6
Report of the Panel on Theory, Computation, and Data Exploration



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports 6 Report of the Panel on Theory, Computation, and Data Exploration

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports SUMMARY The Panel on Theory, Computation, and Data Exploration was charged with surveying two separate branches of astronomy and astrophysics: theoretical astrophysics and data exploration. “Theoretical astrophysics,” in the convention of this report, includes both analytic theory and numerical simulation. The term “data exploration” is introduced to describe the newly emerged discipline of mining insight from large and complex astronomical databases using sophisticated modeling tools. The panel reviews the status of these two branches of astronomy and astrophysics and provides separate sets of recommendations for prioritized initiatives and policy directives. THE SCOPE OF THEORETICAL ASTROPHYSICS Unlike many astronomy communities, which identify themselves by wavelength or mission, the community of theoretical and computational astrophysicists defines itself by synthetic tasks that cross disciplinary boundaries: Defining the frontier. The community invents concepts that create frontiers and a framework for observational discovery—new ideas about the universe—from the extremes of space-time to physics in exotic environments to the new universe of captured digital knowledge. Model building. It creates intelligible descriptions of physical systems with precise quantitative connections to reality, including both sophisticated simulations that incorporate a comprehensive range of physical processes and compact mathematical constructions that identify and represent the key physical effects. Synthesizing a world view. It knits physical science into a coherent narrative of our place in the universe, one that is accessible, interesting, and edifying to society at large. This scientific view of the universe competes in the free market of ideas; theorists tell and sell the astronomers’ story of the universe. In these tasks, theoretical and computational astrophysicists combine leadership for and service to the larger astronomical community. Their specific activities are defined by both their core intellectual values and their interactions and collaborations outside their community. Three important themes recur in this report:

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports The rapid pace of discovery in this golden age of observational astronomy has created for the first time a comprehensive view of what is happening in the universe over much of observable space-time and is quickly expanding the complexity and range of accessible phenomena. The continuing advance in digital technology is redefining and expanding the character of knowledge. In astronomy, the digital revolution is creating explosive growth in the quality and quantity of data and our ability to model complex phenomena, resulting in a new “digital universe.” Paradoxically, the cultural gap between the frontier science community and much of the rest of American society continues to widen at precisely the time that new technologies allow us to create tools for broader and more rapid dissemination of knowledge outside the science community. We must work to resolve this paradox. The panel believes that theory defines the context within which most of astronomy operates. Observers may answer the questions what and where, but theory addresses the how and why and seeks explanation and synthesis. No modern observation would make sense, or could be properly interpreted, without the pioneering theoretical work that gave us white dwarfs, neutron stars, black holes, atomic physics, relativity, radiative transfer, hydrodynamics, mechanics, statistical physics, high-energy radiation processes, and so on, and without the theorists who are the developers and users of these concepts and tools. One need only think of Penzias and Wilson without Gamow, Dicke, and Peebles or Bell and Hewish without Gold and Wheeler to understand the centrality of fundamental theory. Given this history, the panel asserts that the theoretical core of astronomy must be nurtured and strengthened in the next decade in order to optimize the scientific return from the coming explosion of astronomical discoveries. THEORY INITIATIVES PROPOSED BY THIS PANEL The theoretical foundation of any discipline is most secure when the broadest possible range of research is supported. The panel proposes three new initiatives for theoretical astrophysics, believing that they will bring significant, tangible benefits to the entire astronomy community. One of these initiatives advances a new model for supporting theory

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports efforts aligned with observational missions. The other two initiatives are designed to enhance the current broadly based theory programs that support the inspirational theoretical research that complements mission-oriented research. The panel summarizes its initiatives and then supplies details of each in a later section. THEORY CHALLENGES TIED TO PROJECTS AND MISSIONS The most important output of the Astronomy and Astrophysics Survey Committee is its prioritized list of facilities and missions for the next decade. The panel proposes that most of these prioritized initiatives should be accompanied by, and continuously interact with, one or more coordinated theory challenges. The challenges should describe theoretical problems that are ripe for progress and either relevant to the planning and design of the mission or key to the interpretation and understanding of its results in the broadest context. The theory challenges should be planned and budgeted as an integral part of the project or mission. The funds should be allocated through open competition in the national community rather than as addons to observational or instrumentation grants or contracts, and under no circumstances should they divert funds from existing grants programs for broadly based theory. Both individuals and consortia should be supported. Panels drawn from the theoretical community, broadly construed, should select the award recipients. Support should cover the broadest possible range of theory, from the basic theoretical foundations for the mission to detailed modeling. Appropriate challenges will evolve during the life cycle of the project or mission. In the early stages, the challenges contribute to mission definition, identify opportunities, and add enthusiasm and ideas. As the data flow in, theory contributes to the interpretation of the results and sets the context for subsequent initiatives. At the end, theory helps produce a synthesis of the results. The cost of theory challenges might typically be 2 to 3 percent of the project or mission cost, although the scope of the challenges should be determined individually for each project, and much larger fractions—or no theory challenges at all—could be appropriate for some projects. The panel believes that this small cost will be repaid many times over by the contributions of theorists to mission design, analysis of newfound phenomena, and the vision that inspires future missions.

