BASS XXII: The BASS DR2 AGN Catalog and Data

We present the AGN catalog and optical spectroscopy for the second data release of the Swift BAT AGN Spectroscopic Survey (BASS DR2). With this DR2 release we provide 1425 optical spectra, of which 1181 are released for the first time, for the 858 hard X-ray selected AGN in the Swift BAT 70-month sample. The majority of the spectra (813/1425, 57%) are newly obtained from VLT/Xshooter or Palomar/Doublespec. Many of the spectra have both higher resolution (R>2500, N~450) and/or very wide wavelength coverage (3200-10000 A, N~600) that are important for a variety of AGN and host galaxy studies. We include newly revised AGN counterparts for the full sample and review important issues for population studies, with 44 AGN redshifts determined for the first time and 780 black hole mass and accretion rate estimates. This release is spectroscopically complete for all AGN (100%, 858/858) with 99.8% having redshift measurements (857/858) and 96% completion in black hole mass estimates of unbeamed AGN (outside the Galactic plane). This AGN sample represents a unique census of the brightest hard X-ray selected AGN in the sky, spanning many orders of magnitude in Eddington ratio (Ledd=10^-5-100), black hole mass (MBH=10^5-10^10 Msun), and AGN bolometric luminosity (Lbol=10^40-10^47 ergs/s).

Many different optical spectroscopic surveys have been done of AGN in X-ray survey fields. Early work focused on bright sources from the Einstein (e.g. Stocke et al. 1991) and ROSAT surveys (e.g., Voges et al. 1999) typically focused on obtaining basic redshift and counterpart information for tens to hundreds of sources. This was later extended to deeper and fainter sources in Chandra fields (e.g., Green et al. 2004;Szokoly et al. 2004;Eckart et al. 2006;Treister et al. 2009;Silverman et al. 2010) or XMM-Newton surveys (e.g., Menzel et al. 2016). Full spectroscopic completeness was difficult in the deepest surveys owing to the optical faintness of distant targets (Brandt & Hasinger 2005). More recently, these efforts have focused on direct estimates of supermassive black hole (SMBH) masses for X-ray-selected AGN with broad emission lines using virial relations (e.g., Shen et al. 2011).
Larger optical spectroscopic samples of X-ray-selected active galactic nuclei (AGN) now exist within the Sloan Digital Sky Survey (SDSS) footprint, though typically focused on unobscured AGN. A total of 7005 ROSAT sources were cross-matched with the SDSS spectroscopic footprint (Anderson et al. 2007). Due to the soft X-ray sensitivity of ROSAT (e.g. 0.1-2.4 keV), the majority, 89% (6224/7005), were broad-line AGN. Similarly, a study by Mahony et al. (2010) in the southern hemisphere using spectra from the 6-degree-Field Galaxy Survey (Jones et al. 2004(Jones et al. , 2009), a near-infrared-selected redshift survey covering a large area in the Southern hemisphere (∼17000 •2 ), had spectroscopic identifications for 1715 ROSAT sources in the Southern hemisphere. More recently SPIDERS (SPectroscopic IDentification of eROSITA Sources, Dwelly et al. 2017;Comparat et al. 2020) is currently collecting 40,000 spectra of X-ray-selected AGN from ROSAT as well as the XMM slew survey (0.5-12 keV), which is more sensitive to obscured AGN.
Hard X-ray emission (>10 keV) from the corona of the AGN can probe the innermost parts of the central engine of AGN with the advantage over UV/optical/soft X-rays that it can even find AGN in highly obscured (e.g. H > 10 22 − 10 25 cm −2 ) systems. The Swift-BAT survey (Barthelmy et al. 2005), with its all-sky coverage that is insensitive to obscuration up to Compton-thick levels (Koss et al. 2016a), provides the largest, most complete sample of bright, local ( < 0.1), powerful AGN. The spectroscopic coverage in the SDSS for BAT AGN, is, however, only ∼15% . Therefore, a complete optical spectroscopic sample of BAT AGN, including the unobscured to the highly obscured AGN which are largely absent from ROSAT surveys, provides a unique way to fully understand BH growth and its relation to the host galaxy. The BAT AGN survey also provides a bright complement that is more sensitive to obscured AGN compared to the currently ongoing eROSITA satellite mission and its all-sky survey (Predehl et al. 2021).
The goal of the BAT AGN Spectroscopic Survey (BASS) is to provide the largest available spectroscopic sample of Swift BAT ultrahard X-ray (14-195 keV) detected AGN. In the BASS DR1 , mostly archival optical telescope data was used for 641 BAT AGN from the 70-month BAT catalog (Baumgartner et al. 2013) and 102 AGN comosing the NIR DR1 (Lamperti et al. 2017). These data were then used in a variety of scientific studies, such as between X-ray emission and high-ionization optical lines (Berney et al. 2015), ionized gas outflows (Rojas et al. 2020), and radio emission (Baek et al. 2019;Smith et al. 2020). Several works identified the importance of the Eddington ratio in various scaling relations (e.g., Oh et al. 2017;Ricci et al. 2017aRicci et al. , 2018 and links to host galaxy properties such as molecular gas (Koss et al. 2021).
In the BASS data release 2 (DR2), we have identified all AGN among the 1210 sources in the BAT 70-month survey in order to obtain a 100% spectroscopically complete sample of high-quality optical spectra and BH mass estimates for a large fraction of AGN across the entire sky. High signal-to-noise ratio (S/N) and spectral resolution optical spectroscopy with measurements of continuum, emission and absorption lines, over the full optical range (3200-10000Å) provides a large number of important diagnostics. To name a few, these include star formation rates, stellar masses, stellar population ages, dust reddening, metallicities, AGN-driven outflows, and dynamical properties of the galaxy such as velocity dispersions of stellar populations (e.g., Tremonti et al. 2004;Vazdekis et al. 2012;Yates et al. 2012;Shimizu et al. 2018;Rojas et al. 2020). Repeat optical spectroscopy can also probe the time-variable nature of these emission components such as in changing optical type AGN (e.g., Collin-Souffrin et al. 1973;Shappee et al. 2014).
In this first catalog paper of the DR2 release series, we provide an updated list of counterparts and the 1449 optical spectra of 858 AGN among the 1210 sources in the BAT 70-month survey (Baumgartner et al. 2013). We provide an explanation of all the optical spectra obtained, their reductions, and general derived measurements. An overview of the DR2 release and scientific results and a comparison with other surveys are provided in Koss et al. (2022a). Further catalogs of derived measurements will be provided in subsequent papers such as broad-line measurements (Mejía-Restrepo et al. 2022), narrow emission-line measurements from the best available optical spectra (Oh et al. 2022), velocity dispersion measurements from stellar absorption lines (Koss et al. 2022b), and the NIR spectroscopic measurements (Brok et al. 2022;Ricci et al. 2022). We will also provide scientific investigations using DR2 data such as the BH mass and Eddington ratio distribution function for obscured and unobscured AGN Ananna et al. (2022), the ability of the MIR to recover obscured AGN (Pfeifle et al. 2022), and the -★ relation of type 1 unobscured AGNs (Caglar in prep.). Throughout this work, we adopt Ω M =0.3, Ω Λ =0.7, and 0 =70 km s −1 Mpc −1 . To determine extinction due to Milky Way foreground dust, we use the maps of Schlegel et al. (1998) and the extinction law derived by Cardelli et al. (1989) with =3.1.

