Trends in the Conduct of Space Physics
This chapter identifies further trends in the implementation of scientific research in areas of interest to the Committee on Solar-Terrestrial Research (CSTR)/Committee on Solar and Space Physics (CSSP) and relates them to the funding and demographic trends discussed in Chapters 3 to 5. Data have been provided by the funding agencies or assembled by CSTR/CSSP committee members, and individual examples are used where appropriate. Much of the data on past programs were difficult to obtain, especially with respect to parameters defined in hindsight. Nevertheless, trends can be discerned that contribute to a historical perspective.
A vital element of space physics research is the availability of in situ data. Without intending to imply that quantity equals quality, one relevant measure of space physics research opportunities remains the number of launches and experiments. A mainstay of the space physics research program, the National Aeronautics and Space Administration's (NASA) Explorer Satellite Program provides one accessible and well-defined data set1 that illustrates long-term trends in the availability and implementation of space physics research opportunities.
Figure 6.1a shows the number of Explorer launches from 1958 to the present, along with projected launches for programs approved through 1997 (projected launch times should be considered uncertain). The data show a clear and continuing decrease in space-physics-related launches since the 1960s. However, we are not able to conclude from this result alone that the number of research opportunities has decreased. If, for example, satellites increased in size and thus carried more experiments, the number of actual research opportunities (as measured by experiments flown) may not have decreased.
Figure 6.1b shows the mass evolution of the Explorer satellites. Again, the masses shown for future launches should be considered uncertain. The Figure shows a general increase in Explorer size since 1958. Further, it appears to be possible to separate the data into two categories: small and large Explorers. (Interestingly, this implies that at a very early stage, the space physics community saw the need for both small and large missions.)
How do these factors affect the number of experiments flown? Figure 6.2 shows a repeat of the number of space-physics-related Explorer launches since 1958, along with the number of experiments flown onboard those Explorer satellites. The decrease in number of launches from the 1960s through the 1970s is compensated for by the increasing satellite sizes, giving a comparable number of experiments flown in both decades. However, the launch frequency became so
TABLE 6.1 Space-Physics-Related Explorer Launches and Experiments
No. of Launches
No. of Experiments
low in the 1980s that the number of experiments flown experienced a drastic reduction. Table 6.1 summarizes these results.
The preceding Figure and table show that experimental research opportunities for space physics within the NASA Explorer satellite program have become much more scarce since the 1970s.2 This trend was recognized in the early 1980s and discussed in a previous National Research Council report  that argued strongly that NASA return to its earlier philosophy of making available to the scientific community more small and rapidly implemented satellites dedicated to focused scientific problems.
The desire for rapid implementation discussed in the Explorer program report  also represents an important parameter in the conduct of space physics. Ideally, implementation should be on time scales that allow support of contemporary science questions, experimental research teams, graduate students, timely data analysis, and theoretical studies. The CSTR/CSSP has assembled a data base of space-physics-related launches in order to investigate the time required to implement these satellite projects. (The data base is described more fully in Appendix A.)
The implementation time is defined as the time from mission start to satellite launch. The start date is usually set at the date that investigators wrote proposals to place instruments on the spacecraft. This date was chosen because it represents a well-defined starting point that exists in some form for most missions. Where only the proposal year was known, July 1 was used as a start date. In other cases the start dates were obtained directly from the principal investigators. These different start date estimates may vary by a few months, but the variation is small compared to the implementation times obtained.
Figure 6.3 shows the implementation times for the space-physics-related missions in the CSTR/CSSP data base. Arrows showing the effect of a one-year delay are given for projected launches of approved programs. Explorer missions and other NASA space-physics-related missions are identified separately. The Figure shows a striking increase in implementation times, from one to two years in the early 1960s to 10 to 12 years in the late 1980s. All mission types show a steady increase in implementation time.
