In principle, I am suspicious of citations as a measure of importance --- they are affected by so many random factors. I have to concede, however, that there is a pretty good correlation between the most highly cited of my papers and the papers I consider most significant myself, even if the rankings aren't identical. I can thus give a reasonably good guide to my main science contributions by running through a summary of my most highly cited papers, with comments about a few other papers at the end.
I first went through this exercise in October 2010, when it was a list of my 20 most cited papers. When I updated it in 2016 I was reluctant to chop out papers that I had written about previously (and am very fond of!), so I expanded to 25. Now that I'm updating again in 2019, I've decided to adopt a criterion that won't require me to unwrite things I've written: this is a list of my papers with 200 or more ADS citations, ordered with the most highly cited at the top.
But not all of those papers. As a long-time member of the Sloan Digital Sky Survey -- I joined the SDSS as a postdoc in 1992, continued as an external participant after moving to Ohio State in 1995, and became Spokesperson of SDSS-II in 2005 and Project Scientist of SDSS-III in 2007 -- I am a co-author on all of the SDSS data release papers and a number of SDSS technical papers, which collectively dominate the top of my "most cited" list. I have also co-authored many SDSS science papers in which I made a contribution but was not centrally involved in the analysis and writing of the paper. I have omitted all of these from the list below, keeping only those SDSS papers where I played a central role. As of 4/1/2019, this selection gives 40 papers with 200+ citations.
As is immediately evident from this list, I have been very fortunate to work with many terrific collaborators over the years, including a number of great thesis students and postdocs.
Numbers below correspond to the order given by ADS as of 4/1/2019. For those who like h-indices, mine (as of 4/1/19) is 127 if you include all of the SDSS papers, or 90 if you omit SDSS papers in which I am not among the first four authors.
Enough preamble, here is the list:
1. How do galaxies get their gas?
D. Keres, N. Katz, D. H. Weinberg, R. Davé
2005, MNRAS, 363, 2
This paper uses cosmological smoothed particle hydrodynamics (SPH) simulations to argue that galaxies gain a large fraction of their mass by "cold mode accretion" of gas that was never shock heated near the virial temperatures of their parent halos, and that this cold accretion occurs preferentially along filamentary structures feeding low mass (< 1012Msun) halos. One can find precedents for this idea in many earlier analytic papers (as far back as Binney 1977) and numerical simulation papers, including some of our own. Nonetheless, it was the Keres et al. paper and the contemporaneous papers by Birnboim and Dekel (2003, 2006) using high-resolution 1-D simulations that really put "cold accretion" on the map as a key element in understanding the physics of galaxy formation. While this phenomenon is by now robustly established in numerical simulations, the observational evidence for it remains circumstantial.
2. Spectroscopic Target Selection in the Sloan Digital Sky Survey:
The Main Galaxy Sample
M. Strauss, D. H. Weinberg, R. H. Lupton, V. K. Narayanan, J. Annis, M. Bernardi, et al.
2002, AJ, 124, 1810
Despite the qualification above, the second paper on my list is in fact an SDSS technical publication. My most important technical contribution to the original SDSS was my role in designing and testing (on simulated data and, eventually, on real data) the selection criteria for the main spectroscopic galaxy sample. If you hate Petrosian magnitudes, you largely have me (and Petrosian) to blame.
3. SDSS-III: Massive Spectroscopic Surveys of the Distant Universe,
the Milky Way, and Extra-Solar Planetary Systems
D. J. Eisenstein, D. H. Weinberg, E. Agol, H. Aihara, C. Allende Prieto, S. F. Anderson, et al.
2011, AJ, 142, 72
This paper presents an overview of the SDSS-III program and its four component surveys: the BOSS survey of the distant universe aimed at precision cosmological measurements, the SEGUE and APOGEE surveys of the structure and chemical evolution of the Milky Way, and the MARVELS search for extra-solar planets. While the surveys and the tools that enabled them were individually described in many further technical articles, this paper gave a high-level overview of the full project in the early stages of its execution, analogous to the York et al. paper for SDSS-I (which is, as it happens, the most highly cited paper I have co-authored, by a good margin). One of the most important contributions of my career has been serving as Project Scientist of SDSS-III, which in practice has meant a myriad of interestingly varied tasks related to the design, organization, promotion, and execution of the project. One of those tasks was to spearhead the organization and writing of this paper, though Daniel is rightly the first author in recognition of his indispensable and superb work as SDSS-III Director.
4. Cosmological Simulations with TreeSPH
N. Katz, D. Weinberg, L. Hernquist
1996, ApJS, 105, 19
The Keres et al. paper is a recent installment of a program that Lars Hernquist, Neal Katz, and I began in the early 1990s, using SPH simulations to model the formation of galaxies and evolution of the intergalactic medium (IGM) in cold dark matter cosmological models. This paper describes our numerical methods for these simulations and presents numerous illustrations and tests. It therefore underpins many of our own papers, and it has been a useful reference point for many subsequent efforts in cosmological hydrodynamics, especially for its discussions of gas cooling, photoionization, and star formation.
In more recent years, Neal and I have continued this program in collaboration with Romeel Davé and many others, while Lars has continued in collaboration with Volker Springel and many others.
5. Observational probes of cosmic acceleration
D. H. Weinberg, M. J. Mortonson, D. J. Eisenstein, C. Hirata, A. G. Riess, E. Rozo
2013, Physics Reports, 530, 87-255
This article presents a book-length review of the observational and experimental methods being used to probe the origin of cosmic acceleration. It is oriented towards the goals and capabilities of "Stage III" and "Stage IV" dark energy experiments, which include some of the most ambitious cosmological survey projects now being undertaken around the world. After historical and scientific introduction that are intended to make it accessible to any reader who has taken a graduate cosmology course, it provides detailed discussion of theoretical background, observational strategies, statistical errors, and systematic uncertainties for each of the major experimental methods. A central theme of the review is the complementarity of different methods, including extensive forecasts of how they work together and the levels of precision required to remain competitive. I was fortunate to assemble an all-star cast of co-authors for this review, and I think we all learned a lot from each other over the three years it took us to complete it. As a work of scientific literature, it is probably the best thing I have written.
