The Culture of Collaborative Research
While recent technological and conceptual progress makes positive collaborations between biologists and geologists especially promising, organizational frameworks and traditions have the potential to either promote or retard such collaborations. Scientists, whether in academic or agency settings, are sensitive to institutional reward systems and the opinions of their peers as well as new research opportunities.
As new scientific disciplines and subdisciplines emerge, they follow a variety of paths. Frequently, initial efforts focus on the core elements of their new field of inquiry. Scientists record observations, take measurements, and seek patterns and processes that characterize their field. This phase of information and data accumulation tends to focus on the fine points of the discipline, in many cases adopting a reductionist approach that provides great detail about the subject. Then—as a discipline matures—scientists often reach out to adjacent fields to provide new insights and opportunities for investigation. Efforts to organize and synthesize data from disparate observations and experiments also begin to develop. This may start with the search for new tools or with comparisons between different systems, and eventually leads to more conceptual or theoretical considerations about the relationships between issues at hand. This intellectual expansion yields even greater interactions as the web of scholarship expands in all directions.
This same intellectual trajectory happens in many individual scientists over the course of their careers. Initial training usually focuses on collection of data and their analysis. Scientists seek research support for these activities, publish the research results, and generate solutions based
on those results. As they mature individually, some researchers begin to look for patterns and seek the broader implications of their research and thus many senior scientists reach out beyond their own expertise to collaborate with colleagues in allied disciplines, taking advantage of synergistic interactions to seek answers to important questions. A few scientists seek even broader collaborations that synthesize information across extensive areas of expertise in search of comprehensive understanding of the nature and scope of their discipline. Truly exceptional scientists can manage broad syntheses single-handedly, but such individuals are so rare that the scientific community must also develop team-based strategies to yield the insights that emerge from cross-disciplinary syntheses.
Our reward systems tend to promote individual scholarship and its products over the processes required for broader, more synthetic, activities. In many institutions, the order of authorship and even the number of authors is viewed as an important characteristic of published work such that being sole author, first author, or an author among few co-authors is viewed as especially worthy. Individual grants and the number of graduate students in one’s own laboratory are important evaluation criteria rather than the development of multidisciplinary projects with dispersed resources and intellectual capital.
Synthetic, interdisciplinary research requires efforts that can be viewed as unproductive and even onerous (such as increased efforts at communication, data sharing). Furthermore, results emerging from large projects with many authors may be devalued (i.e., divided by the number of authors) in our current reward system, even though it may lead to insights that an individual scientist might not have been able to conceive or might not have had the standing to promote or publish.
These issues arise in many disciplines and are becoming increasingly common as teams of research scientists organize to address issues from different perspectives. Such “team science” requires explicit reconsideration of funding schemes, measures of academic performance, protocols for authorship, and policies for sharing data. For example, the National Institutes of Health (NIH) has established initiatives to facilitate interdisciplinary research,1 NIH’s Bioengineering Consortium has recently reported on the nature and implications of team science,2 and Altshuler and Altshuler (2004) discuss the challenges in combining genome science and human clinical research. Thus, while the issues discussed below are not unique to the integration of biological and geological approaches to understanding biosphere dynamics, they are important to progress on the topic.
PROMOTING THE CULTURE OF COLLABORATION
Opportunities for collaboration begin with the notion that the results of collaborative, synthetic research are important and yield results that would otherwise be difficult or impossible to achieve. Because of the nature of the research we envision here, we must develop a reward system with an appreciation of the collateral responsibilities, time commitments, and costs associated with collaboration—from data sharing to increased respect for large-scale projects that may appear to dilute the role of individuals.
New forms of collaboration, driven by the need to bring together broad expertise to address multidisciplinary questions, must be encouraged and facilitated. Although individual research will always be important, substantial or comprehensive collaborations are needed to address truly large and complex issues. A topic as broad as the biosphere and involving complex dynamics must include consideration of diverse topics requiring associations among researchers from the disciplines of ecology, microbiology, geology, paleobiology, oceanography, climatology, archaeology, geography, and biogeochemistry. Collaborations need to overlap sufficiently to ensure that the full array of issues under consideration are being addressed, but also be broad enough so that each collaborator has distinctive contributions to make.