As so eloquently argued by Freeman Dyson , this large an implementation time (over 10 years) represents ''a terrible mismatch in time scale between science and space missions.'' It does not provide support for contemporary science questions, graduate students, instrument engineering staff, timely data analysis, theoretical studies, or stability of experimental research teams. Because of these and other ramifications (e.g., increased administration, management, planning activities, costs), large implementation times represent one of the leading reasons that space science researchers perceive that too much time is spent on activities other than research.
Increased Planning Activities
The start dates as defined above actually give only a lower limit for the implementation time. This is due to the extended preproposal planning activities that have become normal for most satellite missions. Consider, for example, the Global Geospace Sciences (GGS) element of the International Solar-Terrestrial Physics (ISTP) program.3 The two U.S. GGS satellites are scheduled for launch in 1994 and 1995 and are represented in Figure 6.3 by implementation times (based on proposal submissions in 1980) in the 14- to 15-year range. However, formal planning for this mission actually began in 1977 when NASA established an ad hoc committee to define a program named the Origins of Plasmas in the Earth's Neighborhood, or OPEN. In parallel with this effort, the CSSP developed a research strategy in space physics for the 1980s . The OPEN ad hoc committee issued its report in 1979 , describing a major NASA program that would pursue an important part of the research strategy developed by the CSSP . NASA issued an Announcement of Opportunity in 1979, proposals were written in 1980, and experiments were selected in 1981. Following extensive policy and programmatic planning activities, the OPEN program evolved into the ISTP program, wherein the NASA contribution was sharply limited, and ESA, ISAS, and IKI agreed to provide major contributions. Formal approval of the ISTP program finally occurred in 1988. The net result was that not only were there an additional three years of planning activities prior to proposal submissions, but a large portion of the approximately 12- to 14-year implementation time shown in Figure 6.3 was taken up by planning, policy, and political activities.
Such planning removes resources from the direct support of research, is often spent on missions that are not flown, and rarely reduces the cost of a mission. Indeed, it can be argued that excessive planning increases mission costs. Appendix B presents a detailed case study from the field of solar physics that illustrates these effects. It provides insight into how excessive planning and study activities arise and shows the potentially devastating effects to a research field and its relations with funding agencies.
Reliance on New-Start Approvals
The ISTP program illustrates another trend in space physics programs, namely, the increasing reliance on major programs that require new-start approval on a project-by-project basis by Congress. A recently completed NASA Strategic Plan  for space physics describes a research program through 2010 that is
dominated by such projects. Although divided into major-, moderate-, and intermediate-class missions, all require new-start approval, and together with their mission operations and data analysis costs constitute 70 to 90 percent of the planned NASA space physics program through 2010.
One reason for this increasing reliance on major new-start programs has been the search for additional flight opportunities to compensate for the greatly reduced number of opportunities available to space physics through the Explorer program after 1980. A second incentive for this shift has been the emergence of scientific problems requiring experimental platforms of greater size and complexity than before. The net result of these effects is a space physics program that relies on big science projects to a much greater degree than in the past.
Partly because of budgetary and community pressures, NASA has recently begun to rejuvenate the Explorer program. The Small Explorer Program (SMEX) was initiated with the successful launch of the Solar Anomalous and Magnetospheric Particle Explorer (SAMPEX) mission. Two additional SMEX missions are being built, and NASA expects to continue selecting and launching these missions on a regular basis. Further plans, not yet funded, call for university-class and medium-sized explorers; it is too early to predict the future of these plans. An expanded Explorer program would be a positive and welcome step toward alleviating the serious problem of infrequent access to space.
Solar physics presents a unique opportunity to analyze trends in the implementation of its scientific requirements. It requires large-scale facilities both for its space-based and its ground-based observatories. Although the evolution of solar satellites and instrumentation has moved toward larger, more complex and costly systems, as described in the previous section, this is not true for the ground-based observatories. Large solar observatory projects conducted over 30 years ago were major undertakings fully comparable to present programs. For this reason the comparison of solar satellite and ground-based implementation trends may provide a measure of the relative contributions of technical (e.g., size, complexity) and nontechnical (e.g., administration, management, funding procedures) factors to increasing implementation times.