With Michael Mortonson and Martin White, I have written a (much) shorter review with more focus on recent observational results, for the Particle Data Group Review of Particle Physics. This review is updated every 2 years; the most recent version is available here.
6. Reionization and the Abundance of Galactic Satellites
J. S. Bullock, A. V. Kravtsov, D. H. Weinberg
2000, ApJ, 539, 517
In 1999, Moore et al. and Klypin et al. showed that the cold dark matter scenario generically predicts an enormous population of low mass subhalos within the Milky Way halo, far exceeding the known population of Galactic satellites. This paper proposed a potential solution to this "missing satellite problem": that gas accretion was suppressed in low mass halos by photoionization heating, and that the observed satellites were those that accreted a significant fraction of their mass before reionization. The full explanation of the relation between dark matter subhalos and Galactic satellites is still a matter of debate, but it is now widely believed that the resolution to this puzzle lies in baryonic physics rather than dark matter physics, and (somewhat less) widely believed that photoionization is the critical mechanism that suppresses the luminosity of the low mass subhalos. This paper relies heavily on Thoul & Weinberg 1996, described below. More recently, Koposov et al. 2009 revisited this model in the light of improved numerical underpinnings and, more importantly, the ultra-faint dwarf satellite population discovered by the SDSS.
7. The Halo Occupation Distribution: Toward an Empirical Determination
of the Relation between Galaxies and Mass
A. A. Berlind, D. H. Weinberg
2002, ApJ, 575, 587
This paper, based on work in Andreas Berlind's PhD thesis, proposed to model the relation between galaxies and dark matter in terms of the probability distribution P(N|M) of finding N galaxies of a given type in a halo of virial mass M, together with prescriptions specifying the relative spatial and velocity distributions of galaxies and dark matter within these halos. The direct inspiration for our paper came from Benson et al. (2000), but a number of groups coming at the problem of matter clustering and galaxy bias from different directions had proposed closely related ideas, including Kauffmann et al. 1997 , Jing, Mo & Borner 1998 , Ma & Fry 2000 , Peacock & Smith 2000 , Seljak 2000 , and Scoccimarro et al. 2001 . The distinctive strengths of our paper, I think, were to present halo occupation statistics as a general tool for modeling galaxy bias and to lay out a general strategy for using this tool both to infer relations between galaxies and dark matter and to sharpen cosmological parameter constraints. Reflecting this philosophy, we introduced the now widely used term "halo occupation distribution" or "HOD". We used N-body simulations to show how a variety of galaxy clustering statistics depended on different features of the HOD, with particular attention to the HOD properties required to obtain a nearly power-law two-point correlation function.
8. Baryons in the Warm-Hot Intergalactic Medium
R. Davé, R. Cen, J. P. Ostriker, G. L. Bryan, L. Hernquist, N. Katz, D. H. Weinberg, M. L. Norman, B. O'Shea
2001, ApJ 552, 473
In hydrodynamic cosmological simulations, much of the low-redshift IGM is comprised of gas at moderate densities and temperatures of 105K < T < 107K, which is very difficult to detect. Cen & Ostriker (1999) labeled this gas the warm-hot intergalactic medium, a clunky term with a catchy acronym. This paper used three different numerical codes to demonstrate the robustness of WHIM predictions, showing that it should contain 30-40% of the baryons in the present-day universe. This prediction has inspired numerous observational programs using UV spectroscopy, X-ray spectroscopy, and soft X-ray emission to try to find these "hidden" baryons.
9. The Luminosity and Color Dependence of the Galaxy Correlation Function
I. Zehavi, Z. Zheng, D. H. Weinberg, J. A. Frieman, A. A. Berlind, M. R. Blanton, et al.
2005, ApJ, 630, 1
This paper analyzed a redshift sample of about 200,000 SDSS galaxies, focusing on the projected two-point correlation function of subsamples defined by cuts in luminosity and color. In addition to providing some of the best measurements of these trends (with the other best measurements coming from the analyses of the 2dF Galaxy Redshift Survey by Norberg et al. 2001 and Norberg et al. 2002 ), this paper pioneered the use of Halo Occupation Distribution (HOD) modeling to interpret these trends in terms of the relations between galaxy properties and the masses of their host dark matter halos (see Berlind et al. 2002 and Zheng et al. 2005 below). From the measurements and modeling, we concluded that the luminosity dependence of clustering arose mainly from the trend of increasing halo mass with increasing luminosity for central galaxies, while the color dependence was strongly affected by the larger fraction of red galaxies that were satellite systems in massive halos. We also found a roughly constant ratio of 23 between the minimum halo mass required to host a central galaxy above a luminosity and the halo mass required to host, on average, one satellite galaxy above the same threshold. This rather large gap, a consequence of the rarity of near-equal mass halo mergers, has profound effects on galaxy clustering. We carried out a major update of this work using the final SDSS-II data set (DR7) and taking advantage of subsequent improvements in modeling methods in Zehavi et al. 2011.
10. Galaxy Clustering in Early Sloan Digital Sky Survey Redshift Data
I. Zehavi, M. R. Blanton, J. A. Frieman, D. H. Weinberg, H. J. Mo, M. A. Strauss, et al.
2002, ApJ, 571, 172
This paper presented the first measurements of galaxy clustering from the SDSS redshift survey, specifically the two-point correlation function for a sample of about 30,000 galaxies from the early data release. While there were no hugely surprising results, the measurements showed very clearly the small and large scale signatures of redshift-space distortions induced by galaxy peculiar velocities, the difference in real-space and redshift-space clustering between red and blue galaxies, and scale-independent luminosity bias large separations. It established the basic tools and methodology for subsequent analyses of larger samples (e.g., Zehavi et al. 2005), which led to a number of novel results.
