January/February 2008

Is Bigger Really Better?

I must receive at least 25 eā€‘mails every day assuring me that size matters greatly, and any opportunity to increase it, even of dubious medicinal nature, should be immediately embraced. Well, is this true for science in general and seismology in particular? There certainly seems to be an increasing tendency for Big Science initiatives to take center stage, manifested both in federal funding of consortia, interdisciplinary initiatives, and big equipment items and the pre-eminence of multi-author research projects in major scientific journals. As an active seismological researcher who tends to prefer small-grant, single-investigator research, I look on these trends with mild alarm. Yet, for the past three years I served as Chair of the Board of Directors of the Incorporated Research Institutions for Seismology (IRIS) and as one of the six-member EarthScope Management Team, investing substantial time and energy into the largest consortium funded by the National Science Foundation (NSF) Division of Earth Sciences and the largest NSF major research equipment (MRE) project in the earth sciences, respectively. Given my widely shared concern that such Big Science activities intrinsically come at the expense of Small Science research funding (my own bread and butter), how do we each decide what endeavors to support with our time? Such choices confront all researchers, and now that I am not on a plane every other week, I’ve been giving it some more thought myself.

Clearly, not all Big Science activities are of the same nature. Some involve large financial investments in unusually expensive activities—activities beyond the scope of single-investigator grants—that have very focused goals and products that benefit only a select few in the community. Other large efforts may, on the other hand, have rather open goals and products that serve a large community. I find it much easier to rally behind the latter because it serves my own interests, and IRIS is the quintessential example of a large activity that brings very broad benefits to many researchers. While I do recall some naysayers in the early 1990s, I doubt that any active seismological researcher today would deny the great achievements of IRIS in deploying a new-generation Global Seismic Network, providing easy-touse PASSCAL portable instruments for funded field studies, or breaking down historical attitudes about proprietary seismic data and making IRIS and other data freely and openly available to the global research community through the IRIS Data Management Center. These are outstanding accomplishments that could not have been achieved by Small Science efforts.

IRIS annually receives about $12 million from the NSF EAR (i.e., NSF’s Division of EARth Sciences) instrumentation and facilities budget. While proposals to renew the core IRIS programs have repeatedly fared excellently in the external review process, and NSF is quick to laud the accomplishments of this consortium that counts more than 100 universities as members, the fact is that over the past few years core funding for IRIS has actually diminished. I personally found it stunning that although the NSF budget increased by $25 million each year following the 2004 Sumatra-Andaman tsunami disaster to augment natural hazards research (thanks to successful presentations to Congress by the NSF director that featured displays and analyses of GSN data), not a single dollar more went to IRIS when NSF acted on a favorably evaluated fiveyear renewal proposal. In fact, IRIS received less funding than the year before! If the increased funding had instead gone to EAR small-grants programs, I would have been happy (maybe even happier), but instead it appears to have been largely diverted to operations and maintenance of ocean research vessels in other divisions. The Big Science versus Small Science tension is greatly exacerbated by decisions like this; it would have made more sense to me to float all ships rather than a (very) select few. The fact that the permanent NSF budget increased by more than twice the annual IRIS budget due to an important geophysical event that IRIS played a major role in goes unnoticed by most folks when they see no benefit. Will I be inclined to lift a finger to help NSF the next time they call? Hmm…

If IRIS is Big Science for solid earth activities, then EarthScope is even Bigger Science, with $200 million budgeted for facility construction funding alone from 2003 to 2008 and ongoing annual operational and maintenance costs exceeding $20 million per year.

The very success of IRIS as a community organization and science-serving data facility has made it a role model for other initiatives, and NSF has actively encouraged this. Unfortunately, there has been little funding growth in EAR base programs, so both core small-grants programs and important consortia activities such as UNAVCO, Southern California Earthquake Center (SCEC), and Consortium for Materials Properties Research in Earth Sciences (COMPRES), largely cultivated by NSF’s pleasure with the IRIS model, haven’t been getting the funding they need to achieve their full potentials. Thus IRIS funding is placed in tension with other worthwhile activities, which is fine in principle but frustrating for the community. I am a strong believer that IRIS must continually justify its existence and importance for the broad community, and it should not weaken its shared governance with the research community. There is a real danger that the general enthusiasm for this major and costly seismological activity may diminish if seismologists do not make conscious efforts to communicate their science and results to other areas of earth sciences. Some community activities could be pursued. For example, I personally think that the days of tolerating multiple, competing, global low-resolution tomographic models are passing, and we should synthesize our collective knowledge into a new generation 3D reference model that other disciplines can use. IRIS as a consortium, or other community programs such as Cooperative Institute for Deep Earth Research (CIDER) could play a role in catalyzing such self-sustaining endeavors, because they do not easily or naturally arise from small-grants programs.

