Citizen Science Takes Root

Kayri Havens & Sandra Henderson. American Scientist. Volume 101, Issue 5. Sep/Oct 2013.

In the mid-1800s, Henry David Thoreau sojourned outside Concord, Massachusetts, immersing himself in the world around him. He planted beans, entertained visitors, repaired to his mother’s house in Concord for hearty meals, and wrote Waiden; or, Life in the Woods. Less widely read are the copious observations and measurements he made around the same time, between 1851 and 1858, of plants and animals at Waiden Pond. But more than a century later, his records are drawing attention in their own right among ecologists seeking a view into past climates.

Thoreau was especially interested in the initial leafing and flowering of plants in springtime. He made notes on more than 500 plant species during his time at Waiden. His observations contain valuable reference points for judging how those species have responded to changes in their environment, including changes in temperature and precipitation, over periods for which there would otherwise be little data.

Thoreau thus might be considered an early practitioner of what we now call citizen science. The field, which was given its name in the 1990s by researchers at the Cornell Laboratory of Ornithology, empowers people from all walks of life to participate in the scientific process and help advance knowledge in a wide range of scientific disciplines. Thanks in part to the ease of collaborating with partners and reporting results via the Internet, such projects have proliferated over the past decade, collecting tens to hundreds of thousands of observations and working at continental and even global scales. Their benefits are educational and also scientific: Many studies have found that when reasonable quality control methods are used, the data citizen scientists collect are of publishable quality. The scientific community is beginning to recognize citizen scientist partners as important collaborators, even as ambassadors for science.

Project BudBurst, a large-scale project we have codirected since its inception in 2007, is demonstrating the results such work can have for science education and data collection. In the spirit of Thoreau and other early observers of plant life, the project engages people across the United States in a collaborative effort to gather data on plant life cycles. In the process, participants are broadening their own scientific knowledge-and helping ecologists discover how plants respond to environmental change.

The Science of Appearance

Tracking species’ responses to climate change is notoriously tricky. Without large data sets collected over long timescales, it can be hard to see clearly the’ effects of gradual changes in temperature, rainfall, and other factors. If the data cover only a short period, an anomalous year can have a large effect on observed trends.

Project BudBurst takes a crowdsourcing approach to this problem. In the six years since the project began, participants have contributed tens of thousands of observations on hundreds of plant species-and the number of participants continues to climb. Plants’ life cycles are an attractive subject of study for citizen science because they are relatively easy to discern. In fact, the name for the field in which we work, phenology, literally means “the science of appearance.” The word was coined just a few years after Thoreau made his phenological observations, from the Greek phaino (to show or appear) and logos (to study). Phenology measures

During his time at Waiden Pond, near Concord, Massachusetts, Henry David Thoreau (1817-1862) (facing page, upper left) recorded observations on the timing of events in plants’ and animals’ life cycles. Shown at far right is the spyglass with which he observed birds. He also collected plant specimens, such as twigrush (Cladium mariscoides, near right) and switchgrass (Panicum virgatum), sometimes noting common names and the locations and life stages in which the plants were found.

life cycle events, or phenophases, in all living things. Plant phenophases are associated with leafing, flowering, and fruiting: first leaf, first flower and last flower, among others (see the illustration on the facing page). Bud burst, the phenophase from which Project BudBurst takes its name, refers to the first opening of leaf buds in spring.

The timing of phenological events can be very sensitive to environmental conditions. In a particularly warm spring, bud burst and first flower may occur weeks earlier than usual, whereas in a very cool spring they may be delayed. Such timing tends to vary from year to year in accordance with patterns of weather, climate, and resource availability. Phenological studies are thus a simple and cost-effective way to measure changes in the climate over the long term. They can help in efforts aimed at improving quality of life and economic health-for instance, they are used in maximizing crop production, predicting pollen levels in order to improve seasonal allergy warnings, and anticipating optimal wildflower viewing and fall color conditions.

