Jump to ContentJump to Main Navigation
Citizen SciencePublic Participation in Environmental Research$

Janis L. Dickinson and Rick Bonney

Print publication date: 2012

Print ISBN-13: 9780801449116

Published to Cornell Scholarship Online: August 2016

DOI: 10.7591/cornell/9780801449116.001.0001

Show Summary Details
Page of

PRINTED FROM Cornell University Press SCHOLARSHIP ONLINE (www.cornell.universitypressscholarship.com). (c) Copyright University of Cornell University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in Cornell for personal use (for details see http://cornell.universitypressscholarship.com/page/privacy-policy/privacy-policy-and-legal-notice). Subscriber: null; date: 20 June 2018

Developing a Conservation Research Program with Citizen Science

Developing a Conservation Research Program with Citizen Science

Chapter:
(p.139) 9 Developing a Conservation Research Program with Citizen Science
Source:
Citizen Science
Author(s):

Ralph S. Hames

James D. Lowe

Kenneth V. Rosenberg

Publisher:
Cornell University Press
DOI:10.7591/cornell/9780801449116.003.0010

Abstract and Keywords

This chapter explains how citizen science can be used to develop a conservation research program. It describes a specific case in which “super citizen scientists” used manipulative sampling to gather data implicating acid rain and mercury in forest bird declines, highlighting the advantages of partnerships with governmental and nongovernmental organizations. Focusing on the Birds in Forested Landscapes (BFL) project that was originally developed at the Cornell Lab of Ornithology, the chapter demonstrates how citizen data can help address the effects of pollution on birds over wide regions. It also considers the BFL's collaboration with The Nature Conservancy as well as new research using data from another citizen science project, the Breeding Bird Survey, to develop a program for investigating significant conservation issues for birds and for translating science for policy and management.

Keywords:   citizen science, conservation research program, super citizen scientists, Birds in Forested Landscapes, Cornell Lab of Ornithology, pollution, birds, The Nature Conservancy, Breeding Bird Survey, conservation

Although ecologists are increasingly aware of the role played by wide spread, human-caused change in altering the population dynamics of wild animals, large-scale processes (those covering large geographic extents) remain understudied, mainly because of logistics and costs (Baillie et al. 2000; Caldow and Racey 2000). Understanding the effects of such broad-scale environmental changes requires extensive data collection by volunteers to delineate patterns; elucidating underlying processes requires even more intensive study along with experimentation. In this chapter, we use the Birds in Forested Landscapes (BFL) project as an example of a research program in which a subset of volunteers was recruited to collect additional data and conduct experiments to test specific hypotheses arising from the initial observations and patterns in broad-scale citizen science data.

We have long known that the loss and fragmentation of habitat can have negative impacts on populations of some sensitive species (Andrén 1994; Robinson and Robinson 2001; Saunders et al. 1991). In addition, acid rain (Graveland 1998; Hames et al. 2006; Hames et al. 2002a), resultant depletion of soil pools of calcium (Driscoll et al. 2001; Graveland et al. 1994), and the atmospheric deposition of mercury (Rimmer et al. 2005) have been identified, along with synergies between these multiple stressors (Hames et al. 2006), as potential drivers of population declines of birds in eastern North America (Hames et al. 2002a).

BFL was originally developed at the Cornell Lab of Ornithology to address the question of how forest fragmentation affects breeding birds across North America (Hames et al. 2002b), and BFL data were later used to test specific hypotheses about the ways in which pollution affects forest (p.140) birds. Here we describe how we created a conservation research program based on an intensive, fine-scale study that involved “super citizen scientists” using the BFL protocol (and extensions of it) in a small region of New York, the Catskills. These volunteers worked closely with the Cornell Lab of Ornithology (the Cornell Lab) and The Nature Conservancy (TNC) to help explore the linkage between pollutants, particularly acid rain and mercury, and declines in some species of forest birds.

Here we describe the progression of this research, our partnership with TNC, and new research using data from another long-running citizen science project, the Breeding Bird Survey (BBS), with colleagues at the Environmental Protection Agency (EPA). These current research efforts seek to understand the effects of pollution on birds over wide regions using data on the abundance of territorial Wood Thrushes (Hylochicla mustelina) across the eastern United States and Canada. We are using data gathered by BBS volunteers, whose efforts are administered jointly by the U.S. Geological Survey and the Canadian Wildlife Service (Sauer et al. 2004) to address hypotheses as shown in the path diagram (Figure 9.1) and to identify variation in the outcomes associated with multiple stressors, based on exploratory data analysis (see Chapter 8). Throughout, we highlight the dedication of “super citizen scientists” who allow us to address important environmental

Developing a Conservation Research Program with Citizen Science

Figure 9.1. This path diagram is an explicit hypothesis on the food web pathways that we believe are important in mercury contamination of upland forest birds. Thicker lines represent what we believe are the stronger influences, for which we also have good evidence. Solid lines represent mercury (methyl and other soluble forms); the dashed lines represent acid deposition; the lines with dots and dashes are calcium.