REVISED AGN CATALOG
The initial 70-month catalog (Baumgartner et al. 2013) was composed of 1210 sources, including 822 classified as AGN or associated with a galaxy and likely an AGN, 287 Galactic sources (e.g., high/low-mass X-ray binary, cataclysmic variable, pulsar), 19 clusters, and 82 unknown sources. The counterpart positions and AGN classifications were updated based on WISE and X-ray data for 838 AGN in the BASS DR1 Ricci et al. (see Appendix A, 2017b), which included three dual AGN systems. However, even after the DR1, 44 unknown BAT sources, typically near the Galactic plane (| |<10 • ), had not been associated with counterparts. Here we discuss the BASS DR2 AGN counterparts after an extensive examination of the remaining unknown 70-month catalog sources and important issues for surveys of these AGN.
Given the large (FWHM=19.5 ) Swift BAT point spread function (PSF), it is important to consider cases such as chance alignment of multiple AGN, AGN clustering, and dual AGN for population studies. Due to the scarcity of Swift BAT sources in the sky, the likelihood of chance alignment of any two unassociated sources within the BAT beam is small. Specifically, there are ∼850 sources at | |>5 • , (91% of the sky), so the likelihood of chance alignment of one unassociated source within the BAT beam is very small, ∼0.2%. However, there are some unique cases involving galaxy mergers (e.g., Koss et al. 2018) and galaxy clustering where multiple AGN systems occur. Additionally, at lower fluxes, where AGN are more numerous, there may be some cases of flux boosting where two sources below the sky sensitivity both contribute to be above the detection sensitivity.
Throughout this work, we refer to AGN as Sy1 (with optical broad H ), Sy1.9 (narrow H and broad H ), and Sy 2 (with narrow H and narrow H , including small numbers of LINERs and AGN in H 2 -dominated regions). This nomenclature is used for the sake of simplicity and consistency with previous work, despite the fact that many of our BASS DR2 AGNs may not be considered as Seyfert galaxies, given their high (X-ray) luminosities. Additionally, a small number of Sy2 sources would be better classified as HII or LINERs or composites based on their position in the BPT diagram  or even elusive AGN (Smith et al. 2014), due to a lack of prominent lines. These classifications are discussed in detail in the BASS DR2 paper on narrow emission-line measurements (Oh et al. 2022). Finally, some Sy2 sources with narrow optical lines are known to have polarized broad-lines or NIR broad-lines (e.g., Lamperti et al. 2017).

Newly Identified AGN and Galactic Sources
There were still some unidentified sources listed in the 70-month catalog, typically within the Galactic plane, that were not part of the BASS DR1 sample. We examined them all to ensure that our sample provided a complete census of all AGN detected with Swift BAT across the sky. Further optical spectroscopy found many of them to be stellar in nature.
This particular AGN resides in a region of very high Galactic extinction ( =9.5), and the redshift is based on an He 10830 line measurement from the literature (Fortin et al. 2018).
We uncovered an additional 22 AGN among the remaining unidentified sources (Table 1). These sources were the brightest 2-10 keV sources within the 5 Swift BAT position error circle, and optical spectroscopy confirmed their AGN nature. The total sample then increases from 838 to 858 AGN. Further details on the X-ray modeling of these newly detected AGN will be provided in C. Ricci, et al., (2022 in preparation).
A summary of the 17 sources that are newly classified as Galactic is provided in Table 2. For 10/17 sources, we determine the stellar nature based on optical spectroscopy, with the remaining sources being classified in recent publications.
There are two cases that were thought to be AGN based on only an X-ray and WISE detection in the BASS DR1, but optical follow-up ( Figure 1) found the source to be Galactic. SWIFT J0428.2-6704A, was found to be an eclipsing X-ray binary (Kennedy et al. 2020) which follow-up optical spectra confirmed. We observed SWIFT J1535.8-5749 (aka. IGR J15360-5750), but found the source consistent with a very red star based on the Ca triplet spectral region (8450-8700 Å) and CO-band heads (2.29-2.51 m).
In the other 15 cases, the 70-month X-ray classification was uncertain (Baumgartner et al. 2013), but the sources were found to be Galactic. For instance, SWIFT J0630.9+1129, showed a very red star with H emission at =0. SWIFT J1213.2-6020 (aka IGR J12134-6015) was found to be located within 0. 3 of 2MASS J12132397-6015169 based on Chandra (Karasev et al. 2012 (2): corresponding galaxy counterpart name in NED or SIMBAD based on the WISE positions. Column (3)(4): right ascension and decl. of the IR counterpart of the BAT AGN, in decimal degrees, based on WISE positions. Column (5): telescope and instrument used to observe the Galactic source. Column (6): Galactic latitude, in decimal degrees. Column (7): Galactic classification based on reference or optical spectral features. Star denotes stellar features in optical spectrum at redshift zero, but uncertain classification. Column (8): initial classification in 70-month catalog (Baumgartner et al. 2013). Column (9): optical spectral feature used in redshift measurement. (10) References. The WISE and X-ray counterpart has a GAIA source within 1 with 1.4 ± 0.055 mas parallax, and a spectrum consistent with a star.