Implementation of Solar Satellite Missions
This section presents data on the prelaunch duration (implementation phase) of solar physics space missions in order to study how long scientists prepare for space missions, whether prelaunch duration shows a secular trend, and, if so, what the underlying causes might be.
Only missions primarily devoted to the study of the Sun are considered. Although this criterion excludes some missions that carried solar instruments,
most of those missions are included in the preceding section on satellite observations. We do, however, include the Ulysses mission, which, in its original concept, consisted of two spacecraft, one of which carried an instrument for imaging the solar corona from above the solar poles. More than three years after the release of the Announcement of Opportunity for the mission, the U.S. spacecraft was canceled, leaving only the European spacecraft with its instruments for studying solar fields and particles intact but with no capability to image their solar sources.
The prelaunch phase is taken to begin with an Announcement of Opportunity (AO) that solicits proposals for scientific instruments to be carried on the satellite. This is by no means a general characterization of prelaunch activities; it is simply an attempt to place a lower bound on the period during which a space researcher must devote a substantial fraction of his or her professional effort to a particular space-borne experiment. The AO is usually preceded by a period of study and planning in which scientists are heavily involved. Thus, the prelaunch phase as defined here typically underestimates the period of involvement for selected experimenters, sometimes by years.
The actual AO date was used in most cases. (For some of the Orbiting Solar Observatory (OSO) missions, the AO date was estimated based on other related dates.) The instruments on the Skylab Apollo Telescope Mount (ATM) were originally proposed for the canceled Advanced Orbiting Solar Observatory (AOSO), so the prelaunch phase is taken to begin with the AOSO opportunity. The Orbiting Solar Laboratory (OSL) is discussed extensively in Appendix B, but for the present purpose its origin is taken to be the AO for the Solar Optical Telescope issued in April 1980. The launch date is estimated by assuming that a new start for OSL will not occur before 1997. In the case of missions primarily sponsored by ESA (Ulysses, Solar and Heliospheric Observatory [SOHO]) or the Japanese space agency (Yohkoh), the prelaunch phase is dated from the NASA AO for participation by U.S. scientists.
The results are shown in Figure 6.4. For missions yet to be launched, arrows show the effect of a one-year delay. The early OSO satellites were part of a series, with multiyear funding and a spacecraft design that remained relatively stable. At the other end of the spectrum is the OSL. If this mission eventually flies, it will be both the most delayed and the most ambitious solar physics mission ever, by a wide margin (see Appendix B). Over the entire period, 1970-1992, the trend is consistent with that shown in Figure 6.3 for space physics satellite missions in general. The rise in implementation time has been from a few years in the early 1960s to the present value of over 10 years.
Of course, the duration of the prelaunch phase should really be considered in relation to the overall scope and cost of each mission. We do not attempt to include that level of detail here. We can, however, identify one characteristic that is common to several of the missions with relatively long prelaunch phases: a midcourse change in the scope or conception of the mission itself. As men-
tioned above, Skylab/ATM evolved from the canceled AOSO mission. The character of NASA's participation in Ulysses changed drastically after the AO was issued. In the case of OSL, the AO was issued in anticipation that the mission would be approved; if it is ever approved, over 15 years will have elapsed since the AO date. Experimenters do not play a passive role when missions are redefined or rescoped. To maintain their participation in the mission and, perhaps, to help ensure that the mission flies at all, they may invest as much professional effort as they would have had the mission proceeded on the original schedule.
Because of the way NASA missions are funded, inefficiencies and delays connected with changes in mission concept are linked to the overall cost of the
mission. Large missions require funding over at least five years. When this funding must receive specific congressional approval each year, the program is subject to unpredictable budget fluctuations that, in turn, necessitate continuously evolving plans to accommodate different fiscal scenarios. This shifting-ground effect is even more pronounced when the mission as a whole has not yet been assured new-start approval. Although the political process should ensure responsible government control over large public expenditures, it often has unintended negative consequences; delay and inconsistency lead to inefficient use of human and financial resources.