Minimizing hurdles to collaboration. Institutional structures and processes that inhibit collaboration must be addressed. While some interdisciplinary programs (programs in systems biology to foster collaborative efforts in molecular biology, for example), have flourished in universities, many opportunities are inhibited by department- or institution-specific requirements and accounting systems for both funds and academic credit.
There is no ideal organizational structure for institutions that facilitates collaboration and coordination of activities. However, institutions can develop specific intellectual bridges for interdisciplinary research and actively promote them. Within, and especially among, academic institutions this may mean breaking down the notion that the “winner” in a collaboration is the person or institution that receives the largest share of a project. This means moving toward a system that recognizes that being a contributor to several or many successful projects is at least as valuable as being the major recipient of support for an individual project.
Within the federal system, this may mean ensuring that the most appropriate people are engaged in research projects, regardless of which agency, division, or research facility they are associated with. This requires a means to use or transfer funds that is not easily accomplished in the current system.
Collaborative research requires additional infrastructure, funding,
and effort. Effective collaboration engenders specific additional costs in both time and money (meetings, communications, resolution of misunderstandings or conflicts) that are not associated with individual research. Collaborators, funding sources, and host institutions must recognize that additional communication and coordination results in a net gain in understanding and knowledge that can be greater than the sum of results from smaller projects undertaken over the same amount of time. Participants in large or diverse collaborative projects may need to develop formal or informal management plans to promote effective interaction. Furthermore, funding agencies must be willing to support the extra infrastructure (in the form of salary, personnel, hardware, and travel for face-to-face communication) needed to support collaborative efforts. While all involved may understand the concept of the division of labor, explicit efforts must be made to ensure that scientists are focusing on the intellectual components of a project while being assisted by able project managers, technicians, and other support staff.
Reward systems. Current academic reward systems tend to depreciate the value of collaborative research. The time spent in coordination may be viewed as unproductive, and the activities required for collaboration (e.g., data integration, developing and employing data standards) may not be judged valuable in their own right by peers or administrators. Papers with many authors, or even synthetic results in a book (versus individual research) may be undervalued. While we must continue to place high value on individual research, we should also accept that the activities associated with collaboration and synthesis are especially valid contributions to scholarship and not a simple form of community service.
Our current reward system embodies a complex, but well-known, system of credit. In smaller projects, the role of multiple authors is fairly transparent, and assigning the order of authorship can be straightforward. In highly collaborative projects, there may be many authors with different responsibilities, and their order of precedence may be difficult to ascertain (see Kennedy, 2003). This circumstance is exacerbated by certain journals that arbitrarily limit the number of authors on a paper, further complicating the reward structure of research efforts.
Protecting intellectual property. There is an apparent conflict between making information broadly available and protecting the intellectual property of those who generated it. However, while some information generated by ecologists and geologists might have economic value, most of the value is perceived by scientists as intrinsic to a greater understanding of the processes of nature. In this case, economic value is replaced by the value of recognition and acknowledgement from peers. Nevertheless, just as individuals and corporations may be unwilling to share information with inherent economic value, scientists often are concerned
that if they make their data broadly available before it has been fully explored intellectually, someone else may take advantage of the effort. For example, a scientist might collect deep-time samples from many locations and make the data broadly available. It may be possible for someone, as the project nears completion, to interpret what is already available, thereby pre-empting the primary researcher. The problem is especially acute for new researchers just beginning to build a reputation for originality and productivity. Scientists must be assured that the potential benefits of open access to data outweigh the risks of misuse.