More crucially, these studies help ecologists track the effects of climate change on organisms to make predictions about how the species will respond in the future. Of particular interest is the possibility of phenological mismatches. Plants and other organisms have coevolved and are part of intricate relations, or mutualisms, with insects, birds, and other animals that serve as pollinators and seed disperses. Some plants and their mutualists-organisms with which they interact in order to improve their ability to survive and reproduce-rely on different cues for first emergence and other phenophases. For instance, one species may rely on temperature, the other on day length. In such cases, it is possible that climatic change will disrupt these mutualisms, which may in turn cause some species’ populations to decline.

Human interest in phenology is much older than anthropogenic climate change, however, and certainly older than the citizen science movement. The roots of citizen science can be traced to 1900, perhaps, when the Audubon Society began its Christmas bird counts, or to the late 1800s, when the fledgling National Weather Service culled information from amateur meteorologists (see the timeline below). But the first known phenological records were made by the Chinese in 974 bce. And beginning around the 9th century ce, observations of the timing of peak cherry blossoms in Japan have been recorded.

In Europe, such record-keeping began more recently: In the mid-1700s, the Swedish botanist Carolus Linnaeus systematically noted flowering times for 18 locations in Sweden over many years. He shares the honor of being considered a founder of modern plant phenology with 18th-century British landowner Robert Marsham. Beginning in 1736, Marsham recorded the date of first occurrence of events such as plant flowering and insect emergence on his estate. For generations, Marsham’s family maintained their records, which ended with the death of Mary Marsham in 1958.

Ecologists have worked recently to extend historical records into the present. Several studies have compared Thoreau’s notes on plant species to observations of species found at Waiden Pond today. Abe Miller-Rushing and Richard Primack of Boston University collected observations on hundreds of plant species that Thoreau had carefully tracked. Comparisons between the historical and modern data revealed profound changes in plant phenology. The timing of spring phenophases at Waiden has advanced by an average of one week for many species. These data provide support for a novel hypothesis: Species that can quickly change such timing in response to changes in climate tend to be more common today than those whose timing has not changed significantly since Thoreau’s day.

Contemporary phenologists have the benefit of new technologies. The use of data from satellites and airplanes to track phenology, known as remote sensing phenology, can detect broad-scale events, such as the onset of spring- called green-up-and of autumn, or brown-down, when the leaves fall and the grasses turn brown. Many phenologists also use automatic cameras to assess changes in areas that are difficult to access regularly.

Observations made at the human scale remain an essential component of phenological monitoring, however. Such records are a valuable asset in the environmental sciences. They provide something that human-made instruments do not: integrative measures of organisms’ physical, chemical, and biological environments. In addition, gathering these data offers people both the opportunity for hands-on learning and an excuse to spend time outdoors.

A Nation of Flower Watchers

In the early 2000s, ecologists and other scientists felt the need for a comprehensive plant-monitoring project to understand the effects of climate change- one that could efficiently collect numerous observations from a wide geographic range. At the same time, popular interest in plant phenology was growing, inspired in part by people’s increasing awareness of climate change and its potential effects. The time was right for Project BudBurst. In October 2006, the National Science Foundation funded a meeting at the University of Wisconsin Milwaukee to address the need for a national phenology network. A small working group, of which we were a part, created a draft plan for a project that would enlist the help of the general public in monitoring plants across the United States. Our goals were twofold: first, to aid science education and, second, to advance scientific knowledge. The project would be simple and open to all, with a central database, straightforward protocols, and all necessary materials and resources available online. Shortly after the meeting, we began to develop a pilot version, supported by modest funding from the U.S. Bureau of Land Management, through the Chicago Botanic Garden. Carol Brewer, then associate dean at the University of Montana (now emeritus), worked with us to help make the project ready for its spring 2007 launch. Response to the program greatly exceeded our expectations. It attracted a diverse group of participants from all across the country, as well as widespread attention from blogs, newspapers, and television and radio programs. We were already making progress toward our goal of improving public understanding and awareness of phenology.

