated in the comments or are not relevant for the archive user.
PACS Performance and Calibration
A set of prime stellar flux calibrators, and secondary stellar and solar system flux calibrators were used in the calibration of PACS, with a similar set of objects used by the two sub-instruments. This is explained in the calibration chapters of the PACS Handbook, together with the details of the spectral and spatial calibration. A summary of the calibration uncertainties can be found in the Quick-Start Guides (photometry and spectroscopy). The files of the calibration tree used for the SPG (and the interactive) pipeline processing were updated throughout the mission, and some also up until the end of the post-operations phase (SPG 14). For a listing of the history of of these updates, go to HELL Level 2, where a PDF of the original history updates web-page can be found.
Information about the flux calibration models used by all instruments can be found in the Herschel calibrator page:
The calibration parameters used in the processing pipelines are stored in the so-called calibration tree. The FITS files that make up the PACS calibration tree — ending with those included in the final PACS_CAL_78_0 — are present in any observation downloaded from the HSA, but can also be retrieved in the legacy repository area, where the release notes for each calibration tree update can also be found (both as xxxTImeDependecy...info files and as release-note.html files).
The output of some of the calibration characterisation effort has also been compiled as Ancillary Data Products, and can be found at the corresponding repository.
PHOTOMETER FLUX CALIBRATION
Five late-type giant stars were the prime calibrators for the PACS flux calibration. A number of planets and moons of the Solar system, asteroids and faint stellar standards were used as secondary calibrators. More information on the calibration of the PACS photometer, including the aperture and colour corrections, can be found in the PACS Handbook. The following publications are also prime sources of information:
- Balog et al., The Herschel-PACS photometer calibration, Experimental Astronomy, (2013)
- Müller et al., Herschel celestial calibration sources, Experimental Astronomy, (2013)
- Müller et al., Far-infrared photometric observations of the outer planets and satellites with Herschel-PACS, A&A 588, 109 (2016)
- Nielbock et al., The Herschel PACS photometer calibration, Experimental Astronomy, (2013)
Absolute flux error: 5% in all three filters, with a reproducibility for a non-variable point source of 2%
Additional sources of uncertainty:
- Often the dominant source of photometric error on PACS maps is the scatter in the background. It is recommended that this scatter is measured with apertures of the same size as used for the astronomical source (whether point or extended)
- Uncertainty of mappers: the various mappers used in the photometry pipelines (highpassfilter+photProject, Unimap, Jscanam) have different approaches to cleaning the data timelines and creating sky maps from the scanning data. They therefore have different types of uncertainty. This is explained briefly in the Quick-Start Guide (photometry) and in more detail in the PACS Handbook. In addition, the user can consult the reports on map-making benchmarking that were carried out in 2013/14 can be found in HELL (PACS Level 2)
- Extended emission photometry: three technical reports on a comparison of the extended emission measured from PACS data compared to that from Spitzer/MIPS can be found in HELL (PACS Level 3)
Note that PACS maps have no absolute background level, and all fluxes must be measured with respect to to the local background.
BEAMS AND APERTURE CORRECTIONS
The PSF/beam for any observation depends on the observing parameters (scan speed, filter, and prime vs. SPIRE—PACS parallel mode) and the mapper used. It is hence unique to each observation. Beam images at various various scan speeds—filter combinations have been created from observations of bright point sources. These are explained in the Quick-Start Guide (photometry) and in more detail in the PACS Handbook. The aperture and colour corrections are tabulated in the PACS Handbook. In addition, beam data can be found as Ancillary Data Products:
No 2D beams for the PACS photometer have been provided, for the reason given above. We refer the interested user to the work of G. Aniano (2011: see also http://www.astro.princeton.edu/~ganiano/Kernels.html) for kernels to match the spatial resolution between the PACS bands, PACS and SPIRE, and PACS and various other IR instruments.
Spectrometer flux calibration
For the spectrometer, in addition to the prime and secondary calibrators, the spectrum of the telescope mirror was also used as an intermediate step in the calibration scheme. This is discussed in the PACS Handbook.
Absolute flux error: ±12% in each band independently, with a systematic error of ±2%
Continuum flux reproducibility from observations on HD 161796 is estimated at ±15% (peak-to-peak)
Relative flux accuracies within a band are:
- chop nod observations : ±5% up to 150 μm, and ± 10% beyond
- unchopped observations : ±10% for all wavelengths
Additional sources of uncertainty:
- Continuum level uncertainty for unchopped observations can be as high as ±20 Jy
- Continuum level uncertainty for chop-nod observations: continuum levels < 5Jy should be considered with caution
- For the spectral range >190 μm, products are available only as HPDPs, which were processed with an RSRF that accounts for the order leak in this region. Continuum level is absolutely wrong. Emission line fluxes as measured from the HPDP require an extra multiplication by 1.3 and the uncertainty here is still quite high
- For the spectral range <55 μm only uncalibrated HPDPs are available, as this range cannot be flux calibrated (being formally outside the filter bandwidth)
All cubes contain an “error” or “stddev” extension. This is computed from the scatter in the repeated datapoints that were averaged to create each wavelength bin. An explanation of the data errors for any particular observation can be found in the Spectrometer Data Reduction Guide (sec. 7.6). How to combine errors for emission line measurements is explained in the Quick-Start Guide (spectroscopy).
