9 June 2010 • The days when a new discovery in the atmospheric sciences was first announced on a piece of paper have gone the way of the dodo. Just as the Web enabled the change from paper to electronic journals and digital libraries, the new, more interactive and versatile uses of the Web (often called Web 2.0) are beginning to enable equally profound changes in the entire life cycles of scientific publications—from online collaboration during the writing phases through the changing journals where papers are published to the sites where they reside. Among these developments, one that is currently looming large in many scientists’ minds is open access.
A universal ideal
“Open access is essential, and I fully support it,” says NCAR associate director Greg Holland. “I see it as a natural evolution in scientific publishing.” The ideal that scientific information should be widely available for free within a reasonable time after its creation has always been the norm in the atmospheric sciences, where progress would be halted without global data. So it’s not surprising that Holland’s sentiments are almost universally shared—but with reservations about the cost and consequences of the transition.
In the United States, federal funding for research is one cause of the push toward open access. About a decade ago, the director of the National Institutes of Health voiced the idea that citizens who pay for research through their taxes should not also have to pay to see the results of that same research. This principle worked its way into law in 2008 in the form of a mandate from Congress that research funded by NIH be archived on an open site within a year of publication.
Meanwhile, the idea spread to encompass all federally funded research, gaining extra momentum from the Obama administration’s initiative for more transparency in government. In 2009, a bill called the Federal Research Public Access Act was introduced in the Senate, calling for all federal agencies that fund more than $100 million in research to provide open access to the results. The bill was reintroduced in the House of Representatives on 19 April, and a final version is expected to pass this year.
At the same time, the publishing industry has also been changing. The birth of new scientific subdisciplines, growth of interdisciplinary research, and increase in the volume of data being produced have led to a skyrocketing number of scientific journals. Subscription prices are escalating at a rate well beyond the Consumer Price Index. With more journals costing more money, even the best-funded research libraries can no longer count on being able to maintain a complete collection, and smaller libraries have to weigh each acquisition.
Richard Clark (Millersville University).
UCAR trustee Richard Clark (Millersville University) gives an example: “Our library cannot afford the principal journal in my field, Boundary Layer Meteorology. We get it on microfilm one year later. Through colleagues I’m able to keep current with articles that are related to my particular subdiscipline, but lack of access is a significant hurdle for me to overcome.”
Open archives and open repositories are one way over this hurdle. Among the best known of these is the 18-year-old arXiv.org, which includes some 600,000 preprints and papers in physics, mathematics, computer science, and related disciplines and received 30 million downloads last year. The long lifespan and related subject content of arXiv.org make it a good bellwether for the effects of open access archiving on the atmospheric sciences, which don’t have a subject resource equivalent to arXiv (though OpenSky will be a good start—see below). Holland and Clark are both enthusiastic about that idea. Clark says, “Having open access to journals from the desktop or a laptop would allow me to scan what is going on across the enterprise.”
The cost of free access
If people can read a paper for free shortly after it’s published, it seems obvious that journals’ subscription revenues will drop. The cost of publishing the paper, however, isn’t going to drop in tandem. Scholarly publishers of all sorts are revising their business models to adapt to the challenge.
The largest commercial scholarly publishers are well-funded international corporations. “If I were to generalize,” says NCAR special projects librarian Jamaica Jones, “commercial publishers are more adaptive to open access than the professional societies because they have the luxury to embrace it.” One of these adaptations is to publish open access journals, such as Copernicus Publications’ Atmospheric Chemistry and Physics (see below).
The most common business model for OA journals is known as “author pays,” though it’s usually the funder who foots the bill. The publisher charges a sum in the realm of $1,000–3,000, which covers the cost of producing a peer-reviewed paper and maintaining the site. There’s a certain logic to this model, because it treats the cost of disseminating results as just another part of the cost of doing research. However, a sum that seems reasonable in the United States or Europe may be crippling for scientists in developing countries. “It’s already very difficult for a professor who makes $300 or $400 a month to pay page charges,” says physicist Solomon Bililign (North Carolina A&T State University), who is originally from Ethiopia and taught at a university there last summer.