11. Galaxies in a simulated ΛCDM Universe - I. Cold mode and hot cores
D. Keres, N. Katz, M. Fardal, R. Davé, D. H. Weinberg
2009, MNRAS, 395, 160
This paper is a follow-up of Keres et al. 2005, investigating cold and hot accretion in SPH simulations that had 10 times larger volume than our 2005 simulations and used a different code (Volker Springel's GADGET) with a different implementation of the SPH equations. This study confirmed two key results of the 2005 paper, the major role of cold accretion and the transition to hot gas halos at a dark matter halo mass of roughly 5e11 solar masses, and it provided much better quantification of the evolution of accretion rates onto central and satellite galaxies. The main difference from 2005 was the formation of cores in the hot gas halos of the GADGET simulations leading to much lower rates of hot gas accretion onto galaxies. This difference appears to arise from the difference in the SPH formulations. While a range of numerical methods (SPH, adaptive grid codes, moving mesh grid codes) over a wide range of resolution all agree on the dominance of cold accretion in low mass halos and a transition to shocked gas halos at high masses, they give widely differing predictions for the amount hot gas that cools onto galaxies. This is probably the largest numerical uncertainty in "weak feedback" simulations of galaxy formation, though the implementation of galactic winds powered by star formation or AGN feedback ultimately has a larger impact on predicted galaxy properties.
12. Theoretical Models of the Halo Occupation Distribution:
Separating Central and Satellite Galaxies
Z. Zheng, A. A. Berlind, D. H. Weinberg, A. J. Benson, C. M. Baugh, S. Cole, R. Davé, C. S. Frenk, N. Katz, C. G. Lacey
2005, ApJ, 633, 791
This paper re-examined theoretical predictions of the halo occupation distribution in SPH simulations and semi-analytic models of galaxy formation, first undertaken by Berlind et al. 2003. Here we revisited these models models in light of the HOD description proposed by Kravtsov et al. 2004 based on dark matter subhalos. We found that this description also applied well to our SPH and semi-analytic galaxy populations: the HOD for galaxies above a mass or luminosity threshold is well characterized as the sum of a step-function like component for central galaxies and a power-law component (approximately linear) for satellite galaxies, with the two components offset in mass by a roughly constant factor. We also showed that the conditional luminosity function (see Yang, Mo, & van den Bosch 2003 ), the distribution of galaxy luminosities in halos of fixed mass, is well described by the sum of a log-normal component for the central galaxy and a truncated Schechter function for the satellites, with the global Schechter function emerging only when one sums over the halo population. There was a nice synergy between this theory paper and the contemporaneous Zehavi et al. SDSS paper: when modeling the luminosity and color dependence of the SDSS galaxies, we found that we needed better theoretical models to motivate our parameterizations, and once we developed these improved theoretical models, we found that several trends that had seemed surprising or puzzling in the data were in fact natural predictions of the theory. The HOD parameterizations proposed in this paper are widely used in analyses of observed galaxy clustering and creation of mock catalogs from simulations.
13. Galaxy Clustering in the Completed SDSS Redshift Survey:
The Dependence on Color and Luminosity
I. Zehavi, Z. Zheng, D. H. Weinberg, M. R. Blanton, N. A. Bahcall, A. A. Berlind, et al.
2011, ApJ, 736, 59
As mentioned previously in the discussion of Zehavi et al. 2005, this paper brings to completion the program of measuring the luminosity and color dependence of the galaxy correlation function in the SDSS Main Galaxy Sample, this time from the final data release of SDSS-II (DR7), and interpreting the measurements via HOD modeling. Because of the data set and the modeling techniques, it has become the standard reference on the HODs of different classes of galaxies (especially luminosity-thresholded samples) in the present-day universe. I find myself referring to its Table 3 (and to a lesser extent Table 2) constantly.
14. The Lyman-Alpha Forest in the Cold Dark Matter Model
L. Hernquist, N. Katz, D. H. Weinberg, J. Miralda-Escudé
1996, ApJ, 457, L51
This paper, alongside earlier papers by Cen et al. 1994 and Zhang, Anninos & Norman 1995 , transformed the understanding of the Lyman-alpha forest by showing that it was naturally predicted in hydrodynamic simulations of the CDM cosmological model, as quasar lines of sight traversed the photoionized intergalactic medium and intersected concentrations of neutral hydrogen. While our paper was the third of this group, I think it was particularly influential for advancing the view that the low column density Lyman-alpha forest could really be regarded as a form of the Gunn-Peterson effect arising in a smoothly fluctuating, continuous medium, a view that we developed quantitatively in subsequent work and is now well established. We also pointed out that for standard estimates of the photoionizing background intensity, explaining the observed opacity of the forest required a baryon density at the high end of then-current estimates. (For closely related arguments from an analytic or semi-analytic point of view, see Rauch & Haehnelt 1995, Bi & Davidsen 1997, and Hui, Gnedin, & Zhang 1997 , which are among my favorite Lyman-alpha forest papers.)
15. The Low-Redshift Lyman-Alpha Forest in Cold Dark Matter Cosmologies
R. Davé, L. Hernquist, N. Katz, D. H. Weinberg
1999, ApJ, 511, 521
This paper extended our numerical studies of the Lyman-alpha forest to low redshift, comparing to observations from Hubble Space Telescope. In addition to showing generally good agreement with these observations, we established several key features of IGM evolution: (1) evolution of the Lyman-alpha forest slowed at low redshift as decreasing UV background intensity began to compete with decreasing mean density from cosmic expansion; (2) the mapping between Lyman-alpha column density and gas overdensity shifted at lower redshifts, so that typical 1014 cm-2 absorbers at high z were analogous to weaker 1013 cm-2 absorbers at low z, and the typical absorbers found in low redshift searches were analogous to strong absorbers at high redshift, explaining their stronger correlations with galaxies; (3) at z < 1, a substantial fraction of the IGM is shock heated to temperatures T > 105K and compressed to moderate overdensities (10-100), the component now usually referred to as the WHIM (see Dave et al. 2001 above), though the unshocked, diffuse medium remains a major baryonic component and produces most Lyman-alpha forest absorption. An update of this paper that draws on better simulations and observational lessons learned in the ensuing decade can be found in Davé et al. 2010.
