Skip to main content
Science News

More than a decade ago, ecologist Scott Ollinger helped launch U.S. ecology’s flagship foray into big science. He and other researchers worked to transform the National Science Foundation’s (NSF’s) dream of a continental-scale observatory that would monitor environmental change into a concrete plan. What emerged was the National Ecological Observatory Network (NEON), a unique string of more than 100 data collection stations spread from Alaska to Puerto Rico.

So Ollinger was thrilled when, in 2013, NEON offered him the chance to oversee the network’s expected trove of data on long-term changes in climate, land use, bio-diversity, and invasive species. He arranged for a 3-year leave of absence from his post at the University of New Hampshire in Durham. Then he hit the road to NEON’s headquarters in Boulder, Colorado.

En route, however, Ollinger learned that NSF, which is paying for NEON, had put a hold on an initial $111 million grant to begin operating some of the newly built stations. That meant “I was almost fired the day I arrived,” he says.

The disturbing news was a harbinger of worse to come. Despite his impressive title of observatory director, Ollinger discovered that he had little influence over how NEON was being built, or the day-to-day activities of its growing scientific staff. Soon, “the number of decisions I tried to make that were overruled reached a point where I felt there was no way I could succeed,” he recalls. Frustrated and feeling powerless, Ollinger returned home after less than a year.

Ollinger’s experience reflects management problems that have dogged NEON since its birth and the project’s tense relationship with the community of scientists who will ultimately use its data. This summer those problems came home to roost.

On 3 August, NSF abruptly announced it was scaling back the project in an attempt to prevent an 18-month slip in its schedule and a projected cost overrun of more than $80 million on its $434 million construction budget (Science, 7 August, p. 574). On 8 September, NEON Inc., the nonprofit that manages the project, fired CEO Russ Lea, a former forestry professor and university administrator, after the head of NSF’s biology directorate, James Olds, ordered the corporation to correct “deficiencies in leadership.” And last week Olds told a congressional committee investigating what has gone wrong that NSF would consider replacing NEON Inc. if it doesn’t shape up.

NSF officials say NEON’s “descoping” was prompted by ongoing difficulties in obtaining needed site permits and technical challenges in building NEON’s sensors, some of which take novel approaches to collecting data. NEON’s supporters note that other large, complex science projects that NSF has built have undergone periodic changes in scope and leadership, particularly as they transition from construction to operations. And NEON has been especially challenging because of its complexity and uniqueness, Olds says.

But scientists both inside and outside of NEON say the project’s woes run much deeper. They point to a chronic disharmony among NSF, NEON Inc., and the research community. Ollinger, for instance, is one of five researchers who has held—and then left—NEON’s top scientific post since 2007. This past spring, members of NEON’s chief scientific advisory body even considered a mass resignation.

Now, as NEON regroups, the scientists with whom it has had a love/hate relationship say NSF and NEON Inc. need to turn things around, and fast. “I wish them luck,” says Scott Collins, a plant biologist at the University of New Mexico, Albuquerque, who helped get NEON off the ground as an NSF program manager in the early 2000s. “They need to wake up and change the way NEON operates and get the research community behind the project,” he says. “NSF has invested a ton of money in the infrastructure,” he adds, “and if NEON fails, ecology won’t get another chance.”

When then-NSF director Rita Colwell proposed what became NEON in 2000, she hoped that it would generate questions researchers had never been able to ask—or answer. Although the agency had been funding Long-Term Ecological Research sites across the United States since 1980, those projects tended to focus on hypothesis-driven research by an individual investigator. They weren’t designed to collect and share highly standardized, continental-scale data over many decades.

The move into Big Data is an intoxicating vision to many. “The idea of a community of ecologists coming together to put up a piece of infrastructure as significant as a telescope, atom smasher, or an icebreaker sucked me in,” Lea said last month, explaining why he took the CEO job in early 2012.