Although community standards are emerging on these issues (NRC, 2003b), individual scientists will need to find intellectual satisfactions in shared projects that outweigh the overhead associated with such research. Some solutions will emerge from modified reward systems that nudge us into accepting the costs, and benefits, of collaboration.
OPPORTUNITIES FOR COLLABORATION
Developing new cultural and technological infrastructure for promoting and enhancing interactions will help provide the foundation for increased collaboration. We must also support, and encourage the use of, specific opportunities for training, analysis, integration, synthesis, and widespread sharing of information. These opportunities should be supported as distinct entities whose specific purpose is to facilitate those processes that promote collaboration.
Some existing programs are good models for efforts to promote interdisciplinary, collaborative research projects. The National Science Foundation (NSF) supports collaborative efforts at several points along the research track. For example, the Integrative Graduate Education and Research Traineeship (IGERT) Program has as its goal “to catalyze a cultural change in graduate education, for students, faculty, and universities by establishing new, innovative models for graduate education in a fertile environment for collaborative research that transcends traditional disciplinary boundaries. It is also intended to facilitate greater diversity in student participation and preparation and to contribute to the development of a diverse, globally aware, science and engineering workforce.” These awards support graduate students in interdisciplinary studies across and between campuses, with the goal of producing young scientists who embrace collaboration and synthesis. Breadth in NSF-sponsored research is fostered through cross-cutting initiatives such as Biocomplexity in the Environment, division-scale efforts such as the Division of Earth Sciences (EAR) Biogeosciences Initiative, Directorate for Biosciences (BIO)-funded centers such as the National Center for Ecological Analysis and Synthesis and the planned Center for Synthesis in Biological Evolution,
and the BIO-funded Long Term Ecological Research (LTER) sites. NSF also has a long record of support for ocean drilling, perhaps the foremost geoscience collaborative research activity (see Box 4.1). Proposal development, shipboard participation, and postcruise scientific interaction have enabled generations of young researchers to collaborate closely with senior scientists in an intense, multidisciplinary environment.
The Long Term Ecological Research (LTER) program at the National Science Foundation supports integrated ecosystems-level research at
Almost 40 years ago, testing the seafloor spreading hypothesis was the impetus for a group of four U.S. marine geoscience institutions to collaborate on a proposal to collect seafloor core from the world’s oceans. This proposal resulted in the Deep Sea Drilling Project (DSDP), which existed solely as a U.S. program from its first drilling leg in 1968 until 1973, when the first international partner (Soviet Union) joined. By 1975, DSDP was being supported by five international participants and involved nine U.S. institutions.
In the early 1980s it was clear not only that ocean coring using the DSDP drilling platform—the Glomar Challenger—had been immensely successful, but also that there continued to be considerable potential for additional scientific discoveries from the world’s oceans. DSDP terminated in 1983, but a new program—the Ocean Drilling Program (ODP) with a more advanced drillship, the JOIDES Resolution—commenced in 1985. From this time until it terminated in 2003, ODP drilled almost 1,800 holes at more than 650 sites around the globe, ranging from shallow reef depths (40 m) to the deepest ocean (5.9 km). ODP evolved from an initial scientific collaboration involving 10 U.S. institutions and 17 international partners to a partnership involving 18 U.S. institutions with funding support and scientific participation from 22 international partners. As a consequence, many hundreds of individual scientists from around the world participated in ODP collaborative science. One theme that persisted through this history of strikingly successful national and international scientific collaboration was that the broadest possible participation would be sought to maximize the intellectual stimulus available to the program.
multiple sites in North America and Antarctica (see Box 4.2). The sites emphasize ecological studies across many subdisciplines and regional scales. These projects, which concentrate on ecosystem patterns and processes, have been collaborative, incorporating a wide range of disciplines. Similar programs, focusing on other levels of biological organization (e.g., communities, populations) would expand the high value of programs such as LTER. Furthermore, developing new field research programs that support the technically sophisticated techniques and instruments required for contemporary ecological and geological field research would further promote collaboration between scientists from different disciplines.