Excluded Unknown Sources in the Galactic Plane
The DR2 sample defined here is fully complete for all BAT 70-month BAT AGN and unknown sources above (| |>3 • ) or below =5 mag. Beyond these limits an additional seven sources were excluded (Table 3) in the analysis owing to their very high extinction values (5-43 mag). Many of these sources have been surveyed extensively as part of INTEGRAL surveys (e.g., Tomsick et al. 2008). These sources are very close to the Galactic plane, with many foreground stars and very high extinction levels that make any accurate optical targeting impractical, though they may host AGN in some cases. A follow-up Chandra observation SWIFT J1848.5-0046 (aka IGR J18485-0047) showed a source coincident with strong radio emission consistent with an AGN (Tomsick et al. 2009). SWIFT J1403.6-6146 (aka IGR J14044-6146), was observed with Chandra, but no detection was found after an earlier Swift XRT detection, suggesting variability that may be associated with an AGN (Bodaghee et al. 2012). In one case (SWIFT J2037.2+4151), optical spectral follow-up showed a very red spectrum, but there was no obvious counterpart. SWIFT J2056.8+4939 (aka 4C 49.35), has both radio emission and a likely Fermi detection consistent with a blazar-like beamed AGN. None of the remaining sources have been observed with Chandra, XMM-Newton, or NuSTAR. Further source localization and characterization with Chandra, deeper X-ray observations with XMM-Newton or NuSTAR, radio, and finally NIR spectroscopy will likely be required to identify these sources and their possible AGN nature.
There was also one remaining 70-month transient source outside the Galactic plane, SWIFT J0325.6-0907, which we exclude because it is likely a transient. The source shows a declining significance in the 105-month light curve (Oh et al. 2018), with a drop to zero XMM-Newton flux after 2010 January. A 10 ks Swift XRT exposure finds no sources above obs 2−10 keV =10 −13 erg cm −2 s −1 in flux that could be a likely counterpart. No observations of this source have been performed with Chandra, XMM-Newton, or NuSTAR.   (3)(4): primary and secondary galaxy where the primary galaxy is the one with brighter soft X-ray emission ( in 2−10 keV ). Column (5): total hard X-ray emission detected from BAT for both sources. Column (6): The ratio of the soft X-ray emission of the secondary to the total emission based on intrinsic soft X-ray emission ( in 2−10 keV ). Column (7): the predicted obs 14−195 keV for the secondary given the soft X-ray emission. Column (8): whether or not the secondary galaxy optical spectra is included in the DR2 release. Column (9): references.

Dual AGN
In the case of bright dual AGN, both may contribute significantly to the BAT flux (e.g. Koss et al. 2016b). Dual AGN are typically close together (∼< 30 kpc), show small offsets in redshift (<500 km s −1 ), and often show signs of interaction in imaging. For these sources, we report them as if they would have been individually detected based on the sky sensitivity (Baumgartner et al. 2013), and report the fluxes based on the soft X-ray emission (Ricci et al. 2017b). Additional spectra of merging companions (e.g., Koss et al. 2010Koss et al. , 2012 will be presented in a separate release (Koss, M., et al. 2022, in preparation).
A summary of bright dual AGN in the sample can be found in Table 4. This only includes dual AGN that are both X-ray detected, not lower-luminosity AGN detected only in optical spectroscopy. The galaxy group Arp 318 (NGC 835/NGC 833) and the close dual AGN Mrk 463 are unique in that the sources are individually below the median BAT sky sensitivity, but the combined flux is above the median sky sensitivity. To be complete, we set the limit at obs 14−195 keV >5×10 −12 erg cm −2 s −1 which corresponds to the faintest detected sources in the sky by BAT rather than the median sensitivity obs 14−195 keV >1.03×10 −11 erg cm −2 s −1 . We assume obs 14−195 keV ∝ in 2−10 keV for this calculation, and Γ=1.8, which is consistent at H <10 23.5 cm −2 (Koss et al. 2016a;Ricci et al. 2018). This leads to three sets of bright dual BAT-detected AGN in the DR2. See Ricci et al. (2017b) and Koss et al. (2012) for a further discussion of these sources.
K . N -Column descriptions are the same as in Table 1, unless otherwise noted.
Offset between two BAT AGN in measured redshift ( km s −1 ) and WISE position ( and kpc) at the redshift of the first AGN listed.

Weakly Associated AGN
Previous studies have also found that galaxy clustering is higher around BAT AGN (e.g., Cappelluti et al. 2010;Koss et al. 2010) and this extends to secondary nearby AGN on scales of 70-1000 kpc. These AGN will be separated by significant distances on the sky of several arcminutes, but show small offsets in redshift (<500 km s −1 ) and can often be found with several other galaxies clustering at the same redshift. These two AGN will reside in the BAT beam and may both be bright enough to be individually detected. In other cases, the clustered AGN would not be individually detected, but they are detected by BAT because of flux boosting. See subsection 2.4 for further details on these systems.
Here we provide a list of multiple soft X-ray AGN that are detected within the BAT beam but are weakly associated (∼70-700 kpc; Table 5). This includes seven pairs of AGN. In three cases (SWIFT J0202.4+6824B, SWIFT J0359.0-3015A, SWIFT J1051.2-1704B), one of the AGN is below the median BAT sky sensitivity and should be excluded from population studies.

Multiple Unassociated Faint X-Ray Counterparts
Some sources in the initial 70-month catalog were reported with multiple likely counterparts within the error circle of the BAT beam position (Baumgartner et al. 2013). The details of many of these sources and their fluxes were described further in Ricci et al. (2017b) as part of the BASS DR1. Sources below the median sky sensitivity ( obs 14−195 keV <10.3×10 −12 erg cm −2 s −1 ) should be excluded from population studies based on the all-sky sensitivity because they were likely detected only because of X-ray follow-up (Table 6). Among these, 10 pairs of sources 76% (20/26) are cases of flux boosting that individually are below the median sensitivity. Another six cases are fainter counterparts to a brighter BAT source that is above the median detection sensitivity. There is only one case of a pair of AGN with both AGN above the detection sensitivity (SWIFT J1652.0-5915) which includes NGC 6221 ( =0.0041) and the background galaxy ESO 138-1 ( =0.0091).

Beamed and Lensed AGN
The Swift BAT survey includes beamed, lensed, and unbeamed AGN and it is important to separate them in most scientific analyses. The original DR1 included 105 beamed sources (see, e.g., Table 1 of Ricci et al. 2017b), based on the Roma Blazar Catalog (BZCAT) catalog (Massaro et al. 2009) and DR1 optical spectra . Since this release, a further study by Paliya et al. (2019) used recent Fermi LAT data and SED fitting to identify all the blazars in the 105-month catalog, which includes all DR2 70-month AGN. N -Column descriptions are the same as in Table 1.
For continuity, we provide all of the new beamed AGN, or changes of beamed AGN from the DR1 classifications to unbeamed in DR2 or vice versa, in Table 16 in subsection A.1. In most cases, recent radio observations or SED fitting  revealed beamed AGN to be unbeamed, or conversely, a recent Fermi detection revealed a previously categorized unbeamed AGN to be beamed. In DR2, 13/105 (12%) changed from beamed to unbeamed AGN classification. Conversely, eight unbeamed AGN in the DR1 list were found to be beamed. Finally, five beamed AGN were included because they are part of the expanded list of AGN, that were not in the original list of 838 DR1 AGN.
For DR2, we reviewed all the beamed classifications and found three AGN that were classified as beamed AGN in ), but we now classify them as unbeamed AGN based on further analysis. SWIFT J0312.9+4121 (aka QSO B0309+411), while detected in Fermi, is a broad-line radio galaxy, with double-lobe morphology within a compact structure . For SWIFT J0519.5-3140 (aka PKS 0521-365), Angioni et al. (2019) found with VLBI imaging that the jet of PKS 0521-36 is not highly beamed, with viewing angles larger than 10 • . Finally, SWIFT J1742.1-6054 (aka PKS1737-60), was classified as an FRII radio galaxy ). Further polarimetric observations or detections of compact cores and superluminal motions using high-resolution radio imaging would be needed to further classify these sources.

