A&A 622, A1 (2019) 
https://doi.org/10.1051/0004-6361/201833559 Astronomy
&
 cESO 2019 
Astrophysics <BR>

LOFAR Surveys: a new window on the Universe Special issue 
The LOFAR Two-metre Sky Survey 
II. First data release?,?? <BR>
T.W. Shimwell1,2,C.Tasse3,4,M.J. Hardcastle5,A.P. Mechev2,W.L.Williams5,P.N. Best6,H.J.A. Rtgering2, 
J.R. Callingham1,T.J. Dijkema1,F.de Gasperin2,7,D.N. Hoang2,B. Hugo8,4,M. Mirmont9,J.B.R. Oonk1,2, 
I. Prandoni10, D. Rafferty7,J. Sabater6,O. Smirnov4,8,R.J.vanWeeren2,G.J. White11,12,M. Atemkeng4, 
L. Bester8,4,E. Bonnassieux8,13,M. Brgen7,G. Brunetti10,K.T.Chy˙zy14,R. Cochrane6,J.E. Conway15, 
J.H. 
Croston11,A. Danezi16,K. Duncan2,M.Haverkorn17,G.H. Heald18,M. Iacobelli1,H.T. Intema2,N. Jackson19, 

M. Jamrozy14, M. J. Jarvis20,21, R. Lakhoo22,23, M. Mevius1, G. K. Miley2, L. Morabito20, R. Morganti1,24, 

D. 
Nisbet6,E. Orr1,S. Perkins8,R.F. Pizzo1,C. Schrijvers16,D.J.B. Smith5,R.Vermeulen1,M.W.Wise1,25, 

L. 
Alegre6, D. J. Bacon26, I. M. van Bemmel27, R. J. Beswick19, A. Bonafede7,10, A. Botteon10,28, S. Bourke15, 

M. 
Brienza1,24, G. Calistro Rivera2, R. Cassano10, A. O. Clarke19, C. J. Conselice29, R. J. Dettmar30, A. Drabent31, 

C. 
Dumba31,32,K.L. Emig2,T.A. Enßlin33,C. Ferrari34,M.A. Garrett19,2,R.T. Génova-Santos35,36,A.Goyal14, 

G. 
Gkan18,C. Hale20,J.J. Harwood5,V. Heesen7,M. Hoeft31,C. Horellou15,C. Jackson1,G.Kokotanekov25, 

R.Kondapally6,M.Kunert-Bajraszewska37,V. Mahatma5,E.K. Mahony38,S. Mandal2,J.P. McKean1,24, 

A. 
Merloni39,40,B. Mingo13,A. Miskolczi30,S. Mooney41,B. Nikiel-Wroczy´, 


nski14,S.P. O’Sullivan7,J. Quinn41 

W. Reich42,C. Roskowi´, 
nski37,A.Rowlinson1,25,F.Savini7,A. Saxena2,D.J. Schwarz43,A. Shulevski1,25
S.S. Sridhar1,H.R. Stacey1,24,S. Urquhart11,M.H.D.van derWiel1,E.Varenius15,19,B.Webster11, andA.Wilber7 
(Affiliations can be found after the references) 
Received4June 2018 / Accepted 12 September 2018 
ABSTRACT 
TheLOFARTwo-metreSkySurvey(LoTSS)isan ongoing sensitive, high-resolution 120–168MHz surveyofthe entire northernskyfor which observations are now 20% complete. We present our frst full-quality public data release. For this data release 424 square degrees, or 2% of the eventual coverage, in the region of the HETDEX Spring Field (right ascension 10h45m00s to 15h30m00s and declination 45◦0000000 to 57◦0000000)were mapped usinga fully automated direction-dependent calibration and imaging pipeline that wedeveloped.A totalof 325694 sourcesare detectedwithasignalofatleastfvetimesthenoise,andthe sourcedensityisafactorof ∼10 higher than the most sensitive existing very wide-area radio-continuum surveys. The median sensitivity is S144 MHz = 71µJybeam−1 and the point-source completeness is 90% at an integrated fux density of 0.45mJy. The resolution of the images is600 and the positional accuracyis within 0.200. This data release consists of a catalogue containing location, fux, and shape estimates together with 58 mosaic images that cover the catalogued area. In this paper we provide an overviewof the data release withafocus on the processing of the LOFAR data and the characteristics of the resulting images. In twoaccompanying papers we provide the radio source associations and deblending and, where possible, the optical identifcations of the radio sources together with the photometric redshifts and properties of the hostgalaxies. These data release papers are published together with a further ∼20 articles that highlight the scientifc potential of LoTSS. 
Key words. surveys – catalogs – radio continuum: general – techniques: image processing 
1. Introduction low-frequency observations to facilitate breakthroughs in research areas such as the formation and evolution of mas-
Surveys that probe deeply into new parameter space have enorsive black holes (e.g. Wilman et al. 2008;Best et al. 2014)and mous discovery potential. The LOFAR Two-metre Sky Sur-clusters ofgalaxies (e.g. Cassano et al. 2010; Brunetti&Jones vey (LoTSS; Shimwell et al. 2017) is one example: it is an 2014). However, there are many other important scientifc ongoing survey that is exploiting the unique capabilities of drivers of the survey, and there is active research in areas the LOw Frequency ARray (LOFAR; van Haarlem et al. 2013) such as high redshift radio sources (e.g. Saxena et al. 2017), to produce a sensitive, high-resolution radio survey of the galaxy clusters (e.g. Botteon et al. 2018; Hoang et al. 2017; northern sky with a frequency coverage of 120–168 MHz (see de Gasperin et al. 2017; Savini et al. 2018; Wilber et al. 2018), Fig.1).The survey was primarilymotivatedbythe potentialof 
active galactic nuclei (e.g. Brienza et al. 2017; Morabito et al. 2017; Williams et al. 2018), star forming galaxies (e.g. 
? 

LoTSS. CalistroRiveraetal. 2017),gravitational lensing,galactic radio ?? The catalogue is available at the CDS via anonymous ftp to emission, cosmological studies(Raccanelli et al. 2012), mag-cdsarc.u-strasbg.fr <BR>(130.79.128.5) or via http://cdsarc. <BR>netic felds (e.g. Van Eck et al. 2018), transients and recombi<BR>
u-strasbg.fr/viz-bin/qcat?J/A+A/622/A1 <BR>nation lines (e.g. Oonk et al. 2017). 
Article publishedby EDP Sciences A1, page1of 21 
The LoTSS survey is one of several ongoing or recently completed very wide-area low-frequency radio surveys that are providing important scientifc and technical insights. Other such surveys include the Multifrequency Snapshot Sky Survey (MSSS; Heald et al. 2015), TIFR GMRT Sky Sur-vey alternative data release (TGSS-ADR1; Intema et al. 2017), GaLactic and Extragalactic All-sky MWA (GLEAM; Wayth et al. 2015; Hurley-Walker et al. 2017), LOFAR Low-band Sky Survey (LoLSS; de Gasperin et al. 2019), and the Very Large Array Low-frequency Sky Survey Redux (VLSSr; Lane et al. 2014). However, LoTSS is designed to push further into new territory. This surveyaims to provide a low-frequency survey that will remain competitive even once the Square KilometreArray(Dewdneyetal.2009)isfully operational,and will not be surpassed as a low-frequencywide-area northern sky survey for the foreseeable future. The LoTSS can provide the astrometric precision that is required for robust identifcation of optical counterparts (see e.g. McAlpine et al. 2012) and a sensitivity that, for typicalradio sources, exceeds that achieved in existing very wide area higher frequency surveys such as the NRAO VLA Sky Survey (NVSS; Condon et al. 1998), Faint Images of the Radio SkyatTwenty-Centimeters (FIRST; Becker et al. 1995), Sydney University Molonglo Sky Survey (SUMSS; Bock et al. 1999; Mauch et al. 2003), and WEster<BR>bork Northern SkySurvey(WENSS; Rengelink et al. 1997)and rivals forthcoming higher frequency surveys such as the Evo-lutionary Map of the Universe (EMU; Norris et al. 2011), the APERtureTileInFocus survey(e.g. Rtgeringetal.2011)and the VLA SkySurvey(VLASS1). More specifcally the primary observational objectives of LoTSS are to reach a sensitivity of less than 100 µJybeam−1 at an angular resolution, defned as the full width half maximum (FWHM) of the synthesised beam, of ∼600 across the whole northern hemisphere, using the High Band Antenna(HBA) systemofLOFAR(seeFig. 1). 
In the frst paper of this series (Paper I: Shimwell et al. 2017)we described LoTSS and presented a preliminary data release. In that release the desired imaging specifcations were not reached, as no attempt was made to correct either for errors in the beam models or for direction-dependent ionospheric dis-tortions, which are severe in these low-frequency data sets. However, there has since been substantial improvements in the quality, speed, and robustness of the calibration of directiondependenteffects (DDEs) and imaging with the derived solutions (see e.g. Tasse 2014;Yatawatta 2015;vanWeerenetal. 2016; Tasse et al. 2018).Furthermore, LOFAR surveys of smaller areas of skyhave demonstrated that the desired imaging specifcations of LoTSS are feasiblebymaking useof direction-dependent cal-ibration(e.g. Williamsetal.2016;Hardcastleetal.2016).These newinsightshavefacilitatedthefrstfullqualitypublicdata release (LoTSS-DR1),whichwe presenthereinPaperIIofthis series. 
As part of this series we also attempt to enrich our radio catalogues by locating optical counterparts using a combination of likelihood ratio cross matching and visual inspection (discussed in Paper III of this series: Williams et al. 2019). In addition, where counterparts are successfully located, we pro-vide photometric redshift estimates and hostgalaxy properties (Paper IV: Duncan et al. 2019). In the near future, to improve on the redshifts for manysources, the William HerschelTelescope Enhanced AreaVelocity Explorer (WEAVE; Dalton et al. 2012, 2014)multi-object and integral feld spectrograph will measure redshiftsofoveramillion LoTSS sourcesaspartoftheWEAVE-LOFAR survey(Smith et al. 2016). 
https://science.nrao.edu/science/surveys/vlass <BR>

Fig. <BR>1. <BR>Image rms, frequency, and angular resolution (linearly propor<BR>tional to the radius of the markers) of LoTSS-DR1 in comparison to a selection of existing wide-area completed (grey) and upcoming (blue) radio surveys. The horizontal lines show the frequency coverage for surveys with large fractional bandwidths. The green, blue, and red lines show an equivalent sensitivity to LoTSS for compact radio sources with spectral indices of −0.7, −1.0, and −1.5, respectively. 
In Sects. 2and3we describe the observations, the data pro-cessing procedure for the present data release, and the quality of the resulting images.In Sect. 4 wegivea briefoverviewof the optical cross matching and the photometric redshift estimation. Finally,we outline some upcomingdevelopmentsinSect. 5 before concludingin Sect. 6. 

2. Observations and data reduction 
We describe the status of LoTSS observations in the frst subsection. The second subsection outlines the direction-independent calibration of the data; at present, the main challenge is retrieving and processing the large volume of archived data. The third subsection describes the direction-dependent calibration and imaging, where the focus is on the development and exe-cution of a robust and automated pipeline. The fnal subsection summarises the mosaicing and cataloguing of the DR1 images. 

2.1. Observation status 
The ambitious observational objectives for LoTSS are outlined in Fig. 1. To achieve these objectives at optimal declinations, LoTSS observations are conducted in the hba_dual_inner confgurationwith8hdwelltimesandafrequencycoverageof120– 168 MHz. The entire northern sky is covered with 3168 point-ings. By exploiting the multi-beam capability of LOFAR and observing in 8-bit mode two such pointings are observed simultaneously. As of May 2018, approximately 20% of the data have now been gathered and a further 30% are scheduled over the next two years (see Fig. 2); a total of approximately 13 000h observing time are required to complete the entire survey with the present capabilities of LOFAR. 
As in Shimwell et al. (2017), in this paper we focus on 63 LoTSS data sets (2% of the total survey) in the region of the HETDEX Spring Field that were observed between 2014 May 23 and 2015 October 15. Each8h observationwas bookended by10 min calibrator observations (primarily 3C 196 and 3C 295) and the data are archived witha time resolutionof1s anda frequencyresolution of 16 channels per 195.3 kHz sub-band (SB) 
A1, page2of 21 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 

Fig. <BR>2. <BR>Status of the LoTSS observations as of May 2018. The green dots show the images that are presented in this paper. The red, yellow, and black dots show the observed pointings (but yet unpublished), point-ings presently scheduled for observation between May 2018 and May 2020, and unobserved pointings, respectively. The HETDEX Spring Fieldregionis outlinedin blue.Thevast majorityofthe completedcoverage (20% of the northern sky) and upcoming observations (an additional30%ofthe northernsky)areregionswithlowGalacticextinction. 
by the observatory2. This high time-and frequency-resolution dataiskeptto reducetimeand bandwidth smearingtoalevelthat is tolerable for future studies that will exploit the international baselines of LOFAR (only antennas within the Netherlands are used for the primary objectives of LoTSS). The high spectral res-olution(R∼ 5000−7000 or 22–31kms−1 velocity resolution) of the data is alsofacilitating spectral line(Emig et al. 2019)and spectro-polarimetric studies. 