Many ecological research projects occur on short timescales governed by budget and granting cycles or academic calendars, even though most important ecological processes take place over decades, centuries, and millennia. Recognizing this, the National Science Foundation sponsored a series of workshops in the late 1970s to develop concepts for a program supporting long-term ecological research and promoting forms of collaboration required to sustain a comprehensive research agenda. The program that emerged, the Long Term Ecological Research Network,3 took as it core mission and central intellectual theme the understanding of long-term patterns and processes of ecological systems at multiple spatial scales.
To accomplish this mission, the network initially focused on field research at sites representing many habitats in the United States. Ecological processes at all sites were to be investigated by directing observations, measurements, monitoring, and experiments toward five major areas:
Although all the 26 LTER sites focus on the core objectives of the network, each site has distinctive themes associated with each habitat type. For example, Konza Prairie (KNZ)5 maintains watershed-level manipulation of fire frequency and bison grazing to determine the role these historically
important factors play in maintaining diversity of tallgrass communities. The H.J. Andrews site (AND), managed by the U.S. Forest Service, focuses on successional changes in ecosystems, forest-stream interactions, population dynamics of forest stands, patterns and rates of decomposition, and disturbance regimes in forest landscapes. The two urban sites, Central Arizona6 (Phoenix - CAP) and Baltimore Ecosystem Study (BES)7 are investigating the relationship between ecological, sociological, and economic factors in major urban areas. The other sites contribute equally to core areas while also addressing local and regional issues.
Because the scope of research at LTER sites is quite broad, it is inherently multidisciplinary. This feature is reinforced by the emphasis on the five core areas of research, each of which relies on a wide array of scientists for a comprehensive understanding of the underlying principles. With regard to interactions and collaborations between geologists and biologists, those core areas pertaining to biogeochemistry, soil formation, and surface and subsurface water are particularly reliant on broad field research and intellectual engagement.
As the LTER Network has matured, opportunities have emerged from the core research areas. These include education training and outreach, information management and dissemination, and providing data and ecological insight to the wise management of natural resources. The sites and their network office are also heavily involved with developing research tools, and are actively promoting synthesis between sites.
The LTER Network now represents an important element in the research and intellectual infrastructure of ecology. Ecological research, which has both core elements emerging from theoretical and empirical approaches, and a scope that incorporates many allied disciplines, has flourished in response to the LTER Network. It is clear that this is an effective model that could be embellished or adapted for other forms of collaboration between disciplines.
The proposed National Ecological Observatory Network8 (NEON) would address regional and continental-scale issues in ecology and evolutionary biology through a network of sites focused either on critical environmental challenges or ecosystem types (NRC, 2004). Such a network would certainly facilitate collaboration between ecologists and environmental biologists. However, current plans focus primarily on real-time monitoring and experimentation and do not call for the sort of integration across timescales envisioned in this report. A geohistorical approach could extend NEON’s observational reach to encompass hundreds to thousands of years. This would add value by providing historical context for modern systems, and by providing opportunities for linked modeling, experimental, and geohistorical studies. The availability—or potential for development—of geohistorical records should be considered in selecting NEON study sites, particularly those focusing on climate change, land-use effects, biodiversity, and biogeochemistry.
The National Center for Ecological Analysis and Synthesis
The goal of the National Center for Ecological Analysis and Synthesis (NCEAS)9 is to use existing data from a broad range of disciplines to address important ecological questions. By taking a broad view of ecology, the center has involved many scientists who would not consider themselves ecologists, including geologists, hydrologists, and paleontologists. NCEAS facilitates collaboration by hosting meetings and supporting postdoctoral and sabbatical researchers conducting synthetic research. Furthermore, because the center relies on existing data, it has developed significant research efforts directed toward generic access to data.
Field research in ecology often requires work in distant and isolated locations. One result is that many scientists and students are isolated from intellectual interactions with disciplines peripheral to the particular topic being investigated in the field. Recognizing this, the ecological community has supported the notion of a center for synthesis where ecologists and scientists from allied disciplines could congregate to address important questions.