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports A NATIONAL POSTDOCTORAL PROGRAM IN THEORETICAL ASTROPHYSICS All of astronomy and astrophysics thrives under constant reinvigoration by young researchers. Postdoctoral fellows fill a unique role as innovators, owing to their combination of scientific independence, freedom from administrative and teaching duties, and ambition to establish a personal scientific identity. The presence of talented young people enhances research productivity and cross-field collaborations. Theoretical astrophysics, in particular, has seen its directions and technical methods driven largely by the efforts of postdoctoral fellows, and theory postdoctoral fellows do much of the highly innovative nonprogrammatic research that inspires new missions and research directions far into the future. Nor should we forget the important role of postdoctoral fellowships in training the astrophysicists of the future. The present support mechanisms for theoretical postdocs are inadequate and unstable. Grants to individual theory researchers rarely are sufficient to fund a full-time postdoc. The few who are supported in this way are tied to a specific project, with limited freedom to pursue untested or potentially revolutionary ideas. A handful of U.S. institutions award fellowships that occasionally support theorists. However, several foreign institutions have large and strong theory postdoctoral programs that exceed in scope the programs available at almost any single U.S. institution. Some support for theorists has also been available from the Hubble, Compton, and Chandra fellowship programs; however, this support is programmatically selective and fragile. To meet the need for a healthy postdoctoral research corps in the United States, the panel proposes a national program of freestanding postdoctoral fellowships. The panel envisions the program being administered in much the same way as the successful Hubble postdoctoral program, with postdocs distributed at institutions throughout the country and selected through competitive review. Such a program will provide an indispensable base for fundamental, creative theoretical research; it will identify the most outstanding young theorists and foster their development in a cost-effective way; it can encourage ethnic and gender diversity; and it will enhance the vitality of research at universities across the country and the talent pool available for this research. The panel recommends a program that would award 10 or so 3-year theory fellowships a year, at an annual cost of about $2 million.

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports RIGHT-SIZING THEORY SUPPORT Federal support for theoretical research is central to the continuing health of astronomy. However, despite the best efforts at the National Science Foundation (NSF) and the National Aeronautics and Space Administration (NASA), support for theory has not kept pace with the considerable growth in astronomical data in the last 10 years, creating an imbalance in the funding profile at a time of exceptionally rapid discovery. To help remedy this imbalance, the panel recommends that major observational facilities, projects, and missions share the responsibility of funding both mission-related and broadly based theoretical research. Moreover, because the direct benefits of theoretical research are difficult to quantify, the funding agencies should develop guidelines for its support. The panel believes that a suitable preliminary guideline is that at least 30 percent of the costs for research personnel in grant programs, academic departments, and research institutes should normally be directed toward theoretical research activity. The most cost-effective mechanism to address the challenge of right-sizing support for theory in this era of discovery is the targeted expansion of existing grants programs at the NSF and NASA that support broadly based theoretical research. In particular, the panel recommends a substantial augmentation of NASA’s successful Astrophysics Theory Program (ATP). Enhanced support through such programs would complement the theory challenge program and the national theory postdoc program to establish a thriving and balanced research effort in theoretical astrophysics. DATA EXPLORATION INITIATIVE PROPOSED BY THIS PANEL: THE NATIONAL VIRTUAL OBSERVATORY Astronomy will experience a major paradigm shift in the next few years, driven by large, systematic sky surveys at multiple wavelengths. The panel believes that these digital archives will soon be the astronomical community’s main avenue for accessing data. Systematic exploration and discovery in these databases will play a central role in the day-to-day research activities of most astronomers. This data avalanche—the flood of terabytes of data—is happening, whether or not we plan effectively for it. The first megasurveys are already in progress, including 2MASS, SDSS, and MACHO.

