1 INTRODUCTION
Fast radio bursts (FRBs)—bright millisecond radio pulses—have generated considerable excitement within the astronomy community, with potential uses in cosmology, the intergalactic and interstellar media, coherent emission processes, compact objects, and more (Macquart et al. Reference Macquart2015). To date, the majority of FRB discoveries have been made at the Parkes radio telescope (Lorimer et al. Reference Lorimer, Bailes, McLaughlin, Narkevic and Crawford2007; Keane et al. Reference Keane, Stappers, Kramer and Lyne2012; Thornton et al. Reference Thornton2013; Burke-Spolaor & Bannister Reference Burke-Spolaor and Bannister2014; Petroff et al. Reference Petroff2015; Ravi, Shannon, & Jameson Reference Ravi, Shannon and Jameson2015; Champion et al. Reference Champion2016; Keane et al. Reference Keane2016; Spitler et al. Reference Spitler2016), with additional discoveries made at the Arecibo and Green Bank telescopes (GBTs) (Spitler et al. Reference Spitler2014; Masui et al. Reference Masui2015). With mounting evidence for their astrophysical origin ever more projects dedicated to FRB searches are arising at new and existing telescopes (Law et al. Reference Law2015; Karastergiou et al. Reference Karastergiou2015; Tingay et al. Reference Tingay2015; Chennamangalam et al. Reference Chennamangalam2015). At present, the rate of discovery is still rather slow, but new facilities are under construction which promise an explosion in FRB discovery rates (Macquart et al. Reference Macquart2015). It is therefore timely to catalogue what we know so far and to re-measure quantities for the published FRBs in a uniform and systematic fashion. Furthermore, it is important to identify commonly used derived parameters which are model-dependent and examine the various degrees to which these are uncertain.
Our catalogue is partially presented here for the currently published FRBs in tabular form but is also fully available in the form of an online catalogue and database at http://www.astronomy.swin.edu.au/pulsar/frbcat/ for community use. Where data were publicly available, we have performed a systematic re-analysis with software using the methods outlined in this work.
In the following sections, we describe the online interface and how to use the catalogue. In Section 2, we present the entries in the catalogue used to describe each burst. Section 3 provides information on how to use the catalogue; we conclude in Section 4.
2 PARAMETERS
A list is presented in the following subsections of all the quantities and parameters displayed in the catalogue. These have been divided into three categories: observation parameters, measured parameters, and derived parameters. Observation parameters are related to the telescope and instrument used for observation. Measured parameters relate to the detected burst and quantities obtained through direct processing of the data. Derived parameters related to cosmology and distance are based upon combining the measured parameters with model-dependent numbers for cosmological values (ΩM, ΩΛ, H 0) and the ionised component of the Galaxy (DMGalaxy).
The separation of these three categories is for several reasons. First, they have been separated to draw clear distinctions between quantities that are directly measured, speculative based on FRB position in the telescope beam, and speculative due to a combination of the positional uncertainty and the uncertainty of a model or a models. Errors are provided for measured quantities. Second, this catalogue is intended for use as the FRB population continues to grow. Each observation will also have its own set of measured and derived parameters which will be attached to it. Additional sets of derived parameters can be added when data are re-analysed in new ways.
Measured parameters are included with a reference to the measurements leading to the numbers presented. In the cases where the data are publicly available, two (or more) sets of measured parameters for the original detection observation are included: those published in the discovery paper and those obtained through our systematic re-analysis described here. A reference is included to make note of the measurement method.
The signal to noise ratio (S/N), width, and dispersion measure (DM) for each burst are re-derived using the destroy Footnote 1 single pulse search code, the psrchive Footnote 2 package and scripts made publicly available through the FRBCAT github pageFootnote 3 . These numbers have in turn been independently verified using a python-based fitting routine. For the python-based approach, a filterbank data block of a few seconds around the event was extracted. To remove the effects of narrowband interference, frequency channels with abnormally high or low variance over the data segment were masked. We used the outlier detection method known as Tukey’s rule for outliers (see e.g. Chandola, Banerjee, & Kumar Reference Chandola, Banerjee and Kumar2009). The resultant data block was then de-dispersed at different trial DM values using the smallest DM trial step allowed by the original time resolution of the raw data. In the case of Parkes search-mode data since 2008, ΔDM = 0.0488.