Solar Ground-Based Observatories
This section presents data on the proposal, design, and construction phases of ground-based solar telescopes in order to identify trends and relate them to project cost. Only national or international facilities costing at least $4 million (1991 dollars) are considered. This makes for a homogeneous sample but excludes several major university observatories.
The beginning of a project is taken to be the date of the proposal or, if it can be clearly identified, the date of a study or site survey that led directly to the proposal. In analogy to the launch of a space mission, the end of the project is taken to be "first light," even though significant testing and improvement usually occur for some time after that.4
Costs were converted to 1991 dollars according to the NASA (Code BA) new-start inflation index to reflect the rates of inflation that characterize the technical sector. The cost for the Large Earth-based Solar Telescope (LEST) is taken to be the U.S. share, one-third of the total.
Summary time lines for these observatories are shown in Figure 6.5. The projects divide naturally into an earlier group of three major telescopes and two ongoing projects, the Global Oscillation Network Group (GONG) and LEST.
Figure 6.5 indicates that it takes longer to plan and execute major ground-based projects than it did 20 to 30 years ago. The comparison can be made more directly for these ground-based projects than for space missions because there has not been the same striking evolution in the complexity, scope, and cost of ground-based efforts. The McMath Telescope project, executed in less than five years, was a major undertaking fully comparable to current programs; the McMath and LEST were each designed to be the world's largest solar telescope, and the McMath still is.
Table 6.2 compares some of the characteristics of the McMath Telescope and the LEST project. Figure 6.6 compares their project time lines in more detail. Although the total projected cost of the LEST project is twice as large as the cost of the McMath Telescope, the cost to the U.S. funding agency is smaller for LEST; this is part of the rationale behind international consortia. LEST may also entail a somewhat greater construction task; however, the LEST project before groundbreaking has already taken twice as long as the entire McMath project.
TABLE 6.2 Comparison of Two Ground-Based Solar Telescope Projects
$19M cost (U.S. share of $57M total)
Single proposal (11 pages)
Multiple proposals (>> 500 pages)
Single funding transfer
Multistage, multisource funding
Informal scientific oversight
Scientific and technical advisory committees
<14-year duration (in progress)
Although the international character of LEST has complicated its overall coordination and funding, increased effort in the proposal and advocacy phases can also be identified within each participating country. Table 6.2 shows that the period of active involvement for participating scientists (writing the proposal, advocating the project to funding agencies, and participating in or responding to oversight committees) plays a much more prominent role in LEST.
Probably the single most important factor behind increasing implementation times for major ground-based solar projects is the advent of multistage funding. The McMath, Sacramento Peak Tower, and Kitt Peak Vacuum telescopes were funded with a single transfer of money to the managing organization. From that point the progress of the design and construction phases was limited only by technical issues or practical considerations internal to the project. Basically, the telescopes were built as fast as they could be soundly built.
Figures 6.7 and 6.8 illustrate the effect of multistage funding on the progress of the GONG. Even though the overall project budget and timetable were judged reasonable when the project was approved, in none of the first five years did the actual funding reach the proposed profile. After three years, funds were short by a factor of two. The effect of this mismatch in resources will be a delay of at least three years and an increase of more than 30 percent in total (constant-dollar) cost.
As discussed above, funding decisions are part of a broader, often political, process with many competing demands. However, it is important for decision-makers and the public to understand the true costs, in dollars and morale, of these kinds of project delays and midcourse changes in funding.