16. Hydrodynamic Simulations of Galaxy Formation.
II. Photoionization and the Formation of Low-Mass Galaxies
A. Thoul, D. H. Weinberg
1996, ApJ, 465, 608
This paper used 1-dimensional, spherically symmetric simulations of the collapse of density peaks to investigate the impact of a photoionizing background on the formation of low mass galaxies. We included both collisionless dark matter and baryonic matter with a full treatment of photoionization and radiative cooling. We found that photoionization suppresses formation of galaxies (more precisely, cooled baryonic components) in halos with circular velocity Vc < 30 km/s and has little impact on halos with Vc > 75 km/s, with a steady trend in between. In a cute experiment, we carried out simulations with suppressed cooling but no heating and with heating but unsuppressed cooling to demonstrate that heating of the gas near turnaround was the dominant effect, making the scale of suppression higher than one might naively expect for 104K gas. Recent work using very high resolution 3-d simulations has largely held up this paper's basic results, but it has shifted the characteristic suppression scale somewhat lower, to ~ 25 km/s (Hoeft et al. 2006, Okamoto et al. 2008 ).
17. The Opacity of the Ly alpha Forest and Implications for
OmegaB and the Ionizing Background
M. Rauch, J. Miralda-Escudé, W. L. W. Sargent, T. A. Barlow, D. H. Weinberg, L. Hernquist, N. Katz, R. Cen, J. P. Ostriker
1997, ApJ, 489, 7
This paper drew on the numerical simulations described above (under Hernquist et al. 1996) and on a set of high-resolution Keck Lyman-alpha forest spectra by Rauch, Sargent, and Barlow to infer lower limits on the mean baryon density. The mean opacity of the forest depends on the baryon density (squared) divided by the photoionization rate from the UV background. A lower limit on the latter from the quasar luminosity function then leads to a lower limit on the former, but it requires both a measurement of the mean opacity and a theoretical model of the forest. (However, Weinberg et al. 1997 showed that the conclusion depends on the general features of the model rather than the fine details, making it insensitive to imperfections in the numerical simulations.) At the time of this paper, two groups were getting quite different estimates of the primordial deuterium abundance from Keck spectra, a "high" deuterium abundance implying (via big bang nucleosynthesis) a low baryon density and a "low" deuterium abundance implying a high baryon density. Our paper provided a strong argument for the low deuterium/high baryon density solution, which has been borne out by subsequent observations.
18. Toward a Precise Measurement of Matter Clustering: Lyman-alpha
Forest Data at Redshifts 2-4
R. A. C. Croft, D. H. Weinberg, M. Bolte, S. Burles, L. Hernquist, N. Katz, D. Kirkman, D. Tytler
2002, ApJ, 581, 20
20. Recovery of the Power Spectrum of Mass Fluctuations from Observations of the Lyman-alpha Forest
R. A. C. Croft, D. H. Weinberg, N. Katz, L. Hernquist
1998, ApJ, 495, 44
It's easiest to discuss these two papers together, slightly out of order. Building on the "continuously fluctuating medium" view of the Lyman-alpha forest advocated in Hernquist et al. 1996 and elaborated in Rauch et al. 1997 and Croft et al. 1997(on intergalactic helium), the 1998 paper (listed second) proposed measuring the 1-dimensional power spectrum of transmitted flux in Lyman-alpha forest spectra and using it to infer the power spectrum of the underlying matter distribution. Tests on spectra from our SPH simulations showed that this technique worked remarkably well, providing a novel tool to probe matter clustering in the high-redshift universe. To my knowledge, this paper is the first journal article with the equation generally known as the "fluctuating Gunn-Peterson approximation" or FGPA, a term that I introduced in an earlier conference proceedings. Using this approximate physical picture, we argued that the similarity of shape between the flux and matter power spectra should be robust and, crucially, that measurements of the mean opacity of the forest could effectively determine the "bias factor" of the forest, allowing one to infer the amplitude as well as the shape of the matter power spectrum.
The 2002 paper applied this technique (with some refinements) to a sample of 30 Keck HIRES spectra (obtained for other purposes) and 23 Keck LRIS spectra (obtained in a one-night observing run specifically for this purpose). We obtained what were then much the strongest constraints on matter clustering at redshift z = 2-4 and derived corresponding constraints on cosmological parameters. Consistency of our constraints with estimates from independent data provided a novel test of the Lambda-CDM scenario (inflationary cold dark matter with a cosmological constant) that was then emerging as the standard cosmological model. These papers, together with Weinberg et al. 1998 (a conference proceedings), Croft et al. 1999, and McDonald et al. 2000 established many of the basic principles of Lyman-alpha forest cosmology that are still being used today. Among other things, they showed that moderate resolution spectra, in large numbers, could provide high-precision cosmological constraints, a key insight underlying analyses of the SDSS Lyman-alpha forest ( McDonald et al. 2005) and the BOSS quasar survey of SDSS-III. Major recent developments in the field include a beautiful high-precision measurement of the 1-d flux power spectrum in the BOSS Lyman-alpha forest ( Palanque-Delabrouille et al. 2014) and the detections of large-scale three dimensional clustering and baryon acoustic oscillations in the BOSS Lyman-alpha forest as reported by Slosar et al. 2011, Busca et al. 2013, Slosar et al. 2013, and Delubac et al. 2014.