NCEAS focuses on using existing data to address important ecological questions, and employs three main types of research approaches. The most distinctive are the working groups, teams of up to 20 researchers who conduct meetings lasting three days to three weeks several times over the course of a year or two. Working groups actually conduct research at the center, synthesizing data, developing theories and models,
and testing them with data. Approximately 700 visiting scientists a year participate in working groups. NCEAS also supports four to six center Fellows (sabbatical visitors) each year from around the world. The fellows are in residence for 4-12 months, and typically become engaged in working group activities. Finally, NCEAS supports 15-20 postdoctoral associates for one to three years. About one-third are formally associated with working groups, but even the others, who come to NCEAS to work on their own projects, become involved with one or more working groups. The postdoctoral associates are unusual in that they have no formal mentors. Rather, they interact with the hundreds of visiting scientists at the Center. This yields young scientists with the important skills and appreciation for creative, collaborative, multidisciplinary research. Most projects are initiated by research proposals that are reviewed by a Science Advisory Board composed of 19 members from many disciplines and several countries.
The physical facility consists of meeting rooms and offices for visiting and resident scientists and comprehensive computing capabilities. The atmosphere is promoted by a staff dedicated to minimizing the extra effort required to travel to the center and work collaboratively. A key feature of the center is the breadth of participation. Participants are highly diverse with regard to nationality, academic rank, institutional affiliation, and importantly, scientific discipline. The 3,200 participants belong to more than 180 scientific societies, have published their results in 140 journals, and many come from departments that are not biology-based. The dynamics among the resident scientists (postdoctoral associates and fellows) and the hundreds of visiting scientists each year generate an intellectual ambience that is quite productive (e.g., ~1,000 publications in 10 years, with an average impact factor more than twice that of the top 20 journals in ecology). The NCEAS model is successful in promoting creativity and interdisciplinary research.
While detailed sociological studies are being conducted on NCEAS to characterize how scientists interact effectively, there are also a few obvious features that contribute to collaboration. In the first place, the center was formed at a time when many scientists knew they wanted or would benefit from the opportunities it provides. Perhaps the most important resource the center provides is the opportunity to interact, primarily mediated through specific time set aside to focus on the issues under consideration. Furthermore, scientists are in close proximity and are there explicitly to interact. Another important feature is the center’s pleasant, neutral location; it is not on a campus, freeing participants in unanticipated ways to interact effectively. The facility, logistic support, and comprehensive computing and analytical support lower the activation energy to conduct interdisciplinary research and yield an intellectual ambience that promotes creativity and productivity.
Collaboration, synthesis, and analysis of existing data provide added value to research conducted by individual scientists. The model of collaboration at NCEAS is simple. Attempts are made to choose good projects and productive, creative scientists and then to facilitate their endeavors. It is a highly portable model, and could easily be employed in other circumstances to promote the interdisciplinary efforts envisioned in this report.
Community-wide Databases and Collaborative Research
Science benefits from technological advances that promote information sharing and collaboration. Increased and speedier travel and communication enhancements spurred the sharing of information and development of ideas. Scientists are now approaching the capability to share actual data, even in real time, in a manner that will lead to another quantum leap in the opportunity for integration, synthesis, and understanding. Specifically, we must develop access to the wide array of highly distributed and heterogeneous data that characterize the disciplines of interest—information pertaining to ecological dynamics might range from genetics to global change. This is not only an enormous breadth of data types, but in addition much of the data across this spectrum is gathered, stored, analyzed, and modeled in different ways. The traditional approach to this circumstance is to develop application-specific solutions, particularly databases that are customized for the topics under consideration. While effective to an extent, the data may not be more generally accessible in this form, severely limiting its usefulness. Thus, generic data access and integration tools are essential so that data can be effectively reconfigured for uses other than for which they were originally intended. This requires data management and higher-order access concepts to promote integration across disciplines that have inherently different semantics.
