A History of the International AGN Watch

The Origin of the Consortium

By the mid-1980s, it had become apparent to researchers in the field that emission-line variability was closely tied to continuum variability, which strongly supported the prevailing view that the lines were driven by photoionization from the central compact continuum source, thought to be an accretion disk surrounding a supermassive black hole. It was also recognized that the emission-line variations should follow the continuum variations, but with a time delay that reflects the light-travel time across the broad-line region (BLR). Early attempts at monitoring coordinated continuum and emission-line variations in Seyfert galaxies suggested that the light-travel time was about an order of magnitude smaller than expected on the basis of photoionization equilibrium arguments, but the temporal sampling in these early attempts was not sufficiently good to provide a truly convincing case. The important realization, however, was that in principle well-sampled continuum and emission-line measurements could yield both the size and velocity field of the broad-line region, through the process now known as ``reverberation mapping''.

The NGC 5548 IUE Project (1988-89)

The International AGN Watch was born at a Vilspa-sponsored conference on AGN emission-line variability, organized by Jean Clavel and Danielle Alloin and held in Segovia, Spain, in 1987 October. There was a clear consensus that the only way to obtain the large amounts of data necessary to attack the reverberation problem correctly was through a large-scale coordinated effort. Willem Wamsteker suggested that International Ultraviolet Explorer (IUE) was the logical choice to anchor such a program. Coverage in the space ultraviolet would be necessary since many of the strongest emission lines in AGN spectra are in the satellite ultraviolet (1215 - 3100 Å, and it is also desirable to observe the continuum at the shortest possible wavelengths, close to the ionizing continuum that actually drives the line variations). The bright Seyfert galaxy NGC 5548 was an obvious choice for the target because of its suitability for extended IUE observations (it would remain in the IUE ``observability window'' for as long as eight months at a time), and because there was a substantial body of previous work that suggested that the variability time scales in this source would be appropriate for the kind of monitoring program the consortium felt could be brought to bear.

It was decided to submit a large joint proposal to ESA/SERC (the British Science and Engineering Research Council SERC was the predecessor to PPARC) and NASA, requesting 50% of the total time that would be required from each. Jean Clavel was the ESA/SERC Principal Investigator, and Matt Malkan was the NASA principal investigator. Matt Malkan and Danielle Alloin assumed responsibility for organizing a concurrent ground-based effort, and they quickly recruited Michael Penston (who, sadly, passed away in 1990 December) and Brad Peterson to help with this. Within a few weeks, Brad Peterson had agreed to take overall responsibility for collecting the optical data and preparing it for publication.

Shortly after the Segovia meeting, a US meeting on AGNs, organized primarily by Dick Miller, took place at Georgia State University in Atlanta. The organization of the proposal was settled on at this meeting, and a US committee of responsible parties, consisting of Matt Malkan, Mike Crenshaw, Julian Krolik, Brad Peterson, and Gail Reichert, was established. Plans for the monitoring campaign were announced at the meeting, and participants were recruited. Everyone who was interested in this program was invited to participate in the IUE proposal that was being prepared. Broad participation was a deemed highly desirable for several reasons: First, the radical nature of the proposal (the original proposal was for 18 8-hour shifts (to be scheduled as 36 4-hour half shifts) from ESA/SERC and an identical number of the low-background US1 8-hour shifts from NASA) would require an obvious display of broad community support to be successful. Second, since service observing had not yet been instituted at the IUE Observatory at GSFC, participants had to agree to be present at GSFC to conduct observations for at least two shifts. Third, the broad participation was necessary for the ground-based program, since the minimum number of observations that would be necessary for a successful program was not known. How weather, scheduling difficulties, and instrument failures might affect the program could only be guessed at. A further requirement imposed on participants was that they had to agree to refrain from submitting any directly competing IUE proposals that year.

Twin IUE proposals, entitled ``The International AGN Watch: The Size and Structure of the Broad-Emission-Line Region in NGC 5548", were submitted to ESA/SERC and to NASA on 1987 December 1 for IUE Episode 11; somehow the name ``International AGN Watch'' became associated with the consortium itself rather than the proposal activity, and the name stuck.