2.2. Direction-independent calibration 
The publicly available LOFAR direction-independent calibra-tion procedure was described in detail by vanWeeren et al. (2016)andWilliams et al.(2016)and makes use of the LOFAR Default Preprocessing Pipeline (DPPP; van Diepen&Dijkema 2018) for averaging and calibration and BlackBoard Selfcal (BBS; Pandeyet al. 2009)for calibration. In Paper I we used a pipeline implementation3 of this procedure to process the 63 LoTSS data sets that are described in this publication and we discussed the quality of the images that were produced. This calibration method is not described again in detail in this work, but we developed new tools to maintain a high volume fow of data through this pipeline and we briefy describe these below. 
The LoTSS data are stored in the LOFAR Long Term Archive (LTA4), which is distributed over three sites– 
7 
SURFsara5, Forschungszentrum Jich6, and Poznan´. The archived datavolumeper8hpointingis ∼16 TB, together with ∼350GBfor each10min calibrator observation, which implies an eventual data volume of ∼50PB for the entire 3168 point-ings of the survey (although this will be reduced by implementationof theDYSCO compression algorithm; Offringa et al. 2012). Downloading these large data sets from theLTAsites to localfacilitiesis either prohibitively time consuming orexpensive.Tomitigate this we migrated our direction-independent cal-ibration processing to the SURFsara Gridfacilities. At the time of writing this consists of several hundred nodes of various sizes with a total of ∼7500 compute cores that are linked with a high
2 ∼100of the early LoTSS observations wereaveragedto2s and 
24.4 kHz. 3 https://github.com/lofar-astron/prefactor <BR>using commit dd68c57. 
4 https://lta.lofar.eu/ <BR>5 https://www.surfsara.nl <BR>6 http://www.fz-juelich.de <BR>7 http://www.man.poznan.pl/online/pl/ <BR>
speed connectionof200Gbits−1 peak network traffic to the Grid storage, where the SURFsaraLTAdata are housed. The implementation of the direction-independent calibration pipeline, and other LOFAR pipelines, on the SURFsara Grid is described in detail by Mechev et al. (2017)and Oonk et al. (in prep.) and summarised briefy below. 
The LoTSS data are archived as 244 single SB fles and in our SURFsara implementation of the direction-independent cal-ibration pipeline each SB of the calibrator is sent to an available compute node where it is fagged for interference with AOFLAGGER(Offringa et al. 2012), averaged to two channels per 195 kHz SB and 8s, and calibrated using a model of the appropriate calibrator source, which has a fux density scale consistent with that described in Scaife&Heald (2012).We note that the Scaife&Heald (2012)fux density scale is con-sistent with the Perley&Butler (2017)scale to within∼5%but that there are larger discrepancies(∼5–10%) when comparing with the Baars et al. (1977)scale (seeScaife&Heald 2012 and Perley&Butler 2017 for details). Usinga single compute node the resulting 244 calibration tables are combined and used to derive time-independent amplitude solutions, XX and YY phase offsets, and clock offsets for each station. Similarly, on separate compute nodes, the 244 single SB target fles are each fagged, corrected for ionosphericFaraday rotation8, calibrated using the calibrator solutions, and averaged to a resolution of two chan-nels per 195kHzSB and8s.In the fnal stepof the direction-independent calibration pipeline, the data for each contiguous 10-SB block are sent to different compute nodes where they are each combined to a single fle that is phase calibrated against a sky model for the target feld, which is generated from the TGSS-ADR1 catalogue(Intema et al. 2017). This produces 25 10-SB measurement sets for the target feld,but the six highest frequencySBs are empty because there are only 244 SBs in the highest frequencymeasurement set. 
For the bulk processing of data on the SURFsara facilities we made use of PiCaS9, a CouchDB based token pool server for heterogeneous compute environments. The PiCaS server allows millions of tasks to be scheduled on heterogeneous resources to monitor these tasks viaa web interface and to provide easy access to logs and diagnostic plots, which helps ensure that our data quality is high. Examples of these diagnostic plots for the HETDEX Spring Field data are shown by Shimwelletal. (2017).Wealsomakeuseofarchivingand distri<BR>butionfacilities at SURFsara, allowing us to store the direction-independent calibrated data products (which are reduced from 16TB to ∼500GB per pointing) and freely distribute these amongst LoTSS team members for analysis and further processing. 
The SURFsara Grid processing facilities enable high-throughput processingof large data sets stored on the localLTA site, however the LoTSS data sets are disseminated to all three LTA sites. Since the LTA sites are not linked to each other with a high bandwidth connection, the transfer speed to download data from theForschungszentrum Jich and Pozna ´
nLTA sites to SURFSara(∼200MBs−1)is currently a bottleneck in our processing.We are thereforeworking on implementing the direction-independent calibration pipeline on computefacilities localtoeachoftheLTAsites. 
8 https://github.com/lofar-astron/RMextract <BR>9 http://doc.grid.surfsara.nl/en/latest/Pages/ <BR>Practices/picas/picas_overview.html <BR>
A1, page3of 21 


2.3. Direction-dependent calibration and imaging 
Arobust,fast, and accurate calibration and imaging pipeline is essential to routinely create high-fdelity LoTSS images with a resolution of 600 and a sensitivity of 100 µJybeam−1. However the necessity to correct DDEs, which are primarily ionospheric distortions and errors in the station beam model of the HBA phased array stations, adds signifcant complications to this procedure. These DDEs can be understood in terms of Jones matrices(Hamaker et al. 1996)and to correct for these matrices, which not only depend on directionbut also on time, frequency, and antenna, they must be derived from the visibilities and applied during imaging. Various approaches have been developed to estimate the DDE (e.g. Cotton et al. 2004; Intema et al. 2009; Kazemi et al. 2011; Noordam&Smirnov 2010;vanWeerenetal. 2016;Yatawatta 2015)butfor thiswork wedeveloped KillMS(kms;Tasse 2014;Smirnov&Tasse 2015) to calculate the Jones matrices and ddfacet(Tasse et al. 2018) to apply these during the imaging. Our software packages and the pipeline are publicly available and documented10. Below we briefy outline the calibration and deconvolution procedures before describing the pipeline in more detail. 
2.3.1. Calibration of direction-dependent effects 
One of the main difficulties in the calibration of DDE is the large number of free parameters that must be optimised for when solving for the complex-valued Jones matrices. The consequences of this are that fnding the solutions can become prohibitively com-putationally expensive and that ill-conditioning can introduce systematics in the estimated quantities, which have a negative impact on the image fdelity. 
To tackle the computational expense, Salvini&Wijnholds (2014),Tasse(2014), and Smirnov&Tasse (2015)have shown that when inverting the Radio Interferometeric Measurement Equation (RIME; see e.g. Hamaker et al. 1996;Smirnov 2011) the Jacobian can be written using Wirtinger derivatives. The resulting Jacobian is remarkably sparse, which allows for short-cuts to be used when implementing optimisation algorithms such as Levenberg–Marquardt (see for example Smirnov&Tasse 2015). In particular, the problem can become antenna separa-ble, and to solve for the Jones matrices associated with a given antenna in kms, only the visibilities involving that antenna are requiredateach iterativestep.The computationalgaincanbeas high as n2 (where na is the number of elementary antennas). 
a 
To reduce ill conditioning,kms uses theWirtinger Jacobian together with an Extended Kalman Filter (EKF) to solve for the Jones matrices (Tasse, in prep.). Instead of optimising the least-squares residuals as a Levenberg–Marquardt (LM) proce-dure would, the EKF is a minimum mean-square error estima-tor and is recursive (as opposed to being iterative). In practice, the prior knowledge is used to constrain the expected solution at a given time. While an LM would produce independent “noisier” estimates, the EKF produces smooth solutions that are more physical and robust to ill-conditioning. 
To further improve the calibration, kMS produces a set of weights according to a “luckyimaging” technique in which the weights of visibilities are based on the quality of their calibration solutions(Bonnassieuxetal.2018),so visibilitieswiththeworst ionospheric conditions are weighted down in the fnal imaging. 
10 See https://github.com/saopicc <BR>for kms and ddfacet, and https://github.com/mhardcastle/ddf-pipeline <BR>for the associated LoTSS-DR1 pipeline. 

2.3.2. Wide feld spectral deconvolution 
The ddfacet imager(Tasseetal. 2018)uses thekms-estimated direction-dependent Jones matrices and internallyworks on each of the directions for which there are solutions to synthesise a single image.To do this, several technical challenges had to be overcome. For example, the dependence of the Jones matrices on time, frequency, baseline, and direction, together with time-and frequency-dependent smearing, lead toa position dependent point spread function(psf). Therefore, although ddfacet synthesisesa single image, eachfacet has itsown psf that takes into account the DDE and time and bandwidth smearing whilst ensuring that the correct deconvolution problem is inverted in minor cycles. 
Furthermore, to accurately deconvolve the LoTSS images, which have a large fractional bandwidth and a wide feld of view, spectral deconvolution algorithms must be used to esti-mate the fux density and spectra of modelled sources whilst taking into account the variation of the LOFAR beam through-out the bandwidth of the data. The computational cost of this deconvolution can be high and therefore throughout our processing we makeexclusive useof the subspace deconvolution(ssd) algorithm, an innovative feature of ddfacet (see Tasse et al. 2018 for a description). As opposed to clean and related algorithms, wherea fractionofthefux densityis iteratively removedateach major iteration, ssd aims at removing all the fux density at each major cycle. This is done in the abstracted notion of subspaces– in practice islands–each representing an independent deconvolution problem. Each one of these individual subspaces is jointly deconvolved(allpixelsare simultaneously estimated)byusinga genetic algorithm (the ssd-ga favour of ssd), and parallelisation is doneover hundredsto thousandsof islands.Astrengthof ssd is that we can minimise the number of major cycles, by always recycling the skymodel from the previous step. In practice the skymodel generated in the preceding deconvolution step of the pipeline is then used to initialise the skymodel in the next deconvolution. In other words, a proper dirty image is only formed at the very frst imaging step and, thanks to ssd, the LoTSS-DR1 pipeline can work only on residual images and update the spec-tral skymodel at each deconvolution step. 


2.3.3. The LoTSS-DR1 pipeline 
The LoTSS-DR1 pipeline has many confgurable parameters including resumability, taking into account time and bandwidth smearing, bootstrapping the fux density scale off existing sur-veys, correctionoffacet-based astrometric errors, user specifed deconvolution masks, and substantial fexibility in calibration and imaging parameters. The pipeline is suitable for the analysis of various LOFAR HBAcontinuum observations, including interleaved observations or those spanning multiple observing sessions. The entire pipeline takes less than fve days to image one LoTSS pointing when executed on a compute node with 512GBRAM(the minimum requiredforthe pipelineis192GB) and four Intel Xeon E5-4620 v2 processors, which have eight cores each (16 threads) and run at 2.6 GHz. 
The pipeline operates on the direction-independent calibra-tion products which, for each pointing, are 25 10-SB (1.95 MHz) measurement sets witha time and frequency resolutionof8s and two channels per 195 kHz SB. The pipeline frst removes severely fagged measurement sets (those with ≥80% of data fagged) and selects six 10-SB blocks of data that are evenly spaced across the total bandwidth for imaging. This quarter of the datais self-calibratedto graduallybuildupa modelofthe 
A1, page4of 21 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 

Fig. <BR>3. <BR>Self-calibration loop of LoTSS-DR1. From left to right and top to bottom: 60 SBs without any DDE correction, 60 SBs after applying DDE phase calibration,60SBs after applying DDE phase and amplitude calibration, and 240SB image after applying DDE phase and amplitude calibration.Thecolourscalesare proportionaltothesquarerootofthenumberofSBsandtheblacklinesshowthefacetsusedbykms and ddfacet. 
radio emission in the feld, which is then used to calibrate the full data set.Abrief outline of the steps of LoTSS-DR1, which are shownin Fig. 3,is as follows: 
Step1 Direction-independent spectral deconvolution and 
imaging (6 × 10 SB); Step2 Skymodel tesselationin45 directions; Step3 Direction-dependent calibration (6 × 10 SB, kms with 
EKF); Step4 Bootstrapping the fux density scale; Step5 Direction-dependent spectral deconvolution and imag-
ing (6 × 10 SB, phase-only solutions, three major cycles); Step6 Direction-dependent calibration (6 × 10 SB, kms with EKF) 
Step7 Direction-dependent spectral deconvolution and imag-ing (6 × 10 SB), one major cycle, amplitude, and phase solutions); 
Step8 Direction-dependent calibration (24 × 10 SB,kms with EKF); 
Step9 Direction-dependent spectral deconvolution and imag-ing (24×10SB,twomajorcycles, amplitude,andphase solutions); Step 10 Facet-based astrometric correction. 
The ddfacet is used in Step 1 to image the direction-independent calibrated data using the ssd algorithm, which allows us to rapidly deconvolve very large images. The present implementation of ssd requiresadeconvolution mask and we use ddfacetto automatically generateonebasedona thresholdof15 timesthe local noise, whichis re-evaluatedateverymajorcycle. The mask created during the deconvolution is supplemented with a mask generated from the TGSS-ADR1 catalogue to ensure that allbright sourcesinthefeldaredeconvolvedevenwhen observing conditions are poor and automatically masking the sources is challenging. The image produced from the 60SB data set con-sists of 20 000 × 20000 1.500 pixels,hasa restoring beamof1200 , andthenoisevaries between0.25mJybeam−1 and2mJy beam−1 depending on the observing conditions and source environment. From this image a refned deconvolution mask is created and used to reduce the number of spurious components in the ssd component model of the feld by fltering out those that lie out-side the region within the refned mask. 
AtStep2the resultingskymodelisusedto defne45facets that cover the full 8.3◦× 8.3◦ region that has been imaged. The ssd component model is used for the frst direction dependent 
A1, page5of 21 