AMS executive director Keith Seitter.
In the nonprofit world, Keith Seitter, executive director of the American Meteorological Society, gives an example of how critical subscription income is: “Our journal program is the biggest part of our financial base. If we were in a position where we had to draw funds from other programs to subsidize the AMS journals, we would just collapse.”
Precedent shows, however, that OA archiving has much less effect on professional societies than on commercial publishers. For example, in the field of high-energy physics, the American Physical Society reported last year that its journal Physical Review D (the largest journal in the field) had experienced no decrease in subscription revenue that could be attributed to arXiv, despite the fact that 97% of papers in PRD also were posted at arXiv.
Seitter expects that a new AMS/UCAR agreement on open access (see below) will have an equally negligible impact. “No one, including the people who pay for our journals, wants us to be in the position where we can’t do good things for our community because we don’t have that revenue. It comes down to making sure that institutions still have a really strong reason to subscribe to the journals.”
OA archiving also comes at a cost. Cornell University Library has hosted arXiv.org since 2001, shouldering the cost of salaries, server charges, upgrades, etc. That cost has grown to a crippling $400,000 annually. In February, the library began asking institutional users to contribute voluntarily to the site’s upkeep on a sliding scale from $4,000 for the top 100 downloading institutions to $2,300 for the lightest users. (There’s still no charge to individual users or submitters.) According to the March issue of Physics Today, most of the heaviest users have already agreed to pay. The library has stated, however, that this voluntary-donation model will continue for no more than three years, and it is seeking input on a permanent business model.
As open access journals grow in number, the most common question scientists have about them is whether the peer review process is as rigorous as at conventional journals. The short answer is either yes or yes-plus. Some OA journals, including Advances in Meteorology, use exactly the same review process as conventional journals. Other journals add more reviews, not less.
At Atmospheric Chemistry and Physics, incoming papers go through the usual initial evaluation by an editor and are assigned anywhere from two to five reviewers. But then they are immediately posted online, and when the assigned reviewers’ comments arrive, those are also posted. What makes the process different from a conventional peer review is that not only the assigned reviewers but anyone else who has registered on the site may post a comment. Assigned reviewers post anonymously, and other commenters may also post anonymously or by name. Once the two-month discussion phase is over, the paper goes through the same revision process as in the conventional model. The final paper appears on a different page.
NCAR’s Eric Apel, who is currently going through ACP’s review process, says that 90% of the open comments he and his coauthors receive are anonymous. “It is a bit scary at first, putting something out there on open access for anyone to shoot at, including the referees. However, referees are encouraged to offer comments that are constructive and not demeaning.”
It’s too early to judge the value of open review, and particularly of the additional comments. A 2006 trial of open review by Nature found that, while authors ranked the open comments as useful or very useful, the editors said that there wasn’t a single case where an open comment made a difference in a publication decision. However, the fact that all reviewer comments are posted for anyone to read could eliminate the concerns about peer review that emerged during Climategate. Apel sums up, “It’s an interesting concept, but I haven’t decided whether I prefer it to a conventional review or not.”
Apel points out one distinct advantage of open review: the paper can be read at least two months earlier than it would be with a conventional review. “We thought about that when we were deciding where to publish. We liked the idea that you can get the information out there and people are looking at it immediately.”
NCAR special librarian Jamaica Jones.
News from OpenSky
OpenSky, a new institutional repository serving UCAR, NCAR, and UCP, will be launched this September. It was developed in support of the organization’s recently passed open access policy. “The policy was the first of its kind to be adopted by any NSF-funded FFRDC [federally funded research and development center],” says NCAR Library director Marlino, “and it stands as a solid step toward open access to research in the atmospheric and geosciences.” OpenSky will allow people both inside and outside the institution to freely access peer-reviewed publications and other works by NCAR and UCAR authors.