It took NEON’s planners a decade and several tries, however, to draw a blueprint acceptable to NSF’s oversight body and Congress. The final plan called for dividing the United States into 20 ecological domains (see map below). Each domain would host two “core” observing stations chock-full of standardized sensors and sampling sites (see graphic, below). One core site would focus on a terrestrial ecosystem such as a forest or grassland, the other on an aquatic environment such as a stream or lake. In addition, the domains would support a total of 56 “relocatable” stations that researchers could move a few times during the 30-plus years that NEON is expected to operate. The original plan also included a long-term experiment, called STREON (STReam Experimental Observatory Network), which would simulate abrupt environmental change in aquatic ecosystems by adding nutrients—phosphates and nitrogen—and removing some organisms at 10 sites.

NEON takes shape

Scientists will be collecting information from both terrestrial and aquatic sites located in each of 20 distinct ecosystems, called domains.

All

Aquatic sites

Terrestrial sites

All

Core sites

Relocatable sites

Descoped sites

Descoped urban sites

Show domains

Alaska

Hawaii

Puerto Rico

Notes: Map of core aquatic sites includes eight sites that were to host the now-canceled STREON experiment. There are two core aquatic sites in Florida (Domain 3, Southeast). There is no core aquatic site in Domain 20 (Hawaii). Map includes a planned core terrestrial and a planned relocatable terrestrial site in California (Domain 17, Pacific Southwest).

Since 2011, project managers have completed construction on 48 sites—fewer than half of what was in the original plan—and spent approximately two-thirds of NEON’s construction budget. The descoping preserves the 40 core sites, but eliminates 15 of the 56 relocatable sites—
including seven dedicated to studying urban ecosystems. NEON also pulled the plug on two terrestrial instruments: sensors to measure fluxes of nitrogen oxides and methane, and fiber optic cables for collecting video of underground root growth. And it dropped the STREON experiment (although NSF officials emphasized that they would welcome new STREON-like proposals to another NSF funding program).

The loss of STREON was the latest defeat for aquatic scientists, who had long been unhappy with what they regarded as NEON’s inattention to its river and lake sites. In June, several prominent scientists petitioned NEON to invest more in completing the aquatic observatories. Sensing that STREON was in danger, they also asked to be consulted on any decision to drop STREON.

NEON managers rebuffed both requests, saying that “we cannot make one component of the observatory a higher priority than others.” But the descoping does exactly that, argues ecologist Walter Dodds of Kansas State University, Manhattan, who organized the petition and who has championed STREON. “It’s terrible news for aquatic scientists.”

It’s not unusual for a federal agency to adjust its plans for a major scientific facility, such as a telescope or spacecraft, after construction is underway. But those changes are usually the product of discussions between scientists and project managers. On a typical NASA mission, for instance, “the job of the chief scientist is to understand the high-level science requirements of the mission and to engage in respectful conflict with the project manager to make sure that the best outcome occurs,” says David Schimel, NEON’s first CEO and later its first chief scientist. “They succeed or fail together.”

That give-and-take has not been the norm at NEON, Schimel and others say. In late 2007, for example, geophysicist Michael Keller left his job as a project scientist for a NASA-funded program in the Amazon, bought a house in Boulder, and moved his family in preparation for what he expected to be the crowning achievement of his career: chief of science at NEON. “We had a golden dream that was going to make this incredibly difficult thing happen,” he recalls. “That idealism was our calling card.”

Keller’s first task was to reach a consensus on the scientific requirements for the observatory. “Then we converted those questions into what we were going to measure and how we would report them as products” that scientists could use, he says. The result, he says, was “a very respectable final design.”

That’s when things headed south. “We fully expected to have to adapt what we were doing on a site-by-site basis,” Keller recalls. But that’s not how NSF saw things. “NSF’s model is that you do the science up front,” he says. “And once you come up with the final design, it’s up to the project manager to execute it.” The message from NSF was clear, he says: “Once we had designed it, [scientists] were somewhat obsolete.”

Ollinger says that approach may work well when building a single large facility with a clear and compelling scientific objective—he calls it a “north star”. But NEON lacks that north star, he says. Instead, its fundamental objective is to generate high-quality data that scientists will use to answer a wide array of questions.

After about 3 years at NEON, Keller “decided it was probably time for me to move on.” In late 2010 he returned to Brazil to manage a sustainable development project funded by the United States and Brazilian governments.