calibrationofthe60SBdataset(Step3).This calibrationisdone using kms, which creates an amplitude and phase solution for eachofthe45facetsevery60sand1.95MHzof bandwidth,and the data are reimaged. Throughout the pipeline, in order not to absorb unmodelled sky emission into thekms calibration solu-tions(in particularfaintextended emission seenbya small number of baselines), we always calibrate the visibilities using only baselines longer than1.5km (corresponding to scales of∼4.50). 
After this initial direction-dependent calibration we boot-strap the LoTSS-DR1 fux densities in Step 4 following the procedure described by Hardcastle et al. (2016). This not only improves the accuracy of our fux density estimates but also decreases amplitude errors that can occur owing to imperfections in the calibration across the bandwidth. In this step each of the six 10-SB blocks imaged in the previous step is imaged separately at lower resolution (2000)usingddfacet which applies the direction dependent phase calibration solutions.Acatalogue is made from the resulting image cube using the Python Blob Detector and Source Finder (PyBDSF; Mohan&Rafferty 2015) where sources are identifed using a combined image created from all the planes in the cube and the source fux density mea-surements are extracted from each plane using the same aperture. Sources within 2.5◦ of the pointing centre that are at least 10000 from anyother detected source and have a integrated fux density exceeding 0.15Jy are positionally cross matched with the VLSSr and WENSS catalogues using matching radii of 4000 and 1000, respectively. The WENSS catalogue used has all the fux densities scaledbyafactorof 0.9 which, as describedby Scaife&Heald (2012), bringsit intooverall agreement with the fux density scale we use. Correction factors are then derived for each of the six 10-SB blocks to best align the LoTSS-DR1 integrated fux density measurements with VLSSr and WENSS assuming the sources have power-law profles across this frequency range (74 MHz–325 MHz). During the ftting, sources that are poorly describedbyapowerlaw areexcludedto remove, for example incorrect matches or sources with spectral curvature. From the 70 ± 14 matched sources per feld the correction factors derived for each of the six 10-SB blocks are typically 
0.85 ± 0.1 and these are extrapolated linearly to the entire 25 10-SB data set. The six 10-SB 2000 resolution images are also stacked to provide a lower resolution (2000)image that has a higher surface brightness sensitivity than the higher resolution images. This image is used to identify diffuse structures that are prevalentinLOFAR images,butmaynotbe detectedatsufficient signifcance in the higher resolution imaging. These extended sources are then added to the mask to ensure that theyare deconvolved in later imaging steps. Sources are classifed as extended sources if theyencompass a contiguous region larger than 2000 pixels with all pixels having a signal above three times the local noise of the image. 
After the bootstrap derived correctionsfactors are applied the 60 SBs of data are imaged with the direction-dependent phase solutions applied in ddfacet in Step 5. As explained above, for efficiency reasons ssd is initiated with the ssd components from the direction-independent imaging, which allows us to deconvolve deeply with three major ssd iterations. The image size and resolution are the same as in Step 3 but the input mask is improved because it is a combination of that obtained from the direction-independent imaging, the mask generated from the TGSS-ADR1 catalogue, and the low-resolution mask created from the bootstrapping; at this point the auto-masking threshold is also lowered to ten times the local noise. Again, once the imaging is complete the image is masked and the mask is used to reduce spurious entries in the ssd component model. The noise levels in this second imaging step range from 130µJybeam−1 to 600 µJybeam−1. In Step6 this new model is input intokms whichcalculates improved direction-dependent calibration solu-tions for eachof the45facetsevery60s and 1.95MHzof band-width. 
A third imaging step is performed on the 60 SBs of data (Step 7), this time applying both the phase and amplitude direction-dependent calibration solutionsbut otherwise following the same procedure as before. This produces images with noise levels ranging from 100 µJybeam−1 to500 µJybeam−1 and a fnal ssd component model that is used to calibrate the entire 240SBsof the data set withkms (Step 8). 
The full bandwidth is imaged at both low and high resolution in ddfacet with the newly derived phase and amplitude solutions applied (Step 9). The low-resolution image has a resolution of 2000 anda signifcantly higher surface brightness sensitivity than when imaging at higher resolution. In this low-resolution image ssd is not initiated with a previously derived model because the uv-data used in the imaging are different as an outer uvcut of 25.75km is applied.To deconvolve deeply we perform three separate iterations of the low-resolution imaging, each time improving the input mask and lowering the automasking threshold. The noise level of the fnal 2000 resolution images ranges from 100 µJybeam−1 to 400 µJybeam−1, which corresponds toa brightness temperatureof9K–35K. 
The full bandwidth high-resolution imaging is performed witha resolutionof600. The deconvolution mask that has been graduallybuiltup throughthe self-calibrationofthe60SB data set, as well as that from the lower resolution imaging from the full bandwidth, and an auto masking threshold in ddfacet of fve times the local noise allow for a very deep deconvolution. This is performed with two separate runs of ddfacet with a masking step in between to ensure that the local noise is well estimated andfaint sources (signal-to-noise(S/N) ≥5) are masked. The resulting high-resolution images have noise levels that vary from 60 µJybeam−1 to 160 µJybeam−1. Once the deconvolution is complete the images are corrected for astrometric errors in ddfacet which can apply astrometric corrections to each of thefacets independently (Step 10). The astrometric corrections applied vary from 0.000 to 4.400 with a median of 0.800 and are derived from cross-matching the LOFAR detected sources in eachfacet with thePan-STARRS catalogue(Flewellingetal. 2016). The errors on the derived offsets vary from 0.100 to 4.800 with a median 0.200 . 
During the cross-matching a histogram of the separations between allPan-STARRS sources within 60arcsec of compact LOFAR sources is made for eachfacet. This typically consists of ∼140Pan-STARRS sources per LoTSS-DR1 source and an averageof190 radio sourcesperfacet.Ifall sourcesinthefacets are systematically offset, then this histogram should have a peak at the value of the offset between the LoTSS-DR1 and Pan-STARRS sources. To search for the location of this peak and estimate the RA and Dec offsets and their corresponding errors in each facet, we use a Markov Chain Monte Carlo (MCMC) method and uninformative priors. In this procedure the emcee package(Foreman-Mackeyet al. 2013)is used to draw MCMC samples from a Gaussian function plus a background where the initial parameter estimates are derived from the observed LOFARandPan-STARRS positionoffsets. The likelihood function is calculated using a gamma distribution with a shape parameter defned by the observed LOFAR and Pan-STARRS position offsets. The posterior probability distribution is calcu-lated taking into account the uninformative priors (background, offset, and Gaussian peak greater than zero and a Gaussian 
A1, page6of 21 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 
500) 
standard deviation less than that are put on the offset Gaussian function and background level. 
The pipeline is very robust and with no human interaction the processingfailed for only5of the63 feldsin the HETDEX Spring Field region, thus providing 58 images in this region. One (P2) of these failures was due to exceptionally bad ionospheric conditions and the other four (P31, P210+52, P214+52, and P215+50) were due to the proximity of very bright sources (3C 280 and 3C 295). 



2.4. Mosaicing and radio source cataloguing 
The LoTSS pointings tile the sky following a spherical spiral distribution(Saff <BR>&Kuijlaars 1997);theyare typically separated by 2.58◦ and have six nearest neighbours within 2.8◦.With the FWHM of the hba_dual_inner station beam being 3.40◦ and 4.75◦ at the top (168 MHz) and bottom (120 MHz) of the LoTSS frequency coverage, respectively, there is signifcant overlap between the pointings.To produce the fnal data release images, a mosaic has been generated for each of the 58 pointings that was successfully processed.For each pointing the imagesof the (up to six) neighbouring pointings are reprojected to the frame of the central pointing using the astropy-based reproject code and then all seven (or fewer) pointings are averaged using the appropriate station beam and the central image noise as weights in the averaging. During the mosaicing of the high-resolution images, facets with uncertainties in the applied astrometric corrections (derived as described in Sect. 2.3.3)larger than 0.500 areexcluded to ensure that the fnal maps have a high astrometric accuracy. This criterion is also a good proxy for image quality and allows us to identify and remove any facets that diverged during the processing due to poor calibration solutions. Once the images of the neighbouring pointings are combined the mosaiced map is blanked to leave just the pixels that lay within the 0.3 power point of the station beam of the central pointing. An example region froma mosaicisshowninFig. 4andthe noisemapofthe entire mosaiced regionis shownin Fig. 5. 
To produce a catalogue of the radio sources we performed source detection on each mosaic using PyBDSF. The sources were detected with a 5σ peak detection threshold and a 4σ threshold to defne the boundaries of the detected source islands that were used for ftting. The background noise variations were estimated across the images usingasliding box algorithm, where a box size of 30× 30 synthesised beams was used except in the regions of high S/N sources(≥150) where the box size was decreased to just 12 × 12 synthesised beams; this box size was tuned to more accurately capture the increased noise level in these regions. The PyBDSF wavelet decomposition functionality was also utilised to better characterise the complex extended emission in the images. The resulting catalogues of the individual mosaics were combined and duplications were removed by onlykeeping sources that are detected in the mosaic to which they are closest to the centre. 
In the concatenated catalogue the columns kept from the PyBDSF output are the source position, peak brightness, inte-grated fux density, source size and orientation, and the statistical errors from the source ftting for each of these. In addition we keep the source code which describes the type of structure ftted by PyBDSF (see Table 1 caption for the defnition of these) and the local root mean square noise estimate. We append columns that provide the mosaic identity, number of pointings that contribute to the mosaic at the position of the source, fraction of those in which the source was in the deconvolution mask, and whether or not the source is believed to be an artefact(see Williamsetal.2019fora descriptionof artefact identifcation). The fraction of the source in the deconvolution mask is calculated by fnding the mask value (1 or 0) at the centre of each Gaussian component for every source in all of the contributing pointings and using the effective integration times to calculate the weightedaverage.To fnd the masked fraction for a source that consists of multiple Gaussian components, we use the integrated fux densities of each component as weights and assign the weighted average of the masked fraction of these components to the source. These fnal parameters, together with the mosaiced residual images, which are also provided, allow users to assess the quality of the deconvolution for sources. This is particularly important forfaint sources that may not be in the masks and also for extended sources where, because of the inte-gral of the dirty beam exceeding that of the restoring beam, the apparent fux density in dirty images is substantially larger than in deconvolved images. Example entries from the catalogue are showninTable 1andaselectionof someofthe more spectacular sourcesin our images are representedin Fig. 6. 


3. Image quality 
The observations used in this data release were conducted between 2014 May 23 and 2015 October 15 and the vary-ing observing conditions signifcantly impact the image qualityeven after direction-dependent calibration, which reduces the impact of ionospheric disturbances. In this section, we assess the derived source sizes, astrometric precision, fux-density uncertainty, dynamic-range limitations, sensitivity, and completeness, and briefy discuss some remaining calibration and imaging arte-facts. 

3.1. Source extensions 
Identifying unresolved sources using the PyBDSF-derived measurements is complicated by several factors. For example, astrometric errors in the mosaiced images cause an artifcial broadening of sources, the varying quality of calibration blurs the sources by differing amounts, time averaging and bandwidth smearing can artifcially extend sources, and the extent to which a sourceis deconvolved impactsits measured size.To accurately quantify all this would require realistic simulations in which compact sources are injected into the uv-data taking into account DDEs. Furthermore, as the precise criteria for distinguishing resolved sourcesvaries betweenfacets and observations,a pro-hibitively large number of these simulations would need to be performed. Our calibration and imaging pipelines are continuing toevolveand hencesuchalarge undertakingisbeyondthe scope of this present study.An alternativeapproachwouldhavebeento inject point sources into our maps and use these to characterise the source fnding algorithm; however, such a simulation would not account for distortions in source morphologies caused by calibration inaccuracies. Instead we attempted to assess whether or not sources are resolved by looking at the extensions of real sources that we assert are unresolved and we used these to defne an average criterion with which additional unresolved sources can be identifed across the entire mosaic. 
To create a sample of unresolved sources the LoTSS-DR1 catalogue was frst fltered to contain only isolated sources, which we defne as being sources with no other LoTSS-DR1 source within 4500. Any sources that were not in the deconvolution mask in every pointing in which they are detected were also excluded. From the remaining entries we then selected only sources that are classifed as “S” by PyBDSF; this source 
A1, page7of 21 Fig. <BR>4. <BR>Top fgure: example of a LoTSS-DR1 image; bottom fgures: same region in NVSS(left)and FIRST(right). The black and red circles overlaid on the FIRST image show FIRST and TGSS-ADR1 sources, respectively. In this region there are 689 LoTSS-DR1 sources, 71 FIRST sources, 46 NVSS sources, and 16 TGSS-ADR1 sources. The resolution of the LoTSS-DR1 image is 600 and the sensitivity in this region is approximately 70 µJybeam−1.Thisfeldis dominatedbythe spectaculargalaxyNGC4258,whichinthe LoTSS-DR1imagehasanextentofover 3000 synthesised beams, together with the smaller edge-on spiralgalaxy NGC 4217. 


