All-Quasar Multi-Epoch Spectroscopy’


The All-Quasar Multi-Epoch Spectroscopy (AQMES) program aims to temporally resolve a wide variety of astrophysical phenomena related to supermassive black hole (SMBH) accretion. This includes changes in the broad line region (BLR) on the dynamical time scale (of order months to years), variability in quasar outflows on similarly long time scales, accretion state or other abrupt transitions (such as changing look quasars), binary BH signatures, etc.

AQMES conducts repeat spectroscopy of a large sample of confirmed quasars from SDSS in order to capture and measure their UV/optical spectral features. It is sensitive to typical variations of order 10% in flux density and resolves timescales from months to decades. When combined also with observations from the BHM reverberation mapping (RM) program and observations from past SDSS programs (and other surveys), the time sampling will allow one to explore the widest reasonably accessible range of temporal baselines, from a few days to a two decades, i.e. from the light-crossing time of the inner BLR, to the dynamical time of the outer BLR, and including key accretion-related timescales.

The quasar sample targeted by the core components of AQMES (> 20,000 quasars) and cadence distribution (2–10 new epochs added to those already in the SDSS archive) should enable significant advances over past time-domain spectroscopic studies. The resulting data set will allow statistical analysis, including averaging in sub-samples defined by basic parameters such as luminosity, black hole mass, Eddington ratio. Moreover, the SDSS-V sky coverage and spectral quality should allow comparison to extant large quasar archival spectral datasets.

(Adapted from a figure provided by Yue Shen)

Observational goals and requirements

The SDSS-V/BHM AQMES core program seeks to gather high signal-to-noise ratio (S/N) repeat spectrophotometry of known, luminous broad-line active galactic nuclei (AGN) in order to quantify their variability properties. The faintest quasars (𝑖 = 19.1) will yield spectra with S/N ≈ 10 in a typical 1-hour exposure while quasars with 𝑖 = 18 will yield spectra with S/N ≈ 20 in the same exposure time. These spectra will allow measurements of the centroids of the broad emission lines to 0.1% or better and radial velocity changes to 65 km/s or better. The basic shape properties of line profiles (widths and velocity dispersions) will be measurable to 10% or better, while higher order moments should be measurable to 20%.

One of the primary goals of AQMES is to resolve time scales on the order of the dynamical time of the broad line region (BLR) for a range of black hole masses and luminosities — time scales which range from a few weeks to a century, given the expected range of accessible masses and luminosities. The AQMES observations combined with past SDSS archival observations of the same quasars will be able to probe time scales between 100 and 6000 days (3 months to 16 years), while the observations from the reverberation mapping (RM) program will cover even shorter time scales. The time resolution needed to achieve the above goal is 0.2 dex or better, across the above range of time scales. These considerations apply to all the science questions that the core AQMES program aims to address, including the variability of emission and absorption lines, abrupt and dramatic spectroscopic changes, radial velocity variations, etc.  Since the amplitude of quasar variability anti-correlates with luminosity and given the wide range of quasar redshifts and luminosities accessible through the AQMES observations, the actual survey planning takes into account the resulting distribution of targets in the luminosity vs. rest-frame time interval plane.

Meeting the science goal of sampling the luminosity–time interval plane, requires (a) a core sample of >2,000 bright quasars observed on month-years timescales (i.e. ~10 times over the course of the 5-year survey), and (b) a sample of >20,000 bright quasars sampled on years-decades timescales (~2 times over the course of the survey). The sizes of these samples are dictated by the need to reduce the uncertainties in variability distribution functions (a.k.a. “structure functions”), so that they are comparable to those of the individual spectroscopic measurements. Hence ~500 “time intervals” (i.e., pairs of spectroscopic observations from which differences can be measured) per structure function bin are needed by the end of the survey.

The figure below shows how bins in the luminosity–time interval diagram will be populated, according to simulations of the AQMES observing cadence. Every row in this diagram (or slice along the time interval axis) represents the data needed for a structure function. When the AQMES results are combined with archival (DR16Q) spectra, the desired range of time scales will be adequately sampled over a wide range of quasar luminosities. Such a sampling would not be possible with archival spectra alone. Moreover, since the bins in  this 2-dimensional diagram are very well populated they can be subdivided into bins along a third dimension, such as black hole mass or Eddington ratio. Therefore, at the end of the survey structure functions can be constructed as a function of at least two fundamental system parameters.

Comparison of the distribution of 𝑖 ≤ 19.1 SDSS quasars in the (rest-frame) time interval vs luminosity diagram before and after AQMES. The time intervals refer to all unique pairs of spectroscopic observations of the same object and are indicative of our ability to probe spectroscopic variability on the corresponding time scale via structure function analysis. The left panel shows this distribution among DR16Q quasars with multiple observations (~21,000 DR16Q quasars in all). The right panel shows the simulated distribution after combining the observations of the quasars targeted by AQMES with older observations of the same quasars from the DR16Q catalog (~22,000 AQMES quasars). The bins in the maps cover 0.1 dex in time interval and 0.13 dex in 𝑖-band luminosity.

Target selection and survey implementation

The AQMES program is implemented as a set of targets grouped into a number of target ‘cartons’. The medium-intensity cadence (10 new epochs) program will observe targets in the ‘bhm_aqmes_med*’ cartons. The wider tier of AQMES (2 new epochs) will observe targets from the ‘bhm_aqmes_wide2*’ cartons. For each AQMES tier there is a core sample of brighter QSOs (16<psfmag_i<19.1 AB), supplemented by a sample of fainter QSOs (19.1<psfmag_i<21.0 AB). The sky areas for these core AQMES programs are selected taking account several factors, including: sky density of bright QSOs, availability of Chandra Source Catalog (CSC) targets, overlap with the SPIDERS footprint, and expected availability of observing time as a function of LST.

In addition we will also attempt to opportunistically re-observe any previously known SDSS DR16 QSO regardless of sky location, if and when spare fiber resources are available. This non-core BHM program is implemented in cartons labelled ‘bhm_aqmes_bonus*’, with separate cartons depending on the optical brightness of the QSOs.

Further details of the algorithms and criteria used to select AQMES target cartons are provided via the links below.

Targeting generation v0.5.3

The v0.5.3 generation of targeting was used during initial SDSS-V/FPS operations, below is a list of BHM AQMES cartons it includes:

Spectroscopic data released in DR18

No new AQMES data is released in SDSS DR18

Back to Black Hole Mapper Science Programs

Back to Black Hole Mapper home/overview page

Back to Top