Keller was succeeded by the man who had hired him: Schimel. A biogeochemist who has been a tireless advocate for NEON, Schimel initially tried to recruit people who understood both ecology and what it takes to build a large scientific facility—before realizing that those two cultures rarely overlap. “It was difficult to find ecologists with experience in large projects,” Schimel says. “It was equally hard to find engineers and project managers with experience in ecology. And by difficult I mean impossible—they didn’t exist.”

Even so, Schimel says he’s proud of the team he assembled during his 5 years at NEON. But eventually he was also pushed aside. “My science role was being increasingly marginalized,” he recalls. “I was losing the authority and access to the systems engineering staff and other expertise I needed to do my job.” Schimel left NEON in 2012 to join NASA’s Jet Propulsion Laboratory in Pasadena, California, where he’s analyzing global carbon data.

Next up was Ollinger, whose year at NEON was equally disheartening. Ollinger found out that he didn’t have the promised authority to make sure that sensors passed muster before they went live at a site. Nor was he allowed to create career paths for NEON’s growing staff of scientists, who could never get a straight answer from project managers about whether they would continue to have jobs once NEON was running (see sidebar, p. 1441). A third role that Ollinger relished—figuring out how outside scientists would access NEON’s data—was impossible to fulfill, he says, because “the data weren’t flowing.”

Ollinger’s successor as observatory director, C. J. Loria, lasted just 4 months. A former Navy test pilot hired for his business acumen, Loria was ousted this past winter at the same time that NEON Inc. eliminated the position of observatory director.

The churn has deepened the rift between scientists and the project by creating a “lack of a scientific presence” at the Boulder headquarters, Lea admitted before his departure. “The community wants a mano-a-mano relationship with a strong scientific leader at NEON on a daily basis,” he said. “Scientists want to talk to their peers.”

Lea’s interim replacement as CEO is Eugene Kelly, a soil scientist at Colorado State University, Fort Collins, who only this summer was hired to be NEON’s visiting chief scientist. Kelly agrees that the research community “feels it has been kept in the dark about NEON for many years.” The low point in NEON’s relationship with the ecology community may have come this past winter, when members of its principal advisory panel, the Science, Technology and Educational Advisory Committee (STEAC), seriously considered disbanding the group.

STEAC “made several explicit recommendations over the years, and those recommendations were either ignored or opposed,” explains the panel’s chair, integrative biologist Todd Dawson of the University of California, Berkeley. “People were saying, ‘There’s no point having an advisory committee if [NEON] is not going to use it.’”

James Collins, chair of NEON’s board of directors, agrees that top management has historically shown a disregard for what scientists can bring to the project, and says that attitude must change. A biology professor at Arizona State University, Tempe, who helped get NEON off the ground as head of NSF’s biology directorate in the late 2000s, Collins says the board expects the next CEO to take a different approach. “The CEO has to set a tone in which people feel they are being treated well and their contributions are valued,” he says.

How NEON will work

NEON will use a standardized suite of autonomous instruments and data collection protocols at dozens of terrestrial and aquatic sites to monitor long-term, continental trends in climate, land use, biodiversity, and invasive species. Core sites are expected to operate for 30+ years; relocatable sites would move every 8-10 years. Sites vary in size, the largest being a 500 sq km swath of the Greater Yellowstone Ecosystem.

Tap to read more

Towers

Towers as tall as 60 meters will provide a platform for sensors, operating at different wavelengths, which will measure sunlight, carbon flux, and a host of atmospheric and meteorological parameters at distances up to 1500 m. A camera at the apex will record seasonal changes and snow depth every 15 minutes.

Soil sampling

A string of sensors up to 200 m long and buried 2 m in the soil will monitor changes in temperature, nutrient movements, and carbon dioxide.

Utility hut

A 6-by-2.5-meter utility hut, 3 meter high, stores scientific equipment and houses computer and communications gear with fiber optic, cellular, or satellite connections. Also contains battery back-up if power grid fails.

On the water

A suite of instruments, including a buoy in middle of lake, will periodically measure various chemical and physical attributes. Other sampling protocols will track changes in populations of microbes, algae, invertebrates, and fish.

On shore

Sensors will record the movement of ground water and collect various meteorological and hydrological data.

Relations between NEON Inc. and NSF also need to improve, say scientists both within and outside the project. The funding delay that sabotaged Ollinger, for example, was the result of a festering disagreement over when a site is ready to be commissioned.