Both the ESA/SERC and NASA were awarded time, 30 half-shifts from ESA/SERC and 24 half-shifts from NASA. The final program consisted of 60 4-hour half-shifts shifts, with the time lost from the NASA allocation made up by generous donations from other approved IUE programs by Jean Clavel, Paul Barr, Matt Malkan, Martin Gaskell, and others (?).

Once the IUE proposals had been accepted, planning for the ground-based program began in earnest. A special evening session was held at IAU Symposium 134 in Santa Cruz, California, in 1988 July. Invitations to participate in the ground-based program were extended to all the attendees at this meeting - anyone who could provide usable data would be included as a co-author on the data paper that was expected to result from this program. Additional participation was solicited through electronic mail (which was new at the time and greatly enhanced the efficiency with which the project was organized), personal contacts, and an IAU Circular (No. 4664).

Many of the members who would ultimately comprise the AGN Watch central committee met at the Santa Cruz meeting to work out some of the organizational guidelines that characterized the operation of the consortium. It was agreed, for example, that the consortium as a whole would publish only the basic data and the most obvious sorts of interpretation (which, operationally, was later agreed to include emission-line cross-correlation lags, but little else), and that detailed analysis and interpretation would be left to individuals or teams of investigators. However, participants agreed to honor an embargo on the data in order to make sure that the consortium data paper could be published in a refereed journal. This meant specifically that any papers that used the data or results in any way would not be submitted to a refereed journal prior to acceptance of the consortium data paper. Mostly for the peace of mind of the participants rather than on account of any serious concern, Danielle Alloin, Jean Clavel, Matt Malkan, and Brad Peterson co-signed letters to the editors of the major refereed journals informing them of the embargo on the data, and requesting that they hold for later publication any papers using the data that were submitted prior to the consortium data papers. Despite the irregularity of the request, the editors agreed to cooperate.

The IUE program began on 1988 December 14, and continued until 1989 August 7. Each shift 4-hour half-shift began with an optimally exposed SWP (1150 - 1975 Å) image, followed by an optimally exposed LWP (1900 - 3000 Å) image. If time remained in the shift, a redundant SWP exposure was taken to fill the rest of the time. The redundant SWP exposures turned out to be important in estimating the uncertainties in the continuum and emission-line measurements. Observations were made once every four days for eight months, split between the low-background US1 shifts from GSFC and Vilspa shifts from Madrid.

The results of the experiment were prepared for publication by Jean Clavel. There were predictably a number of scientific disagreements among the organizers that had to be worked through. Where consensus or compromise did not seem possible, alternatives were presented. For example, the spectra were extracted from the original images in two separate ways, using the familiar IUESIPS software and also a flux-weighted extraction method known as GEX (for Gaussian EXtraction) that improved the pixel-to-pixel signal-to-noise ratio. Some of the organizers preferred IUESIPS as it was at the time the standard procedures for reduction of IUE data. Other preferred GEX, which seemed to yield higher signal-to-noise spectra. Since consensus was not achieved, measurements from both extraction methods were published in the consortium data paper, which was to be Paper I in the series ``Steps Toward Determination of the Size and Structure of the Broad-Line Region in Active Galactic Nuclei'' (Clavel et al. 1991, ApJ, 368, 64). This paper contained light curves for 5 ultraviolet continuum bands, for the optical band, as measured by the IUE Fine-Error Sensor (FES), and for the Ly-alpha 1215, C IV 1549, C III] 1909, N V 1240, Si IV 1402, and He II 1640 emission lines.