A1, page8of 21 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 

Fig. <BR>5. <BR>Noise image of the LoTSS-DR1 where the median noise level is 71 µJybeam−1. As described in Sect. 3.4 manyof the regions with high noiselevels are causedby dynamic-range limitations. Sources fromtherevised3C catalogueof radio sources(Bennett 1962)areoverplottedas black circles to show the location of potentially problematic objects. 
code corresponds to sources that are the only objects within a PyBDSF island and are well ft with a single Gaussian. Finally, as described below, we imposed a cut on the major axis of the LoTSS-DR1 sources to limit the maximum extent of the low-frequencyemission. 
We emphasise that, owing to imperfect calibration, most truly unresolved sources in the LoTSS-DR1 catalogue do not have an integrated fux density to peak brightness ratio of 1.0 ora ftted major axis sizeof600 (i.e. a size equal to the restoring beam).Forexample, the approximately50 seemingly compact, bright (S/Nin excess of 500) sources that meet the above criteria all have measured sizes in the FIRST catalogue of less than500 and we can therefore assert these are either unresolved or barely resolved. However, in the LoTSS-DR1 catalogue these sourceshavea median ratiooftheintegratedfux densitytopeak brightness equal to 1.12 with a median absolute deviation of 
0.04. Furthermore, for seemingly compactLoTSS-DR1 sources that are detected with a lower S/N there is signifcantly more variation in the measured integrated fux density to peak brightness ratio.To characterise this, and separateextended from compact sources, we derived an envelope with the functional form 
Sint  Speak 
= 1.25 + A&#1;B, which encompasses 95% of the LoTSS-
Speak rms 
DR1 sources that meet the above criteria (see Fig. 7). Thefactor of 1.25was derived from the median plus three times the median absolute deviation of the integrated fux density to peak brightness ratio of the seemingly compact highS/N(≥500) sources. We used this envelope to defne a boundary between compact and extended sources. 
The ftted envelope is dependent upon the cut used on the major axis of the LoTSS-DR1 sources and we explored the impact of this by varying that selection criterion from 1000 to 2000 (see Fig. 7). We fnd that this has little impact on the classifcation of sources with S/N of more than 100 as either extended or compact; however, it has a much larger impact on sources with lower S/Ns. Whilst there is no definite value to use for this cut, we chose a 1500 limit on the LoTSS-DR1 major axis, which gives a best ft envelope of 
Sint  Speak &#1;−0.53 
= 1.25 + 3.1. There are a total of 280 000 
Speak rms 
LoTSS-DR1 sources within this envelope and we defne these as compact. As a cross check we note that 19500 of these sources correspond to entries in the FIRST catalogue and in that catalogue 88% of them are less than500 in size, indicating that they are also compact at higher frequencies. 


3.2. Astrometric precision 
The astrometry of our images is originally set by our phase cal-ibration based on the TGSS-ADR1 catalogue. However, dur-ing direction-dependent calibration the astrometry can shift between regions because of the varying precision of the cali-bration models that are built up in different facets. For example, after direction-dependent calibration of a LOFAR data set Williams et al. (2016)found∼100 offsets that varied systematically across their feld,but they were able to correct these using the positions in the FIRST catalogue to provide a LOFAR HBA image with a standard deviation in the RA and Dec offsets from FIRST of just 0.400. In our processing we also refned the astro-metric accuracy after the self-calibration cycle is complete by correcting eachfacet independently using positions in thePan-STARRS optical catalogue. Furthermore, during the mosaicing wedonot includefacetsthathavean uncertaintyinthe estimated astrometric correction of greater than 0.500 to ensure high astro-metric accuracy(see Sect. 2). 
To determine the resulting astrometric accuracy of our mosaic catalogue we performedasimple nearest neighbour cross match in which we took the closestPan-STARRS, WISE, and FIRST counterpart that lies within 500 of each of the compact LoTSS-DR1 sources that were identifed using the procedure described in Sect. 3.1. We then created histograms of the RA and Decoffsets and ft these witha Gaussian, where the location of the peak and the standard deviation correspond to our system-atic position offset and the total uncertainty; these total errors are a combination of errors in the LoTSS-DR1 positions from the source fnding software, the real astrometric errors in the LoTSS-DR1 positions, and the errors in the positions of objects inthe cross-matchedsurveys(whichwere selectedowingtotheir high astrometric accuracy). The astrometryof thePan-STARRS 
A1, page9of 21 
A1, page 10 of 21 
Table 1. Example of entries in theLoTSS-DR1 source catalogue. 
SourceID RA Dec S peak S int Maj Min DC_Maj DC_Min PA rms Type Mosaic Number Masked Arte
( 00 )( 00 )( 00 )( 00 ) 
(mJybeam−1) (mJy) ( ◦ ) (mJybeam−1) pointings fraction fact 

ILTJ140629.07 211.6211◦ 54.0190◦ 0.3 0.6 9.0 7.3 6.7 4.1 89 0.08 S P209+55 2 0.00 × +540108.3 ±1.2 00 ±0.8 00 ±0.1 ±0.1 ±2.9 ±2.0 ±2.9 ±2.0 ±64 ILTJ113833.90 174.6413◦ 52.4434◦ 0.4 0.5 7.7 5.2 0.0 0.0 87 0.06 S P12Hetdex11 4 0.00 × +522636.3 ±0.6 00 ±0.3 00 ±0.1 ±0.1 ±1.3 ±0.6 ±1.3 ±0.6 ±19 ILTJ114532.75 176.3864◦ 47.9582◦ 0.4 0.4 7.3 5.8 4.2 0.0 61 0.06 S P15Hetdex13 4 0.00 × +475729.5 ±0.6 00 ±0.5 00 ±0.1 ±0.1 ±1.5 ±0.9 ±1.5 ±0.9 ±36 ILTJ114443.86 176.1827◦ 46.5350◦ 1.4 2.2 8.7 6.2 6.3 1.7 168 0.11 S P15Hetdex13 2 0.77 × +463206.0 ±0.2 00 ±0.4 00 ±0.1 ±0.2 ±0.8 ±0.5 ±0.8 ±0.5 ±12 ILTJ134204.70 205.5196◦ 49.5226◦ 1.0 1.1 6.8 6.0 0.0 0.0 87 0.06 S P42Hetdex07 3 1.00 × +493121.4 ±0.2 00 ±0.2 00 ±0.1 ±0.1 ±0.5 ±0.4 ±0.5 ±0.4 ±21 ILTJ111452.35 168.7181◦ 54.7318◦ 0.5 0.4 5.9 5.4 0.0 0.0 121 0.08 S P169+55 3 0.71 × +544354.6 ±0.4 00 ±0.3 00 ±0.1 ±0.1 ±0.9 ±0.8 ±0.9 ±0.8 ±70 ILTJ122721.08 186.8378◦ 50.8748◦ 0.6 3.0 17.3 10.2 16.2 8.3 18 0.06 S P26Hetdex03 4 0.60 × +505229.1 ±0.4 00 ±0.7 00 ±0.1 ±0.1 ±1.8 ±0.9 ±1.8 ±0.9 ±10 ILTJ121124.15 182.8506◦ 47.0435◦ 1.9 8.0 27.5 5.5 0.0 0.0 177 0.11 M P19Hetdex17 3 0.93 × +470236.5 ±1.8 00 ±0.2 00 ±0.1 ±0.3 ±4.2 ±0.4 ±4.2 ±0.4 ±6 ILTJ143026.91 217.6121◦ 48.0969◦ 0.5 0.6 7.3 6.2 4.2 1.7 135 0.06 S P217+47 3 0.70 × +480548.8 ±0.4 00 ±0.4 00 ±0.1 ±0.1 ±1.0 ±0.7 ±1.0 ±0.7 ±35 ILTJ124531.14 191.3797◦ 53.6874◦ 0.4 0.7 10.6 6.2 8.8 1.7 46 0.06 S P191+55 4 0.59 × +534114.7 ±0.7 00 ±0.7 00 ±0.1 ±0.1 ±2.1 ±0.8 ±2.1 ±0.8 ±17 
√ 
ILTJ144630.58 221.6274◦ 52.8271◦ 0.6 7.4 25.7 17.0 25.0 15.9 61 0.30 S P223+52 3 1.00 +524937.5 ±3.6 00 ±2.8 00 ±0.2 ±0.2 ±9.2 ±5.7 ±9.2 ±5.7 ±47 ILTJ115435.71 178.6488◦ 49.2769◦ 0.3 0.4 9.9 5.0 0.0 0.0 102 0.06 S P18Hetdex03 5 0.00 × +491636.8 ±1.0 00 ±0.4 00 ±0.1 ±0.1 ±2.5 ±0.7 ±2.5 ±0.7 ±15 ILTJ134314.24 205.8093◦ 51.0778◦ 0.3 0.6 8.5 7.7 6.0 4.8 166 0.06 S P206+52 4 0.00 × +510440.1 ±0.6 00 ±0.7 00 ±0.1 ±0.1 ±1.8 ±1.5 ±1.8 ±1.5 ±87 ILTJ150957.26 227.4886◦ 54.5732◦ 0.4 1.9 15.8 11.4 14.6 9.7 99 0.14 S P227+53 2 0.00 × +543423.6 ±2.5 00 ±1.6 00 ±0.1 ±0.2 ±5.9 ±3.8 ±5.9 ±3.8 ±56 
Notes. Theentirecataloguecontains325694sources.Theentriesinthecatalogueareasfollows:sourceidentifer(ID),J2000rightascension(RA),J2000declination(Dec),peakbrightness(S peak), integratedfux density (S int), major axis (Maj), minor axis (Min), deconvolved major axis(DCMaj), deconvolved minor axis(DCMin), position angle (PA), local noise at the position of the source(rms),typeofsourceasclassifedbyPyBDSF(Type–where “S” indicates an isolated source that isft with a single Gaussian; “C” represents sources that areft by a single Gaussian but are withinan island of emission that also contains other sources; and “M”isusedforsourcesthatareextendedandfttedwithmultipleGaussians),themosaicidentifer(Mosaic),thenumberofpointingsthatare mosaiced at the position of the source (Number pointings), the fraction of pointings inwhich the source is in the deconvolutionmask (Masked fraction), and whether or not the entry is believedtobeanartefact(Artefact).Only2590entrieshavebeenidentifedasartefacts(see Williamsetal.2019).The errorsinthecataloguearetheuncertaintiesobtainedfromthePyBDSF sourceftting.Additional uncertainties on the source extensions, astrometry, andfux scale are described in Sect. 3. 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 

Fig. <BR>6. <BR>Selection of resolved sources in the LoTSS-DR1 images with the colour scale and contours chosen for display purposes. The synthesised beam is shown in the bottom left corner of each image. 
catalogue was determined using a combination of 2MASS and Catalog for sources detected at high signifcance, and the FIRST Gaia positions and the typical standard deviation of the offsets surveyhas astrometric uncertainties of 0.100 with respect to the from Gaia positions is less than 0.0500 (Magnier et al. 2016).The absolute radio reference frame(White et al. 1997). WISE catalogue has a positional uncertainty of 0.200 (Cutri et al. We cross-matched a total of 7100 sources from the LoTSS2012)in RA and Dec with respect to the 2MASS Point Source DR1 catalogue to all three comparison sources and we found 
A1, page 11 of 21 Fig. <BR>7. <BR>Ratio of the integrated fux density to peak brightness as a func<BR>tion of S/N for sources in the LoTSS-DR1 catalogue. All catalogued sources are shown in red and the sources we used to defne an envelope that encompasses 95% of the compact sources are shown in blue (see Sect. 3.1). The impact of varying the limit on the major axis size of LoTSS-DR1 sources is shown with the triangles, crosses, and diagonal crosses corresponding to 1000, 1500, and 2000 limits, respectively. Each of these is ftted with an envelope and the fnal selected envelope of 


Speak 
Sint = 1.25 + 3.1 &#1;−0.53 was derived from the 1500 limit. 
rms 
Speak 
that,for these sources, thereisasystematic positionaloffset from Pan-STARRS of less than 0.0300 and the standard deviation of the offsets is less than 0.200 in both RA and Dec (see Fig. 8). Simi<BR>larly,in comparisontoWISE,wefoundthe same sourceshavea systematic offset of less than 0.0100 and a standard deviation of less than 0.2700 in both RA and Dec. When comparing to FIRST, the systematicoffsets are less than 0.0200 and the standard deviation is approximately 0.300 in both RA and Dec. The direction of the derived systematicoffsetsvaries when comparing the LoTSS positions with the three different surveys.We alsoexamined the astrometric accuracy of our mosaic catalogue as a function of the LoTSS-DR1 peak brightness. We checked the accuracy of the catalogue to better estimate the real astrometric errors in the LoTSS-DR1 positions as bright(≥20 mJy), compact sources typ-ically have errors in their derived positions of less than 0.0500 . For the compact LoTSS-DR1 sources above 20 mJy the ftted standarddeviationtoa GaussianoftheRAandDecoffsets from Pan-STARRS, and hence the approximate absolute astrometric accuracy of LoTSS-DR1, is less than 0.200. The standard deviation gradually increases to 0.500 for the faintest LoTSS-DR1 sources(≤0.6 mJy) where the uncertainty in position from the source ftting can be as high as 1.000 . 
To assess thevariationin the astrometric accuracyofvarious pointings prior to mosaicing the same analysis was performed on the catalogues derived from the LoTSS-DR1 images of the individual pointings. We only used similar sources to the previous analyses by frst cross-matching the catalogues derived from the individual pointings with the LoTSS-DR1 compact source catalogue (see Sect. 3.1). The resulting catalogue was then cross-matched with the Pan-STARRS catalogue. In addition we also imposed cuts on the catalogues from each LoTSSDR1 pointing to include only sources within the 0.3 power point of the station beam, which is where the primary cut is made during the mosiacking. Furthermore, we only used sources classifed by PyBDSF as “S” type sources in the pointing catalogues and those locatedinfacets where the uncertaintiesin the Pan-STARRS dervied astrometric corrections of less than0.500 . 
A1, page 12 of 21 Fig. <BR>8. <BR>Residual RA and Dec offsets for LOFAR detected sources matched with their Pan-STARRS counterparts. The histograms show the number of sources at various RA and Dec offsets and the ellipse shows the peak location (less than 0.0200 from the centre in both RA and Dec) and the FWFM(σ ≈ 0.200)of the Gaussian functions that are ftted to the histograms of the offsets. Similar plots showing the same LoTSS-DR1 sources cross-matched with WISE or FIRST sources show comparable systematicoffsets and standard deviations of less than 0.2700 and 0.300, respectively. 