(p.141) questions at a hierarchy of geographic extents that is not possible with small, intensive studies carried out only by professional scientists.

Using BFL Data to Address Impacts of Large-Scale Environmental Change

Design and Implementation

BFL (Hames 2001; Hames et al. 2002b) and its precursor, Project Tanager (Rosenberg, Lowe, et al. 1999), are large, extensive, volunteer-based projects with protocols that were designed to address specific questions, such as, what are the effects of habitat fragmentation on forest-breeding songbirds and raptors? These projects incorporate more complicated and rigorous protocols than some projects mentioned in this book to ensure that detailed data on important influences on the abundance and distribution of the focal species are collected using the same methods everywhere. Standardized protocols enable comparisons between points sampled by different participants over large distances. Detailed, rigorous, and simultaneous data collection at several hierarchical geographic extents also allows us to focus on one or more factors influencing the response to fragmentation or pollution, while statistically adjusting for other influences (Hames et al. 2002a). Further, our survey protocol, including passive listening period, playback of conspecific vocalizations, and playback of small birds mobbing a predatory owl (www.birds.cornell.edu/bfl), results in a predicted detection rate for target species of 0.93, measured with the software PRESENCE (Hines, 2008). The high detection rate ensures that our BFL volunteers are unlikely to miss a breeding bird that is actually present, an important consideration in analyzing environmental effects on bird populations.

Compared with Project Tanager, which used a relatively simple protocol to study three species of tanager, BFL’s more complex protocol produced improved data collection on more than forty species, with better estimates of vegetation structure, landscape configuration, and site location. Each BFL volunteer thus collected large amounts of detailed data at each site, including patch-specific data and landscape-level data. Each study site was evaluated for breeding success by the focal species using behavioral cues from breeding bird atlas codes (Smith 1990). The volunteers who persevered with this new, somewhat formidable, protocol were a self-selected group who weren’t intimidated by a demanding research plan.

To date, BFL participants have returned data from approximately 3800 study sites across North America, yielding over 30,000 (species by year by location) records. Some sites were sampled in only 1 year, but many sites were resampled for up to 12 years. The improved location data available now for each study site allow for “post-processing” by matching the location to new environmental data sets that are becoming increasingly available (p.142) on the Internet. We therefore have the ability to address new questions that were not contemplated when the data were first collected.

BFL Results on Fragmentation and Acid Rain

To answer specific questions about environmental change, we linked BFL data with several environmental monitoring data sets. Specifically, BFL data, originally collected for fragmentation studies (see Hames et al. 2002b), were used retrospectively to address the combined effects of pollution and habitat fragmentation, especially combined atmospheric deposition of acidifying ions and mercury (Hames et al. 2006; Hames et al. 2002a), using data from the National Atmospheric Deposition Project (NADP/NTN/MDN) (Lamb and Bowersox 2000). Additionally, we derived soil pH and other properties using the USDA Natural Resources Conservation Service’s STATSGO and SSURGO soil databases (USDA/NRCS 1994). Finally, the BBS provided information on the regional, or background, abundance of the focal species (Sauer et al. 2004). Our ability to combine data from a number of sources greatly increased our understanding of broad-scale, human-caused change because it allowed us to both cross-validate and leverage BFL data with other data, including data that are hard for volunteers to collect, such as the environmental monitoring data sets mentioned above (Link and Sauer 2002; Sauer and Link 2002; Takashi 2004; Thogmartin et al. 2004).

For example, all of the above data sources were used in analyses to test the hypothesis that population declines in a forest songbird, the Wood Thrush, were related to both forest fragmentation and acid deposition (Hames et al. 2002a). We demonstrated that the probability of Wood Thrushes attempting to breed at a site declined with increasing acid deposition, after controlling for a number of other factors such as elevation, distance to patch edge, patch size, types of vegetation, and fragmentation measures. Further, we showed synergies between acid deposition and habitat fragmentation in the form of an interaction that made the negative effects of acid rain approximately 3 times greater at sites with even moderate forest fragmentation (Hames et al. 2002a). Further analyses showed negative population trends associated with low pH acidic soils as well as acid and mercury deposition (Table 9.1).