Every de-dispersed time series was then separately searched for pulses. A running median of length 0.5 s was subtracted to mitigate the effects of low-frequency noise, then the time series was normalised to zero mean and unit variance by carefully excluding outlier values (i.e. time samples containing the FRB signal). Finally, a standard pulse search algorithm was used, which consists in convolving the time series with a set of top-hat pulse templates (Lorimer & Kramer Reference Lorimer and Kramer2004), here with widths covering all values from 1 to 400 time samples, trying to maximise the response:
$$\begin{equation}
\text{S}/\text{N} = \frac{1}{\sqrt{W}} \sum _{\text{j}=i}^{i+W} t_{\text j},
\end{equation}$$
where W is the width in number of bins of the top-hat pulse template, i the trial starting sample index of the pulse, and t j the j-th bin of the time series. The maximum S/N value points to the optimal FRB parameters: DM, pulse width, and arrival time.
Both the published and re-derived parameters are included on the page for an individual burst. If available, the re-derived values are the ones presented on the catalogue homepage to maintain consistency across the sample where possible, as some search codes used for the initial discoveries have been found to under-report S/N (Keane & Petroff Reference Keane and Petroff2015).
Values for some of the following parameters for the published FRBs are presented in Table 1. Values for cosmological parameters, as derived by cosmocalc (Wright Reference Wright2006), are also included in the catalogue.
Table of select observation parameters (left) and measured parameters (right) from the catalogue. Discovery observations for each burst are listed first. Where a re-analysis of the data has been performed for this work, it is presented as a second entry for the burst. The references are [1] Burke-Spolaor & Bannister (Reference Burke-Spolaor and Bannister2014), [2] This work, [3] Keane et al. (Reference Keane, Kramer, Lyne, Stappers and McLaughlin2011), [4] Lorimer et al. (Reference Lorimer, Bailes, McLaughlin, Narkevic and Crawford2007), [5] Champion et al. (Reference Champion2016), [6] Thornton et al. (Reference Thornton2013), [7] (GBT Burst), [8] Spitler et al. (Reference Spitler2014), [9] Ravi et al. (Reference Ravi, Shannon and Jameson2015), [10] Petroff et al. (Reference Petroff2015), [11] Keane et al. (Reference Keane2016). The FWHM for the telescopes include in this table are 15′ (Parkes), 7′ (Arecibo), and 16′ (GBT). Question marks denote values that were not specified in the original publication or were not available publicly.
a Errors in gl and gb refer to the pointing accuracy of the given telescope.
b Originally incorrectly published as FRB 011025.
c Originally incorrectly published as FRB 110627.
2.1. Observation parameters
Telescope —Telescope used to take the observation.
Receiver —Receiver system on the telescope used to take observations. At Parkes, the primary instrument is the 13-beam multibeam receiver (MB20; Staveley-Smith et al. Reference Staveley-Smith1996), at Arecibo the 7-beam ALFA receiver (Spitler et al. Reference Spitler2014), and at the GBT the 800 MHz receiver (Masui et al. Reference Masui2015).
Backend —The data recording system used for the observation. At Parkes, two primary data recording systems have been used: the analogue filterbank (AFB) and Berkeley Parkes Swinburne Recorder (BPSR; Keith et al. Reference Keith2010). The Mock spectrometers and the GUPPI backends have been used at Arecibo and the GBT, respectively.
Beam —The telescope beam number in which the FRB was detected. For FRBs detected with single-pixel feeds, this value is set to 1. For detections with an interferometer, this may be the number designation of the tied-array beam or fan beam.
Sampling time —The length of a time sample for the observation, in milliseconds. Sampling times between 1 ms and 64 μs are common for observations in which FRBs were detected.
Bandwidth —The observing bandwidth in MHz. In the case of observations using BPSR, the system bandwidth is 400 MHz but the usable bandwidth (after excision of channels rendered unusable by interference) is 338.281 MHz; the latter is quoted in this catalogue as it is the relevant number for calculations of S/N and flux density.
Centre frequency —The centre frequency of the observation in MHz.
Number of polarisations —The number of polarisations used to record the total signal.
Channel bandwidth —The bandwidth of the individual frequency channels in MHz.
Bits per sample —The number of bits recorded for an individual time sample in the final data product.
Gain —The telescope gain in units of K Jy−1.
System temperature —The receiver system temperature in Kelvin. Throughout this analysis, the Parkes MB20 system temperature is taken to be 28 K, as given in the Parkes Telescope Users GuideFootnote 4 . We note that this is up to 5 K higher than the system temperature used in many publications and that true system temperature also depends on observing elevation.
Reference —The journal reference for the burst discovery paper where the event was first reported.
2.2. Measured parameters
RAJ—The right ascension in J2000 coordinates of the best estimate of the FRB position. This may correspond to the pointing centre of the detection beam or the location determined from an interferometric detection. Error on pointing accuracy of the telescope is also listed.