19. Black Hole Masses and Eddington Ratios at 0.3 < z < 4
J. A. Kollmeier, C. A. Onken, C. S. Kochanek, A. Gould, D. H. Weinberg, M. Dietrich, R. Cool, A. Dey, D. J. Eisenstein, B. T. Jannuzi, E. LeFloc'h, D. Stern
2006, ApJ, 648, 128
This paper applied virial black hole mass estimators to quasar spectra from AGES (the AGN and Galaxy Evolution Survey, led by Kochanek, Eisenstein, and the second half of the above author list) to show that the distribution of Eddington ratios (the ratio of the quasar luminosity to the Eddington limit for its central black hole) is approximately log-normal with a mean of 1/4 and a width of less than 0.3 dex, with an inequality rather than an estimate because the observed width also includes observational errors in the black hole mass estimates, whose magnitude is poorly known. (Indeed, one somewhat surprising result of this paper is to limit those observational errors to less than 0.3 dex rms.) This result holds for several different bins of luminosity and redshift, with a slight trend towards lower mean ratios at lower redshift. Crucially, for some ranges we were able to show that it holds in bins of black hole mass, and that the drop in the distribution towards low Eddington ratio was not a consequence of sample selection. The emergence of a characteristic Eddington ratio rather than a distribution rising continuously (or staying flat) to low values suggests that, in the quasar regime, mass accretion rates are controlled by black-hole self-regulation rather than by galaxy scale dynamical events. While I am proud to be on this paper, I have to say that my role was really as a kibbitzer, and that I deserve a tiny fraction of credit relative to the first two authors, who conceived the project and did nearly all of the analysis, and relative to the authors who actually collected the data set in the first place.
21. Feedback and recycled wind accretion: assembling the
z = 0 galaxy mass function
B. D. Oppenheimer, R. Davé, D. Keres, M. Fardal, N. Katz, J. A. Kollmeier, D. H. Weinberg
2010, MNRAS, 406, 2325
Our papers on cold and hot mode accretion (Keres et al. 2005 and 2009) used simulations that did not have an explicit recipe for driving galactic winds, with the result that they produced overly massive galaxies and ejected little enriched material to the circumgalactic or intergalactic medium. This paper returned to similar topics using the formulations of galactic winds that Oppenheimer and Davé had developed in earlier papers focused on intergalactic metals. The key result of this paper is that "recycling" -- the reaccretion of material ejected in winds -- plays a major role in shaping the mass function of galaxies, becoming the dominant mode of mass growth for massive galaxies at low redshift. Below the characteristic mass M*, recycling improves the agreement with the observed galaxy mass function, but above M* it makes galaxies even more overmassive. A plausible solution to this conundrum is AGN feedback, not included in the simulations here.
The linked topics of accretion, winds, recycling, and the circumgalactic medium (CGM) are often referred to collectively as "the baryon cycle," and this paper played a significant role in framing this way of thinking. Our SPH collaboration has subsequently devoted a lot of effort to more detailed modeling of galactic wind launch and, more importantly, to the evolution of gas once it enters the CGM, as this will have a major impact on growth by recycling.
22. Wide-Field InfrarRed Survey Telescope-Astrophysics Focused Telescope
Assets WFIRST-AFTA 2015 Report
D. Spergel, N. Gehrels, C. Baltay, D. Bennett, J. Breckenridge, M. Donahue, et al.
The Wide Field Infrared Survey Telescope (WFIRST) is the highest priority large space initiative recommended by the Astro2010 decadal survey, in part thanks to input from the panel and working group reports I helped to write. I have served on a succession of WFIRST science definition teams starting in 2011, and I am currently (2019) a member of the WFIRST Formulation Science Working Group and the Cosmology with the High Latitude Survey science investigation team (PI O. Doré). This report presented a detailed mission concept based on a 2.5-m telescope (AFTA) transfered to NASA by another agency; the current (2019) mission concept remains close to the one described in this report, but with a number of refinements and changes. I wrote a lot of the dark energy section of this report (together with Chris Hirata, Yun Wang, Saul Perlmutter, Charlie Baltay, and Jason Rhodes) and contributed to other sections. As background for Astro2020, I recently spearheaded a white paper that briefly summarizes the current mission concept, titled The Wide Field Infrared Survey Telescope: 100 Hubbles for the 2020s.
I am an enthusiast for WFIRST in part because of its potential for revealing the origin of cosmic acceleration, which I consider the single most important problem in contemporary cosmology, but also because it marries the revolutionary strengths of the Sloan Digital Sky Survey and the Hubble Space Telescope. If WFIRST survives the fiscal challenges ahead and makes it to launch, I hope that my contributions to its success will be a bookend to my contributions to the SDSS.
23. The Large, Oxygen-Rich Halos of Star-Forming Galaxies Are a Major
Reservoir of Galactic Metals
J. Tumlinson, C. Thom, J. K, Werk, J. X. Prochaska, T. M Tripp, D. H. Weinberg, et al.
2011, Science, 334, 948
The installation of the Cosmic Origins Spectrograph (COS) during the Hubble Space Telescope 5th servicing mission opened a new opportunity to probe the metal and gas content of galactic halos through sensitive UV spectroscopy of background quasars. Jason Tumlinson put together a team to carry out an ambitious systematic survey of 42 galaxies, dubbed COS-Halos, including our SPH group as theory collaborators. This paper presented the first major results from COS-Halos, showing abundant OVI absorption out to distances of 150 kpc around star-forming galaxies implying an amount of oxygen in the CGM comparable to or larger than that in the galaxies' interstellar medium. This paper also showed substantially less OVI in the gaseous halos of passive galaxies, though active and passive galaxies show similar amounts of neutral hydrogen detected in Lyman-alpha absorption. Many simulation groups including ours (e.g., Ford et al. 2016) have carried out detailed comparisons to the COS-Halos results, and COS-Halos has itself inspired many similar studies of different populations of galaxies. A synthesis of the final results from COS-Halos is presented by Werk et al. 2014.