Figure 6.3 showed that implementation times for solar satellite programs have increased from two to three years in the early 1960s to the present value of well over 10 years. In the solar ground observatory program, where there has been a much less dramatic evolution in complexity than in the satellite program,
we also see a major increase in implementation times. Much of this has been due to changing managerial and funding procedures. Much more time is spent in study, planning, selling, and oversight activities, all of which add to the final cost. Funding is apportioned on an incremental basis that usually falls short of planning expectations.
This comparison indicates that administrative procedures have had at least
as large an impact on increasing implementation times as have the elements of project size and technical complexity.
The NASA suborbital program supports scientific experiments carried out on airplanes, balloons, and sounding rockets. The experiments come from disciplines in astrophysics, earth sciences, microgravity research, solar physics, and space plasma physics. This section is limited to discussion of sounding rockets in space physics research.
Sounding rockets provide unique capabilities not easily attained by other means. For example, sounding rockets are the only vehicles that can launch payloads to observe space phenomena from unique geographic locations, altitudes, and times. Thus, sounding rocket experiments can accomplish specific scientific goals and have been especially valuable in obtaining information on small-scale and rapid temporal features that are difficult to obtain from rapidly moving spacecraft.
NASA currently offers 15 configurations of sounding rockets to provide the scientific community with different capabilities. Scientific requirements and the payload weight dictate which rocket to use. The average weight of sounding rocket payloads has been growing steadily since 1960, as shown in Figure 6.9.
Sounding rocket payload weight has more than quadrupled in the last 20 years, from about 125 pounds in 1970 to 600 pounds in 1990. This increase has come about as the scientific fields matured and experiments became more demanding and complex. It has also had a direct impact on the number and types of rockets launched. Figure 6.10 shows that the total number of rocket launches has de-
creased over the past decade. However, the percentage of large rockets included in those launches has increased markedly, from 24 percent in 1984 to 64 percent in 1991.
Sounding rocket grants are awarded for a three-year period to cover the experimenter's expenses for fabrication, launch, and data analysis. The funding level per experiment and the number of experiments supported by the Space Physics Division were fairly steady through the 1980s (Figure 6.11). Note the change that occurred in 1990: the funding level per grant was reduced from the 1989 average of $200,000 to $150,000. This occurred because the total number of grants supported increased from 29 to 41, without an increase in the overall funding level.
NASA's suborbital program serves an extremely useful function to the scientific community. It provides the opportunity for (1) research groups in industry, university, and government laboratories to develop space-borne instruments for orbital missions; (2) university research groups to train graduate students in experimental methods in space physics; and (3) research groups to conduct inexpensive research programs that continue to yield original results. About 261 papers were published in refereed journals based on results obtained from sounding rocket experiments during the period 1986-1990.
The overall NASA funding for rocket sciences increased by approximately 20 percent (in constant-year dollars) from 1979 to 1991. The number of rocket
launches decreased by approximately 45 percent, but payload weight increased substantially. The demand for sounding rocket support has been increasing steadily since the mid-1970s due to decreased opportunities in orbital missions. Recently the support level per experiment in solar and space physics has decreased substantially in order to fund more of these projects. Increasing university overheads (described in Chapter 5) further decrease the funds directly available to the rocket researcher. Finally, rocket experiments continue to be managed
by individual principal investigators and on fixed cost and time schedules despite their increasing complexity.
NASA's balloon program has historically provided a mechanism to obtain exposures to energetic particle radiation. When high-altitude balloon technology was developed in the late 1940s, balloons were used to discover the major constituents of the steady flux of cosmic rays bombarding the Earth. Many of the scientists who went on to play a major role in NASA's flight programs learned how to operate a payload in the harsh space environment by doing thesis or postdoctoral research using a balloon-borne payload.
This section draws on data about scientific ballooning as a whole, and on a small sample from a very specific and narrow discipline area, namely studies of high-energy cosmic rays. The trends in the two data sets are consistent. Three cosmic-ray groups provided information about flight rates and implementation times: the California Institute of Technology, University of Chicago, and Goddard Space Flight Center. Figure 6.12 shows the number of NASA balloon flights per year. The trend toward decreasing flight frequency is very clear. Figure 6.13 shows the trend toward increasing payload complexity, as reflected
by payload mass, during the same time period. Although payload weight increased, the reliability has remained relatively constant (Harvey Neddleman, Balloon Projects Branch, Wallops Flight Center, personal communication, 1993), except for periods of specific difficulty in the late 1970s discussed further below.