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports This transition is driven by advances in technology. The last decade witnessed a thousandfold increase in computer speed, a significant increase in detector size and performance, a dramatic decrease in the cost of computing and data storage capabilities, and widespread access to high-speed networks. Despite these advances, the environment to exploit these huge data sets does not exist today. In order to handle terabytes of data efficiently, one needs database engines with fast input/ output speeds and advanced query engines that can access databases spread throughout the country. Existing analysis tools do not scale to terabyte data sets. In combination, these factors make a new initiative, the National Virtual Observatory (NVO), both feasible and compelling. The NVO will help the astronomer preparing for the next observing run, the theorist analyzing large-scale structure, or the phenomenologist searching for extremely rare objects. The NVO will link the major astronomical data assets into an integrated—but virtual—system that enables a qualitatively new type of astronomical research: automated multiwavelength and multiple-epoch exploration and discovery among all known catalogued astronomical objects. The NVO will initially provide access to tens of terabytes of catalog and image data, growing to multiple petabytes by the end of the decade. It will influence all disciplines of astronomy and astrophysics, from x rays to optical and infrared through the radio wavelengths, and it will be essential for confronting sophisticated theories with observations. The NVO is “national” because it serves the needs on a national scale; it is “virtual” because it supports observations on digital representations of the sky, and it is an “observatory” because it is a general-purpose facility, just like a traditional observatory. The four major elements of the NVO are (1) integration of major data archives, (2) advanced services for the astronomical community, (3) standards and tool development, and (4) education. The NVO will involve a coordinated—but distributed— effort of universities and national centers to develop an integrated data architecture for astronomical research. Such standards and coordination will play a key role in linking the multiple archives and service providers; without them, astronomy will be unable to exploit its data fully and efficiently. The NVO will be a powerful resource for public education and outreach at many levels. A digital representation of the sky, easily accessible via the Web, has the potential to excite the imagination of future scientists and allow them to participate in the astronomical discovery process. Educators will draw on the NVO to develop educational

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports materials, and public institutions such as planetariums will use the NVO to develop presentations and displays. The software and standards developed as the core of the NVO will be relevant to many other fields of research and should attract the attention of researchers in computer science, statistics, bioinformatics, earth sciences, and other fields. The NVO supports many of the goals of the recent federal Information Technology for the 21st Century Initiative (IT2) and is particularly appropriate for multiagency funding. The panel envisages that the management of the NVO will be similar to that of the NSF Science and Technology Centers or NASA’s Astrobiology Institute. The NVO should be a multistaged effort, consisting of definition, demonstration, development, and deployment phases. Cost estimates for the NVO project are preliminary. Firm estimates will require a broad consensus on how the project is to be organized as well as on the scope and schedule of the project. Current projections for the definition and demonstration phase in years 1 to 3 are around $5 million total, with development scoped at $10 million total and deployment in years 3 to 5 at around $30 million total. Public outreach, grants to observers, and research in related areas of computer science and astronomy, including theory, could amount to an additional $15 million over 5 years. SUMMARY OF PANEL FINDINGS AND RECOMMENDATIONS The specific findings and recommendations of the Panel on Theory, Computation, and Data Exploration are summarized in this section. Details and supporting arguments are found, in each case, elsewhere in this report. PROPOSED INITIATIVES The panel’s principal recommendations take the form of three initiatives in theoretical astrophysics and one in data exploration: The panel recommends that most prioritized projects or missions recommended by its parent committee, the Astronomy and Astrophysics Survey Committee, be accompanied by one or more coordinated theory challenges. The theory challenges should be budgeted and programmed as an integral part of the project or mission. However, the funds should