The ground-based data were obtained at the same time. The average spacing between observations, which came from 20 separate sources, was about 2.5 days, and extended in time somewhat beyond the IUE experiment. The first paper on the ground-based results was submitted for publication shortly after the IUE paper was accepted by The Astrophysical Journal. The original spectroscopic data were presented in Paper II of the "Steps" series (Peterson et al. 1991, ApJ, 368, 118), which gave optical continuum (4850 Å) and Hbeta 4861 emission-line light curves. Further subsequent analysis produced emission-line light curves for Halpha 6563, Hgamma 4340, He I 5876, and He II 4686 (Paper IV = Dietrich et al. 1993, ApJ, 408, 416) and for the ``small blue bump'', a blend of UV Fe II multiplets and the Balmer continuum (Maoz et al. 1993, ApJ, 404, 576). Optical continuum measurements based on CCD imaging were presented by Romanishin et al. (1995, ApJ, 455, 516).

The ground-based program was continued for a second year, i.e.. from 1989 November through 1990 October. This provided continuing Hbeta and optical continuum light curves (Paper III = Peterson et al. 1992, ApJ, 392, 470). In this paper, the continuum measurements were made at a ``cleaner'' location in the optical spectrum, at 5100 Å. Re-measurement of the first-year's spectra were also presented in an Appendix to this paper.

The principal results of the NGC 5548 campaign were the following:

A summary of our state of understanding following the NGC 5548 project is given by Peterson (1993, PASP, 105, 247).

The ground-based program on NGC 5548 continued, and two more years of data, covering 1990 October through 1992 September, has also been published (Paper VII = Peterson et al. 1994, ApJ, 425, 622).

The NGC 5548 HST Project (1993)

The original NGC 5548 left a number of important unresolved questions: At a special session of the 1990 September Heidelberg meeting on AGN variability, it was decided that the AGN Watch would submit a proposal for Hubble Space Telescope Cycle 3 (i.e., the last cycle before the COSTAR servicing mission) to monitor NGC 5548 at high signal-to-noise and high time resolution (one day). Brad Peterson was the Principal Investigator for this project. In addition to the HST proposal, a proposal to use IUE for contemporaneous observations was submitted, with Danielle Alloin as the ESA/SERC Principal Investigator and Brad Peterson as the NASA Principal Investigator. The purpose of the IUE program was (1) to provide an independent check of the pre-COSTAR HST flux calibration and (2) to extend the temporal baseline of the UV spectroscopic monitoring; both of these turned out to be important to the final success of the program, even though the IUE data were not of the hoped-for quality on account of the scattered-light problem (known to IUE users as the ``FES streak'') that limited photometric accuracy.

The original proposal was to use the GHRS, but loss of the ``blue side'' of the instrument shortly after the proposal submission required a change to the FOS.

The original time request was based on Monte Carlo simulations that were designed to determine the minimum number of observations that would be required to distinguish between emission-line lags of zero days and two days with 90% confidence. For daily observations with GHRS or FOS, the simulations indicated that 40 days of observations would be sufficient.

Because of the complex data reduction that would be required for this project, a full-time postdoctoral research associate was hired for the project. This position was filled by Kirk Korista. Three additional astronomers with significant experience in observations of AGNs with HST - Ross Cohen, Ian Evans, and Simon Morris - were added to the central committee (as the group of organizers became known) for this experiment.

Ian Evans, one of the STScI staff members responsible for scheduling HST recommended that we use one of the 40 allocated visits to NGC 5548 to test the acquisition procedures. This turned out to be excellent advice - the ``binary acquisition'' procedure, used in a test acquisition on 1993 February 3, failed on account of the combination of the poor pre-COSTAR point-spread function and the relatively weak contrast between the nucleus and the host galaxy. A more complex but reliable acquisition procedure was therefore indicated, but this would require a 15.5 more hours of spacecraft time, in addition to the 60.6 spacecraft hours already awarded to the project. STScI Acting Director Pete Stockman agreed to provide this extra time, without which the project almost certainly would have failed.

The FOS monitoring program began on 1993 April 19, and terminated after 39 daily ``visits'' on 1993 May 27. Each visit required three orbits, the first for slewing and target acquisition, the second for the short-wavelength exposure (with the G130L grating, 1155 - 1605 Å), and the third for the long-wavelength exposure (1574 - 2330 Å, with the G190L grating). While there were various serious photometric calibration problems, all 39 spectra turned out to be useful.