SURVEY, OBSERVATIONS, AND DATA REDUCTION
Here we provide an overview of DR2 survey and observations. The DR2 targeting criteria goals were to provide the largest possible sample of BH mass measurements from either broad Balmer lines or stellar velocity dispersion measurements, as well as the broadest possible spectral coverage (e.g. 3000-10000 Å) for emission-line measurements for the entire catalog of 858 AGN. In practice, outside of echelle instruments, this required multiple spectra with broad wavelength coverage with lower-resolution K .  (2): instrument. Column (3): total number of DR2 spectra observed with this setup. Column (4): grating listing the lines/mm, if applicable. For instruments with a blue and red side two numbers are listed associated with each grating. Columns (5-9): wavelength range (Range), pixel dispersion, resolution (Res.), and resolving power ( ). These represent typical values for this setup and may have small differences within individual spectra depending on observing conditions. These quantities may also be wavelength dependent in some cases, and so the values are given at 5000 Å and 8500 Å depending on the spectral range. Two values are listed when the instrument had both a blue and red arm with different settings. Column (10): method used to determine the instrumental resolution, either with sky lines, telluric features with molecfit, with arc lines, fitting stellar templates to stars, or based on resolutions provided within the survey (e.g. SDSS). The measurements are for 5000 or 8500 Å or both depending on the spectral range. See subsection 3.4 for further details.
or higher-resolution gratings with narrower wavelength coverage. Repeat observations were done if the S/N of the broad Balmer lines (H or H ) was too low for measurements (Mejía-Restrepo et al. 2022) or the low S/N and/or spectral resolution of the stellar absorption features resulted in a failed measurement of velocity dispersion. We did not reobserve targets with acceptable spectra and measurements from the SDSS. The requirement for high S/N, high spectral resolution, and broad wavelength coverage, combined with queue mode observing approved months in advance, sometimes resulted in duplicate (or more) observations of the same source. Example spectra for different instrumental setups are provided in Appendix B. All spectra associated with DR2 for each AGN will be provided at the BASS website.
A summary of all the observational setups used is shown in Table 7. We did not specifically exclude sources with high Galactic extinction in spectroscopic targeting if there was an obvious optical counterpart to the WISE counterpart of the soft X-ray emission. This resulted in observations of 12 AGN with very high extinctions (between =5 and 10 mag) close to the Galactic plane (0 < < 3 • ), primarily to determine the first redshift and AGN type.
The data reduction and analysis of DR2 spectra maintain the uniform approach described in the initial DR1 paper . All new spectra are processed using the standard tasks for cosmic-ray removal, 1D spectral extraction, wavelength, and flux calibrations, in either IRAF or the ESO/Reflex environment for the Very Large Telescope (VLT) instruments. The spectra are flux-calibrated using standard stars, which were typically observed two to three times per night. The spectra are corrected for Galactic reddening. Finally, a telluric absorption correction is applied to the spectra with the software molecfit.

Master Observing Table
We provide the following key parameters when possible for each individual spectrum (subsection 8): 1. BAT ID: Catalog ID in the BAT survey.
2. Telescope, diameter, instrument: Name of observatory, its diameter, and instrument used.
3. File and File red: Name of associated fits spectral file. For telescopes with both a blue and red side, two spectra are listed.

4.
Flags: Any associated flags with calibration or spectral extraction. Star: indicates foreground stellar contamination, that a very nearby star (<2 ) contributed to the emission despite a very small extraction region. Red: indicates that only the red side is extracted because the Galactic extinction was so high (e.g. >3) that no source is detected in the blue. Calibration: indicates that the object was observed under poor conditions or the standard star was observed on a different night, so spectral calibration may be more uncertain than usual. Tellurics: indicates that the spectrum suffers from worse-than-usual telluric correction or that the molecfit correction was unsuccessful. Shortblue: the setup has a shorter than normal blue wavelength coverage due to a reduction issue.

5.
Date: UT date of observation.
6. Spectral Range: Range of the spectra in Å. For telescopes with both a blue and red side, two spectral ranges are listed with the blue side first.

7.
Grating: Name of associated grating or grism. If the instrument had both a blue (shorter wavelength) and red camera (longer wavelength), two gratings are listed with the blue side listed first.
8. R and Res FWHM: Instrumental resolution and FWHM in Å. For telescopes with both a blue and red side, two numbers are provided with the blue side first.
9. Slit: Slit width in . For telescopes with both a blue and red side, two widths are listed with the blue side first.
10. Slit length: Extraction length along the slit in . For telescopes with both a blue and red side, two lengths are listed with the blue side first. If multiple exposures were combined with optimal extraction (e.g. Palomar/DBSP), the average value is listed.
K . 11. Angle: Position angle in degrees, measured east of north. In most cases the sources were observed at parallactic unless a nearby galaxy was observed in the same slit.
12. Seeing: Recorded seeing of observations. When possible we use the average seeing. We have not corrected the seeing observations to the observed air mass.
13. Exposure: Total exposure from all combined observations for the individual spectra.
14. CDELT: Pixel dispersion in Åper pixel. For telescopes with both a blue and red side, two widths are listed with the blue side first. Only included for spectra with linear dispersions (e.g. not the SDSS).
16. BC: Barycenter correction km s −1 needed for the Earth's motion based on observation time and observatory location. The computed correction should be added to any observed velocity to determine the final barycentric radial velocity. As this correction is small (e.g. < 30 km s −1 ), it has not been applied to any catalog measurements in the DR2.
We note that all of these observing parameters are not available for every spectrum, including most of the archival sample, but we provide them when possible.

Overview of Samples
Here we provide a list of the telescopes used and their respective data reductions. A plot of the number of spectra from each telescope is provided (Figure 2) as well as summary plots of typical observing conditions and resolution ( Figure 3). The redshift range of observations can be found in Figure 4.

Palomar Telescope
The largest sample of targeted sources, 402, was observed with the Palomar Double Spectrograph (DBSP) on the 200-inch Hale telescope. These AGN were observed as part of a dedicated program on BAT AGN (P.I. M. Urry or M. Powell) or as part of the NuSTAR program (P.I. F. Harrison and D. Stern), where they served as backup targets to faint NuSTAR serendipitous sources.
The observations were performed between 2012 October and 2020 November. The majority of observations were taken with the D55 dichroic, with the 600/4000 and 316/7500 gratings using a 1. 5 slit, This setup provided resolutions of ∼4.4 Å and ∼5.8 Å FWHM, over the ∼3150-5650 Å (blue) and ∼5400-10500 Å (red) region, respectively, providing full spectroscopic coverage of the optical region. The size of the aperture used for extraction along the slit depended on the specific observing run, but was either fixed at 1.5-2 or matched to the extended profile of the source in the IRAF APALL task. We note that there are some flux calibration issues at some of the grating edges due to loss of sensitivity specifically at 5400-5650 Å on the blue side and 5400-5600 Å on the red side and at 10000-10500 Å (see examples in Appendix B). In some cases, the region between 5400 and 5650 Å was not extracted. Also, we did not force the spectra to be flux-calibrated in the overlapping regions, and differences of 10-20% in flux may occur between the blue and red sides.
A smaller set of 66 narrow-line AGN and Sy1.9 AGN were also observed using the higher-resolution 1200 lines/mm grating for higher spectral resolution velocity dispersion measurements of the 3960-5500 Å and 8150-9600 Å regions, respectively.