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports be allocated through periodic peer-reviewed open competitions in the national community rather than as add-ons to project grants or contracts. Both individuals and consortia should be supported. Support should cover the broadest possible range of theory, from speculative scenario building to detailed modeling. The panel finds that present support mechanisms for theoretical postdoctoral fellows are inadequate and unstable, and it proposes a national program of “portable” theoretical postdoctoral fellowships. The program would be administered much like the successful Hubble postdoctoral program. Postdocs will be distributed at institutions throughout the country and selected through competitive review. The program will identify the most outstanding young theorists and foster their development in a cost-effective way. It will enhance the strength and vitality of university-based research and encourage ethnic and gender diversity. The panel believes that there is an imbalance between the considerable growth in astronomical data in the last 10 years and the support of theory that is essential to a proper understanding and context for that data. The panel recommends that steps be taken to correct this imbalance. Specifically, the panel recommends that at least 30 percent of the support for research personnel in grant programs, academic departments, and research institutes normally should be directed toward theoretical research activities; that major observational facilities, projects, and missions should fund both “harvest” and “seed-corn” theoretical research; and that funding agencies should develop overall guidelines for right-sizing their levels of theory support. A high priority should be given to the expansion of NASA’s Astrophysics Theory Program. The panel recommends the creation of a National Virtual Observatory to accomplish the integration of the nation’s priceless astronomical data, to facilitate its archiving, and to lead in the development and dissemination of tools for new scientific discovery from archival data. The NVO will undertake activities in standards, archive services, basic analysis tools, and advanced analysis tools. In toto, these efforts will require resources comparable to those of a small satellite mission. Very roughly, a 3-year definition and prototyping stage will be budgeted at $5 million. Development and testing would pick up in the second year and cost $10 million over 4 years. Deployment and operations would ramp up from the second through fourth years, with a 5-year cost of $30 million. Public outreach, grants to observers, and research in related areas of computer science and astronomy, including theory, would be

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports budgeted at $15 million over 5 years. Unless specifically rechartered, the project would terminate or descope after 5 years of operation. OTHER RECOMMENDATIONS DOE support of theoretical astrophysics. While the survey committee is not specifically tasked to examine Department of Energy (DOE) support of astrophysics, it would be impossible to give a complete picture of U.S. theoretical astrophysics without doing so to some extent. The panel offers the following suggestions for optimizing the effectiveness of DOE’s contribution to astrophysics: DOE’s Office of Science should leverage its efforts in low-energy nuclear physics with appropriate research in nuclear astrophysics; a particular example is radioactive ion beam research. Research in cosmology and matter under extreme conditions is similarly relevant to major new DOE facilities such as the Relativistic Heavy Ion Collider (RHIC) and the Continuous Electron Beam Accelerator Facility (CEBAF), the Stanford Linear Accelerator Center’s (SLAC’s) B-factory, the European Laboratory for Particle Physics’ Large Hadron Collider (LHC), and the Fermi National Accelerator Laboratory’s (FNAL’s) collider and fixed target programs. DOE’s Defense Programs should recognize more explicitly the close synergy between their national security missions and research in astrophysics and (noting past high returns on investments made, particularly in terms of quality personnel brought into the defense program) should support with programmatic funds certain areas of relevant theoretical research in astrophysics. The ASCI program, in particular, is both the beneficiary and the benefactor of research in theoretical astrophysics, and this connection could be strengthened. Institutes with opportunities for visiting theorists. While conferences allow sharing of research results, they seldom provide opportunities for actual collaborative work. The panel endorses the support by funding agencies of the several institutes that provide opportunities for extended working visits by scientists away from their home departments. The panel recommends that such institutes continue to receive healthy support. High-performance computing. The panel notes that several national initiatives in computing will allow for the participation of, and

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports allocation of resources to, computational astrophysicists—but only if every effort is made within NASA and NSF to be responsive to these initiatives at an agency level. The panel recommends that the funding agencies position themselves to benefit from such initiatives by supporting algorithm development, Grand Challenge applications, and the development and dissemination of community codes and by conveying at agency levels the message that they support the continued health of the national supercomputer centers. DESCRIPTION OF THEORETICAL ASTROPHYSICS The initiatives that the panel recommends represent only the part of theoretical astrophysics that readily matches the AASC process. The panel represents a much broader community than can be adequately served by “initiatives.” More than any other area of astrophysics, theory depends on continuous healthy support of broadly based science rather than large discrete investments, although history has shown that theory has a remarkably strong influence on large scientific investments. To help the reader understand the field of theoretical astrophysics, the panel describes here its vision for the field, and in doing so looks at who theoretical astrophysicists are, where they have been, and where they are going. THE NEW THEORIST Traditionally, a theoretical astrophysicist studied astronomical phenomena primarily by analytical and conceptual (pencil-and-paper) reasoning. Today’s theorist performs on a larger stage. In particular, the 1980s saw the emergence of numerical modeling and simulation as a distinct and powerful subdiscipline of theoretical astrophysics. In a wide variety of astrophysical problems—cosmological structure formation, globular-cluster evolution, star formation, supernova explosions, accretion disks and jets, galactic dynamics, formation and evolution of planetary systems—the synergy of numerical modeling and analytic theory led to far greater progress than either tool could achieve on its own. Numerical modeling has also blurred the distinction between theory and observation: to properly understand and model selection biases, experiments and observations are now often numerically simulated, and