It’s not a minor issue. NEON managers argue that a site should be considered operational once all the equipment works and the instruments start to generate data. Any delay in commissioning, they note, forces NEON to use construction dollars for operating costs, such as power and maintenance. That leaves less money to complete new sites.

NSF’s position, however, is that a site cannot be commissioned until its data are available online, says Elizabeth Blood, the agency’s longtime project manager for NEON. That process could add months to the commissioning process, she concedes, adding that NSF has no intention of changing its criteria.

NSF has the authority to decide the issue. But NEON’s position got a strong endorsement this past July from a high-powered panel of scientists from both inside and outside the project, which reviewed NEON’s future shortly before NSF announced the descoping. Some of NEON’s cost overruns were due to “delayed transition to operations,” the panel concluded. Its recommendation was unequivocal: “The cost of carrying field operations on the construction project is unjustified,” and NSF needed to start paying the operating costs.

Even NEON’s critics are willing to cut NSF some slack, however, because they recognize that the foundation has little political margin for error. The Republican-led science committee in the U.S. House of Representatives has repeatedly questioned whether NSF has been a proper steward of taxpayer dollars, and NEON’s missteps have provided some ammunition for those attacks. In recent months the committee has held hearings to berate NSF officials for allowing NEON Inc. to use $150,000 of its management fees on what the agency later admitted were inappropriate activities, including a Christmas party.

Last week, the panel grilled the agency on its oversight of the entire project, and Olds made it clear that the corporation is on shaky ground. “By December 1 NSF will have enough information to make a determination as to whether NEON Inc. has made sufficient improvement to successfully complete construction,” Olds told the panel. Pressed by one legislator whether that could mean replacing the current contractor, Olds hemmed and hawed before concluding, “Yes, that is an option.”

NSF has already shortened the leash. Olds said NSF is taking a closer look at the project’s financial books, and an advisory committee to the biology directorate is examining whether the descoping will affect NEON’s scientific goals. The National Science Board, NSF’s oversight body, has formed a NEON task force. Olds was also critical of NSF’s performance to date. “We could have done a better job,” he admitted.

Despite its many problems, NEON has made considerable progress. Managers said last month that 33 sites in 15 domains are now ready for operations. By September 2016, Lea predicts it will have “upward of 60% of final capability” at the 81 sites currently planned. The final goal, he says, is “100% of capability by the end of 2017.”

Getting to 100%, however, will require NEON to fully resolve longstanding permitting problems. NEON doesn’t own any of its sites, so before it can do any work it must obtain the permission of the landowner—whether a federal or state agency, an environmental nonprofit, a university, or private individuals. Construction also has to go through numerous environmental reviews. It all has taken much longer than anyone anticipated. “We’ve needed probably five to 10 times more permits than was originally thought,” Lea says. “It’s become a huge drain on time and resources.”

A tower studded with sensors keeps watch at a NEON core terrestrial site near Front Royal, Virginia.

STREON posed an especially high permitting hurdle that NEON never cleared. “Dropping pollutants into a reach of streams for 30 years was a hard thing for most people to swallow,” Lea says.

NEON officials also have had to deal with everything from protests by local residents to a pair of murders that ultimately doomed an urban site in Puerto Rico (see sidebar, p. 1440). In Hawaii and Alaska, the permitting process has been so problematic that this past summer NSF officials proposed dropping those two states, plus Puerto Rico, from NEON. Scientists reacted with horror, pointing out that Hawaii alone provides 25% of the climate variability across NEON sites and that, together, the three locales double the amount of biodiversity being monitored. The idea, which NSF’s Blood says was simply a trial balloon, was eventually abandoned.

The descoping offers outside scientists a golden opportunity to reconnect with the project, Kelly says. Last month the Ecological Society of America issued a supportive letter from 16 present and past presidents. “We remain excited about the potential new science that could emerge” from NEON, they wrote, asking NSF and NEON Inc. “to re-engage with the ecological community.”

But NEON board chair Collins acknowledges that the project’s checkered history means it has a lot of ground to make up. “It still needs to prove itself to the community,” he says.

Additional Credits

Video header: Christy Steele

Video: Christy Steele

Design and development: Chris Coleman

DOI: 10.1126/science.aad4620