24. Self-Consistent Models of the AGN and Black Hole Populations: Duty Cycles, Accretion Rates, and the Mean Radiative Efficiency
F. Shankar, D. H. Weinberg, J. Miralda-Escudé
2009, ApJ, 690, 20
One of the most elegant results in the study of AGN is the argument due to Soltan and Paczynski that links the integrated emissivity of the quasar population to the mean present-day mass density of relic black holes. In this paper we extended this technique to predict the full shape of the relic black hole mass function based on the full shape and redshift evolution of the quasar luminosity function. This calculation constrains the mean radiative efficiency of quasar accretion and the Eddington ratios and duty cycles of the accreting black holes. I think this paper has been influential in part because we explained these connections clearly and in part because Francesco Shankar devoted enormous effort to compiling and comparing the relevant observational results. Observational uncertainties remain a challenge in this line of investigation, and Francesco remains a leader in reconciling a multitude of results into a physically coherent picture. My own involvement in this field traces back to my 2001 paper with Paul Martini on quasar clustering and its connection to quasar lifetimes.
25. On the Mass-to-Light Ratio of Large-Scale Structure
J. L. Tinker, D. H. Weinberg, Z. Zheng, I. Zehavi
2005, ApJ, 631, 41
This paper, a chapter of Jeremy Tinker's PhD thesis, was our first attempt to derive cosmological parameter constraints from HOD modeling. We showed that if we fit the observed SDSS galaxy correlation function (Zehavi et al. 2005) in a cosmological model with the then-standard parameter values Omegam=0.3 and sigma8=0.9, we predicted mass-to-light ratios of clusters higher than observed. From a joint fit we inferred (sigma8/0.9)(Omegam/0.3)0.6 = 0.75 +/- 0.06, perfectly anticipating the change in preferred parameter values that came with the transition from the 1st-year WMAP CMB data to the 3rd-year data (mainly because WMAP1 had overestimated the electron scattering optical depth as a result of underestimated polarization foregrounds). Along with papers making closely related arguments by Vale & Ostriker and van den Bosch, Mo, & Yang, this paper was one of the early cosmological successes of halo-based approaches to galaxy clustering. We revisited this method with many improvements in the data and the modeling in Tinker et al. 2011.
26. Cosmological implications of baryon acoustic oscillation measurements
E. Aubourg, S. Bailey, J. E. Bautista, F. Beutler, V. Bhardwaj, D. Bizyaev, et al.
2015, Phys Rev D, 92, 13516
This paper synthesized baryon acoustic oscillation (BAO) measurements from the SDSS-III Baryon Oscillation Spectroscopic Survey (BOSS) into constraints on a wide range of cosmological models, in combination with cosmic microwave background (CMB) and Type Ia supernova (SN) data. I particularly liked the approach in this paper, devised largely by Anze Slosar, for combining BAO measurements with a 3-parameter data vector summarizing the CMB constraints most relevant to cosmic geometry, which allowed us to give intuitive explanations rather than "black box" results. Among our key results: using only the sound horizon calibration from the CMB, BAO and SN data combine to give a tight "inverse distance ladder" constraint of H0 = 67.3 +/- 1.1 km/s/Mpc even with very flexible dark energy models; a curvature constraint Omegak = -0.003 +/- 0.003 assuming constant dark energy; a constraint w = -0.97 +/- 0.05 for a flat wCDM cosmology; a 0.5 eV upper limit on neutrino mass from expansion history alone; tension at the 2-2.5 sigma level between the BOSS Lya forest BAO measurements and any dark energy model that fits the CMB and galaxy BAO data; and tension between dark energy models constrained by the expansion history data and many measurements of matter clustering at low redshift. This paper represents the collective efforts of a large and talented collaboration; Slosar and I did a lot of the wrangling and writing. The analysis of the final BOSS galaxy clustering measurements by Alam et al. 2017, which includes constraints from the broad band power spectrum and redshift space distortion as well as BAO, supersedes this paper in terms of cosmological constraints, but the intuitive explanations make Aubourg et al. my personal favorite among the BOSS cosmology papers.
27. Percolation Galaxy Groups and Clusters in the SDSS Redshift Survey:
Identification, Catalogs, and the Multiplicity Function
A. A. Berlind, J. A. Frieman, D. H. Weinberg, M. R. Blanton, M. S. Warren, K. Abazajian, et al.
2006, ApJ Supp, 167, 1
In Berlind & Weinberg (2002) we concluded that the multiplicity function of galaxy groups would be a powerful tool for constraining the HOD and breaking degeneracies between galaxy bias and cosmology. As a postdoc at Chicago, Andreas Berlind set forth to construct a group catalog from the SDSS and measure the multiplicity function. This paper presents Andreas's group catalog and careful investigation of contamination and incompleteness biases with matched analysis of simulations; the group catalog has been used by many investigators over many years for many purposes. A lovely recent paper by Sinha, Berlind, et al. (I am an appreciative reader but not a co-author) finally realizes the vision of constraining the HOD jointly with the multiplicity function and projected correlation function, and it shows why doing this properly was much harder than we thought it would be in 2002 ...
28. The Halo Occupation Distribution and the Physics of Galaxy Formation
A. A. Berlind, D. H. Weinberg, A. J. Benson, C. M. Baugh, S. Cole, R. Davé, C. S. Frenk, A. Jenkins, N. Katz, C. G. Lacey
2003, ApJ, 593, 1
This paper, based on Andreas Berlind's thesis research, investigated physical predictions for the halo occupation distribution from SPH simulations and from semi-analytic models of galaxy formation. The first key finding was that the HODs predicted by the two methods agreed remarkably well once we corrected for the fact that they predicted different galaxy stellar mass functions. Other key findings were the form of the HOD: a sharp rise from zero to one over a narrow range of halo mass, a slowly rising plateau from one to two over a factor of ~20 in halo mass, and a roughly linear relation between halo mass and galaxy number at high masses. Fluctuations about the mean are sub-Poisson at low masses, which has a crucial impact on clustering statistics, and the HOD depends strongly on the age of the stellar population, with older galaxies preferentially residing in high mass halos. We connected these trends to the physics in the galaxy formation models. Zheng et al. 2005 was the successor to this paper.