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports The NSF and NASA were important sources of support for astrophysics theory during the 1980s and 1990s, with the programs of the Division of Astronomical Sciences at the NSF and the Astrophysics Theory, Long-Term Space Astrophysics, and Supporting Research and Technology programs at NASA providing the bulk of the individual investigator grants. The introduction of NASA’s Astrophysics Theory Program (ATP) in the 1980s was a major shot in the arm for theoretical astrophysics, so that the NSF and NASA programs now provide comparable support for theory. However, despite the best efforts of program officers and administrators at both NSF and NASA, support for theoretical research has not kept pace with the impressive growth in astronomical data over the last 10 years. A few numbers taken from Federal Funding of Astronomical Research serve to illustrate the current problems. For the ATP, the oversubscription by number increased from ~3.0 in 1987 to ~4.8 in 1997; the oversubscription by funds was higher. The per-grant award (in 1997 dollars) decreased from a high of $190,000 in FY1987 to $85,000 in FY1997, reflecting the trend away from group grants and the attempt to maximize the number of principal investigators supported within a limited budget. The annual awards declined from a peak of $5.8 million (1997 dollars) in 1994 to $3.1 million in 1997. At the NSF, the oversubscription rate for individual proposals increased steadily over the past decade and has recently been ~5.0, while the average award has remained flat in constant dollars. This unhappy situation is exacerbated by other circumstances. The best students are increasingly turning away from theory to observation, because that is where the money and the exciting new facilities lie. In addition, many senior theorists are joining guest observer programs to obtain funding. Although the panel applauds closer connections between theorists and observers, these should be motivated by science rather than funding imperatives. Whether or not theorists have obtained access to MO&DA money in this way, during budget squeezes support for theory is always the first to be cut. The panel’s concerns about theory support are echoed in Federal Funding of Astronomical Research, which concludes that theory and instrumentation are two specific areas in which “support has not been adequate to support the dramatic scientific discoveries of the last decade…. [T]his has severely limited the field’s ability to understand and interpret the wealth of new data.” By far the most direct and cost-effective single contribution to right-

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports sizing support for theory would be the long-overdue expansion of the individual investigator grants program at the NSF and NASA. In particular, the panel recommends a substantial augmentation of NASA’s ATP. Such expansion would benefit broadly based theoretical research and thus would complement the more directed initiatives of the theory challenge program and the career-development initiative of the national postdoctoral fellowships to establish a balanced and thriving effort in theoretical astrophysics research. Not to address the growing crisis in theory research will jeopardize the future health of both observational astronomy and theoretical astrophysics. The astronomy community must decide whether theory will be given the modest resources it requires to flourish or whether through benign neglect and inflexibility it will be driven into decline. As a path forward, the panel suggests that NSF’s Division of Astronomical Sciences and NASA’s Office of Space Science study the issue of right-sizing theory support, perhaps through the mechanism of the NRC’s Committee on Astronomy and Astrophysics, and certainly in consultation with NSF’s Division of Physics, which exemplifies the separate theory program. INSTITUTIONAL ISSUES FOR THEORETICAL ASTROPHYSICS UNIQUE ROLE FOR THE DEPARTMENT OF ENERGY The DOE supports astrophysics and cosmology at universities and at its national laboratories in areas where there is intellectual overlap between astrophysics and the missions of the DOE. Much of this work involves theoretical and computational astrophysics; in addition, there is a vigorous and growing interest in astronomical data exploration at the national laboratories. The panel briefly highlights some of this work and then makes recommendations. DOE-supported university research programs in nuclear and particle physics often contain elements of nuclear or particle astrophysics. This work is frequently only part of the research effort of one scientist in a group; support for purely astrophysical research is uncommon. One outstanding historical example of DOE-sponsored research is the discovery of cosmological inflation. Substantial astrophysics research programs are found at the national laboratories. There are programs at both the Office of Science laborato-