We found that the standard deviation of the Gaussian ftted to a histrogram of the RA and Dec astrometric offsets fromPan-STARRS varied from 0.3100 to 0.5400 with an average of 0.3900 and that the peak of the ftted Gaussian functions were displaced by between 0.0500 and 0.1200. These numbers give an indication of the varying astrometric accuracy across the HETDEX Spring Fieldregion.Wenotethat,aswasfoundinthe mosaicedimages, these astrometric errorsvary with theS/Nof the detections and thisexplainswhytheindividual pointingshave apparentlylarger astrometric errors than the mosaiced images. 


3.3. Accuracy of the fux density scale 
Owing to inaccuracies in the existing LOFAR beam models, transferring amplitude solutions derived from calibrators to the target feld data does not generally result in an accurate fux density scaleforthetarget feld.Forexample, Hardcastleetal. (2016)found the errors in the fux density scale to be up to 50%. To correct thisHardcastleetal. (2016)deviseda bootstrapping approach to align the fux density scale of their LOFAR images with the fux density scales of other surveys whilst also providing more reliable in-band spectral index properties.We applied this technique early in the LoTSS-DR1 processing pipeline to ensure consistencywith the VLSSr and WENSS fux density scales (see Sect.2).To assess whether the fux density scale remains consis-tent throughout the processing we performed the same bootstrap-ping calculation with our fnal images. From our fnal images, the recalculated correctionfactors range from 0.8 to 1.3 witha meanof1.0anda standarddeviationof 0.08.Wedidnotapply these recalculatedfactorsin this data releasebut they indicate 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 

Fig. <BR>9. <BR>LoTSS-DR1 to TGSS-ADR1 integrated fux density ratio as a function of integrated fux density(left panel)and for sources with a integrated fux density higher than 100mJy asa functionof distance from the nearest LoTSS pointing centre(right panel). Below 100mJy the completeness of TGSS-ADR1 drops below 90% and, as a consequence, there is signifcant scatter in the integrated fux density ratio for sources below this limit. In the right panel we show that the 835 compact sources above this integrated fux density limit have a median integrated fux density ratioof0.94anda standarddeviationof0.14(blue points)anda medianpeak brightness ratioof0.83anda standarddeviationof0.13(red points). The thicker symbols show the median within bins indicated by horizontal error bars and the vertical error bars show the 95% confdence interval of the derived median value estimated by the bootstrap method. The bins are chosen to contain equal numbers of sources, which is 500 and 170 for the left and right panels respectively. The vertical dashed line shows the median distance between LoTSS pointings and manyof the measurements at greater distance are due to the edges in the LoTSS-DR1 mosaic. 
that in some circumstances the fux density scale can drift dur-ing the processing; however, 60% percent of the felds remain within 5% of the original bootstrapped derived values. 
For furtherverifcationofthefuxdensityscalewecompared the catalogued integrated fux density in the compact source LoTSS-DR1 catalogue to those in the TGSS-ADR1 catalogue. The TGSS-ADR1 measurements were not used during the boot-strapping to allow for this comparison. Furthermore, the TGSSADR1 fux density scale is not tied to the fux density scales of VLSSr or WENSS as the survey was instead calibrated directly against bright, well-modelled sources, on the Scaife&Heald (2012) fux density scale. For the 835 compact sources with LoTSS-DR1 integrated fux densities in excess of 100 mJy the median ratio of the integrated LoTSS-DR1 fux densities to the integrated TGSS-ADR1 fux densities is 0.94 and the standard deviation of 0.14 (see the left panel of Fig. 9). However, at integrated fux densities below 100 mJy, where the point-source completeness of the TGSS-ADR1 catalogue decreases to less than90%and detectionsarenotalwaysatveryhigh signifcance, there is substantially more scatter in the ratio of TGSS-ADR1 to LoTSS-DR1 integrated fux densities with a standard deviation of 0.27. 
Part of the scatter in the TGSS-ADR1 and LoTSS-DR1 inte-grated fux density ratios is from variations in the quality of the imagesofvarious LoTSS-DR1 pointings.Toexamine the con-sistency of our measurements we compared the integrated fux density of compact sources in catalogues derived from each of the pointings used in LoTSS-DR1 with the TGSS-ADR1 catalogue. The median ratio of the LoTSS-DR1 integrated fux densities to the TGSS-ADR1 integrated fux densities varies from 0.75 to 1.15 with an median of 1.0 and a standard deviation of 0.08. The discrepancy between this median integrated fux density ratio, which is derived from individual LoTSS-DR1 pointings, and corresponding value for the entire mosaic (0.94), 
appears to be a consequence of the mosaicing. Sources with apparently low LoTSS-DR1 integrated fux densities more often reside in pointings with apparently lownoise levels that are more highly weighted during the mosaicing procedure. Furthermore, we made use of the large overlap between pointings to examine fux density scale variations and found that the standard deviation of the median ratio of the integrated fux density between pointings is 0.2 and, whilst the maximum discrepancyin the inte-grated fux density measurements is 55%, 80% of the ratios are within 20% of unity. 
We also searched for trends between the source integrated fux density measurements and the distance from the LoTSS pointing centres (see Fig. 9). Using the 835 bright compact sources in the mosaic catalogue that were cross-matched with TGSS-ADR1 we found no strong dependence of the ratio of the LoTSS-DR1 integrated fux density to the TGSS-ADR1 inte-grated fux density on the distance from the closest LoTSS point-ing; the inner bin has a ratio of 0.95 and the outer has a ratio of 
0.92.For the peak brightness the radial dependence is slightly stronger with the inner bin at 0.86 and the outer bin at 0.81. To assess the impact at further distances we look at the peak brightness to integrated fux density ratio of compact sources in the LoTSS-DR1 catalogues derived from individual pointings. Giventhatourdataareaveragedtotwo channelsperSBand8s, it maybe expected that time-averaging and bandwidth-smearing effects are non-negligibleinthe LoTSS-DR1 mosaics;forexam-ple, we estimate using the formulasgivenby Bridle&Schwab (1989)that at600 resolution the time-averaging and bandwidth smearing are as shown in Fig. 10. However, ddfacet hasafacet-dependent PSF which, for deconvolved sources, accounts for the impact of smearing. As a result the ratio of the peak brightness to integrated fux density in our LoTSS-DR1 images does not have as strong a dependence on distance from the nearest point-ing centre as found in other studies that used imagers that do not 
A1, page 13 of 21 


1.0. The effects of time and bandwidth smearing are taken into account during deconvolution in ddfacet. The red points show the median ratios within bins of distance; the 95% confdence intervals are ∼0.02 and were estimatedby the bootstrap method. The horizontal errors barsgive the bin width and the vertical dashed line shows the median distance between LoTSS pointings. 
correct for this.We note that thereis stilla small radial dependence.Thismaybe becausefacets furtherfromthe pointing centre are generally larger and, as a consequence, the ionospheric calibration in those regions is not as precise. Overall, whilst there are variations in the accuracy of the fux density scale across the mosaic, we place a conservative uncertainty of 20% on the LoTSS-DR1 integrated fux density measurements. 


3.4. Dynamic range 
The dynamic range in our images is limited and bright sources haveanimpactontheimagenoise propertiesina non-negligible fraction of the area that has been mapped. Whilst there are many factors that impact the dynamic range, our testing of the data pro-cessing procedure has indicated that the amplitude normalisation scheme that we used certainly playsasignifcant role. Other con-tributors include the layout and sizeof thefacets and the quality of the models that arebuiltup during the self-calibration proce-dure. 
To assess the dynamic-range limitations we examined pixels on mosaics of the fnal ddfacet residual images in500 wide annuli around compact LoTSS-DR1 sources that were identifedin Sect. 3.1.Aprofleofthepixel standarddeviation within every annulus was determined for each of these sources out to a radius of 50000. Each profle was ft with a Gaussian function plus a constant, which we assume is the level of the noise in the surrounding region and we used this to normalise the mea-surements. Within each distance bin, we averaged together all normalised noise measurements of sources within a given inte-grated fux density ranger and the mean and standard deviation 
A1, page 14 of 21 
was determined to create an average noise profle as a function of distance. These average noise profles for various integrated fux density ranges are shown in Fig. 11. 
The area in square degrees of sky that surrounds bright sources and has a noise level more than 15% higher than the noise in the wider region depends on the source integrated fux density according to approximately 0.1(e−0.007S − 1) − 0.002, where S is the integrated fux density in mJy. Fromthis equation, and removing overlapping regions, we calculated that the noise is limited by the dynamic range of our maps (i.e. the noise is more than 15% higher than the noise level in regions uncontaminatedby bright sources)for32 squaredegreesofthe424 square degrees that were imaged, i.e. 8% of the total area of the survey. Similarly, we calculated the area witheven more enhanced noise levels of 50% and 100% higher than the noise level in uncontaminated regions as 3% and 2%, respectively. 


3.5. Sensitivity 
The latitude of LOFAR is 52◦5403200, putting the HETDEX Spring Field region, which has a declination ranging from 47◦ to 55◦, close to the optimal location where the projected area of the HBA dipoles and hence the sensitivity of the array is 
600 
at its highest. The entire LoTSS-DR1 resolution mosaic of the HETDEX Spring feld region covers an area of 424 square degrees and the median noise level across the mosaic is 71µJybeam−1;65%, 90%, and 95% of the area has noise levels below 78 µJybeam−1, 115µJybeam−1, and 147µJybeam−1, respectively (see Fig. 12). These variations are due to varying observing conditions, telescope performance (e.g. missing stations or a higher level of interference), pointing strategy, and imperfections in the calibration and imaging procedure. The impact of the calibration and imaging procedure is particularly evidentaroundbright sourcesinwhichthenoiseis limitedbythe dynamic range, as discussed in Sect. 3.4. The variations due to the observing conditions are also signifcant and the noise level on images of the individual pointings varies from 60 µJybeam−1 to 160 µJybeam−1. The sensitivity variations due to the mosaicing strategy in this region are much smaller. We fnd that the average mosaic noise as a function of distance from the clos-est pointing centre (just including regions covered by more than one pointing) only varies from 72 µJybeam−1 to 78 µJybeam−1 with a minimum at ∼1◦ from a pointing centre and a maximum at ∼1.6◦ from the nearest pointing centre. By comparison, the LoTSS-DR1 2000 resolution mosaic has higher noise levels due to the uv-cut applied in the imaging step. In this case the median noise level is 132 µJybeam−1, and 65%, 90%, and 95% of the area has noise levels below 147 µJybeam−1, 223µJybeam−1, and 302 µJybeam−1, respectively. 
The contribution of confusion noise to the total noise level that is measured on our 600 resolution images is also small. To quantify this we followed the approach of Franzen et al. (2016)and injected a broken power-law distribution of point sources convolved with a 600 Gaussian into a blank image. As in Franzen et al. (2016) the power law used for sources with an integrated fux density in excess of 6mJy was dN = 
dS 
6998S−1.54Jy−1sr−1 in agreement with Euclidean normalised differential countsat154MHz derivedby Intemaetal. (2011), Ghosh et al. (2012), and Williams et al. (2013). For fainter sources we ftted a power law of dN = 82S−2.41Jy−1sr−1 to 
dS 
the deep 150MHz counts presentedin Williamsetal. (2016) and, whilst these counts reach a depth of 700 µJy, for simplicity we assumed theyhold to an integrated fux density limit of 10µJy. Given that the counts are thought to decrease towards 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 

such low fux densities (e.g. Wilman et al. 2008) this should result in a conservative estimate for the confusion noise. From the pixel values in the simulated image we derived the probability of defection [P(D)], which is highly skewed with an interquartile range of 18 µJybeam−1. Whilst this distribution is not Gaussian, to approximate the confusion noise this can be converted to a crude estimate of the sigma by dividing the interquartile range by a factor of 1.349, which gives a confusion noise estimate at 600 of 14 µJybeam−1, which is signifcantly lower than the rms levels obtained. Our lower resolution images, however, are much more severely impacted by confusion noise and when repeating the analysis at 2000 our confusion noise estimate is 85 µJybeam−1.We note that theveryfaint sources do not have a large impact on the sigma for the P(D) distributions; for example assuming the counts instead extend to 1µJy assumes 5.1 million sources rather than 200 000 sources per square degree but increases the 2000 resolution confusion noise estimate by only 5% to 89µJybeam−1. The power-law indices assumed in the calculations, however, play a more signifcantrole;forexample,againfollowing Franzenetal. (2016), if for the sources between 10 µJy and6mJy we assume dN = 
dS −1 
6998S−1.54, 1841S−1.8, 661.8S−2.0 or 237.9S−2.2Jy−1srwe estimate 2000 resolution P(D) sigma values of 1 µJybeam−1, 10 µJybeam−1, 24µJybeam−1, and 47µJybeam−1. 
Several of the early LoTSS observations were conducted in a manner in which two neighbouring pointings were observed simultaneously, including 10 observations (thus 20 pointings) in this data release. In these circumstances a minor impact on the sensitivity in the overlapping regions of the simultaneously observed pointings is correlated noise. In an attempt to quantify the impact we examined pixel values in the overlapping regions of pointings by reprojecting the images to a common frame and ignoring regions containing sources (defned as those withvalues more than 1σ). The Pearson correlation coefficient calculated from these noise pixels was generally found to be 0.03–0.13 for pointings observed simultaneouslybut typically less than 0.03 for pointings observedat separate times.Wealso compared noise levels in mosaiced regions that contained data from two simultaneously observed pointings with regions where all contributing pointings were observed at different times. We found that regions where simultaneously observed pointings contribute to the mosaics have a median noise level that is ∼2% higher than other regions. The LoTSS observations of neighbouring point-ings have not been conducted simultaneously since these very early observations. 