29. On Departures from a Power Law in the Galaxy Correlation Function
I. Zehavi, D. H. Weinberg, Z. Zheng, A. A. Berlind, J. A. Frieman, R. Scoccimarro, et al.
2004, ApJ, 608, 16
One of the conclusions of Berlind & Weinberg 2002 was that it is relatively hard to get a power-law galaxy correlation function, and that the combination of cold dark matter cosmology and HOD models generically predicts an inflection in the correlation function at scales of ~ 1-2 Mpc, the transition from dominance by galaxy pairs in a single halo to galaxy pairs in separate halos. At the simplest level, the physics behind this inflection is that galaxies orbit within halos, so this scale marks a transition from "single stream" cosmological flows to a "multi-stream" regime where galaxies can encounter other galaxies that have changed direction. As the SDSS redshift survey grew beyond the early sample analyzed in Zehavi et al. 2002, the increased measurement precision gave us a chance to test this prediction. We showed that there is a statistically significant departure from a power-law correlation function for SDSS galaxies that is naturally explained in the HOD framework. This paper also required developing the HOD modeling calculations themselves to the level of accuracy demanded by the precise measurements. These deviations from a power-law were confirmed and examined as a function of galaxy properties in Zehavi et al. 2005.
30. The COS-Halos Survey: Physical Conditions and Baryonic Mass in
the Low-redshift Circumgalactic Medium
J. K. Werk, J. X. Prochaska, J. Tumlinson, M. Peeples, T. M. Tripp, A. J. Fox, et al.
2014, ApJ, 792, 8
This paper presents the synthesis interpretation of results from COS-Halos, using the Hubble Space Telescope and ground-based (Keck Telescope) absorption line measurements of multiple ionization states of multiple elements to constrain CGM properties galaxy by galaxy, using equilibrium CLOUDY models fit to the data. The results demonstrate that typical galaxies in COS-Halos have massive amounts of cool gas in the CGM, at least 2e10 Msun and probably half or more of the baryon content expected for the halo mass. They also show that this cool gas is relatively low density, far lower than would be required for pressure equilibrium with a virial temperature hot gas halo. There are some unavoidable uncertainties in the modeling assumptions, but this paper presents striking evidence for a massive cool CGM, which is not easy to explain in simple models of galaxy formation.
31. Cold dark matter: Controversies on small scales
D. H. Weinberg, J. S. Bullock, F. Governato, R. Kuzio de Naray, A. H. G. Peter
2015, Proc Natl Acad Sci, 112, 12249
In 2013 I organized a panel for the National Academies' Sackler Symposium on Dark Matter, with a focus on the challenges to the cold dark matter (CDM) model from small scale observations -- the inner rotation curves and satellite populations of the Milky Way and other disk galaxies. The resolution to these challenges could lie in new dark matter physics, such as self interactions, or it could lie in baryonic physics effects that redistribute cold dark matter on kpc scales. This short review article summarizes our panel discussion. I think it has been particularly useful in encapsulating the observational situation for physicists whose expertise is in dark matter phenomenology rather than galaxy properties.
32. Hierarchical Galaxy Formation and Substructure in the Galaxy's Stellar Halo
J. S. Bullock, A. V. Kravtsov, D. H. Weinberg
2001, ApJ, 548, 33
For the dwarf satellite study in Bullock, Kravtsov, & Weinberg 2000, James and Andrey developed a semi-analytic model to track the mergers and orbits of subhalos into a Milky Way mass parent halo, populate them with stars, and decide which ones survived. When early papers from the SDSS showed evidence of multiple large substructures in the Galactic halo (e.g., Ivezic et al. 2000, Yanny, Newberg, et al. 2000, we decided we should check whether the disrupted satellites in this model could produce the kinds of structures turning up in the SDSS and other halo surveys. To our great surprise, we found that the disrupted satellites in this model could produce not just structure in the halo but the entire stellar halo itself, including its total mass and average radial profile. This idea is similar to that proposed by Searle & Zinn 1978, based on empirical arguments, but here it is shown to be a natural consequence of galaxy formation theory in the cold dark matter scenario. We presented simple visualizations and statistics that could be used to test the hypothesis, showing, in particular, that substructure would remain more evident in the outer halo because of the smaller number of contributing systems and the longer mixing times. James Bullock, Kathryn Johnston, and their collaborators have since extended this model into a much more sophisticated and powerful framework for predicting properties of the stellar halo (e.g., Bullock & Johnston 2005, Johnston et al. 2008), and the first statistical tests against substructure observed in the SDSS look encouraging (Bell et al. 2008, Xue et al. 2010).
33. Quasar Clustering and the Lifetime of Quasars
P. Martini, D. H. Weinberg
2001, ApJ, 547, 12
This paper proposed a novel method to constrain the lifetime of high redshift quasars: since the number density of quasars is observed, if they are long-lived then their host dark matter halos must be rare and should be strongly clustered, while if they are short-lived then their host dark matter halos are common and they should be less clustered. Quasar clustering measurements at the time were too uncertain to derive much of a constraint, but the concepts developed here have played a big part in interpreting subsequent clustering measurements from the SDSS and other quasar surveys. If we were writing the paper now we would probably frame the discussion in terms of duty cycles (like Shankar et al. 2009) instead of lifetimes. I no longer clearly remember, but I think this paper must also have had a significant effect on my early thinking about HODs, which took root in the thesis of Martini's graduate school classmate Andreas Berlind.