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports ries (which are discussed first) and the Defense Programs laboratories (see below). At Fermi National Accelerator Laboratory (FNAL), there is an active theoretical astrophysics group. At Oak Ridge National Laboratory, there is extensive research in nuclear astrophysics, and much of the work on the development of new radioactive-ion-beam facilities is motivated theoretically by astrophysical considerations. Similarly, much of the motivation for the RHIC at Brookhaven National Laboratory and CEBAF at the Jefferson Laboratory is to study the quark-hadron phase transition in the early universe. Cosmology is also an important consideration in the experimental programs at SLAC’s B-factory, the LHC, and FNAL’s collider and fixed target programs (e.g., search for supersymmetry, neutrino mass, and charge-parity violation). Basic research at the Defense Programs laboratories, especially at Los Alamos National Laboratory (LANL) and Lawrence Livermore National Laboratory (LLNL), is concentrated in areas that support the core national security mission of the labs. This has been interpreted to include a wide variety of astrophysical problems, including modeling the conditions interior to stars and supernovae, modeling explosive processes such as supernovae, and computing the synthetic spectra of high-temperature gases. The national laboratories have been in the vanguard of large-scale data exploration. The MACHO project originated at LLNL and has now gathered and processed over 6 TB of data. The data processing for SDSS is performed at FNAL. Two projects designed to detect optical counterparts at gamma-ray bursts, ROTSE at LANL and LOTIS at LLNL, routinely image the entire night sky and have gathered several terabytes of image data. These various efforts are intellectually successful, but their visibility and impact could be greatly enhanced if the DOE were to recognize their value more explicitly and encourage research in astrophysics as a policy at DOE headquarters level. Specifically, the panel offers the following suggestions: The Office of Science should support research in nuclear astrophysics to complement its efforts in low-energy nuclear physics and, in particular, to support its new programs in radioactive-ion-beam research. The Office of Science should support research into cosmology, especially where relevant to its major new facilities (e.g., RHIC, CEBAF) or to theoretical work in particle physics. Defense Programs should recognize the close synergy between its national security missions and research in astrophysics and should

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports support with programmatic funds certain areas of basic theoretical research in astrophysics at Defense Programs laboratories. Examples include the physics and computational methods of stellar structure and evolution, the theory and computation of magnetic accretion disks and jets, the modeling of relativistic blast waves, and data-mining techniques for large astrophysical databases. The ASCI program, in particular, is both the beneficiary and the benefactor of research in theoretical astrophysics, and this connection could be strengthened. Finally, the panel stresses what is perhaps the most important contribution of DOE to the day-to-day life of astronomers: the celebrated Los Alamos preprint server, which provides a fully automated e-print archive for astrophysics and many other subjects. The Los Alamos server now archives a significant fraction of all new astronomical literature, distributes new e-prints around the world less than 24 hours after submission, and has virtually eliminated the practice (and significant expense to funding agencies) of distributing paper preprints. INSTITUTES FOR VISITING THEORISTS There is no substitute for face-to-face interaction and collaboration. Even though physicists pioneered new technologies for electronic collaboration (such as preprint archives and browsing software), personal contact adds an important dimension to creative work. This is true in all of science but especially in conceptually challenging theory. Astrophysics is a big subject covered by relatively few researchers. While faculty groups are the norm in many large disciplines of physics, astrophysical theorists are thinly spread among institutions. Conferences allow sharing research results but seldom allow for collaborative work or collective exploration of ideas. The panel therefore applauds the success of several institutes that provide opportunities for extended working visits by scientists away from their home departments, allowing them an opportunity to work closely with their collaborators and providing the infrastructure required for research. One example is the Institute for Theoretical Physics (ITP) in Santa Barbara. It organizes topical programs lasting for several months and encourages and financially supports long visits. The science program is anchored by a strong core of long-term participants. Another example is the Aspen Center for Physics. The summer program is organized around a series of thematic workshops lasting several weeks, with 3-week-long