DECJ—The declination in J2000 coordinates of the best estimate of the FRB position. As with RAJ, this is the pointing centre of the detection beam or location determined from an interferometric detection. Error on pointing accuracy of the telescope is also listed.
gl —The Galactic longitude, in degrees, of the best estimate of the FRB position.
gb —The Galactic latitude, in degrees, of the best estimate of the FRB position.
Positional uncertainty —This parameter serves as the best estimate of the positional uncertainty of the burst as a detection is most likely to lie within this area.
DM—The DM of the FRB in units of cm−3 pc. The integrated electron column density along the line-of-sight to the burst obtained either with a pulse fitting algorithm or by the search code. For bursts with a published DM produced with a detailed fitting code such as the one described in Thornton et al. (Reference Thornton2013), this DM is used throughout the re-analysis.
S/N—The S/N of the burst.
W OBS—The observed width of the FRB in ms obtained either with a pulse fitting algorithm or by the search code. The width reported here is not the intrinsic width.
S PEAK,OBS—The observed peak flux density of the burst in Jy calculated using quantities above via the single pulse radiometre equation (Cordes & McLaughlin Reference Cordes and McLaughlin2003). Note that this flux density is derived from observed values and is not necessarily the true peak flux density that would be measured if the burst occurred at beam centre or was perfectly localised; this value should be taken as a lower limit on the true flux density.
F OBS—The observed fluence of the FRB in units of Jy ms calculated as F obs = S peak, obs × W obs. Again, the observed fluence should be taken as a lower limit on the true fluence due to the likely off-axis detection or imperfect localisation of the burst.
DM index —The DM index of the burst α such that DM ∝ ν−α obtained with a pulse fitting algorithm. The DM index for the propagation of waves through a cold plasma is α = 2 (Lorimer & Kramer Reference Lorimer and Kramer2004).
Scattering index —The evolution of pulse width as a function of frequency due to scattering such that W ∝ ν−β obtained with a pulse fitting algorithm. The index for the propagation of radio waves through an inhomogeneous turbulent medium is β = 4 (Lorimer & Kramer Reference Lorimer and Kramer2004).
Scattering time
—A measure of the fluctuations in electron density along the line-of-sight contributing to scattering of the pulse obtained with a pulse fitting algorithm. The number presented in the catalogue is the scattering time for a radio pulse at 1 GHz in ms. Scattering time scales with frequency as
$\tau (\nu ) = \tau _{\text s} (\nu /\nu _0)^{-\beta }$
, where τs and ν0 are at the reference frequency and β is the scattering index.
Linear polarisation fraction
—If polarised data were recoreded for the FRB, the fractional linear polarisation is reported with errors. The total linear polarisation is the quadrature sum of Stokes Q and U such that
$\sqrt{Q^2+U^2}/I$
.
Circular polarisation fraction —If polarised data were recorded for the FRB, the fractional circular polarisation is reported with errors. The total absolute value of circular polarisation is given by |V|/I.
z phot—Photometric redshift of host galaxy, if available.
z spec—Spectroscopic redshift of host galaxy, if available.
2.3. Derived parameters
DM Galaxy —The modelled contribution to the FRB DM by the electrons in the Galaxy. The Galactic DM contribution is derived using the NE2001 Galactic electron density model (Cordes & Lazio Reference Cordes and Lazio2002) and should be taken as an estimate as the free electron content of the Galactic halo is not well constrained (Dolag et al. Reference Dolag, Gaensler, Beck and Beck2015).
DM EXCESS—The DM excess of the FRB over the estimated Galactic DM. This is calculated as DMEXCESS = DMFRB - DMGalaxy.
z —The estimated redshift of the FRB based on DMEXCESS. The redshift is calculated as z = DMEXCESS/1200 pc cm−3 from estimates of the intergalactic medium (IGM) electron density from Ioka (Reference Ioka2003). This relation approximates the full expression to better than 2% for z < 2. This uncertainty is much less than the line-of-sight variation expected (McQuinn Reference McQuinn2014). This value for redshift assumes that any host galaxy or surrounding material contributes nothing to the DM and should be taken as an approximate upper limit on the true redshift.
D COMOV—Comoving distance in units of Gpc derived using the cosmology calculator CosmoCalc (Wright Reference Wright2006). One should note that this parameter is highly uncertain as, unless an independent redshift measurement is made for an FRB, the comoving distance depends on models of electron density in the Galaxy and IGM as well as the chosen cosmological parameters.
D LUMINOSITY—Luminosity distance in units of Gpc calculated as D L = D COMOV × (1 + z). Again, this value is an upper limit based on the upper limit on redshift and comoving distance.