Very Large Telescope
A total of 211 were observed with X-shooter, a multiwavelength (3000-25 000 Å) echelle spectrograph with medium spectral resolution = 4000 − 18 000 (Vernet et al. 2011). Two dichroics are used to split the incoming light into the three arms for efficient observation of all three arms simultaneously. The UVB arm (3000-5595 Å), VIS arm (5595-10240 Å) and NIR arm K . were used (10240-24800 Å). In some cases the NIR range only extended to 21010 Å rather than 24800 Å. More information on the NIR reductions and scientific results can be found in Brok et al. (2022). The majority of observations were observed with 1. 6, 1. 5 in the UVB and VIS arms, respectively, in the NODDING mode and extracted with a 4 aperture along the slit. The focus of the program was on obscured AGN (e.g. Sy1.9 and Sy2) to use the broad wavelength coverage and high spectral resolution to measure BH masses from velocity dispersion. These observations were performed as part of a filler program sometimes during bad weather conditions; however, the median seeing was still 1. 02 owing to the inherently good conditions at the VLT. In the case of VLT/X-shooter the spectra were first reduced using the standard pipeline in the ESO reflex software (v2.3.0; Freudling et al. 2013).  Another 19 sources were observed as part of the LLAMA sample (Davies et al. 2015) of low-redshift, luminous BAT AGN ( <0.01). These sources were observed in the IFU-offset mode with a field of view (FOV) of 1. 8×4 with resolution ∼8400 and ∼13200 in the UVB and VIS arms, respectively. The spectra were reduced using the ESO X-shooter pipeline v2.6.0. The spectra were corrected for telluric absorption using telluric standard stars. A more detailed description of the VLT/X-shooter data processing is given in Burtscher et al. (2021).
Finally, three spectra were part of the Science Verification data from VLT/X-shooter in 2009 and lack header information necessary to be processed in the standard way using ESO/Reflex. Two of these spectra, obtained in the SLIT, NODDING mode, were processed using optimal extraction and telluric standard stars following the reductions procedure of Becker et al. (2019). The UVB arm was binned to 15 km s −1 pixels. VIS and NIR arms were binned to 10 km s −1 . A final IFU spectrum of NGC 7319 was extracted with a 4 by 2 region following the LLAMA sample.
There were also 69 observations with VLT/FORS2 done in a single observing semester in 2017 (099.A-0403A) and were focused on Sy1 or Sy1.9 AGN. For FORS2, the majority were observed with the 600B grating, with 1 slit, covering 3400-6100 Å. A smaller subset of higher-redshift sources ( >0.8) was done with the 300I grating covering 6100-11 000 Å. All sources were reduced with v5.3.32 of the pipeline. Optimal extraction was used with typical extractions along the slit of 9 .

Southern Astrophysical Research Telescope
We observed 153 sources at the Southern Astrophysical Research (SOAR) telescope using the Goodman instrument as part of six programs between 2017 and 2020 (P.I. C. Ricci). Observations were acquired in two lower-resolution setups focused on unobscured AGN with the 400 lines mm −1 grating and GG455 blocking filter or 600 lines mm −1 and GG385 blocking filter. We performed higher resolution observations of obscured sources focused on the calcium triplet (CaT: 8498, 8542, and 8662 Å) using the 930 lines mm −1 or 1200 lines mm −1 grating. Nearly all observations were done with a 1. 2 slit with just two sources done with a 0. 45 slit, because of the very low velocity dispersions. All sources were extracted optimally, with typically slit lengths of 4. 8. K .

du Pont Telescope
Over 11 nights in 2016, 119 AGN were observed with the du Pont telescope with the Bollens & Chivens spectrograph (P.I. C. Ricci). All sources were observed with a 1 slit, the 300 lines mm −1 grating covering 3000-9070 Å. The sources were extracted with an optimal extraction with typical lengths along the slit of 6. 6. The sources were typically unobscured AGN, due to the relatively low resolution (FWHM∼8.7 Å).

Keck Telescope
Some AGN were also observed with the Keck telescopes associated with observations of NuSTAR-observed AGN and mergers (e.g., Koss et al. 2016b). A total of 21 observations were carried out using the Low Resolution Imaging Spectrometer (LRIS; Oke et al. 1995) on the Keck I Telescope. The setup used the blue (600 lines −1 ) grism and the red (400 lines mm −1 grating, with the D560 dichroic. The majority of observations were done with a 1 slit with a handful done with a 1. 5 slit.

Magellan Telescope
We performed 12 observations with the Magellan Echellete (MagE) spectrograph (Marshall et al. 2008) on the Magellan Clay telescope. We used a 1 slit for observations with the slit angle set to parallactic. The data were processed with the MagE pipeline, which is part of the Carnegie Python Distribution (CarPy, v. 1.4.2). The wavelength coverage was ∼3300-10000 Å though there was significant detector fringing above 8280 Å, resulting in strong instrumental issues above this range. Typical exposure time was 1 hr and targeted higher-redshift type Sy1.9 and Sy2 AGN ( >0.08) for velocity dispersion measurements.