Inclusivity and ease of participation likely helped in our early success. To participate, volunteers choose a plant, observe it, and report a variety of data about it, noting the timing of phenophases such as first flower, first fruit, and all leaves withered. We ask for the plant’s common name and scientific name as well as details about the habitat in which it is growing: amount of shade, amount of irrigation, and distance to human development. The unique location is tracked via a participant-chosen site name, latitude and longitude, and city and state. Volunteers can submit data via regular reports, in which they observe a plant throughout the seasons and note the timing of phenophases, and single reports, in which they send a one-time observation of a plant’s status on a given day.

To help participants identify and choose plants to monitor, we created a master plant list that highlights more than 250 species, selected for their wide geographic distribution, ease of identification, and scientific interest. Our website offers a resource page for each species, which includes an identification guide, data sheets for observations, a live map, and a data viewer. There is no cost to participate, and the reporting interface is simple enough that, with assistance, even elementary school-age children can help out. We also made a 10-most-wanted plant list, which encourages observations that will help build a more comprehensive data set for these target species. Early on, however, we received many observations of species not on our lists, so we decided to support reports on any plant in order to keep participation as inclusive as possible.

With feedback from participants, along with funding from a variety of sources, we have made the BudBurst website and supporting resources much more comprehensive over the past six years. We have created campaigns—Cherry Blossom Blitz, Summer Solstice Snapshot, and Fall into Phenology—that promote the project and highlight seasonal phenophases. A new project, the Floral Report Card, has created gardens of identical species in 10 sites around the country, with the goal of monitoring plants across a wide geographic and climatic range. This past summer our team developed a mobilefriendly version of the project website to make things even easier; now participants can report field observations from their smartphones or tablets.

We have also developed online courses to support educators who want to use Project BudBurst in their classrooms. In less than a year, more than 400 educators have enrolled in these self-paced courses. In 2014, we plan to launch a curriculum series for grades 5-12, Climate Change in My Backyard, that uses Project BudBurst participation and NASA climate data to teach about climate change and its effects on the environment.

It can be a challenge to measure the achievement of broad educational goals, but we find encouragement in several sources. A Google search for Project BudBurst in mid-2013 returned close to 23,000 entries. The project has been covered in more than 300 articles and interviews. And since the beginning, that high level of interest has made it easy to recruit volunteers: To date we have more than 15,500 registered participants making regular contributions.

Project BudBurst has thrived on a partnership model in our educational and scientific efforts. We work with agencies such as the U.S. Fish and Wildlife Refuge System, the U.S. National Park System, science centers, and natural history museums to identify plants of unique local interest. Data gathered by the project are freely available for scientists and citizens to use. The result is a robust nationwide citizen science program-one in which anyone can step into Thoreau’s or Marsham’s shoes.

Educational Science

Project BudBurst’s goals of aiding science education and advancing scientific knowledge are mutually supportive: Volunteers gain satisfaction from the knowledge that their observations are contributing to the project, and scientific publications that use project data can feed interest (and help in fundraising). BudBurst data have been used in several scientific papers, including one by Uran Chung and her colleagues, who modeled the timing of cherry-blossom peak bloom for trees in the Washington, DC, region under various scenarios of climate change. The ecologists found that under the Intergovernmental Panel on Climate Change’s A2, or high, emissions scenario, the trees may bloom 29 days earlier by 2080.

The results have also been used in several regional historical studies. For example, in 2012 we compared Project BudBurst data collected in the Chicago region with observations published by botanists Floyd Swink and Gerry Wilhelm in their 1994 book Plants of the Chicago Region. Swink and Wilhelm collected phenology observations from the mid-1950s to the early 1990s. Of the 15 local species for which we had good contemporary and historical data, 13 had an earlier first flower in one or more years between 2007 and 2012 than was ever observed by Swink and Wilhelm. Data on these timescales thus begin to reveal the effects of climate change on plants.

In another study, Elizabeth Wolkovich, of the National Center for Ecological Analysis and Synthesis, and Eisa Clelând, of the University of California-San Diego, asked whether the success of invasive introduced species was due, at least in part, to a “seasonal priority” effect. If a normative plant species leafs out earlier than native species, it may be able to exploit resources first and shade out its competitors. The authors’ analysis of Project BudBurst data revealed that introduced species leafed out on average 3.74 days earlier than their native neighbors, suggesting that climate change is accelerating the threat of these species.