Based on limited information from the cosmic-ray groups, Figure 6.14 shows the trends in implementation time (i.e., conception to first successful flight). Understanding these data requires some guesswork based on knowledge of the history involved. In the 1960s there was a major expedition each summer to Ft. Churchill, Manitoba, known as the Skyhook program, through which most of the balloons, all with relatively small payloads, were flown. In the 1970s, balloons that could reliably lift a few thousand pounds came into being, bigger experiments were possible, and the flight frequencies declined. This growth in balloons and payloads finally ran into technology limits in the last half of the 1970s. Difficulties with balloon materials and flight reliability began to manifest themselves and were not resolved until the mid-1980s. Meanwhile, the size and complexity of payloads steadily increased; it now can take several years of funding to build and fly a complex balloon payload. Even with these factors, the time from conception to successful flight recovered in the late 1980s and is as good now as it was in the late 1960s and early 1970s—a testimonial to the
attention paid to balloon problems by NASA. This is a success story that has led to demands for more and longer-duration flights.
Data on total funding and funding per grant for cosmic ray balloon payloads are shown in Figures 6.15a and b. (Unfortunately, these data only to back to 1979, and further historical data were difficult to obtain.) The number of grants and the number of projects are shown in Figure 6.16. The increase in payload complexity has led to multiinstitution (and hence multigrant) collaboration on individual projects.
The data support the recollections of some researchers that in the 1950s and 1960s one could rely on steady funding as long as meaningful results were forthcoming. Programs proposed were multiyear ones and began returning results within a year or two of conception. In the 1970s the time taken to accomplish research seemed to increase: payloads became bigger and more complex, taking longer to build and fly. The late 1970s and early 1980s were a period of poor balloon reliability, when a failure usually meant a year's delay. Following the technical problems of the early 1980s and the corresponding dip in funding, the number of groups flying balloons, as well as the funding, began to increase again.
Ballooning is a viable method for conducting galactic cosmic-ray studies. The long-duration balloon capability currently emerging has promise for the
1990s. The trend toward increasing payload complexity has generally been absorbed by the space physics community without loss of reliability, with the exception that occasional balloon reliability problems arise and must be solved. The increase in complexity has not itself given rise to an increase in implementation time for the balloon programs surveyed.
There has been a trend toward increasing numbers of small grants, mitigated in part by an increase in multiinstitution (and hence multigrant) collaborations. These collaborations have been managed by allowing several institutions to prepare a single proposal, accompanied by separate institutional endorsements.
The time scales for ballooning have remained compatible with the education of students, especially those familiar with space hardware. This program is, relatively speaking, a success story and may contain lessons for others.
Progress in science results from the interaction between its theoretical and experimental (or observational) branches. This interdependence holds as much for space physics as it does for the rest of science. About a third of all space physicists consider themselves to be theorists or modelers, according to one survey (see Chapter 4, Figure 4.6). This percentage is larger for younger physicists (40 percent for those under age 40). Both aspects of space physics science have evolved over the past few decades. As discussed earlier in this chapter and in Chapter 2, experimental space physics has been getting steadily ''bigger,'' starting in the early 1950s with sounding rockets and progressing to larger and more expensive satellite programs. The way in which theoretical work is carried out also has changed over the years.