SDSS and Archival Data
We also included additional spectra from archival sources. We used spectra from the SDSS (York et al. 2000), with 151 sources from data release 16 (DR16; Ahumada et al. 2020), which were observed in the legacy survey with 3 fiber at ∼3800-9200 Å coverage or 2 fiber from BOSS or eBOSS with ∼3600-10200 Å coverage, respectively.
We also include 90 additional archival spectra of AGN that were acquired after the DR1. These include AGN observed with earlier surveys of ROSAT AGN that overlap with BASS of unpublished and published (Grupe et al. 2004) sources, from the Palermo surveys of Swift BAT AGN (Rojas et al. 2017), or as part of an atlas of low-redshift AGN (Ho & Kim 2009).
Finally, we include 6 archival spectra from VLT/MUSE when our existing BASS spectra were insufficient. This includes the nearby Circinus galaxy and NGC 3393, which was too bright to observe using our standard setups and was not part of the DR1 archival sample. In addition, the dual AGN NGC 6240N and NGC 6240S have a VLT/MUSE spectrum for each AGN obtained in adaptive optics (AO) mode owing to their close separation. Finally, the relatively distant Sy 2 sources, BAT ID 1204 (z=0.6) and ID 209 (at z=0.09) were included in order to enable velocity dispersion measurements. We use the processed data from the ESO Science archive. We extract a 2 -radius aperture at the WISE position, except for NGC 6240N and NGC6240S, for which we use an 1 aperture owing to their closeness.
In molecfit, model spectra are fitted to the observed spectra to derive the best-fit atmospheric parameters by iteratively computing transmission curves using a simple Earth atmosphere structure at the time of observation. Global weather data are combined with local weather data to provide a likely best fit (humidity, pressure, temperature). We use four regions with strong atmospheric features (O 2 at ∼ 6800 < /Å< 6900, O 2 at ∼ 7460 < /Å< 7560, telluric regions at ∼ 9100 < /Å< 9200, and telluric regions at ∼ 9400 < /Å< 9500) to determine the best-fit atmospheric parameters. An example is provided in Figure  Figure 5. We mask regions with strong AGN emission-line features from the fitting. We use a Gaussian kernel variable with wavelength (varkern=1).
Telluric corrections with molecfit were applied to all spectra with coverage above 7500 Å except archival observations (which lacked local weather data), Magellan/MAGE observations (because of significant fringing), and SDSS spectra (which have already had this correction applied). An example of the telluric correction and its importance for narrow-and broad-line measurements is provided in Figure Figure 6. We note that the ability to recover the intrinsic spectra is dependent the ability to measure foreground atmospheric absorption lines in the source, and thus very faint intrinsic spectra, such as those with very high Galactic extinction have little or no correction. While most spectroscopic regions were adequately corrected, the 9300-9700Å region suffers very high extinction, and emission-line fitting should be approached with caution.

Overlap with the DR1
Initial targeting priority focused on sources without BH mass measurements or spectra from the archival data in the DR1, but this was later expanded to complete the whole sample. The DR2 includes all of the SDSS spectra in the DR1 as well as 35 early Palomar spectra that were reprocessed to include corrections with molecfit. Beyond this there are still 42 AGN in the DR1, with no new DR2 spectra.

Instrumental Resolution
We determine the instrumental resolution and line-spread function (LSF) FWHM for each spectral setup and provide the best estimate in Table 7. A description of how this quantity was measured is given in subsection 3.4.
In real observations, the slit width is not the only factor in determining the instrumental resolution. If the slit is large, the image quality at the entrance slit can be smaller than the slit width. The observed spectral resolution is then better than what is measured from the slit width for arc or sky lines which fully fill the slit, but is instead determined by the sharpness of the image of the object at the entrance slit. Thus, the spectral resolution in real observations of AGN may be somewhat smaller than when measured using sky lines or arc line spectra that fully fill the slit.
The measurements from molecfit, which trace the absorption of telluric lines, can provide an additional estimate of the effective resolution. We note, however, that the ability to fit the absorption profile is difficult in faint sources and only done for setups that include coverage of telluric features (e.g., >9000 Å). In several setups without telluric features, only a few AGN were observed, and we use sky or arc lines for resolution measurements.
An independent estimate of the spectral resolution and the line spread was also done with the penalized PiXel-Fitting method (pPXF, Cappellari & Emsellem 2004;Cappellari 2017) by fitting stellar absorption lines to individual Galactic stars that were observed during the observations. We use two template libraries depending on the resolution. We use the X-shooter data release 2 library, which was obtained at much higher resolutions than typical observations (e.g. ∼ 10000). For spectra obtained at higher resolutions ( >3000) based on sky lines, we use the PHOENIX theoretical spectral library (Husser et al. 2013) as a template which has much higher resolutions ( ∼ 500, 000). We fit the 3880-5500 Å region and 8350-8730 Å to determine the LSF in the blue and red ranges, respectively, to target stellar absorption features.
We find that the spectral resolution as measured from individual Galactic stars during real observations or telluric absorption lines in the AGN galaxies tends to indicate spectral resolutions somewhat sharper (∼ 20%) than those of sky lines. For instance, for the largest sample observed using the 1.5 slit with Palomar/Doublespec, the average of stars observed on different nights is 4.7 ± 0.3 Å in the CaT, compared to the average of the telluric absorption with molecfit which is 4.9 ± 0.6 Å. The average from fitting sky lines for these observations is 5.8±0.3 Å. In this case, we use use the molecfit average, though it is statistically similar K . 9531. The bottom row shows the CaT region before (left) and after (right) molecfit correction when being fit with a galaxy template for a velocity dispersion measurement (Koss et al. 2022b). The model fit is shown in red, with residuals from the fit shown in green dots below the spectra. The TAC spectrum is able to recover a larger region of the Ca triplet spectral region (8450-8700 Å) which is redshifted into telluric features at >0.04 with no significant increase in residuals.
to that from stars. For the 3880-5500 Å region, molecfit was not used because of a lack of telluric features, but the pPXF fits were 4.1 ± 0.2 Å vs. 4.8 ± 0.4 Å for the sky lines within each observation, and the average of the pPXF fits was used.

SURVEY MEASUREMENTS
The spectroscopic release provides spectra for 95.1% (816/858) of DR2 AGN. When combined with the DR1 spectra, all AGN have spectra (e.g. 100%, 858/858), representing a complete census of luminous hard-X-ray-selected AGN over nearly the entire sky outside of a small region on the Galactic plane (94.8%,| |>3 • ). Here we describe the main survey measurements (e.g. , AGN type, BH ). Further derived measurements (broad-line widths, narrow-line widths, velocity dispersions) will be provided in subsequent papers.
For all the AGN in our sample, we provide the following key parameters when possible for each AGN

5.
Dist and z ind : Distance assumed based on redshift or redshift independent distance measurements in Mpc. 6. Best MBH and MBH Meth . The best BH mass measurement and the method used for the measurement. We do not report errors from either broad-line fitting or velocity dispersions as the they are less than 0.1 dex and the errors are dominated by the intrinsic spread of virial and ★ based BH mass estimates of order 0.5 dex ).
7. bol and / Edd : Measurement of the AGN bolometric luminosity and Eddington ratio based on BH mass.

AGN Type
For classification, we first split the sources into 752 unbeamed AGN and 105 beamed AGN (and 1 lensed AGN) following the spectroscopic classification of the Roma Blazar Catalog catalog (BZCAT; Massaro et al. 2009). The beamed AGN were split into three categories based on their optical spectral properties, specifically, based on the presence of broad lines (BZQ), only host galaxy features lacking broad lines (BZG), or traditional continuum-dominated blazars with no emission lines or host galaxy features (BZB).
Overall 72/105 (69%) of the DR1 beamed AGN maintained the same classification with further optical spectroscopic study. 20/105 (19%) changed specific beamed AGN classification based on additional DR2 optical spectra (e.g., BZG to BZQ, based on the detection of broad H or any other broad line such as H ).
We provide an unbeamed AGN type based on the presence of broad Balmer lines from visual inspection after fitting with host galaxy templates (Koss et al. 2022b). These include sources with broad H (Sy1), sources with narrow H , but broad H (Sy1.9), and sources with only narrow optical lines (Sy2). Further classification of broad-line AGN (e.g. Sy 1.2, 1.5, 1.8 etc.) is provided in subsequent studies (Mejía-Restrepo et al. 2022) and also of narrow-line AGN such as LINERs (Oh et al. 2022). When spectra are not available in DR2 for AGN type, we use the DR1 AGN type.
Overall, there are 168 AGN in DR2 for which we derive a revised or first classification of AGN type based on our measurements. A detailed comparison of DR2 measurements compared to the DR1 is provided in Appendix A.