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports visits encouraged. The science program is anchored by the members of the center, chosen from leaders in many areas of physics, and by workshop organizers, chosen for each summer’s program. Housing subsidies are funded on a graded scale that strongly favors young scientists. Both institutes also support shorter conferences organized along the themes of their long-term programs. There are plenty of examples where subfields of astrophysics were given a significant boost by programs such as these; for example, the basic elements of the Cold Dark Matter paradigm for structure formation were substantially worked out during a program at ITP, which is generally acknowledged to have accelerated the development of the subject by at least 2 years. The panel recommends that institutes such as these continue to receive healthy support. HIGH-PERFORMANCE COMPUTING What is in store for high-performance computing and communications in the first decade of the 21st century? Continuation of exponential growth. The National Technology Roadmap for Semiconductors of the Semiconductor Institute of America foresees that exponential growth in computing power will be sustained by current complementary metal oxide semiconductor (CMOS) technologies and manufacturing processes through 2006. Beyond that, new short-wavelength lithography techniques will be required, as well as overcoming formidable device design and packaging challenges. The industry has historically met such challenges, and the panel believes it will do so again. March to the petaflop. Spurred on by the DOE’s Accelerated Strategic Computing Initiative (ASCI), the NSF supercomputing centers have deployed 1 Tflop computers and are pursuing aggressive upgrade paths to deploy 10 and 100 Tflop computers by 2003 and 2007, respectively. It is expected that NASA will follow suit. These levels of performance are achieved through massive parallelism (~103 CPUs) and advances in commodity microprocessor architectures that should yield 1 Gflop chips early in this decade. Affordable, ubiquitous computing. For a few thousand dollars, every researcher will be able to afford desktop computers only about 1000 times less capable than the high-end supercomputers described

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports above. This puts the supercomputer performance of the mid-1990s on everyone’s desktop within this decade. Proliferation of commodity PC clusters. Research groups and departments are currently building high-performance clusters based on commodity PCs that rival the capabilities of high-end supercomputers for far less money. While these systems have distributed memories with slower interprocessor communications than tightly integrated supercomputers, their number, performance, and scientific impact will grow in this decade. A key policy issue will be striking the appropriate balance between federal investments in national supercomputing centers and in group and departmental resources. Information grids, computational grids. Both the NSF and NASA have programs to link nationally distributed high-value resources such as supercomputers, data archives, and scientific instruments via high-speed networks into what is referred to as a grid. The NSF Partnerships for Advanced Computational Infrastructure Program will probably be in place until 2007. NASA’s Information Power Grid project has a more uncertain tenure. Computational astrophysics and remote and interactive observing are key application drivers. Information technology. The President’s Information Technology Advisory Committee, recognizing that IT will be a key factor driving American progress and economic competitiveness in the 21st century, has recommended a broad-based, long-term program of basic research and development in IT across federal agencies, including NSF (lead agency), NASA, DOE, the Department of Defense, the National Institutes of Health, and the National Oceanic and Atmospheric Administration. One of the novel recommendations is the creation of interdisciplinary R&D virtual centers of computer and application scientists making bold assumptions about the future and then asking what information technologies will be required to get there. How can astronomy and astrophysics take advantage of these technological and societal trends? Because the Information Technology for the 21st Century Initiative (IT2) initiative is broad based, there is no guarantee that the programs that fund astronomers and astrophysicists will benefit. Therefore, every effort must be made to ensure that existing programs within NASA and NSF are responsive to this and similar follow-on initiatives. Areas that are particularly ripe for increased funding include but are not limited to the following:

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports Algorithmic development and Grand Challenge applications. To take advantage of these computational advances, the panel recommends that funding be provided for both algorithmic development and scientific applications. This funding should support both small consortia and Grand Challenge efforts. Development and dissemination of theoretical simulation software (community codes). Such software should increasingly come to be viewed as a standard tool. Just as it has proven centrally useful to the astronomy community to have available standardized software such as FITS, IRAF, and AIPS++, so too would it be productive for theorists to have access to repositories of well-tested, flexible, expandable, and documented production codes. However, the best numerical codes that are the natural products of many modern astrophysical investigations generally are not made available to the broader community. The panel recommends support for code documentation and standardization to facilitate public dissemination of commonly used software. Support for the national supercomputer centers. The national centers provide a unique resource beyond the capabilities of any single institution for state-of-the-art calculations. The panel recommends that support for these centers be continued. The ASCI program, in particular, is both the beneficiary and benefactor of research in theoretical astrophysics, and this connection can and should be strengthened. Although not strictly within its scope as a panel on theory, the panel nevertheless notes the importance of experimental efforts that verify and validate the underlying physics. The goal of simulation is reality, not virtual reality. ACRONYMS AND ABBREVIATIONS 2MASS —Two Micron All Sky Survey ADS —Astrophysics Data System (NASA) AGN —active galactic nuclei AIPS++ —Astronomical Image Processing System, software used for image processing and data analysis ALMA —Atacama Large Millimeter Array ASCI —Accelerated Strategic Computing Initiative (DOE) AST —Advanced Solar Telescope ATP —Astrophysics Theory Program (NASA)