As with the original NGC 5548 program, a massive concurrent ground-based effort was undertaken contemporaneously. Unfortunately, the FOS program started about two weeks later than originally scheduled on account of the HST spacecraft going into a ``safe mode'' when the SADE-1 unit (Solar Array Deploy Electronics, which control the angle of the solar arrays relative to the telescope axis) failed. As a result, the densest part of the optical sampling ended before the end of the HST program.

The HST/IUE and ground-based data were published together in AGN Watch Paper VIII (Korista et al. 1995, ApJS, 97, 285). The principal results of this program were:

The NGC 3783 Project (1991-92)

At the same 1990 Heidelberg meeting that established the HST program on NGC 5548, it was decided to undertake a program similar to the original NGC 5548 program on another Seyfert 1 galaxy, again using IUE. The main goal of this program was to see how general the NGC 5548 results were and to begin to provide data to test relationship between the broad-line region radius R and the nuclear luminosity L, which making naive assumptions, should have a form R = C L² where C is a constant. The southern hemisphere Seyfert galaxy NGC 3783 was selected (a) because historical data indicated that it varied on time scales appropriate for this kind of experiment, (b) it is one of the brightest AGNs in the UV sky, and (c) it was intermediate in luminosity between two of the best-studied variable Seyfert 1 galaxies, NGC 4151 and NGC 5548.

Again joint ESA/SERC and NASA proposals were submitted, with Jean Clavel as PI of the ESA/SERC proposal and Brad Peterson as PI of the NASA proposal.

The ground-based program was difficult to organize on account of the relatively few candidate observatories in the southern hemisphere. However, the efforts of Danielle Alloin and Giovanna Stirpe at ESO and Jack Baldwin, Miriani Pastoriza, Brad Peterson, and Thaisa Storchi-Bergmann at CTIO led to large allocations for this project from both of these major observatories. These large programs were carried out in heroic fashion by Cláudia Winge at CTIO and by Bruno Altieri at ESO. Additional spectra were obtained with the 2.2-m Complejo Astronómico El Leoncito (CASLEO) telescope, with the 1.0-m Vainu Bappu Observatory (VBO) telescope, and with the 1.8-m Perkins Telescope at Lowell Observatory. Optical and near-IR photometry was obtained at the South African Astronomical Observatory (SAAO), Observatorio Astronómico Felix Aguilar (OAFA), and VBO.

The IUE program began on 1991 December 21. As with the original NGC 5548 program, the observations were made in 4-hour half-shifts, half at Vilspa and half at GSFC. The exposure pattern was also as before, an optimally exposed SWP followed by a optimally exposed LWP, with any remaining time used to obtain a redundant SWP image. The interval between observations was again nominally four days until 1992 June 10, after which the sampling interval was decreased to 2 days until the end of the program on 1992 July. The original proposals requested a switch from 4-day to 2-day intervals half way through the program; the motivation for this was to provide better constraints on the line-transfer functions than had been obtained in the NGC 5548 experiment. The reason for putting the denser sampling at the end of the campaign is because emission-line from the beginning of the campaign are of relatively little value because the lines are still responding to continuum events that occurred prior to the beginning of the program. It was not possible to carry out this strategy in practice because the time allocations were insufficient.

An operational change that occurred during this program was due to the appearance of the ``FES streak'', apparently sunlight scattered onto the FES (which was used both for guiding and as an unfiltered optical photometer), probably by loose insulation (IUE was already some eight years beyond the most optimistic prediction of its useful lifetime at this time, and various spacecraft components were showing signs of wear and tear). This made acquisition and guiding much more difficult than it had been previously, and thus the GSFC observations were turned over to only experienced IUE observers, usually on-site Computer Sciences Corporation personnel whose work was overseen and coordinated by Mike Crenshaw.

Optical spectroscopy began at ESO on 1991 December 3 and at CTIO on December 17. The ground-based program ended on 1992 August 9.