Investigating the Relationship between Acid Rain, Calcium, and Invertebrates

We then wished to further explore the mechanisms, or processes, that led to the observed patterns of declines in the Wood Thrush, using the efforts of both citizen and professional scientists. Much was already known, based (p.143)

Table 9.1. Performance of pollution variable cutoffs or thresholds at predicting declining population trends for the Wood Thrush within the focal 10 × 10 km GIS cell

Landscape matrix

Cutoff

e/N

Sensitivity

Specifi city

False positive

False negative

% correct

Odds ratio

Kappa

Soil

≤5.6 pH

0.56

0.75

0.65

0.35

0.25

0.72

5.56

0.38

Acid rain

≤4.7 pH

0.56

0.61

0.71

0.29

0.39

0.64

3.85

0.28

Hg dep.

≥10 ng/m2

0.56

0.66

0.40

0.60

0.34

0.58

1.33

0.06

(p.144) on the work of other scientists in Europe and North America. Acid rain, for example, and leaching of soil calcium, with concomitant declines in calcium-rich invertebrates, had been linked to population declines in several species of birds, particularly in Europe (Drent and Woldendorp 1989; Graveland 1996, 1998). We knew from the literature (Bures and Weidinger 2000; Graveland 1996; Graveland and Drent 1997; Graveland et al. 1994) and our own intensive studies (Hames et al. 2006) that the number of calcium-rich invertebrates decreased as soils became more acidic and contained less calcium.

The nonbreeding diet of most insectivorous and granivorous bird species contains too little calcium to lay a clutch of eggs, thus females must consume supplemental calcium before laying (Graveland 1995; Johnson and Barclay 1996; Poulin and Brigham 2001; Tilgar et al. 2002). Parents must also provide calcium supplements to growing nestlings for development of robust skeletal systems (Bures and Weidinger 2000, 2003; Graveland 1995; Poulin and Brigham 2001). We wished to discover whether acid rain was negatively impacting Wood Thrush populations, which have been declining precipitously across their breeding range, but especially at higher elevations in the eastern part of the United States and Canada (Sauer et al. 2004), by causing decreases in the abundance of calcium-rich invertebrates.

Measuring Forest Health by Engaging Super Citizen Scientists

To investigate these complex processes, we required a simple, repeatable protocol that addressed acidification’s effects on breeding birds using simple materials available to all volunteers. We therefore developed a protocol for BFL volunteers to use in sampling calcium-rich invertebrate prey available to birds foraging in the leaf litter (Hames et al. 2006).

We tested several methods to enumerate snails and other calcium-rich invertebrates in the leaf litter and chose the “cardboard trap” method (Hames et al. 2006), a simple method based on the malacological literature (Hawkins et al. 1998; McCoy 1999; Oggier et al. 1998). We first tested the methodology for one field season at forty sites in New York. At each site, we placed squares of dampened cardboard on the leaf litter, left these “traps” overnight, and returned to count the invertebrates the next morning (Hames et al. 2006).

We also needed an index of bioavailable calcium, based on invertebrate biomass, to investigate the mechanisms through which acid precipitation may negatively impact the Wood Thrush (Hames et al. 2006). We developed and tested a categorical method to estimate the calcium available to birds as invertebrate biomass in the forest leaf litter. These data were based on trapping and sorting of invertebrates into five broad taxa (isopods, millipedes, snails, slugs, and earthworms), and four 10-mm size classes (p.145) (small, medium, large, and extra large). Our categorical index showed a close and highly significant agreement between estimated and actual biomass of calcium-rich invertebrates, ranging from r2 = 0.89 to 0.97 (Hames et al. 2010). Further, the number and biomass of calcium-rich invertebrates were significantly related to soil properties, including acidity and calcium content (Hames et al. 2006).

By recruiting BFL participants across the East to use simple invertebrate sampling techniques, which resulted in counts and estimated biomass of calcium-rich invertebrates, we were able to test two critical predictions of the hypothesis that acid precipitation’s effects on bird populations are mediated through depletion of calcium-rich invertebrates from the litter. Hames et al. (2006) ultimately demonstrated that the probability of patch occupancy by a territorial Wood Thrush declines to zero with declines in the mass of millipedes at a site; these thrushes are absent in the absence of millipedes. Such knowledge is crucial for conservation, because human-caused changes to the environment can be mitigated to some degree—but only if their effects on the biota are understood.