3.6. Completeness 
To thoroughly estimate the completeness of the survey, sources of varying fux densities and positions should be injected into simulated data sets that include realistic DDEs. However, in the absence of such simulations, we instead inject sources into the direction-independent calibrated data sets taking into account the direction-dependent corrections that are applied in that specifc direction to correct for the ionospheric and beam errors. After these data sets are processed with the pipeline and the injected sources are catalogued and their properties are compared to the parameters of the sources that were injected.We note that this procedure assumes that the direction-dependent corrections, with which thefake sourcs are injected, accurately describe the real DDEs. Given the computational cost of our calibration and imaging and that our pipelines will be improved for future data releases, we only performed this analysis for 10 SBs of data from one pointing following the procedure outlined below: 
Step1 Obtain the fnal direction dependent calibration solutions froma 240SB runof the LoTSS-DR1 pipeline; Step2 Createa simulated imageof 300 delta functions drawn 
from a power-law distribution(dN ∝ S−1.6) and use 
dS 
ddfacet to predict the visibilites for this model, cor
rupted by the same direction dependent distortions and 
add these to real direction independent calibrated data in 
the 10-SB data set; Step3 Execute the LoTSS-DR1 pipeline on the 10-SB simulated data set; 
For comparison, following the approach described in Heald et al. (2015), we also estimated the completeness by injecting 300 point-like sources with integrated fux densities drawn from a power-law distribution(dN ∝ S−1.6) into 
dS 
the fnal restored image of1 of the 10SB runs produced in Step 3.To improve the statistics the realistic simulations were 
A1, page 15 of 21 



Fig. <BR>13. <BR>Estimated point-source completenessfora10SB(1/24th of the data) for a single LoTSS-DR1 pointing. The red line shows the com-pleteness above a given integrated fux density and the blue line shows the fraction of sources detected at a specifc integrated fux density value. The solid lines show the results of the simulation in which point sources are injected into PyBDSF residual images and the dashed lines show results from when delta functions corruptedby realistic direction-dependent errors are injected into the uv-data before it is run through LoTSS-DR1. The error bars give the Poisson errors. 
repeated8timesgivinga totalof 2400 simulated point sources and the injection of sources into the fnal image was repeated 50 times giving a total of 15 000 sources. For both simulation types we ran PyBDSF on the simulated images and classifed the injected sources as detected if they are recovered within 7.500 of the injected location and with a measured inte-grated fux density within 10 times the error on the integrated fux density uncertainty. The fraction of the simulated sources that were detected as a function of integrated fux density, and the derived completeness, for both methods are shown in Fig. 13. 
Whilst the injection of distorted point-like sources into the uv-data gives a much accurate understanding of the true com-pleteness that we obtain from LoTSS-DR1 it is computationally expensive to perform such simulations for the full bandwidth of each of the data sets in the surveywith the full bandwidth of data. 
A1, page 16 of 21 Fig. <BR>14. <BR>Estimated point-source completeness of the LoTSS-DR1 cat<BR>alogue. The red line shows the completeness above a given integrated fux density and the blue line shows the fraction of sources detected at a specifc integrated fux density value. Because of the large number of sources injected during the simulation the Poisson errors are negligible but the errors bars refect the standard deviation of the measurements as a function of position across the mosaic. 

However, performing such simulations with 10 SBs of data from a single pointing suggests that the shape of completeness curves derived from realistic simulations is similar to that obtained from injecting sources into calibrated images (Fig. 13). Therefore, to approximate the completeness of the entire LoTSS-DR1 we only used the less computationally expensive approach of injecting point sources into residual images. 
From each of the 58 mosaic images a residual image is gen-erated using PyBDSF as a byproduct of the LoTSS-DR1 catalogue creation. Into each of these residual maps we inject 6000 sources with integrated fux densities drawn from a power-law distribution( dN ∝ S−1.6)and ranging from 0.1mJy to 10Jy. 
dS 
This procedure is repeated 50 times for each of the mosaiced imagesto ensureastatisticallyrobust measurement.The fraction of sources recovered above an integrated fux density limit, or the point-source completeness, varies with integrated fux den-sityasshowninFig. 14andis65%at 0.18mJy,90%at 0.35mJy, and 95% complete at 0.45mJy. However, we emphasise that, as shown in Fig. 13, the real integrated fux density level for the completenesslevelsis likelyafactorof ∼1.3 higher (thus 90% at 0.45 mJy). 


3.7. Image artefacts 
In the LoTSS-DR1 mosaics there are several different types of artefacts. The low-level positive and negative haloes are particularly prominent around some sources; these haloes can be diffcult to distinguish from real emission and make it challenging to precisely characterisefaint diffuse emission. These artifcial haloes are believed to be a product of having a minimum uvdistance on the baselines used in the calibration; we suspect this expedient, implemented to avoid modelling out extended emission, can cause the amplitude solutions of the antennas with more short baselines to become slightly discrepant from the more remote antennas.For comparison with our images, we note that several diffuse objects within the region coveredbythis data release have been processed using a different direction dependent calibration algorithm(Facet Calibration; vanWeerenetal. 2016).This procedure does not usea large minimumuv-distance 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 
in the calibration and the images do not suffer from artifcial haloes; see the maps presented in, for example Brgen et al. (2018),Saviniet al. (2018), and Wilberet al.(2018). 
In some felds there are also clear amplitude calibration artefacts that are primarily a consequence of the amplitude nor-malisation scheme that we used during the direction-dependent calibration. Some felds that were observed in bad conditions also have clear phase errors that are dependent on both our calibration solution interval (1min) and the size of the facets. Finally, whilst we attempted to ensure that our masks encompassextended sources, there are instancesin whichfaintdiffuse emission has still not been fully deconvolved. 
As described in Sect. 5.1, in future data releases we plan on improving upon each of these issues. However, for this data release, to aid with the identifcation of artefacts, we provided mosaics of the fnal residual maps to accompany our deconvolved continuum images along with the artefact fag resulting from the source (dis-)association and hostgalaxy identifcation workofthe companion paper Williamsetal.(2019). 



4. Public data release 
In this section, we summarise the products that form the frst LoTSS public data release11. These products consist of the mosaiced images that have been described in this paper in addition to the catalogue that we derived from the direct application of PyBDSF to the mosaiced 600 resolution images. In some cases, PyBDSF does not perfectly represent the radio source population: large extended radio sources may be split across several different catalogue entries in the PyBDSF catalogue, or alternatively two closely separated but physically distinct radio sources may be merged into a single catalogue entry by PyBDSF. Therefore, to enhance the scientifc value of the released LoTSS-DR1 catalogues we attempted to associate or deblend the catalogued components of radio emission into actual radio sources where necessary, and also to identify the optical counterparts of all sources. If an optical counterpart has been located we also estimated its photometric red-shift.For completeness, these projectsthataddvalueto publicly released LoTSS-DR1 catalogue are briefy summarised below, but for a full description see Papers III and IV in this series (Williams et al. 2019;Duncan et al. 2019). 
4.1. Mosaics and raw PyBDSF catalogue 
We released both the 600 and 2000 resolution 120–168 MHz mosaiced images that were created following the direction dependent calibration procedure described in Sect. 2. These mosaics cover 424 square degrees in the region of the HETDEX Spring Field (see Fig. 5)andhave the quality shown in Figs. 4 and 6 and described in detail in Sect. 3. We released 600 and 2000 mosaiced residual images to help assess the reliability of the morphology of extended structures; these images show the quality of deconvolution and properties of the background noise. 
The raw PyBDSF catalogue that was released was created from the600 resolution mosaiced images; this catalogue is described in Sect. 2.4. This catalogue contains 325 694 radio sources, has a source density of 770 sources per square degree and a point-source completeness of 90% at an integrated fux densityof 0.45mJy (see Sect. 3.6).To aid the interpretationof the catalogue completeness we released the PyBDSF derived noise mapsof the600 mosaics. 
11 https://lofar-surveys.org <BR>

4.2. Source (dis-)association and optical counterparts 
For most radio sources theexpected hostgalaxy positionis well defned by the properties of the radio source and it is therefore appropriate to use a statistical method to identify the counter-parts in Pan-STARRS and WISE. For this we employ a likelihood ratio method (e.g. Richter 1975; de Ruiter et al. 1977; Sutherland&Saunders 1992).However, fora numberof com-plex sources, such methods are either not possible or unreliable, so we employ a human visual classifcation scheme based on the Zooniverse12 framework. Sources in the raw PyBDSF catalogue are frst sorted based on their catalogued characteristics and selected either for visual (dis-)association and identifcation or for likelihood ratio cross-matching by means of a decision tree. The details of how these decisions are made and full details of the likelihood ratio and visual classifcation methods aregivenby Williamsetal. (2019). Using this procedure, counterparts were identifed for 71% of the radio sources. These source characteristics and visual inspection procedure are also very useful in fagging probable artefacts in thePyBDSF cata-logue.Again, details aregivenby Williamsetal.(2019),butthe fnal columnofTable 1providesafag highlightingthePyBDSF sources identifed as probable artefacts based on that work. 

4.3. Photometric redshift estimation 
Knowing the redshift of a source is a fundamental requirement for extracting key physical properties from continuum radio observations, such as luminosity or physical size, and for under-standing the hostgalaxy (e.g. its stellar mass). Although future optical spectroscopy campaigns such as WEAVE-LOFAR13 (Smith et al. 2016)will target more than 106 150 MHz-selected sources and provide high-precision spectroscopic redshifts and accurate source classifcations for a large portion of the LoTSS population, existing spectroscopic redshifts, largely from SDSS, are available only for a very small subset of sources. Therefore, photometric redshifts (photo-zs) are a vital method for identifying the physical properties of radio sources and we produced photo-z estimates for all plausible counterparts in the combined Pan-STARRs/All-WISE catalogue thatwas used for host-galaxy identifcation in the previous section. Full details of the photo-z estimation, which combines template-based and machinelearning estimates, are presented in a companion release paper (Duncan et al. 2019). 



5. Future prospects 
In future data releases we will not only present maps from a signifcantly larger fraction of the sky, but there is also active development to improve many aspects of the LoTSS data pro-cessing; in the survey we observed almost 20% of the northern sky and in this work we only presented 10% of these data or 2% of the northern sky. For example, to tackle the large LoTSS data rates we areworking with the LOFAR e-infra group to implement our direction-independent calibration pipeline on theForschungszentrum Jich and Pozna´nLTA sites. Furthermore, the observatory is beginning to utilise Dysco compression(Offringa 2016) to reduce the size of the archived data sets by a factor of approximately four. To improve the accuracyof the direction-independent calibration pipeline, amongst other things, the accuracy of the derived amplitude and clock 
12 www.zooniverse.org <BR>13 http://www.ing.iac.es/weave/weavelofar/ <BR>
A1, page 17 of 21 

solutions are being increased. In the direction-dependent cali-brationandimagingpipelinethereis signifcantworktoimprove the fdelity of the images and to implement the pipeline on the SURFsara Grid.Tofurther enhancethe scientifc potentialof our data products thereis also activeworktoexploitthe polarisation 
(e.g.Van Eck et al. 2018),wide fractional bandwidth, the longest baselines provided by the international stations, and to use the data for spectral line studies(Oonk et al. 2017;Salas et al. 2018; Emig et al. 2019)and for searches for transient sources. 
Discussing all of these future prospects in detail is beyond the scope of this article; however, in the following subsections we provide some details on several prospects, namely improving the direction-dependent calibration and exploiting the fractional bandwidth of LoTSS. 
5.1. Reducing image artefacts 
The LoTSS-DR1 processing strategy has produced sensitive and good quality LOFAR images, however it failed for 8% of the felds and, as described in Sect. 3.7, the fnal mosaics contain several different types of artefacts. Therefore, in an attempt to improve the images the development of the pipeline has been ongoing. The latest tests that use a refned recipe that still makes use ofkms and ddfacet for calibration and imaging, respectively, have shown that by removing the minimum uv-distance in the calibration and instead smoothing the amplitude solutions with a low-order polynomial function and ftting the phase solutions with a function proportional to ν−1 (which is, to frst order, the phase behaviour introduced by free electrons in the ionosphere) the artifcial haloes and holes can be effectively removed. Fur-thermore, these changes, together with other enhancements such as turning off the amplitude normalisation, improving the sky modelsusedfor calibrationbyincreasingthedepthofthe deconvolution, and refning the direction-independent calibration by making use of accurate models derived from the direction-dependent imaging, have allowed us to decrease thefailure rate of the pipeline, improve the dynamic range, and increase the number of sources detected. A demonstration of the improvements that are a result of these recent developments is shown in Fig. 15 and a refned version of the LoTSS processing pipeline will be fully described in a future publication. 