In 2012 we also plotted trends in first flowering date from Project BudBurst data (see the illustration on page 383 for one example). We found a trend toward earlier flowering in all of the graphs we made-likely due in part to the extraordinarily early spring we saw in much of the country in 2012.

With data sets collected over relatively short timescales, an anomalous year can have a large effect on the observed trend. As we gather data over longer and longer periods, such effects diminish, and the ecological effects of climate change grow more distinct. Over time, BudBurst participants will be able to see more and more clearly that their individual observations have global importance.

Lessons for Scientists

Successful citizen science projects work to create connections between the intimate stories that unique observations tell and the bigger story of the project data set. These programs, such as Project FeederWatch, Monarch Watch, and CoCoRaHS (a network that measures precipitation), take advantage of activities that people like to do already, such as birding, gardening, or taking nature walks, and allow participants to share and archive their observations as well as contribute to scientific studies. Project BudBurst is fortunate to be part of a worldwide trend: Work by several researchers shows that the design and implementation of citizen science projects is improving as practitioners learn from one another.

For our part, we have learned several key lessons. Properly attributing citizen scientists’ contributions to publications and other projects resulting from the data is essential. Our website lists the names of everyone who has contributed data. We upload each year’s data set, with a short summary, to the site each year, and the data are freely available for anyone to use. We request that people who rely on this data cite their use of it-and recognize the volunteers with a link to our attribution page.

New technologies are facilitating data collection, assisting with quality control, and improving data management. For instance, many citizen science projects are adding game components that broaden the appeal of the project and help motivate and retain volunteers. BudBurst has joined this trend, collaborating with BioTracker (a mobile-phone application for biological data collection) to create a “floracaching” game. The idea is similar to geocaching (in which people use GPS to find objects or clues hidden at precise locations): Players focus on locating plants and making phenological observations.

We are always working to improve quality control. An optional step in the registration process asks people to note their level of expertise in making observations of plants. Our mobile applicaOak tion automatically geotags the location at which each observation is made, eliminating the possibility of error in determining the latitude and longitude of the site. We hope to add a feature that allows participants to upload photos of the plants they observe so that we can confirm reports of phenophases at unusual times. We are also taking steps to address the issue of spotty geographic coverage-the majority of reports we receive are from cities. Many of our more interesting findings are coming from places where we have continuous observations of the same species over multiple years. To spur local involvement, we plan to add data visualization tools that will allow users to explore their observations in the context of climate data sets.

Perhaps our most important challenge is the need to sustain both participation by volunteers and the project as a whole over time. Initially, the Project BudBurst website was housed at the University Corporation for Atmospheric Research (UCAR) in Boulder, Colorado. In fall 2010, it moved to its permanent home at the National Ecological Observatory Network (NEON), where it has extensive support for website maintenance. We are thus well positioned to maintain and build on current programs for decades to come.

Our future sustainability will rest on a foundation of hard work and thought from our colleagues and from participants. We are thankful for the participation and support of our colleagues Kirsten Meymaris, Dennis Ward, Sarah Newman, Leah Wasser, Paul Alaback, Pati Vitt, and Jennifer Schwarz. And we thank the thousands of volunteers who have submitted their observations to Project BudBurst, without whom it would be impossible to improve our knowledge of plant phenology on such a scale.

On April 2, 1856, Thoreau wrote in his journal of venturing out to find “evidences of spring.” On discovering a particularly early skunk cabbage (Symplocarpus foetidus) in bloom, he wrote, “Perhaps it is always earlier than I have thought, if you seek it in a favorable place.” Later in the same entry, he wrote, “It will take you half a lifetime to find out where to look for the earliest flower.” Little did Thoreau know that his solitary searches would spawn scientific studies aided by the careful observations of a group of school children in Seattle, a retiree hiking in the Rocky Mountains, an educator monitoring plants on a Colorado ranch, a botanist in Chicago, and more than 15,000 other citizens across the country. Citizen scientists today can choose to spend half a lifetime in search of earliest flowers, or they can spend an afternoon-but in any case, they will have the satisfaction of knowing that they and their observations are in good company.