Traditionally, theoretical work has been carried out at an individual level or
in very small groups. Most support for theoretical work comes from small, short-term grants from NSF and from NASA's Supporting Research and Technology (SR&T) program. Consider the theory and modeling component of NASA's Space Physics Division SR&T funding for FY 1991 (excluding the balloon and suborbital programs). The total number of SR&T grants was 288, of which 140 (or 48.6 percent) were for theory and modeling. The total division SR&T budget was $17.046 million, of which theory and modeling accounted for $7.614 million (or 44.7 percent). (This does not include the Space Physics Theory program, which is discussed below.) Thus, the average SR&T grant size was $59,000 for the division as a whole, and $54,400 for theory and modeling. This average grant size has remained the same for about a decade, thereby falling significantly behind the typical 5 percent rate of inflation. Proposal pressure on the SR&T program was high at the end of FY 1991. For example, in the solar
branch, $2.5 million was available for new grants for FY 1992, but proposals totaling $11 million were submitted.
There has been some tendency for theoretical work to be put into larger packages. Part of this is associated with advances in numerical simulations of space plasma phenomena, which require a larger infrastructure than traditional theory. One example of this is NASA's Space Physics Theory program (previously the Solar Terrestrial Theory program). One of the purposes of this program is to assemble "critical masses" of theorists to work on certain key problem areas. A large number of these groups emphasize numerical plasma simulations. Currently, there are 17 such groups, each with several senior scientists, several junior scientists, and graduate students. The total FY 1992 funding is $4.3 million. The program started in FY 1980 with 13 groups and $2.27 million. The overall program grew at a rate of 5.5 percent per year, thus approximately keeping pace with inflation, but the rate of increase per group was only 3.1 percent. These groups have been very productive and important for space physics, but most theoretical work still takes place elsewhere.
Some theoretical work has also been supported by large NASA missions under various guises: (1) as theorists included on instrumental proposals as coinvestigators and (2) as interdisciplinary scientists. The NSF also supports theory and modeling via relatively small grants, comparable in size to NASA's. However, over the past few years "new" money at NSF has gone not to the "base" program but to new initiatives, such as the magnetospheric GEM pro-
gram and the upper-atmosphere CEDAR program. Both GEM and CEDAR have significant theoretical and modeling components but are "managed" programs, in the big science mold.
Numerical simulation of space plasma phenomena using computers has become an important theoretical method over the past two decades. Examples include particle-in-cell simulations, hybrid simulations, and three-dimensional magnetohydrodynamic modeling. Numerical models have increased in size and complexity over time, taking advantage of technological developments in computers, such as the CRAY-YMP and other supercomputers. Large computational facilities have been created, such as the four NSF supercomputer centers, the National Center for Atmospheric Research's Scientific Computing Division, and government laboratories such as NASA's Ames and Goddard centers and the Los Alamos and Livermore national laboratories. The largest space (and nonspace) plasma simulations have been run at the national laboratories, where large blocks of computer time are available. In this sense, too, also theory has become "big" science.
Some relatively large groups have grown to support the development and running of these plasma simulations and models, but overall these groups are still smaller than the large experimental teams that are put together to design, build, and use space-based instruments. In fact, even large modeling efforts such as 3-D magnetohydrodynamic simulations are frequently undertaken by only one or two senior scientists. A recent trend running counter to the dependence on supercomputers is the increasing use of powerful workstations that permit all but the largest simulations to be run locally.
Theory is an important part of space physics, and numerical plasma simulations have played an increasingly important role in this field. The emphasis on simulation has resulted in some increase in the scale of operations for theory over the past couple of decades, but theoretical space plasma physics seems to have found a balance between big and little science.
The chief difficulty for theory and modeling lies in the small grant sizes ($54,000 on average), which are not keeping up with the rate of inflation, as well as the decreasing probability of a proposal getting funded and rising overhead costs. As discussed more generally in Chapter 5, these factors lower the effectiveness of the Space Physics Theory and Modeling program.
The reason for flying scientific space missions is to obtain and analyze new data. NASA currently funds mission operations and data analysis (MO&DA) efforts separately from spacecraft development, in part to protect the postlaunch funds from being used to solve development problems or overruns. This effort at protection has not always been successful, however.