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports CEBAF —Continuous Electron Beam Accelerator Facility (DOE) CMB —cosmic microwave background CMOS —complementary metal oxide semiconductor COBE —Cosmic Background Explorer, a NASA mission launched in 1989 to study the cosmic background radiation from the Big Bang Con-X —Constellation-X Observatory DOE —Department of Energy DPOSS —Digitized Palomar Observatory Sky Survey EVLA —Expanded Very Large Array EXIST —Energetic X-ray Imaging Survey Telescope, to be attached to the ISS FIRAS —Far Infrared Absolute Spectrophotometer, an instrument on COBE FIRST —European Far Infrared Space Telescope FITS —Flexible Image Transport System; format adopted by the astronomical community for data interchange and archival storage FNAL —Fermi National Accelerator Laboratory GALEX —Galaxy Evolution Explorer, a space ultraviolet imaging and spectroscopic mission Gemini —an international project operating two 8.1-meter telescopes, one located on Mauna Kea and the other in Cerro Pachon, Chile GLAST —Gamma-ray Large Area Space Telescope, a joint NASA-DOE mission GRB —gamma-ray burst GSMT —Giant Segmented Mirror Telescope, a 30-m-class, ground-based telescope HST —Hubble Space Telescope, a 2.4-m-diameter space telescope designed to study visible, ultraviolet, and infrared radiation; the first of NASA’s Great Observatories IPAC —Infrared Processing and Analysis Center (NASA) IR —infrared IRAF —Image Reduction and Analysis Facility, a set of computer programs for working with astronomical images ISS —International Space Station IT —information technology IT2 —Information Technology for the 21st Century Initiative (federal program) ITP —Institute for Theoretical Physics (in Santa Barbara) LANL —Los Alamos National Laboratory (DOE) LHC —Large Hadron Collider (European Laboratory for Particle Physics)

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports LISA —Laser Interferometer Space Antenna LLNL —Lawrence Livermore National Laboratory (DOE) LOTIS —Livermore Optical Transient Imaging System, the primary purpose of which is to search for simultaneous optical counterparts of gamma-ray bursts LSST —Large-aperture Synoptic Survey Telescope MACHO —massive compact halo object MAP —Microwave Anisotropy Probe mission MHD —magnetohydrodynamic MO&DA —mission operations and data analysis NASA —National Aeronautics and Space Administration NED —NASA/IPAC Extragalactic Database NGST —Next Generation Space Telescope, an 8-m infrared space telescope NSF —National Science Foundation NVO —National Virtual Observatory, a virtual sky based on enormous data sets NVSS —NRAO/VLA Sky Survey (uses the VLA in producing radio images) R&D —research and development RHIC —Relativistic Heavy Ion Collider (at DOE’s Brookhaven National Laboratory) ROSAT —Röntgen Satellite, an orbiting x-ray telescope launched in 1990 and named after the German scientist W.Röntgen, the discoverer of x rays; a German-U.S.-U.K. collaboration ROTSE —Robotic Optical Transient Search Experiment, designed and operated by a collaboration of astrophysicists from the Los Alamos National Laboratory, Lawrence Livermore National Laboratory, and the University of Michigan SDO —Solar Dynamics Observatory, a successor to the pathbreaking SOHO mission SDSS —Sloan Digital Sky Survey SIM —Space Interferometry Mission SIMBAD —Set of Identifications, Measurements, and Bibliography for Astronomical Data, created and maintained by the Strasbourg Astronomical Data Center SIRTF —Space Infrared Telescope Facility, NASA’s fourth Great Observatory, which will study infrared radiation SLAC —Stanford Linear Accelerator Center (DOE)

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports TPF —Terrestrial Planet Finder, a free-flying infrared interferometer designed to study terrestrial planets around nearby stars VERITAS —Very Energetic Radiation Imaging Telescope Array System VLA —Very Large Array, a radio interferometer in New Mexico consisting of 27 antennas spread over 35 km and operating with 0.1 arcsec resolution

OCR for page 275
Astronomy and Astrophysics in the New Millennium: Panel Reports This page in the original is blank.