Collaboration with The Nature Conservancy to Measure Forest Health

The Cornell Lab has been collaborating with The Nature Conservancy (TNC) in the Catskill Mountains since 2006 on research designed to develop volunteer-based measures of forest health, to be used in their own conservation work as well as the conservation work of other organizations. Originally, TNC approached the Cornell Lab because they wanted a rigorous, citizen science–driven data-collection protocol, such as BFL’s, to use in setting up a monitoring plan for the Catskill Park. The areas of interest were large contiguous forest blocks, some roadless. TNC also sought to develop a rigorous, quantitative method for measuring forest health to use in the monitoring program. They were interested in using data derived mainly from BFL sites for their “measures of success,” a quantitative methodology that allows them to set numeric goals for conservation actions and assess whether the goals are reached (Tear et al. 2005).

Lab scientists, meanwhile, sought to address the effects of acid rain, soil-calcium depletion, and mercury deposition on breeding forest birds. The Catskills have some of the highest atmospheric deposition rates in the Northeast, including high rates of acidifying ion deposition (NADP NTN 2001) and high rates of mercury deposition (Miller et al. 2005). Mercury deposition, like acid precipitation in the rural East, arises mainly from the combustion of coal (Cohen et al. 2004). In its organic form, mercury is a potent neurotoxin, affecting adult birds and their offspring, causing behavioral, motor, and sensory deficits in aquatic bird species (Evers et al. 2008).

(p.146) Whereas mercury in the food web is well studied for aquatic birds, little work has focused on mercury in terrestrial, upland food webs (but see also Rimmer et al. 2005). Hames et al. (unpublished data) showed low-level mercury contamination (< 1.0 ppm) linked to soil acidity to be ubiquitous in sampled upland forest birds from five regions of New York. We hypothesized that acid rain’s action on soils leads to an export or leaching of calcium from the system, resulting in shortages in soil pools of calcium and decreases in the abundance of calcium-rich invertebrates. This decline, and the paucity of calcium that results, increases the uptake of trace metals in breeding birds that are likely already calcium stressed (Scheuhammer 1991, 1996). Further, some leaf-litter invertebrates are themselves at risk for mercury contamination, as they have the metabolic pathways to absorb metals, because of their high calcium demand (Beeby 1990). Additionally, some of the invertebrates in the forest leaf litter may slow their intake of food, and hence their growth, when fed contaminated leaf litter (Beeby 1990). These hypothesized effects of mercury contamination raise the specter of breeding forest birds searching more intensively to find the calcium they require to breed from invertebrates that are smaller and less abundant and that may also be contaminated with the bioavailable form of mercury, methylmercury (Figure 9.1).

Integrating Research on Mercury Deposition at BFL Study Sites in New York

To collect the data necessary to test the hypothesized effects of mercury contamination on forest birds, Lab scientists selected intensive acid rain/mercury deposition study sites arrayed along a gradient of soil pH and mercury content in the Finger Lakes region of New York. Together, TNC and the Cornell Lab also developed a monitoring project for forests in the Catskills, collecting data on soils, invertebrates, vegetation, and birds.

TNC and Lab scientists divided tasks for this new project and worked in a way that made the most of our respective institutional strengths. First, the Catskill Mountain Program of TNC recruited local volunteers, divided them into two-person teams, provided field equipment for each team, and selected the forest blocks of interest. A group of at least five BFL study sites was located in each of six large forest blocks in the Catskills, and a team of two BFL volunteers was assigned to each group of five study sites. The first site was placed in a deciduous forest block, then four subsequent sites were placed approximately 300 m apart while remaining in deciduous or mixed forest.

Because of their location, most Catskills BFL sites (which were defined as a circle with a radius of 150 m and an area of ca. 7 ha) were in 1000 ha landscapes that were more than 95% forest and showed little or no fragmentation. We were therefore able to ignore fragmentation effects and (p.147) focus our investigation on acid deposition, mercury contamination, and soil properties, as well as the abundance and biomass of calcium-rich invertebrates. Lab scientists provided expertise in training volunteers on the goals and hoped-for outcomes of the project, as well as background information on the effects of forest fragmentation. We took all twenty-five Catskill Project BFL volunteers into the field to demonstrate the protocol and met with volunteers and TNC staff at the beginning, midpoint, and end of each field season to answer questions, give guidance, and show the results of analysis of data derived wholly, or in part, from their efforts.