The repository recently reached a milestone: a database of citations has been completed for 15,000 peer-reviewed publications from 1960 to 2010. Special projects librarian Jamaica Jones is quick to point out that these are just citations, not the papers themselves. “We collected citations from the Web of Science, old NCAR Annual Scientific Reports, and the publications database maintained by MMM [NCAR’s Mesoscale and Microscale Meteorology Division],” she says. “We recognize that it is probably incomplete, but it’s fairly robust, and we’re proud of it.”
Although OpenSky is ready for submissions, the format will change between now and the official launch. Currently, submitted papers are supported through the MMM publications database, developed by Jun Akiyama, Pat Waukau, and Kay Sandoval. Jones and her team are adapting the structure of that database to capture more and different kinds of information. “For example, rights information must be captured and shared,” says Jones. “That’s been a creative challenge.”
Another challenge is getting scientists into the habit of submitting their papers, a feature of publications reporting that is required by UCAR’s open access policy. “Behavioral changes are always difficult,” Jones points out. To make sure that future submitters are informed and comfortable with the process, she has been meeting with groups from all labs and divisions throughout the year. “Their response has been that it sounds great and makes perfect sense,” she says. “But we’ve also had a lot of questions. Authors don’t want to violate their copyright agreements, and they’re especially concerned about how OpenSky is going to affect our relations with our academic societies.”
With 50% of papers by NCAR and UCAR authors being published either by the American Meteorological Society or the American Geophysical Union, this concern has been foremost in Marlino’s mind as well. “They are extremely valuable partners; clearly, we were not going to make any moves without consulting them,” she says.
In December 2009, following discussions with Marlino and AGU, Keith Seitter of AMS announced a change in AMS copyright policy specifically for NCAR and UCAR authors of papers published in AMS journals, who may now deposit their articles in OpenSky six months after the official publication date. Seitter explains that AMS generalized this agreement to apply to other archives. “We’d been approached by Harvard and MIT separately, wanting us to come up with some compromise on what we needed versus what they wanted,” he says. “We told them we would work with UCAR and NCAR first and come up with a policy that works for all institutions.”
This spring, UCAR also completed an agreement with Springer Publishing, the world’s largest commercial publisher of scientific and technical journals. Like many other publishers, Springer currently allows authors to designate their papers as open access immediately upon publication for a fee of $3,000. The publisher has agreed to waive that fee for NCAR and UCAR authors—whether first authors or coauthors—for five years. During that time, Marlino says, “We have agreed to work together on a joint research project to look at the impact of open access on the advancement of the science. There’s a significant body of evidence that open access translates into a greater citation rate, but we want to go beyond that: What is the impact on the science?” Marlino is now contacting schools of library science around the country in hopes of involving them in the project.
Solomon Bililgn (North Carolina A&T State University).
Access in the developing world
“Open access will help if scientists really don’t have to pay,” says physicist Solomon Bililgn cautiously. However, he points out that connectivity is the main problem. Teaching in Ethiopia last summer, he found that the lack of high-speed Internet connections and unpredictable losses of power crippled his students’ access even to free information. “The American Physical Society allows three free journal downloads a month, but the system is extremely slow. Sometimes you’ll sit there for half an hour waiting to open an e-mail.” His students came to school at 6:00 a.m. to download material for class before the Internet traffic became too high.
For scientists looking to publish, however, cost is probably the biggest barrier. Choosing a journal is a balancing act, Bililign says. “They want the prestige of publishing in a reputable journal from America or Europe, but it’s very expensive.” NCAR’s Andrew Gettelman, who did a study on barriers to information in developing nations, notes that scientists there may prefer to publish in commercial journals that don’t require page charges. AMS’s Keith Seitter points out that most nonprofit publishers will waive charges if necessary; AMS itself waived $500,000 worth of page charges last year.
by Carol Rasmussen