My primary field of interest is the formation of structure in the universe. Observations of the cosmic microwave background indicate that the early universe was smooth and nearly homogeneous; the goal of theory is to explain the transition from this smooth early universe to the hierarchically structured universe that we observe today. The basic questions of the field include the following: Is the standard hypothesis, that structure in the universe developed by gravitational instability from small-amplitude, primordial fluctuations, correct? If so, what were the properties of the primordial fluctuations, and what physical process produced them? What are the properties of the dark matter, and what role does it play in the formation and clustering of galaxies? What is the average density of matter in the universe, and does it imply a cosmos that will expand forever or one that will eventually halt and recollapse? What are the relations between the large-scale distribution of matter in the universe and the distributions of directly observable tracers, such as galaxies, galaxy clusters, and the ``Lyman-$\alpha$ clouds'' that absorb the light of distant quasars?
A typical theory of structure formation answers the first four questions a priori, specifying the properties of primordial fluctuations and the material contents of the universe. Numerical simulations play a key role in working out the observable consequences of such theories. Using cosmological simulations that combine smoothed particle hydrodynamics with gravitational force computations, my collaborators and I can go from initial conditions to the evolved distribution of mass, galaxies, and intergalactic gas in a representative volume of a theoretical universe. The simulations thus answer the last of the above questions in the context of a particular theoretical model, and they illuminate the physical processes that govern the growth of structure. Because these simulations incorporate a more detailed physical treatment than traditional, purely gravitational calculations, we can compare their results to a much broader range of observations in order to test the underlying theoretical models; indeed, the simulations themselves often suggest new ways of analyzing or interpreting observational data. In our most recent work, we have studied the origin and predicted properties of Lyman-$\alpha$ absorbers in the ``cold dark matter'' model. Our simulations reproduce the observed characteristics of these systems remarkably well. Lyman-$\alpha$ studies will be a focus of our efforts over the next two years, as we look at other theoretical models and make more detailed comparisons to the data emerging from Hubble Space Telescope and the Keck 10-meter telescope. For a semi-popular account of this work, from the San Diego Supercomputer Center's magazine ``Gather/Scatter,'' click here. I have also put together a collection of color plots illustrating some of the simulation results here.
The second major element of my research is the analysis of galaxy redshift surveys, which provide three-dimensional maps of the nearby universe. I am now trying to extend and combine two approaches developed in my earlier papers: reconstruction analysis, which works backward from observed structure to plausible initial conditions, then evolves these forward to produce a detailed model of the input data; and clustering anisotropy analysis, which uses the directional dependence of clustering in a redshift survey to obtain a statistical measurement of galaxy motions. Reconstruction can test the hypothesis that primordial fluctuations were Gaussian, as predicted by the simplest theories of the early universe. Clustering anisotropy constrains a combination of two parameters that describe, respectively, the average mass density and the ``bias'' between galaxies and mass. A combined method should greatly improve the constraints from existing redshift samples, for reconstruction controls the statistical noise caused by random orientations of structures in the survey, leaving only the systematic anisotropy produced by galaxy motions. By extending the anisotropy technique, I aim to break the degeneracy between density and bias and thus obtain separate constraints on these two crucial cosmological parameters.
I will test these methods on simulations and apply them to current redshift surveys, but their ultimate target is the million galaxy redshift sample of the Sloan Digital Sky Survey. I have been a member of the SDSS collaboration for several years now, working on various tasks related to software testing, selection of galaxy targets, and survey strategy, and attending lots and lots of meetings. The project has recently (5/98) carried out its first test observations, and I am looking foward to the enormous fun of analyzing the data. For information on the Sloan Survey, from overview to technical details, check out the Black Book, an online version of the project's NASA proposal.