Like most seismologists, I want to have both strong, core small-grant research programs and centralized facilities that provide me with bounties of data.

If IRIS is Big Science for solid earth activities, then EarthScope is even Bigger Science, with $200 million budgeted for facility construction funding alone from 2003 to 2008 and ongoing annual operational and maintenance (O&M) costs exceeding $20 million per year. I came into EarthScope management by virtue of my service on the IRIS Board, without having been involved much in the prior development of the program. I honestly was not a fan; I found the science plans for the three components of EarthScope—San Andreas Fault Observatory at Depth (SAFOD), Plate Boundary Observatory (PBO), and USArray—all rather lackluster and felt that while a big infusion of new equipment funds into EAR geophysics was intrinsically a good thing, this was not clearly the best way to spend the money. As a Big Science project, EarthScope had many danger signs of being inefficient, poorly managed, and with the potential to wreak havoc on the budgets of small-grants programs nearer and dearer to my heart. After a rough year or two in which some of my worst fears were realized, management problems were eliminated, the intrinsic strengths of the groups operating each component showed through, and the facility was built with great efficiency, now approaching an on-time, on-budget completion. As yet, the small-grants programs have not been savaged (at least not by EarthScope, although other questionable NSF EAR decisions are affecting funding for areas that I care about).

The general NSF model for MRE projects like EarthScope is that capitalization costs are provided by new money but ongoing O&M costs must be accommodated by the host division. This model seems predicated on there being some earlier MRE project that is phasing out as the new one phases in, so existing O&M money can be redirected. But EAR had no earlier major project (certainly not IRIS!) that it could gracefully phase out. The potential remains for deleterious impacts on small-grants support; unless EAR budgets grow, the large O&M costs of EarthScope and the limited additional funding for research that uses the facility will really squeeze small-grant core programs. Significant NSF budget growth or internal NSF reprioritization to invest preferentially in areas where first- MRE investments have been made are the only means I see of avoiding significant harm to EAR small-grants programs.

I’m still on the fence as to whether EarthScope is going to justify itself in terms of major transformative research impact. I sense that the potential is there; already USArray data are stimulating many unprecedented applications of a continental-scale array along with enhancing resolution of the upper mantle structure as expected. PBO’s geodetic data are rapidly bringing focus to distributed deformation in the complex plate boundary region, and new faulting phenomena are being revealed. SAFOD successfully brought up 44 m of precious core from a ~ 2.5-km-deep transect through a creeping section of the San Andreas fault. By sustaining easy and open access to EarthScope data there is no question that many more important results will come. Still, would it have been better to pump up the small-grants programs by $20 million per year (as if that were even possible without a sexy new initiative to justify it)? My Small Science bias still holds, so I would answer Yes. You get an awful lot of bang-for-the-buck from well-administered small-grants programs, and with EAR proposal success rates now plummeting to below 20%, many good projects are not getting funded. Big projects tend to have high overhead and sometimes less than cuttingedge excitement on a day-by-day basis. However, EarthScope is a success by most measures (particularly amongst NSFfunded MRE projects, which have a checkered history at best), and it will be here for a long time, so making it as successful as possible is in all earth scientists’ best interests. I still wish there were a way to convince NSF to enhance funding of its core programs rather than always relying on big initiatives and cross-cutting redefinitions of existing programs that tend to come with unfunded mandates. But I am assured by many that this is just about impossible given the budgetary appropriation process, though the reasons they give still ring hollow in my ears.

Like most seismologists, I want to have both strong, core small-grant research programs and centralized facilities that provide me with bounties of data. Thus, I can comfortably wear both hats when times are good and big projects are not cannibalizing smaller ones (I do not think the opposite ever happens). While I am strongly supportive of IRIS, and now to a lesser extent of EarthScope, I am not blindly so, and over the years I have forcefully voiced my dissent when I took exception to management decisions or actions that I felt did not represent my or the broader community’s opinions or priorities. Because both of these Big Science activities have built themselves around community-based organizations, active community participation in their shared governance is the best way to maintain a healthy balance of Big Science versus Small Science to the extent that we can. So, is bigger really better? Fortunately for me, the answer is: Not always. 

Thorne Lay
University of California, Santa Cruz
Email - thorne [at] pmc.ucsc.edu


To send a letter to the editor regarding this opinion or to write your own opinion, you may contact the SRL editor by sending e-mail to <lastiz [at] ucsd.edu>.




Posted: 18 January 2008