True generic access requires a distinct data management model that provides access to information in place of a uniform database structure. Rather than standardizing inputs and outputs, the solution lies in developing ontologies (complex synonymies of fields and attributes) and metadata (information about the data itself). Information, once linked through ontologies or characterized with metadata, can be made accessible in context-dependent ways that are both comprehensive and efficient. With a community-based web portal to the database, hundreds of thousands of data resources, from small compilations (doctoral dissertations, individual researchers’ results) to large, well-known data sources (weather, ocean temperatures, remotely sensed data) can be incorporated in research, planning, and resource management. The key tools are efficient and flexible ways to describe and enter ecological and geological information, powerful searching capability, and tools for data visualization and analysis.
Effective collaboration requires more than access to pertinent data. Collaborating scientists must be able to quickly analyze data, and transport it into appropriate visualization tools that make the data understandable to the array of collaborators from several fields. Once user-friendly tools to acquire and analyze data are available, it will be possible to quickly find, download, and analyze disparate datasets to test an idea much in the way we now use abstracts of articles to get a sense of whether a new idea warrants further consideration.
Existing community-wide database efforts. An extraordinary amount of ecological, environmental, geochronological, and paleobiological information is becoming available electronically. Many individuals, research groups, and institutions are providing access to vast amounts of data online in the form of customized datasets generated to address specific questions or issues. These efforts to turn data into datasets add value that can be useful in unanticipated ways. We note here some ongoing efforts that will be important to the successful analysis of the geologic record of ecological dynamics.
The Paleobiology Database10 is an NSF-funded project that provides global, collection-based occurrence and taxonomic data for marine and terrestrial animals and plants of any geological age, as well as web-based software for statistical analysis of the data. This project currently has 133 participants from 57 institutions in 11 countries and promotes collaborative efforts to answer large-scale paleobiological questions by developing a useful database infrastructure and bringing together large datasets. It presently contains 409,210 occurrences (documented presence of a taxon at a geographic-stratigraphic site) among 39,721 collections.
FAUNMAP11 is a relational database that documents the temporal and spatial distribution of mammals for the last 2 million years in Canada and the United States (Faunmap Working Group, 1994). Entries in the database are based on collections in a public repository with precise geographic location and a fine-scale chronological framework. FAUNMAP is linked to a Geographic Information System equipped with ARC/INFO software that allows for the geographic distribution of more than 200 individual species to be mapped for 11 distinct time periods. Statistical parameters derived from manipulation of the database and other treatments can also be mapped at a variety of scales.
The Chronos Project12 is an NSF-funded effort to assemble, integrate, analyze, and disseminate geologic and paleontological data relating
to geologic time with a platform that links to a variety of chronostratigraphic, paleogeographic, paleoclimatic, paleontological, and geochemical databases. Chronos will be a “federation” of databases. Chronos will also provide individual researchers with web-based tools to search for information relevant to research questions, Geographic Information System and Time Information System capabilities, correlation routines that can be remotely run on the San Diego Supercomputer, and the visualization tools necessary for the effective display and integration of information about the history of life on Earth. Still in its early stages of development, Chronos is also designed to provide outreach to students and the general public.
EARTHTIME13 is an NSF-funded project involving geochronologists, paleontologists, and stratigraphers to produce the highly resolved geological timescale necessary for the rigorous analysis of evolutionary and geological rates. Current geochronological information, particularly in deep time, is insufficient to adequately constrain rates and thus processes of change. This effort will work closely with Chronos and other projects.