The IUE observations were published by Reichert et al. (1994, ApJ, 425, 582) and the optical observations by Stirpe et al. (1994, ApJ, 425, 609). The results of this program were basically consistent with those obtained in the NGC 5548 program; specifically,

The results were less spectacular than those obtained on NGC 5548 because (a) the emission-line variations were even more rapid than anticipated so that the responses were rather poorly resolved, and (b) there was evidence for unresolved continuum variability, especially (ironically enough) during the higher-intensity monitoring during the last 50 days of the program.

Another important feature of this program was the multiwavelength snapshot of NGC 3783 that was obtained during this campaign. Under the auspices of the ESA-sponsored World Astronomy Days program, Willem Wamsteker initiated this program to obtain a very broad-band continuum snapshot of NGC 3783. It turned out to be impossible to arrange simultaneous observations by all telescopes on a single day, so the ``snapshot'' was actually obtained on two days, 1992 June 29 and 1992 July 27. Data for this program were obtained by CGRO (0.05 - 0.25, 0.15 - 0.50, and 0.50 - 1.0 MeV), ROSAT (0.1 - 2.4 keV), Voyager 2 (92.5 - 112 nm), IUE (115 - 300 nm), HST (116 - 215 nm), the ESO 1.5-m telescope (480 - 840 nm), the CTIO 1.5-m telescope (370 - 820 nm) the SAAO 1.9-m telescope (1.25, 1.65, 2.20, and 3.40 µm) and the Very Large Array (20 cm, 3.6 cm, and 2 cm). The observations obtained in this program are described by Alloin et al. (1995, A&A, 293, 293).

The NGC 4151 Multiwavelength Project

Several members of the Central Committee attended the 33rd Herstmonceux meeting on "The Nature of Compact Objects in Active Galactic Nuclei'' in Cambridge, England, in 1992 July. At that time, observations for the NGC 3783 project were nearly completed, and intensive preparations were underway for the NGC 5548 HST program, which had been accepted the previous January. Much of the informal discussion among members of the Central Committee involved what might be done next with IUE. Among various members, there was strong enthusiasm for two different projects: (1) more critically testing the predicted R = C L1/2; relationship through monitoring of a higher-luminosity AGN, such as Fairall 9, and (2) extending the kind of program that had been done already to a different type of AGN, perhaps a broad-line radio galaxy such as 3C 390.3. Both of these programs were eventually implemented, but the consensus that arose in Cambridge was for a highly intensive multiwavelength monitoring program on NGC 4151, the brightest AGN in the sky. This was first suggested by Rick Edelson, who had been one of the key figures in a similar project on the bright blazar PKS 2155-304 which had been carried out with spectacular success the previous November. The behavior of non-blazar AGN continua on time scales of hours was really unexplored territory, although it was generally believed that any continuum variations on time scales shorter than the 4-day interval between NGC 5548 observations in 1989 must be no larger than around 5% or so. There was even less information on how the variations at different wavelengths might be related; there were a few contemporaneous UV/X-ray observations of a few sources that suggested that variations in the two bands were somehow correlated, but perhaps not very strongly, and the previous NGC 5548 monitoring had shown that the UV and optical continua varied in phase to within about 1-2 days. Another technical question that had arisen concerned the response of emission lines to continuum variations on the shortest time scales. Some reconstructions of emission-line transfer functions from the NGC 5548 data suggested that there was little, if any, emission-line response (at least for some lines) at zero time delay - this implied an absence of responding gas along our line of sight to the continuum source, either because there is no material there (e.g., a face-on disk) or because the line-emission is anisotropic and directly primarily back towards the central continuum source.

An intensive IUE program was envisaged as the way to address these problems. With IUE and ground-based data alone, it would be possible to determine to high accuracy the time delay between UV and optical continuum variations, and to explore for the first time small-scale continuum variations (``flickering'') in a non-blazar AGN. It was thus envisaged that the IUE program would serve as a cornerstone for a larger multiwavelength effort.

There were two important technical questions that had to be addressed in preparation of the proposal. First, could we achieve sufficiently high signal-to-noise ratios with IUE, and could the exposures be short enough to provide sufficiently high time resolution? Second, could useful IUE observations be made during the relatively high background US2 shifts? There was also some concern about whether or not a non-blazar AGN might be completely quiescent during a 10 to 15-day program and show no continuum variability at all!