5.2. Exploiting the large fractional bandwidth of LoTSS 
With a fractional bandwidth of approximately 33%, LoTSS has thethirdlargest fractional bandwidthofanyverywide arearadio continuum survey produced to date. Only MSSS(Heald et al. 2015)and the GLEAM(Wayth et al. 2015;Hurley-Walker et al. 2017)surveyhave observed the skywith larger fractional band-widths, but both have signifcantly poorer angular resolutions and sensitivities (see Fig. 1).To demonstrate the scientifc poten<BR>tial of the spectral information that can be derived from LoTSS, for a test feld we divided a direction-dependent calibrated LoTSS data set into three parts, each with a width of 16 MHz, and generated a three-channel image with ddfacet. The inte-grated fux density measurements in each part of the bandwidth and the source association between the three images was done using PyBDSF. An example of some observed spectra, with comparison to other surveys, is shown in Fig. 16. In this demon-stration feld we were able to accurately derive (with 10% uncertainty or less) in-band spectra for compact, isolated (no other source within 10000 of the LoTSS position) sources with inte-grated fux densities ≥10mJy, where the uncertainty estimateof the derived spectral indexes was obtained by comparing with 
A1, page 18 of 21 
spectral indexes measured from ftting to VLSSr and NVSS (≥50mJy)orto TGSS-ADR1and NVSSforthefainter sources (≥10 mJy). 
Low-frequency spectral information is valuable for many science cases, such as for identifying low-luminosity peaked-spectrum sources(Callinghametal.2017)andinvestigatingthe energy distribution of electrons in the emitting region of radio sources (e.g. Bonavera et al. 2011). For example, the source shown in the left panel of Fig. 16 is a peaked-spectrum source with a radio luminosity <1025WHz−1, which is two orders of magnitude fainter than the median radio luminosity of previous peaked-spectrum samples (e.g. O’Dea 1998). Probing this population of low-luminosity peaked-spectrum sources could potentially identify sources powered by a short-lived outburst of the central activity that might not able to escape from the host galaxy(Czernyet al. 2009). Such sources could be the short-lived precursors needed to account for the overabundance of peaked-spectrum sources relative to the large-scale radiogalax-ies(Kunert-Bajraszewska&Labiano 2010). 
The right panel of Fig. 16 is the spectrum of a source that shows a signifcant deviation from a standard power law. If such a deviation is not taken into account, it leads to orders of magnitude uncertaintyinthe estimateofthe energy storedbythe lobes of the radio galaxy (Duffy&Blundell 2012; Harwood et al. 2017). 
Therefore, the spectral information that can be supplied by LoTSS will have diverse scientifc impact, providing internal spectral index information to fux densities below the levels possible by cross-comparison with existing sky survey data. As a consequence of processing constraints this spectral information is not included with this current release,but we plan to include it in future releases. 



6. Summary 
In this publication we have described the frst full quality LoTSS data release, which is available online14. We outlined how we managed the large LoTSS data rate and we introduced the com-pletely automated direction-dependent calibration and imag-ing pipeline that we used to produce 120–168 MHz continuum images. The high-resolution (600)images we present cover 424 square degrees in the region of the HETDEX Spring Field and contain 325 694 sources that are detected with a signifcance in excessoffve timesthe noise. This source densityisafactorof at least ten higher than any existing very wide area radio continuum survey. As described in companion papers(Williams et al. 2019;Duncan et al. 2019)the LoTSS-DR1 catalogue has been enhancedbyidentifyingthe optical counterpartsof radio sources and estimating their photometric redshifts. Finally, this data release is published together with ∼20 articles to highlight the scientifc potential of LoTSS. 
The LoTSS-DR1 images have a median sensitivity of 71µJybeam−1 with approximately 10% of the mapped area being dynamic-range limited. For point sources, the survey is 90% complete at a peak brightness of 0.45 mJy beam−1. We examined the fdelity of our images and found that the astrometric accuracyis approximately 0.200 in both RA and Dec. The fux density scale is in overall agreement with other radio surveys and the uncertainty on the integrated fux density measurements is ∼20%. 
There are manyopportunities to enrich the LoTSS data prod-ucts through, for example polarimetric measurements or full 
14 https://lofar-surveys.org <BR>

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 

√ 

± 1, 2, 4,... × 5σ levels where σ = 120 µJybeam−1. The LoTSS-DR1 image of this cluster suffers from artifcial haloes around the extended structures and a low dynamic range. The improved LoTSS image has a higher fdelity and is in good agreement with the independently processed image presentedin Wilberetal.(2019). 

exploitation of the longest baselines in the international LOFAR array.We briefy demonstratedafew such possibilities including improvements to the calibration and imaging and the measurement of the in-band spectral index. 
Acknowledgements. We thank the anonymous referee for his/her comments. This paper is based on data obtained with the International LOFARTelescope (ILT) under project codes LC2_038 and LC3_008. LOFAR(van Haarlem et al. 2013)is the LOw Frequency ARray designed and constructed by ASTRON. It has observing, data processing, and data storage facilities in several countries, which are owned by various parties (each with their own funding sources) and are collectively operated by the ILTfoundation under a joint scientifc policy. The ILT resources have benefted from the following recent major fund-ing sources: CNRS-INSU, Observatoire de Paris and Université d’Orléans, France; BMBF, MIWF-NRW, MPG, Germany; Department of Business, Enterprise and Innovation (DBEI), Ireland; NWO, The Netherlands; The Science and Technology Facilities Council (STFC), UK. Part of this work was car-ried out on the Dutch national e-infrastructure with the support of the SURF Cooperative through grant e-infra 160022&160152. The LOFAR software and dedicated reduction packages on https://github.com/apmechev/GRID_LRT <BR>were deployed on the e-infrastructure by the LOFAR e-infragroup, consisting of J. B. R. Oonk (ASTRON and Leiden Observatory), A. P. Mechev (Leiden Observatory) andT. Shimwell (ASTRON) with support fromN. Danezi (SURFsara) and C. Schrijvers (SURFsara). This work made use of the Univer-sity of Hertfordshire high-performance computing facility(http://uhhpc. <BR>herts.ac.uk)and the LOFAR-UK computing facility located at the Univer-sity of Hertfordshire and supported by STFC [ST/P000096/1]. The Data Center of the Nançay Radioastronomy Station acknowledges the support of the Conseil Régional of the Région Centre Val de Loire in France. We thank Forschungszentrum Jich for storage and computing. This research made use ofAstropy, a community-developed core Python package for astronomy (AstropyCollaboration 2013)hosted at http://www.astropy.org/, and of the astropy-based reproject package(http://reproject.readthedocs.io/ <BR>en/stable/).We are grateful to Thomas Robitaille for support in adapting the reproject packageto supporttheverylargeimagesusedinthispaper.HR,DNH, KJD, and RJvW acknowledge support from the ERC Advanced Investigator programme NewClusters 321271. AB acknowledges support from the ERC-Stg DRANOEL, no 714245.AOC gratefully acknowledges support from the European Research Council under grant ERC-2012-StG-307215 LODESTONE. RM and MB gratefully acknowledge support from the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013) /ERC Advanced Grant RADIOLIFE-320745. RJvW acknowledges support from the ERC VIDI research programme with project number 639.042.729, which is fnanced by the Netherlands Organisation for Scientifc Research (NWO). 
A1, page 19 of 21 

KLE acknowledges fnancial support from the Dutch Science Organization (NWO) through TOP grant 614.001.351. MJH and WLW acknowledge support from the STFC ST/M001008/1. PNB and JS acknowledge support from the STFC ST/M001229/1. JHC and BM acknowledge support from the STFC ST/M001326/1and ST/R00109X/1. CL, RK, RKC, and BW acknowledge support from STFC studentships. LA acknowledges support from the STFC through aScotDIST IntensiveData Science Scholarship. LKM acknowledges the support of the Oxford Hinzte Centre for Astrophysical Surveys, which is funded through generous support from the HintzeFamily CharitableFoundation. LKM is also partly funded by the John Fell Oxford University Press (OUP) Research Fund. GJW gratefully acknowledges support from theLeverhulmeTrust. VHM thanks the University of Hertfordshire for a research studentship ST/N504105/1. AD acknowledges fnancial support from German Federal Ministry for Education and Research (BMBF,Verbundforschung, projects D-LOFAR III and IV, grants 05A15STA and 05A17STA). S.P.O acknowledges fnancial support from the DeutscheForschungsgemeinschaft (DFG) under grant BR2026/23.AGacknowledges full support from the Polish National Science Centre (NCN) through the grant 2012/04/A/ST9/00083. MJ acknowledges support from the Polish National Science Centre under grant no. 2013/09/B/ST9/00599. Grudziacka 5, 87-100 Toru´
n, Poland. MKB acknowledges support from the Polish National Science Centre under grant no. 2017/26/E/ST9/00216. IP acknowledges support from the INAF SKA/CTAPRINproject “FORECaST”.EB,MA,andOSare supportedby the South African Research Chairs Initiative of the Department of Science and Technology and National ResearchFoundation. GGU acknowledges OCE Postocdoral fellowship from CSIRO. SM acknowledges funding through the Irish Research CouncilNewFoundations schemeandthe Irish Research Council Post-graduate Scholarship scheme. This publication has emanated from research supportedin partbya research grant from ScienceFoundation Ireland (SFI) under the Grant Number 15/RI/3204. 