The North American Pollen Database14 (NAPD) is a cooperative database of Quaternary pollen data from North America and is one of several continental databases that together constitute the Global Pollen Database15 (GPD). The NAPD is funded by the National Oceanic and Atmospheric Administration (NOAA), whose primary goal is to facilitate the application of fossil pollen data to paleoclimatic inference and model testing. However, the NAPD has been widely used in ecological and biogeographic studies (e.g., Williams et al., 2001, 2004). The NAPD, GPD, and other paleontological databases administered by the NOAA’s World Data Center for Paleoclimatology (e.g., North American Plant Macrofossil Database, Western North American Packrat Midden Database, Insect Database, International Multiproxy Paleofire Database) deserve continued support, particularly from agencies charged with supporting research in biosphere and ecological history. NOAA’s Paleoclimatology program has recently been targeted for elimination or severe downsizing; such action would be disastrous for these vital databases. Increasing research emphasis on biosphere responses will place more demands on paleoclimatic reconstructions, and in many cases will require development of paleoclimate records and analytical capabilities that are specifically geared to biological questions. Although the NOAA paleoclimate databases are
managed specifically for paleoclimatic applications, many are of major importance for paleoecology (e.g., pollen, plant macrofossils, tree rings).
The Geosciences Network (GEON) research project,16 supported by NSF, is a collaboration between information scientists and geoscientists to develop ontologies and portals to link datasets and provide analytical and visualization tools to the geosciences community at large. The central Appalachians and the Rocky Mountains are being used as “test beds” to demonstrate how existing geophysical, structural, geochemical, paleontological, and paleogeographic databases can be linked to examine the tectonic, sedimentary, and biotic evolution of the two regions. Integration of the heterogeneous datasets in these two areas will provide a model for the even broader range of geosciences datasets necessary to understand the complex dynamics of the earth system.
Deep Time17 is a collaborative project between paleobotanists and molecular systematists focused on the evolutionary history of plants. It is funded through the NSF’s Division of Biological Infrastructure and involves more than 100 participants as well as postdoctoral support. The project addresses common interest areas related to molecular clocks, integration of molecular and morphological data, examples of cladistic analysis of particular datasets, and the agreement and discordance of the molecular and paleontological datasets. These interest areas often cross boundaries between plant- and animal-based data, so both are included.
The development of these community databases provides some important general lessons for the future:
The most successful databases appear to be those that are motivated by closely related scientific questions generated at the community level. Such grassroots efforts ensure the “buy-in” of contributors, the rapid generation of research results, and the continued growth and evolution of the effort.
Development and management of databases involves interactions between natural scientists and information scientists, whose research strategies and objectives may not always converge. The problems that may stymie natural scientists may be routine and uninteresting to information scientists, and vice versa. Communication between these cultures is essential, and creative ways to facilitate it are needed.
Databases require commitments from one or more funding sources to ensure that they continue to be maintained and to assimilate data as
long as they are scientifically useful. They also require long-term commitments from scientists and institutions to house them. Research museums may be particularly well suited as homes for databases because curatorial and archiving activities are routine functions in these institutions.
Databases involve trade-offs between posterity-driven archival functions and the immediate needs of scientific users. There is a risk that shortcuts taken in data formatting, database architecture, and metadata assimilation may facilitate short-term and ongoing applications at the expense of future applications that may have different requirements.
The various models discussed in this chapter (e.g., IGERT, LTER, NCEAS, and community-wide databases) promote collaboration through training, research, and analysis and synthesis. Although most of these particular examples are supported by NSF, they—and other models that encourage collaborative research—are easily exportable to other research and education entities, from individual campuses to government agencies. Indeed, successful large-scale multidisciplinary efforts in genetic sciences (e.g., National Center for Biotechnology’s GenBank,18 the Human Genome Project19) and NASA’s Astrobiology Institute20 demonstrate that effective collaborations among disciplines can be stimulated and supported though the creation of appropriate programs, centers, and other community-based efforts. Future progress in understanding the geologic record of ecological dynamics will require not only new and better data but also better capacity for analyzing and synthesizing the data that we already possess. Although the collective cultures of the relevant disciplines are evolving in this direction, much more can be done to facilitate this evolution at levels ranging from individual institutions to funding agencies.