Redshifts
To determine the best redshifts for the sample Table 10, Table 9 is published in its entirety in the machine-readable format for 858 AGN. A portion is shown here for guidance regarding its form and content. See subsection 3.1 for a description of each data column.
For the Sy1.9 and Sy2 AGN, or any broad-line AGN where the fitting failed, we fit the [O ] emission line in our sample using PySpecKit, an extensive spectroscopic analysis toolkit for astronomy, which uses a Levenberg-Marquardt algorithm for spectral fitting (Ginsburg & Mirocha 2011). We fit the [O ] emission line using a single Gaussian. Finally, for 10 sources that are highly reddened and often in the Galactic plane where an [O ] line was not detected we use the H or [S ] 6717, 6731 or He 10830 line. We note that for NGC6240N and NGC6240S, the sources were too close together to resolve in our spectra, so we provide a single measurement.
There are 30 beamed AGN that are strongly continuum-dominated blazars (BZB), with only weak stellar features, and hostgalaxy-dominated blazars (BZG), where no [O ] 5007 emission lines are measured. For 25 of these sources, we can measure redshifts using pPXF with the Ca H+K lines or Ca absorption features. Further details, including a full list of redshifts from stellar absorption features from galaxy template fits for Sy1.9 and Sy2 AGN can be found in Koss et al. (2022b).
We used the BASS DR1 data for the 33 AGN in DR2, which did not have a new spectrum at all or were missing one that covered the [O ] 5007 line.
For the remaining sources we rely on NED or SIMBAD or past publications for redshift measurements. Of the remaining six BZB sources we could not detect strong host galaxy features for a redshift, there are five sources with an existing redshift from NED or SIMBAD, which we use as the redshift measurement. There is one very high-extinction AGN ( =9.5) 2MASS J10445192-6025115, which was found to have a redshift of =0.047 (Fortin et al. 2018) based on He 10830 line. The beamed and lensed AGN PKS 1830-21, was measured using H in the NIR at =2.507 (Lidman et al. 1999) due to it's high extinction ( ) from being in the Galactic plane.
Overall the redshift completion is extremely high, 99.8% (857/858) for the full sample. Of these AGN redshifts, 47 (Table 11) are found for the first time (Figure 7). The only AGN without a redshift is a continuum-dominated blazar (BZB). The blazar, B3 0133+388, was first discovered in the third Bologna sky survey of 408 MHz radio objects (Ficarra et al. 1985) and is also shows bright gamma-ray emission above 1 GeV in Fermi. The source shows faint Ca H+K lines at redshift zero in two different Palomar spectra (and also in a Keck/LRIS spectra shown in Aliu et al. (2012)). However, given the radio and Fermi detection, the source is unlikely to be Galactic but maybe a blazar with a foreground star.

Distance and Luminosity
The AGN host galaxies span a large range of redshifts down to very nearby (<50 Mpc) systems. These AGN can have substantial peculiar velocities compared to velocities of the Hubble flow where a simple assumption of uniform expansion would lead to large errors. High-quality redshift-independent distances to nearby galaxies such as through using the tip of the red giant branch (TRGB) are now available (e.g., McQuinn et al. 2017). Further compilations such as the Extragalactic Distance Database (EDD, Tully et al. 2009) or the Cosmicflows-3 project (Courtois et al. 2017) have now compiled motions of many thousands of local galaxies.
We follow the approach of Leroy et al. (2019), which performed a careful analysis of which compilations to adopt at different distances based on statistical uncertainties. Specifically, we limit our search to <50 Mpc (or 3500 km s −1 ) galaxies to adopt redshift-independent distances as beyond this the typical uncertainties are larger than those in the Hubble flow. We focus on using EDD, Cosmicflows, and NED for adopting distances. We adopt a TRGB and "quality" distances from EDD whenever available. When this is not available, we take the Cosmicflows-3 value. If none of these are available, we use the most recent redshift-independent estimate from NED. This results in redshift-independent measurements for all 59 galaxies (Table 12) below 3500 km s −1 in our survey.

Black Hole Mass, Bolometric Luminosity, and Eddington Ratios
We also provide the best BH mass measurements for each AGN in our catalog outside of continuum-dominated blazars (BZB). A small number of sources have direct (or higher-quality) measurements of BH masses, either from reverberation mapping ( =48), OH megamasers ( =10), or high-quality IFU observations of gas or stars ( =12) which we have adopted and tabulated when available.
Some AGN may have multiple BH mass measurements from broad lines and velocity dispersions, so we select the best measurement (Table 13) based on the following ordered scheme: 1. Literature measurements with megamasers, reverberation mapping, or stellar and gas dynamics. -Restrepo et al. (2022). The conversion of broad-line measurements to BH masses is given by Trakhtenbrot & Netzer (2012). K . A summary of the number of best BH mass measurements is found in Figure 8.  Koss et al. (2022b) for more details about the individual observations, calculations, and methodologies.
The bolometric luminosity is calculated from the intrinsic luminosity in the 14-150 keV range as shown in (see Table 12, Ricci et al. 2017b). This analysis was done using the 0.3-150 keV range by combining the 70-month average Swift BAT spectra with data below 10 keV from Swift XRT, XMM-Newton, Chandra, Suzaku, and ASCA using detailed spectral models. Here we calculated the bolometric luminosity using a 14-150 keV bolometric correction of 8 based on the factor of 20 for the 2-10 keV range (Vasudevan & Fabian 2009) and assuming Γ=1.8. We prefer this rather than using the direct calculation from the 2-10 keV range because the corrections are less dependent on H for CT AGN. The 14-150 keV emission is also integrated over 70 months, so it is more likely to be representative of the average value.
For Eddington ratios we assume an Eddington luminosity consistent with solar metallicity: Edd = 1.5 · 10 46 erg s −1 10 8 We note that more complicated procedures than a simple bolometric corrections from the intrinsic X-ray flux such as in terms of the Eddington ratio ( / Edd ) are sometimes used (e.g., Marconi et al. 2004;Lusso et al. 2011). However, we prefer this simple approach, similar to what was done in the BASS DR1, that can be reliably applied to all AGN.
We do not list individual errors for each BH , bol , and / Edd measurement, as they are dominated by the systematic uncertainties in the scaling relations rather than the emission-line fitting or velocity dispersion measurements which are typically <0.1 dex. Errors in BH are of order 0.4-0.5 dex owing to systematic uncertainties in virial and * -based scaling relations (e.g., McLure & Dunlop 2002;Vestergaard & Peterson 2006;Ricci et al. 2022). For bol , the scatter between BAT 14-195 keV luminosity and 5100 was 0.46 dex in the DR1  Efforts are currently underway for future BASS surveys to better calibrate the bolometric correction with AGN source properties and estimate its intrinsic reliability. A large HST program (>100) AGN is currently underway obtaining high spatial resolution near UV (<3000 Å) imaging of the AGN emission, combined with simultaneous measurement of the AGN emission in the X-rays and UV/optical from Swift, with ground based imaging in . A summary of the survey completeness in BH mass measurements for unbeamed AGN is provided in Table 14 separated by AGN type. Overall the completeness is slightly higher for Sy1 and Sy1.9 (>96%) than for velocity dispersion measurements (>93%). Outside of the Galactic plane (| |>10 • ) the survey completeness rises to 98% for all unbeamed AGN because of the typically lower extinction in these regions. Finally, for beamed AGN with broad lines (BZQ) the measured BH masses (Table 14) and completeness are somewhat lower, but still the majority (91%).
A summary of the typical BH masses, bolometric luminosities, Eddington ratios, and X-ray column densities is provided in Table 14 and Table 15 for unbeamed and beamed AGN, respectively.