One consequence of these concerns was that there was never any question about the target - NGC 4151 is by far the brightest known AGN in the UV, and it had been well-studied with IUE in the past. It was known to be variable on short time scales and the emission-line lags were known to be of order days. It was also argued that the IUE cameras would produce exceptionally good data with uninterrupted observations of a relatively faint source. Furthermore, the opportunity to achieve very high-quality spectra was enhanced by the appearance of Tom Ayres' new spectral extraction software which minimized the microphonics that had limited IUE signal-to-noise ratios in the past (this extraction method, known as TOMSIPS, was used in the NGC 4151 and NGC 7469 projects).

As mentioned earlier, one of the original principal goals of this project was to determine whether or not the emission lines had little response at zero time delay, as had been suggested by some of the attempts to reconstruct transfer functions for the NGC 5548 data. Monte Carlo simulations (mostly by Keith Horne and Kirk Korista) indicated that we would be able to resolve this question with the proposed IUE program.

The proposed IUE program was for 13 Vilspa shifts, 9 US1 (low-background) shifts and 19 US2 (high-background) shifts. The requested sampling plan was to use one-half of a Vilspa or US2 shift (4 hours) per day for 12 days, followed by 9.6 days of continuous monitoring, followed by one half of a Vilspa or US2 shift for 12 more days. Unfortunately, the actual allocation was smaller, and in the end precluded any progress on the emission-line variability. (This explains, incidentally, why the NGC 4151 papers constituted a series numbered separately from the Steps Toward Determination of the Size and Structure of the Broad-Line Region...; the shortened program yielded little new information on the emission-line regions). The program as implemented consisted of one-half of a US2 shift for 5 days, followed by 9.6 days of continuous monitoring, and ending with one-half of a US2 shift each day for 5 days.

The approved IUE program was used as the cornerstone upon which a very large multiwavelength program was built. Intensive ground-based observations were also obtained, anchored by daily or better monitoring with the Wise Observatory 1.0-m telescope and the Perkins 1.8-m telescope. Proposals for monitoring the high-energy spectrum of NGC 4151 were submitted for observations with ROSAT, ASCA (known at the time as Astro-D), and Compton Gamma-Ray Observatory.

Organization of this project was considerably more difficult than the previous projects had been on account of the larger number of spacecraft involved, each with its own complex pressures and operational difficulties. Problems were exacerbated by the absence of NASA funding for IUE observations. The complex nature of the project required extra attention from the Central Committee, and many of the organizational problems were ironed out at a meeting in Geneva in 1993 September. (Several of the project principals met again at the 1994 November AAS HEAD meeting in Napa, California, to iron out final details on how the data papers would be organized.)

Despite the many operational difficulties encountered, the program was successfully executed beginning in late 1993 November. Spacecraft problems limited the ASCA program to only four observations, and the ROSAT observations, which were to be made once every 12 hours during the intensive part of the IUE campaign, missed four days due to a spacecraft problem. CRGO were made almost daily for 12 days. In all bands, the variations were of lower amplitude than in previous monitoring campaigns, primarily on account of the relatively short duration of the monitoring program. No significant variations were detected by CGRO, and the optical data base had to be restricted to only the highest-quality observations for the variations to be detectable.

The results of this multiwavelength program were published in four consecutive papers in the 1996 October 10 issue of the Astrophysical Journal: Paper I (Crenshaw et al. 1996, ApJ, 470, 322) on the IUE observations, Paper II (Kaspi et al. 1996, ApJ, 470, 336) on the optical ground-based observations, Paper III (Warwick et al. 1996, ApJ, 470, 349) on the high-energy (ROSAT, ASCA, and CGRO) observations, and Paper IV (Edelson et al. 1996, ApJ, 470, 364) on multiwavelength analysis. The basic results of this program are:

More to follow...


Copyright © Bradley M. Peterson, 1997.

[OSU Astronomy Home Page] [AGN Watch Home Page]
Updated 24 November 1997
peterson@astronomy.ohio-state.edu