References 
AstropyCollaboration (Robitaille,T.P., et al.) 2013, A&A, 558, A33 Atemkeng, M. T., Smirnov, O. M., Tasse, C., Foster, G., & Jonas, J. 2016, 
MNRAS, 462, 2542 Baars,J.W.M., Genzel,R.,Pauliny-Toth,I.I.K.,&Witzel,A.1977, A&A,61, 99 Becker,R.H., White,R.L.,&Helfand,D.J.1995, ApJ,450,559 Bennett, A. S. 1962, MmRAS, 68, 163 Best,P.N. 2004, MNRAS, 351,70 Best,P.N.,Arts,J.N., Rtgering,H.J.A.,etal.2003, MNRAS,346,627 Best,P.N.,Ker,L.M., Simpson,C., Rigby,E.E.,&Sabater,J. 2014, MNRAS, 445, 955 Bock,D.C.-J.,Large,M.I.,&Sadler,E.M.1999, AJ,117,1578 Bonavera, L., Massardi, M., Bonaldi, A., et al. 2011, MNRAS, 416, 559 Bonnassieux, E.,Tasse, C., Smirnov, O.,&Zarka,P. 2018, A&A, 615, A66 Botteon, A., Shimwell,T.W., Bonafede, A., et al. 2018, MNRAS, 478, 885 Bridle,A.H.,&Schwab,F.R. 1989, ASPC,6,247 Brienza, M., Godfrey, L., Morganti, R., et al. 2017, A&A, 606, A98 Brgen, M., Rafferty, D., Bonafede, A., et al. 2018, MNRAS, 477, 3461 Brunetti, G.,&Jones,T.W. 2014, IJMPD, 23, 1430007 Callingham,J.R., Gaensler,B.M.,&Ekers,R.D.2015, ApJ,809,168 Callingham,J.R.,Ekers,R.D.,&Gaensler,B.M.2017, ApJ,836,174 Calistro Rivera, G., Williams, W. L., Hardcastle, M. J., et al. 2017, MNRAS, 469, 3468 Cassano,R., Brunetti,G., Rtgering,H.J.A.,&Brgen,M.2010, A&A,509, A68 Ciliegi,P., Zamorani, G., Hasinger, G., et al. 2003, A&A, 398, 901 Condon,J.J. 1997, PASP, 109, 166 Condon,J.J., Cotton,W.D., Greisen,E.W.,etal.1998, AJ,115,1693 Condon,J.J., Cotton,W.D.,Fomalont,E.B.,etal.2012, ApJ,758,23 Cotton,W.D., Condon,J.J.,Perley,R.A.,etal.2004, SPIE,5489,180 Cutri, R. M., Wright, E. L., Conrow,T., et al. 2012, Explanatory Supplement to the WISE All-SkyData Release Products Cutri, R. M., Wright, E. L., Conrow,T., et al. 2013, Explanatory Supplement to the AllWISE Data Release Products 
Czerny, B., Siemiginowska, A., Janiuk, A., Nikiel-Wroczy ´
nski, B.,&Stawarz, Ł. 2009, ApJ, 698, 840 de Gasperin, F., Intema, H. T., Shimwell, T. W., et al. 2017, Sci. Adv., 3, e1701634 de Gasperin,F.,Dijkema,T.J., Drabent,A.,etal.2019, A&A,622,A5 (LOFAR SI) de Ruiter,H.R.,Willis,A.G.,&Arp,H.C.1977, A&AS,28,211 Dalton,G.,Trager,S.,Abrams,D.C.,etal.2012, Proc.SPIE,8446, 84460P Dalton,G.,Trager,S.,Abrams,D.C.,etal.2014, Proc.SPIE,9147, 91470L Dewdney,P.E., Hall,P.J., Schilizzi,R.T.,&Lazio,T.J.L.W. 2009, IEEEP,97, 1482 
A1, page 20 of 21 
Duffy,P.,&Blundell,K.M. 2012,MNRAS, 421, 108 
Duncan, K., Sabater, J., Rottgering, H., et al. 2019, A&A, 622, A3 (LOFAR SI) 
Emig,K.L.,Salas,P.,de Gasperin,F.,etal.2019, A&A,622,A7 (LOFARSI) 
Flewelling, H. A., Magnier, E. A., Chambers, K. C., et al. 2016, ArXiv e-prints [arXiv:1612.05243] 
Fleuren, S., Sutherland,W., Dunne, L., et al. 2012, MNRAS, 423, 2407 
Foreman-Mackey,D.,Hogg,D.W.,Lang,D.,&Goodman,J.2013,PASP,125, 306 
Franzen,T. M. O., Jackson, C. A.,Offringa, A. R., et al. 2016, MNRAS, 459, 3314 
Ghosh, A., Prasad, J., Bharadwaj, S., Ali, S. S., & Chengalur, J. N. 2012, MNRAS, 426, 3295 
Hamaker,J.P. 2000, A&AS, 143, 515 
Hamaker,J.P.,Bregman,J.D.,&Sault,R.J.1996, A&AS,117,137 
Hardcastle,M.J., Gkan,G.,vanWeeren,R.J.,etal.2016, MNRAS,462,1910 
Harwood, J. J., Hardcastle, M. J., Morganti, R., et al. 2017, MNRAS, 469, 639 
Heald,G.H., Pizzo,R.F.,OrrE.,etal.2015, A&A,582,A123 
Hill,G.J.,&Lilly,S.J. 1991, ApJ,367,1 
Hill,G.J., Gebhardt,K.,Komatsu,E.,etal.2008, ASPC,399,115 
Hoang,D.N., Shimwell,T.W.,Stroe,A.,etal.2017, MNRAS,471,1107 
Hurley-Walker,N., Callingham,J.R., Hancock,P.J.,etal.2017, MNRAS,464, 1146 
Intema,H.T.,vanderTol,S., Cotton,W.D.,etal.2009, A&A,501,1185 
Intema,H.T.,vanWeeren,R.J., Rtgering,H.J.A.,&Lal,D.V.2011, A&A, 535, A38 
Intema, H.T., Jagannathan,P., Mooley, K.P.,&Frail, D. A. 2017, A&A, 598, A78 
Jackson, N.,Tagore, A., Deller, A., et al. 2016, A&A, 595, A86 
Jarrett,T. H., Chester,T., Cutri, R., et al. 2000, AJ, 119, 2498 
Kazemi, S.,Yatawatta, S., Zaroubi, S., et al. 2011, MNRAS, 414, 1656 
Kunert-Bajraszewska,M.,&Labiano,A.2010,MNRAS,408,2279 
Lane,W.M., Cotton,W.D.,vanVelzen,S.,etal.2014, MNRAS,440,327 
Mahony, E. K., Morganti, R., Prandoni, I., et al. 2016, MNRAS, 463, 2997 
Mauch,T.,Murphy,T., Buttery,H.J.,etal.2003, MNRAS,342,1117 
Magnier, E. A., Schlafy, E. F., Finkbeiner, D. P., et al. 2016, ArXiv e-prints [arXiv:1612.05242] 
McAlpine,K., Smith,D.J.B., Jarvis,M.J., Bonfeld,D.G.,&Fleuren,S. 2012, MNRAS, 423, 132 
Mechev, A., Oonk, J. B. R., Danezi, A., et al. 2017, Proc. Int. Symp. on Grids and Clouds (ISGC) 2017,2 
Mohan, N., & Rafferty, D. 2015, Astrophysics Source Code Library [record ascl:1502.007] 
Morabito,L.K., Deller,A.T., Rtgering,H.,etal.2017, MNRAS,469,1883 
Noordam,J.E.,&Smirnov,O.M.2010, A&A,524,A61 
Norris,R.P., Hopkins,A.M., Afonso,J.,etal.2011, PASA,28,215 
O’Dea,C.P. 1998, PASP, 110, 493 
Offringa, A. R. 2016, A&A, 595, A99 
Offringa,A.R.,vande Gronde,J.J.,&Roerdink,J.B.T.M. 2012, A&A,539, A95 
Oonk,J.B.R.,vanWeeren,R.J., Salas,P.,etal.2017, MNRAS,465,1066 
Pandey,V.N.,van Zwieten,J.E.,de Bruyn,A.G.,&Nijboer,R. 2009,ASPC, 407, 384 
Perley,R.A.,&Butler,B.J.2017, ApJS,230,7 
Prestage,R.M.,&Peacock,J.A.1988, MNRAS,230,131 
Raccanelli, A., Zhao, G.-B., Bacon, D. J., et al. 2012, MNRAS, 424, 801 
Rengelink,R.B.,Tang,Y.,de Bruyn,A.G.,etal. 1997, A&AS,124,259 
Richter, G. A. 1975, Astron Nachr., 296, 65 
Rtgering, H., Afonso, J., Barthel,P., et al. 2011, JApA, 32, 557 
Saff,E.B.,&Kuijlaars,A.B.J. 1997, Math. Intell.,19,5 
Salas,P.,Oonk,J.B.R.,vanWeeren,R.J.,etal.2018, MNRAS,475,2496 
Salvini,S.,&Wijnholds,S.J. 2014, A&A,571,A97 
Savini,F., Bonafede, A., Brgen, M., et al. 2018, MNRAS, 474, 5023 
Saxena,A., Rtgering,H.J.A.,&Rigby,E.E.2017, MNRAS,469,4083 
Scaife,A.M.M.,&Heald,G.H.2012, MNRAS,423,L30 
Schlegel,D.J., Finkbeiner,D.P.,&Davis,M.1998, ApJ,500,525 
Shimwell,T.W., Rtgering,H.J.A.,Best,P.N.,etal.2017, A&A,598,A104 
Smirnov, O. M. 2011, A&A, 527, A106 
Smirnov,O.M.,&Tasse,C.2015, MNRAS,449,2668 
Smith, D. J. B., Dunne, L., Maddox, S. J., et al. 2011, MNRAS, 416, 2857 
Smith,D.J.B.,Best,P.N., Duncan,K.J.,etal.2016, SF2A-2016: Proceedings of the Annual meeting of the French Society of Astronomy and Astrophysics, 271 
Stewart,A.J., Fender,R.P., Broderick,J.W.,etal.2016, MNRAS,456,2321 
Sutherland,W.,&Saunders,W. 1992, MNRAS, 259, 413 
Tasse, C. 2014, ArXiv e-prints [arXiv:1410.8706] 
Tasse, C., Hugo, B., Mirmont, M., et al. 2018,A&A, 611, A87 
van Diepen, G., & Dijkema, T. J. 2018, Astrophysics Source Code Library [record ascl:1804.003] 

T.W. Shimwelletal.:TheLOFARTwo-metreSkySurvey–DR1 
Van Eck, C. L., Haverkorn, M., Alves, M. I. R., et al. 2018,A&A, 613, A58 van Haarlem,M.P.,Wise,M.W.,Gunst,A.W.,etal.2013, A&A,556,A2 vanWeeren,R.J.,Williams,W.L., Hardcastle,M.J.,etal.2016, ApJS,223,2 Wayth,R.B., Lenc,E., Bell,M.E.,etal. 2015,PASA,32, e025 White,R.L., Becker,R.H., Helfand,D.J.,&Gregg,M.D. 1997, ApJ,475,479 Wilber, A., Brgen, M., Bonafede, A., et al. 2018,MNRAS, 473, 3536 Wilber,A., Brgen, M., Bonafede, A., et al. 2019,A&A, 622, A25 (LOFAR SI) Williams,W.L., Intema,H.T.,&Rtgering,H.J.A.2013,A&A,549,A55 Williams,W.L.,vanWeeren,R.J., Rtgering,H.J.A.,etal.2016,MNRAS, 
460, 2385 Williams, W. L., Calistro Rivera, G., Best, P. N., et al. 2018, MNRAS, 475, 3429 Williams, W. L., Hardcastle, H. J., Best, P. N., et al. 2019, A&A, 622, A2 
(LOFAR SI) Wilman, R. J., Miller, L., Jarvis, M. J., et al. 2008,MNRAS, 388, 1335 Yatawatta, S. 2015,MNRAS, 449, 4506 
1 ASTRON, The Netherlands Institute for Radio Astronomy, Postbus 2, 7990 AA Dwingeloo, The Netherlands e-mail: shimwell@astron.nl 
2 Leiden Observatory, Leiden University, PO Box 9513, 2300 RA Leiden, The Netherlands 3 GEPI, Observatoire deParis, Université PSL, CNRS,5Place Jules Janssen, 92190 Meudon, France 4 Department of Physics&Electronics, Rhodes University, PO Box 94, Grahamstown 6140, South Africa 
5 Centre for Astrophysics Research, School of Physics, Astronomy and Mathematics, University of Hertfordshire, College Lane, Hat-feld AL10 9AB, UK 
6 SUPA, Institute for Astronomy, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ, UK 7 University of Hamburg, Hamburger Sternwarte, Gojenbergsweg 112, 21029 Hamburg, Germany 8 SKA South Africa, 3rd Floor, ThePark,Park Road, Pinelands 7405, South Africa 9 Ampyx Power B.V. Lulofsstraat 55–Unit 13, 2521 AL The Hague, The Netherlands 10 INAF–Istituto di Radioastronomia, Via Gobetti 101, 40129 Bologna, Italy 11 School of Physics and Astronomy, The Open University, Walton Hall, MiltonKeynes MK7 6AA,UK 12 RAL Space, The Rutherford Appleton Laboratory, Chilton, Didcot, OxfordshireOX11 0NL,UK 
13 LESIA, Observatoire deParis, PSL, CNRS, Sorbonne Universités, UPMC Univ.Paris 06, Univ.Paris Diderot, SorbonneParis Cité,5 place Jules Janssen, 92195 Meudon, France 
14 Astronomical Observatory,Jagiellonian University, ul. Orla 171, 30244 Krak, Poland 
15 Department of Space, Earth and Environment, Chalmers University ofTechnology, Onsala Space Observatory, 43992 Onsala, Sweden 
16 SURFsara, Science Park 140, 1098 XG Amsterdam, The Netherlands 
17 Department of Astrophysics/IMAPP, Radboud University Nijmegen, PO Box 9010, 6500 GL Nijmegen, The Netherlands 
18 CSIROAstronomy and Space Science,PO Box 1130, Bentley,WA 6102, Australia 
19 Jodrell Bank Centre for Astrophysics, School of Physics and Astronomy, University of Manchester, Manchester M13 9PL, UK 
20 Astrophysics, University of Oxford, Denys Wilkinson Building, Keble Road, OxfordOX1 3RH,UK 
21 Physics and Astronomy Department, University of the Western Cape, Bellville 7535, South Africa 
22 Oxford e-Research Centre, University of Oxford, 7 Keble Road, OxfordOX1 3QG,UK 
23 Wolfson College, University of Oxford, Linton Road, OxfordOX2 6UD, UK 
24 Kapteyn Astronomical Institute, University of Groningen, Postbus 800, 9700AVGroningen, The Netherlands 
25 Anton Pannekoek Institute for Astronomy, University of Amsterdam, Postbus 94249, 1090 GE Amsterdam, The Netherlands 
26 Institute of Cosmology and Gravitation, University of Portsmouth, Dennis Sciama Building, Burnaby Road, Portsmouth PO1 5AR, UK 
27 Joint Institute for VLBI ERIC, PO Box 2, 7990 AA, The Netherlands 
28 Dipartimento di Fisica e Astronomia, Università di Bologna, Via 
P. Gobetti 93/2, 40129 Bologna, Italy 29 Center for Astronomy andParticle Theory, University of Nottingham, NG7 2RD Nottingham, UK 30 Astronomical Institute, Ruhr-University Bochum, 44780 Bochum, Germany 31 Thinger Landessternwarte, Sternwarte5, 07778Tautenburg, Germany 32 Mbarara University of Science & Technology, PO Box 1410 Mbarara, Uganda 33 Max Planck Institute for Astrophysics, Karl-Schwarzschild-Str. 1, 85741 Garching, Germany 
34 Laboratoire Lagrange, Université Ce d’Azur, Observatoire de la Ce d’Azur,CNRS, Blvd de l’Observatoire, CS 34229, 06304 Nice Cedex 4, France 
35 Instituto de Astrofsíca de Canarias, 38200 La Laguna, Tenerife, Canary Islands, Spain 36 Departamento de Astrofísica, Universidad de La Laguna (ULL), 38206La Laguna,Tenerife, Spain 37 Toru´
n Centre for Astronomy, Faculty of Physics, Astronomy and Informatics, NCU, Grudziacka5, 87-100Toru ´
n, Poland 38 CSIRO Astronomy and Space Science, PO Box 76, Epping, NSW 1710, Australia 39 Max-Planck Institute f Extraterrestrische Physik, Giessenbachstr. 1, 85741 Garching, Germany 40 Excellence Cluster Universe, Boltzmann Strasse2, 85748 Garching, Germany 41 School of Physics, University College Dublin, Belfeld, Dublin 4, Ireland 42 Max-Planck-Institut f Radioastronomie, Auf dem Hel 69, 53121 Bonn, Germany 43 Fakultät f Physik, Universität Bielefeld, Postfach 100131, 33501 Bielefeld, Germany 
A1, page 21 of 21