SUMMARY
We have presented an overview of the BASS DR2 survey with 1449 optical spectra, of which 1182 are released for the first time, for the 858 hard-X-ray-selected AGN in the Swift BAT 70-month sample. With this first DR2 catalog release we provide the following: K .  Histogram of the BH mass estimates in the small number of X-ray-obscured AGN with only broad-line measurements. While many of these AGN do have BH masses consistent with the other unobscured distribution (e.g. > 10 7 ) there is a larger fraction with low masses (< 10 7 ), which may be significantly underestimated.
• A revised catalog based on optical and NIR spectroscopic follow-up that identifies all of 858 among unknown sources above (| |>3 • ) or below V =5 mag excluding only 7 unknown sources deep within the Galactic plane at high extinction. We have included new identifications of 17 Galactic sources.
• We have further classified our sources by AGN type based on the presence of broad lines (e.g. Sy1, Sy1.9, Sy2) as well as beamed and lensed AGN. We have further provided important catalogs for population studies including dual AGN, weakly associated AGN, and multiple weak confused sources within the BAT beam.
• A full master catalog summary of the 1449, instrumental settings, their reductions, and observing conditions. With this we have provided a master catalog of redshifts, distances, bolometric luminosities, and BH masses.
• Overall the completion for the survey is 99.9% in redshift outside the extreme regions in the Galactic plane (| |>3 • ). In BH mass, the survey is 98% complete using broad lines and velocity dispersions for unbeamed AGN outside the Galactic plane (| |>10 • ). The final catalog contains 47 new redshift measurements and 790 BH mass measurements.  (2): Total for the whole sample. Column (3): total excluding the Galactic plane region | | < 10 • where high optical extinction makes measurements more difficult. Column (4): median redshift from optical lines. Columns (5)(6)(7)(8): number of unique AGN with BH measurements and excluding the Galactic plane region | | < 10 • where high optical extinction makes measurements more difficult. Also listed as percentages. Columns (9-12): median BH , bol , / Edd , and, log( H /cm −2 ) for the sample.  (2) Total for the whole sample and excluding the Galactic plane region | | < 10 • where high optical extinction makes measurements more difficult. (4) Median redshift from optical lines. (5)(6) Number of unique AGN with BH measurements and percentages. (7-10) Median BH , bol , / Edd , and, log( H /cm −2 ) for the sample.

ACKNOWLEDGMENT
We thank the reviewer for the constructive comments that helped us improve the quality of this paper. BASS/DR2 was made possible through the coordinated efforts of a large team of astronomers, supported by various funding institutions, and using a variety of facilities.
We acknowledge support from NASA through ADAP award NNH16CT03C ( This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France.
A significant part of the BASS observations and work took place during the COVID-19 crisis. We thank the health care experts in communities around the world, for their tireless efforts to keep us all as safe and healthy as possible.

APPENDIX A. COMPARISON TO DR1 AND PAST SURVEYS
Here we provide a comparison to the 641 optical spectroscopic measurements from the BASS DR1 including redshifts, AGN classification, and BH mass measurements as well as past measurements in the literature. ,

A.1. Beamed AGN Changes in DR2
Here we provide a list of all beamed AGN that are newly identified or changed in DR2 in Table 16 compared to the DR1. Example spectra for the various classes (e.g., BZB, BZG, BZQ) are found in Figure 9.

A.2. AGN classification
Overall, there are 168 DR2 AGN that we provided revised or the first classification of AGN type based on our measurements. For comparison, we first look at the 641 DR1 AGN types compared to overlapping DR2 AGN types. There are 10% changes (64/641). This includes 52 reclassifications from Sy2 to Sy1.9 or Sy1, or from Sy1.9 to Sy1 based on the detection of broad lines that were not detected in DR1 spectra. Conversely, 12 spectra change from Sy1 to Sy1.9 or Sy2 or from Sy1.9 to Sy2. This shift largely reflects the higher-quality spectra in terms of resolution and signal-to-noise ratio compared to the archival DR1, which used much smaller telescopes and lower spectral resolutions rather than "bona fide" AGN that have undergone changes. The "bona fide" changing optical type AGN are part of a future study (Temple, M. et al. 2022, in preparation). Among the 216 AGN which were not part of the DR1 release, roughly half have their first or revised classifications (48%, 103/216). This is compared to the most recent 105-month survey (Oh et al. 2018) which includes updates from SIMBAD and NED. From these 103, 72 measurements are to previously unknown AGN without available optical spectroscopy or classification.

A.3. Redshifts
We compare the redshifts from the 599 DR1 to the revised measurements from spectroscopy in DR2. Among the low-redshift sample ( <0.3) the agreement is excellent, with no differences larger than 1000 km s −1 . At higher redshifts (0.3> >1) the median offset increases (| 2 − 1 |= 157 km s −1 ), and finally increases to 1180 km s −1 at >1 owing to the use of the intrinsically broad lines of Mg 2798 and C 1549 for the derivation of the redshift.
K . ID=1164, Magellan/MAGE, Sy1.9, z=0.120 Figure 11. Figure showing examples of high-resolution spectral setups not shown earlier in the text excluding the archival observations. A spectrum taken at the Palomar telescope with the Doublespec instrument with the 1200 lines/mm grating is shown on top, with a spectrum from SOAR with 1200 lines/mm grating using the Goodman instrument in middle, and finally an echelle spectrum from Magellan using the MAGE instrument.