Energy —The estimated FRB energy in units of 1032 joules calculated as
$$\begin{equation}
E_\mathrm{FRB} = F_\mathrm{obs} \; \mathrm{BW} \; D_{\text L}^2 \times 10^{-29} \; (1+z) \; \mathrm{J},
\end{equation}$$
where fluence is in units of Jy ms, bandwidth is in units of Hz, D L is in units of metres, and 10−29 is a conversion factor between Jy ms and joules. The equivalent conversion factor from Jy ms to ergs is 10−22.
3 HOW TO USE THE CATALOGUE
The catalogue can be viewed on the web or downloaded as a plain text file for use offline. The homepage of the catalogueFootnote 5 presents all the FRBs with a few key parameters: telescope name, gl, gb, FWHM, DM, S/N, W obs, S peak, obs, F obs, and reference. For each burst, there is a detailed page containing all the parameters described in Section 2. Individual pages for the bursts contain relevant images and figures such as the dispersion sweep for the burst, polarisation (if available). Images are also available for download from the catalogue page and links to the publicly available data are given on individual burst pages.
Cosmological parameters such as D COMOV, D luminosity, and Energy are derived on-the-fly using the Cosmology Calculator module from Wright (Reference Wright2006) with the input parameters displayed on the individual burst sub-pages. Inputs for ΩM, ΩΛ, and H 0 can be modified on the catalogue webpage; default values are ΩM = 0.286, Ωvac = 0.714, and H 0 = 69.6, Figure 1. Updating these numbers will automatically update the calculated values of comoving distance, luminosity distance, and energy.
Example of the beginning of an FRB entry on the catalogue webpage. Telescope-specific observation parameters have been separated from the measured parameters measured from the available data. Where the data have been re-analysed for the catalogue multiple measurement methods are available.
Alternatively, the catalogue can be downloaded either in CSV, tabular plain-text, or VOTable format from the catalogue homepage. The file generated will have all information contained in the most recent version of the online catalogue. If derived and measured parameters for a burst have been derived using multiple methods (i.e. values from publication and re-analysis), all methods will be included in the downloaded table with the appropriate reference. The value used in any studies of the bursts is the choice of the user.
3.1. Future plans
Currently the total number of FRBs is small, however, a large population boom is expected as FRB-finding initiatives such as the ALERT project at the Westerbork Synthesis Radio TelscopeFootnote 6 , UTMOSTFootnote 7 , and CHIMEFootnote 8 , as well as many others, come on-line. The catalogue is well-equipped to deal with this influx and can accommodate a large number of bursts. At present, the catalogue is maintained and updated by the authors. In the event of a new burst, there is currently no method for external users to upload new events. In the near future, the catalogue will also have the functionality to automatically ingest new bursts in the form of VOEvents from trusted users and be fully VO compatible, i.e. capable of accepting and issuing VOEvents for new FRBs. The catalogue will continue to be curated and managed by the authors for the lifetime of the catalogue, or as long as it continues to be of use to the community.
4 CLOSING REMARKS
In this paper, we present the FRBCAT, an online catalogue of FRBs. This catalogue includes all bursts currently available in the literature and will be updated as new bursts are published in the future. The catalogue presents an overview of all bursts on the main page but also includes a page for each individual burst with a number of parameters that describe the observational setup, the observed burst properties, and model-dependent cosmological parameters. Additionally, all bursts for which the data are public have been re-analysed using a standardised method in an effort to make detections more consistent and directly comparable. The tools for our re-analysis have been made available through github and the data can be processed with freely available software packages and processing tools. It is our hope that data for all future bursts will be made public upon publication and this catalogue will encourage communication and collaboration.
ACKNOWLEDGEMENTS
The Parkes radio telescope and the Australia Telescope Compact Array are part of the Australia Telescope National Facility which is funded by the Commonwealth of Australia for operation as a National Facility managed by CSIRO. Parts of this research were conducted by the Australian Research Council Centre of Excellence for All-sky Astrophysics (CAASTRO), through project number CE110001020. This work was performed on the gSTAR national facility at Swinburne University of Technology. gSTAR is funded by Swinburne and the Australian Government’s Education Investment Fund. The research leading to these results has received funding from the European Research Council under the European Union’s Seventh Framework Programme (FP/2007–2013)/ERC Grant Agreement n. 617199.
The authors would like to thank the SUPERBFootnote 9 collaboration for their help in beta testing the website. The authors would also like to thank the members of the community who have used and tested the catalogue since it has been publicly available and helped us to improve usability particularly Sarah Burke-Spolaor, Fernando Camilo, Jamie Farnes, Jason Hessels, Shri Kulkarni, Maxim Lyutikov, and Vikram Ravi. We thank the anonymous referee for their useful suggestions for improvement of this manuscript and the catalogue.