Beams and point-source corrections
The 3D shape of the spectrometer beam (at the various key wavelengths accross the PACS bands) has a dependence on the observing mode: pointed observations very much undersample the PSF and the shape of the Integral Field Unit (IFU) can be seen in the beam images, while for mapping observations the PSF is better sampled. Beam images at the key wavelengths were created via a set of fine-step rasters on a bright point source. These “beam efficiencies” are maps of the response of each detector on the sky. They describe the (relative) coupling of a point source to each spaxel as a function of its (the source) position in the FOV. These can be downloaded from the HELL PACS Level 2 page: start with this summary document. (Further, but less user-friendly, documents can be found on Level 3.)
In addition, beam data can be found at as ancillary products as Ancillary Data Products:
The measured beams were used to calculate the point-source corrections that must be applied to all point sources to recover the true flux. This can either be done in HIPE or the appropriate product taken from the observation downloaded from the HSA. See Data Processing for more information.
PACS Data Processing
The final pipeline products served in the archive were processed with the pipeline version 14.2 using PACS Calibration Tree PACS_CAL_77_0. The final calibration tree is PACS_CAL_78_0 and contains some extra data on aperture corrections for photometry which is not relevant for the pipeline processing. The final version of the Herschel Interactive Processing Environment (HIPE) was version 15.0.1, but this version was optimised for interactive analysis, and no product regeneration was involved. Reprocessing products with HIPE 15.0.1 should, however, yield the same outcome as those stored in the HSA.
In case you wish to perform data reduction or analysis with HIPE, version 15.0.1 can be downloaded from the following page:
There are still a handful of known issues with HIPE, and those are compiled in the following page:
Interactive pipeline scripts and useful analysis scripts in HIPE
Those users considering fully or partially re-reducing their data in HIPE should first read the Quick-Start Guides (photometry or spectroscopy) for a guide to the need for this. The Launch Pads (photometry and spectroscopy) can also be consulted for a summary of the cases under which re-processing is useful or necessary. When running the pipelines in HIPE, the PACS Photometer Data Reduction Guide and the PACS Spectrometer Data Reduction Guide should be turned to (these can be best accessed within HIPE, so the links to other HIPE documents work).
There are various interactive pipeline scripts available in HIPE that are complementary to the pipeline scripts that were used to reduce the data stored at the HSA. These scripts are also included in the menu of interactive pipeline scripts.
A handful of cube, image, and spectral analysis scripts are available from a Scripts menu in HIPE, including:
- aperture photometry (including photometry errors and aperture corrections)
- making emission line maps from cubes
- creating cubes with different spatial grids (spaxel sizes, combining unrelated but spatially overlapping fields)
- spectral image convolution
- spectroscopy point-source corrections
These are also discussed in the Launch Pads.
Spectroscopy: Recommended post-processing tasks
The Quick-Start Guide (spectroscopy) and the Launch Pad (spectroscopy) contain a summary of the cases were reprocessing and post-processing is necessary or recommended. For a flow-chart guide to which products to use for what type of science, see the product decision trees which can be found on the HELL pages. Here we would like to highlight two cases were post-processing is strongly recommended.
- Point sources: for point sources in PACS spectral cubes, it is not enough to simply add up the spectra in an aperture around the source: aperture corrections are necessary. Corrected spectra for centred point sources from pointed observations are included in the spectral tables downloaded from the HSA. For other cases (faint, offset, or where broad-band features are important), read the product decision trees, and the Quick-Start Guide and Launch Pad.
- Forward modelling tool (mainly extended sources): The forward modelling tool (FMT) has been provided to allow users to estimate the correction necessary to their spectral cubes to account for the uneven illumination of the PACS footprint. This uneven illumination is due to the non-contiguous nature of the spaxels that make up the PACS IFU, and the subsequent loss of flux “between spaxels”. This affects: off-centred point sources, extended irregular sources (i.e. which are not very flat over the observed field), and crowded fields. The FMT documentation is provided on the HELL pages, and the product decision trees outline the cases where using the FMT is recommended; more advice is given in the Quick-Start Guide.