34. Wide-Field InfraRed Survey Telescope-Astrophysics Focused Telescope Assets
WFIRST-AFTA Final Report
D. Spergel, N. Gehrels, J. Breckinridge, M. Donahue, A. Dressler, B. S. Gaudi, et al.
This report is the predecessor to the 2015 report cited above. I spearheaded a summary version of this report, titled WFIRST-2.4: What Every Astronomer Should Know.
35. Galaxies and gas in a cold dark matter universe
N. Katz, L. Hernquist, D. H. Weinberg
1992, ApJ, 399, L109
This is the first paper from our SPH collaboraton, and the first paper to show that incorporating gas dynamics and radiative cooling in a simulation of a cosmological volume led to the formation of dense clumps of cold gas with the approximate masses and sizes of the stellar components of galaxies, and to the formation of groups containing multiple galaxies in a common dark matter halo. While this behavior was anticipated by analytic models of galaxy formation, it was reassuring to see it work out in a full calculation. This paper also presented first calculations of galaxy clustering and galaxy bias from hydrodynamic cosmological simulations, albeit in a very small box.
36. Protecting Life in the Milky Way: Metals Keep the GRBs Away
K. Z. Stanek, O. Y. Gnedin, J. F. Beacom, A. P. Gould., J. A. Johnson, J. A. Kollmeier, M. Modjaz, M. H. Pinsonneault, R. Pogge, D. H. Weinberg
2006, Acta Astronomica, 56, 333
This paper emerged from a discussion at OSU Astro Coffee in which Kris Stanek remarked that all of the long GRBs identified with supernovae in nearby galaxies were in low metallicity systems. Under the tutelage of Andy Gould, OSU astronomers have long been aficionados of small number statistics, and from this discussion we convinced ourselves that GRBs must arise preferentially in low metallicity stars rather than tracing star formation. I did almost none of the work needed to convert this lively conversation into a paper, but I guess I deserve some credit for the fact that we were having the coffee conversation in the first place. We were given grief by our first referee for drawing strong conclusions from a sample of five, but I think this conclusion has held up pretty well. Doom and gloom papers about GRBs destroying life on earth were popular at the time, hence the planetary safety focus of the title.
37. Cooling Radiation and the Lyα Luminosity of Forming Galaxies
M. A. Fardal, N. Katz, J. P. Gardner, L. Hernquist, D. H. Weinberg, R. Davé
2001, ApJ, 562, 605
The motivation for this paper was the discovery of giant Lyman-alpha emission nebulae ("Lyman-alpha blobs") by Steidel et al. 2000, which did not have evident central AGN to power them by photoionization. We investigated our simulations to see if these objects could be powered by cooling radiation onto forming galaxies. Our answer was a tentative "yes", at least as far as energetics went. In retrospect, the most important implication of this paper was off to the side of its main topic: because so much of the gravitational potential energy was being radiated in Lyman-alpha, it meant that most of the gas cooling onto galaxies was never heated to the halo virial temperature. This realization led us to undertake a more thorough investigation establishing the primacy of cold mode accretion, in the Katz et al. 2003 conference proceeding and the Keres et al. 2005 journal article. The conventional wisdom now seems to be that giant Lyman-alpha nebulae are mostly powered by AGN (sometimes obscured), but gravitational cooling radiation may still play a role.
38. Photoionization Feedback in Low-Mass Galaxies at High Redshift
M. Dijktstra, Z. Haiman, M. J. Rees, D. H. Weinberg
2004, ApJ, 601, 666
This paper is a partial counter-point to Thoul & Weinberg 1996. Dijkstra and Haiman adapted Thoul's 1-d hydro code to model the collapse and cooling of gas in low mass halos at high redshift, near the epoch of reionization, and they found that because of high densities, self-shielding, and a low ionizing background intensity, the suppression of low mass galaxy formation is much weaker in these conditions, allowing galaxies to form in halos with circular velocities down to 10 km/s. This substantially increases the population of low mass galaxies that can contribute to reionization. My direct contribution to this paper was quite small, but I'm pleased to be part of it because it is, surprisingly, my only paper with any of these three co-authors.
39. Chemical Cartography with APOGEE: Metallicity Distribution Functions and the Chemical Structure of the Milky Way Disk
M. R. Hayden, J. Bovy, J. A. Holtzman, D. L. Nidever, J. C. Bird, D. H. Weinberg, et al.
The SDSS-III APOGEE survey drew me into the field of Galactic chemical evolution. This is my favorite of all of the APOGEE-1 papers, mapping the distribution of stars in [alpha/Fe] vs. [Fe/H] (the second measuring overall metal content, the first measuring the ratio of core collapse supernova elements relative to total enrichment) as a function of position in the Galactic disk. At most locations, there are two distinct tracks in this space. The relative number of stars on the two tracks, and the [Fe/H] distribution of those stars, varies strongly with Galactocentric radius and vertical distance from the disk midplane. The shape of the [Fe/H] distribution changes with radius, a likely signature of radial migration of stars. For the first time, APOGEE was able to map these distributions across the disk with large enough numbers of stars to provide excellent statistics. My students and I contributed theoretical interpretation to go along with these beautiful observational findings. We have recently ( Weinberg et al. 2019) used APOGEE-2 data to map a much larger set of abundance ratios across the disk, with further striking results.
40. The Power Spectrum of Mass Fluctuations Measured from the Lyα Forest at Redshift z = 2.5
R. A. C. Croft, D. H. Weinberg, M. Pettini, L. Hernquist, N. Katz
1999, ApJ, 520, 1
After the Croft et al. 1998 paper on power spectrum recovery from the Lyman-alpha forest, we collaborated with Max Pettini to apply the method to an observational data set, a sample of 19 quasar spectra originally collected to study damped Lyman-alpha systems. We significantly improved our understanding of the method itself and how to characterize what we could learn from observations, so although this paper was superseded by Croft et al. 2002, which analyzed a larger sample, it played an important role in the development of our Lyman-alpha forest cosmology program.