There are several requirements for ensuring appropriate data analysis: (1) adequate funding of the principal investigators (PIs) responsible for instrument development and operation and the subsequent reduction and analysis of the data; (2) adequate funding and the commitment to enforce documentation and archiving requirements so that data are available in a format that can be used by scientists other than the original PIs; and (3) broadening the base of researchers who know of, understand, and can use the data. The first two requirements are self-explanatory. This section addresses the third requirement, especially as it relates to the interplay of big and little science.
There has been a recent trend toward broadening the base of researchers working with space data through the use of interdisciplinary scientists and guest investigators. This trend is driven by the increasing breadth of many of the space physics missions and by the abundant correlative data available from both ground-based observations and spacecraft operating in different locations in the solar-terrestrial system at the same time.
One space mission that was particularly successful in attracting the participation of many researchers was NASA's Solar Maximum Mission (SMM). Launched in 1980, SMM was an Earth-orbiting satellite that carried six instruments for studying the Sun. For its time, SMM could be considered a moderate-cost space mission (prelaunch cost of $125 million in FY 1979 dollars). In addition to the six principal investigators and their teams, about 25 Guest Investigators per year were funded, at an average of $34,000 per investigation, for a total cost of roughly $6 million over seven years. About twice that number of investigators obtained SMM data for their own research, either in person or over the phone. SMM also provided modest support for correlative ground-based observations of the Sun. Summaries of selected SMM data were published in broad-circulation periodical data reports to make those outside the program aware of the data resource. In addition, the SMM project sponsored a series of workshops that focused on SMM data and were open to all interested investigators. By 1989 the data obtained by SMM had led to over 700 scientific publications. Over 75 percent of U.S. solar physicists and a large number of non-U.S. researchers have participated in SMM in one way or another.
SMM was not the only successful guest investigator program in space physics. Over the period 1984-1987, for example, NASA's Dynamics Explorer (DE) guest investigator program funded 39 investigations at a total cost of $1.45 million, resulting in roughly two new publications per investigator per year.
In 1991, NASA's Space Physics Division initiated a plan for the archiving and analysis of data from six Explorer missions. The plan had three elements:
documentation and archiving of Explorer data at various investigator facilities and at the National Space Science Data Center to make the data accessible to the scientific community.
phase-out of direct funding for data analysis by the Explorer principal investigator teams, and
initiation of a guest investigator program to which all interested scientists could apply for continued study of these valuable data bases.
The first two elements above were implemented quickly. As for the third element, 41 guest investigators were supported in 1991, but the funding for continuation of this important program is highly uncertain.
The space physics community is united in its support of vigorous guest investigator programs for many reasons: (1) they are very cost effective, (2) they bring different talents and interests to the analysis of costly space data, (3) they provide the PIs and their teams with new collaborative capabilities and opportunities, (4) they help bridge the gap between major missions, and (5) they provide a natural way to encourage the interplay between big and little science that enriches space physics research.
Sometimes individual investigators or small teams find it advantageous to coordinate their data analyses in order to achieve their respective research objectives. As an example, it has been suggested that the time is ripe to undertake a retrospective analysis of existing and complementary sets of space-based and ground-based data dealing with magnetospheric substorms. The relevant data have now been archived, and the theoretical models of substorms have changed considerably since the individual data sets were first analyzed. Another example is the Coordinated Heliospheric Observations (COHO) program, which seeks to support theory, modeling, data analysis, and guest investigations aimed at understanding the heliosphere and its boundaries through coordinated data obtained from spacecraft now widely distributed throughout the heliosphere. The COHO initiative was approved in NASA's Space Physics Division program for FY 1993, but no funds were appropriated to support it.
Data analysis is but one example of the strong synergism that can exist between big and little science. It shows, as discussed in Chapter 2, that both are essential for the advancement of space physics. Furthermore, large and small efforts must be carefully balanced and coordinated to optimize the scientific return.