These “super-volunteers” collected data using a “superset” of the already difficult BFL protocol; in other words, they used all of the steps in the existing protocol and gathered other data as well. In addition to hiking several kilometers, with elevation gains/losses in hundreds of meters, these super-volunteers carried out the protocol for calcium-rich invertebrates with double the number of traps. This meant carrying 4 liters of water to each site. We also worked with these volunteers to develop a protocol to assay coarse woody debris and standing dead trees, both indicators of forest health at their sites.

Through frequent interaction with Lab scientists, volunteers could see the tangible results of their work. Subjectively, we feel, based on our experience with volunteers in conservation projects, that meeting as a group with other volunteers, Lab, and TNC scientists helped keep volunteer interest and participation high. Retention was very good, with 68% returning after the first season and 44% of our initial participants still participating in the third year of data collection at our Catskill sites.

Early results of the detailed studies at sites in New York can be seen in Figure 9.2, which shows the proportion of times a Wood Thrush was detected (out of all visits) against abundance of calcium-rich invertebrates, at all five sites in each area. These results corroborate with greater precision our earlier work showing an association between acid precipitation and calcium-rich invertebrates in areas with Wood Thrush declines.

While large geographic-scale citizen science is usually restricted to monitoring, we have shown that it is also a powerful vehicle for developing question-driven research projects that dissect the critical mechanisms underlying bird declines, especially when combined with small-scale intensive studies conducted by citizen and professional scientists. We believe that BFL is an example of what can be done using an approach that is focused on collection of data that address very specific questions using precise, question-driven protocols. Because the methods still involve citizen scientists scattered broadly over a landscape, as well as intensive studies addressing mechanisms with fieldwork by Lab scientists, the data lend themselves (p.148)

Developing a Conservation Research Program with Citizen Science

Figure 9.2. The total number of calcium-rich invertebrates and proportion of sightings of Wood Thrushes by area (five sites) in the Catskills shows an increase in thrush sightings with increases in invertebrates. Fit with a linear model, the upward trend mirrored in invertebrates and thrush sightings is marginally significant (F1,9 = 5.16, p = 0.049, R2 = 0.36).

to the study of hierarchical processes and mechanisms underlying declines at multiple scales.

We have presented a model of a collaborative research program that is iterative, building on past results, and involves the Cornell Lab of Ornithology’s Conservation Science program in studies designed to answer scientific questions at a variety of scales. Only citizen science could accomplish this research so inexpensively and expeditiously, providing scientific evidence required for environmental decision making in the face of growing human alteration of the biosphere. While large-scale monitoring is invaluable, especially in combination with newly available environmental data obtained through remote sensing, making solid biological sense of data we gathered required a network of super-volunteers willing to head out into the woods and do work equivalent to professional field biologists. We discovered a competent, dedicated set of people whose efforts have allowed us to identify mechanisms in addition to patterns. Their importance to this and other such research would be hard to overestimate.

The Conservation Science program at the Cornell Lab comprises efforts to study significant conservation issues for birds and translate science for policy and management. We have been active in a number of partnerships (p.149) with governmental and nongovernmental organizations. Our efforts in these arenas have relied heavily on use of citizen science data to conduct research and to create management guidelines, species conservation plans, and high-profile reports, such as The U.S. State of the Birds Report (North American Bird Conservation Initiative, U.S. Committee 2009, 2010), and to inform major conservation efforts, such as Partners in Flight. This application of research to conservation policy and management is strengthened by approaches, such as we describe here, that couple large-scale citizen science with fine-scale research aimed at pinpointing causal factors in bird declines.

Acknowledgments

The authors gratefully acknowledge support from the Leon Levy Foundation for our research on forest health and bird populations. The Birds in Forested Landscapes project was developed and supported by the National Fish and Wildlife Foundation. We also are grateful for the support of the U.S. Environmental Protection Agency, a McIntyre-Stennis grant, a New York State Wildlife Grant to the Wildlife Conservation Society et al., New York State Museum’s Biodiversity Research Institute, and the U.S.D.A. Forest Service. We thank our colleagues Diane E. Nacci and Carol Trocki of the EPA ORD AED/NHEERL and Daniel Fink of the Cornell Lab of Ornithology for permission to use the data in Table 9.1. We are humbled by and grateful for the dedication and hard work of thousands of BFL and BBS volunteers.