Ecology, 94(11), 2013, pp. 2645–2652� 2013 by the Ecological Society of America

Using citizen scientists to measure an ecosystem service nationwide

RIIKKA KAARTINEN,1,2 BESS HARDWICK,1 AND TOMAS ROSLIN1,3

1Spatial Food Web Ecology Group, Department of Agricultural Sciences, P.O. Box 27 (Latokartanonkaari 5),00014 University of Helsinki, Finland

2Department of Ecology, Swedish University of Agricultural Sciences, P.O. Box 7044, 75007 Uppsala, Sweden

Abstract. The decomposition of dung constitutes an ecosystem service of massiveproportions. Previous studies addressing how it depends on individual invertebrate taxa havebeen focused on small spatial scales, neglecting the impact of large-scale factors like climate.Here, we use the concept of ‘‘citizen science’’ to quantify taxon-specific contributions to dungdecomposition at the level of a nation. Young people across Finland manipulated thedecomposer communities of cow pats, then measured changes in pat mass over the grazingseason. In southern Finland most (90%) of the cattle dung hitting pastures decomposed in justtwo months, whereas 1100 km to the north the corresponding fraction was smaller (74%). Ofthe total invertebrate-caused decomposition (13% of dung pat mass, independent of latitude),large tunneling dor beetles in the genus Geotrupes account for 61%, hence removing dungtwice as fast as do smaller dung-dwelling beetles and earthworms. Overall, this paperillustrates how ecologists may direct citizen scientists to implement massive ecologicalexperiments. Compared to an approach based purely on professional scientists, we savedthree-quarters of the costs. Ultimately, citizen science may offer a key tool for testing currentecological theories at relevant spatial scales—and for disseminating these theories in theprocess.

Key words: Aphodius spp.; citizen science; dung beetles; dung decomposition; earthworms; ecosystemfunctioning; ecosystem services; experimental study; Finland; Geotrupes spp.

INTRODUCTION

The decomposition of dung is an enormous ecosystem

service. In Finland, roughly one million heads of cattle

will produce an annual total of ;4 billion kg of dung

(Roslin and Heliovaara 2007). In the United States, the

corresponding figure has been estimated at 900 billion

kg, and the value of dung decomposition at US$380

million per annum (Losey and Vaughan 2006). Much of

this service is provided by the insects feeding and

breeding on dung, thereby facilitating dung decomposi-

tion, nutrient cycling, seed dispersal, parasite control,

and soil aeration (Nichols et al. 2008).

Given the overall value of dung decomposition, the

specific contributions of different decomposer taxa have

been the subject of intensive investigation (Holter and

Hendriksen 1988, Gittings et al. 1994, Larsen et al. 2005,

Slade et al. 2007, 2011, Horgan 2008, Rosenlew and

Roslin 2008, O’Hea et al. 2010). Dung beetles in the

superfamily Scarabaeoidea have been identified as a

global key taxon (Nichols et al. 2008), with added

contributions by the earthworms (Oligochaeta; Holter

and Hendriksen 1988, Gittings et al. 1994). Within the

dung beetles proper, groups with different nesting

behaviors have been proposed to play different roles.

For Northern European invertebrates, a recent study

identified large tunneling dor beetles in the genus

Geotrupes as paramount players in dung decomposition,

whereas the researchers hypothesized that earthworms

and dung beetles in the genus Aphodius were of minor

importance (Rosenlew and Roslin 2008).

Importantly, all previous experiments on dung de-

composition have been conducted at a small spatial scale;

the study by Rosenlew and Roslin (2008) consisted of 40

microcosm containers 66 cm in diameter, spread over an

area of less than 1 ha. Any generalization regarding the

relative contribution of individual invertebrate groups to

the level of the larger ecosystem service has then relied on

a forced leap of faith: since implementing the same

experimental manipulations across a large set of sites

would exceed available resources, what occurs in one

place must of necessity be assumed to occur in others.

This precludes the quantification of other impacts as

realized over larger areas—in particular the impact of

climate (Wall et al. 2008).

Quantifying ecosystem services over large spatial

scales will no doubt call for inventive approaches. Here,

Manuscript received 7 July 2012; revised 5 April 2013;accepted 11 April 2013. Corresponding Editor: W. E. Snyder.

3 Corresponding author. E-mail: tomas.roslin@helsinki.fi

November 2013 2645NOTESNovember 2013 2645NOTES

the topical concept of ‘‘citizen science’’ may offer a

solution. Over the last few decades, this idea of involving

the general public in collecting data sets for scientific

purposes (for a definition of citizen science, see Cooper

et al. [2007]) has increasingly been used to gather

observational data sets at large spatial scales (Winkler et

al. 2002, Devictor et al. 2010, Beaubien and Hamann

2011). Yet the idea of citizen science offers scope for

much more than observational data collection. During

the last decade, it has been repeatedly suggested (e.g.,

Dickinson et al. 2010) that volunteer scientists might be

used to implement manipulative experiments across

large areas and numerous sites (for successful examples,

see Jones et al. 1998, Hames et al. 2002). Such a solution

may allow ecologists to efficiently target relevant

questions at relevant scales—since unlike professional

ecologists, citizen scientists are available everywhere, at

a fraction of the cost (Schmeller et al. 2009, Devictor et

al. 2010, Dickinson et al. 2010).

In the current study we use the principle of citizen

science to quantify an ecosystem service at the level of a

full nation. By using local young people living at or near

cattle farms, we are able to replicate an experiment at an

unparalleled level, targeting three objectives. First, we

estimate the contribution of different invertebrate guilds

by experimentally excluding them from selected pats of

cow dung replicated across Finland. Second, we

compare the effect of these biotic agents to effects of

abiotic variation in climate across a north–south

gradient exceeding 1000 km. Third, we test the a priori

prediction of Rosenlew and Roslin (2008) that dor

beetles in the genus Geotrupes are functionally dominant

decomposers across their Finnish range.

MATERIAL AND METHODS

To quantify the impact of different dung decompos-

ers, we targeted 82 cattle farms across Finland (Fig. 1).

This large-scale experiment was conducted in collabo-

ration with citizen scientists from the 4H Federation of

Finland, a youth organization with 79 000 members

distributed across the Finnish countryside. Drawing on

the pre-existing infrastructure of the organization

allowed us to efficiently recruit, hire, and instruct young

people, at a scale unachievable through sampling by

professional scientists (see Appendix A for a detailed

description of methods and costs).

Experimental design

Each citizen scientist was instructed to collect 20 L of

fresh dung from a cattle barn at the onset of the

experiment (30 May–12 July 2011, with the specific start

date progressively delayed with latitude to match local

phenology). The dung was divided into 15 pats of 1.2 L

each, using a standard measure supplied to the

volunteers. This volume corresponds to an initial

average fresh mass of 1165 6 8 g (mean 6 SE) (n ¼ 20

pats from a compound collection of dung). These

experimental pats were placed on pastures where cows

had been grazing for at least three days, or right next to

such pastures.

The pats were randomly assigned to six treatments

(Fig. 1b): Five of the pats (‘‘bait pats’’) were used to

sample dung beetles occurring in the area (see Appendix

B for further details). The remaining 10 pats were

assigned to five further treatments with two replicates of

each: (1) pats exposed to all decomposers present at the

site, (2) pats from which the burying activity of large

tunneling dung beetles was excluded by a coarse metal

mesh plate (mesh size 13 1 cm) placed under the pat, (3)

pats from which all activities by large tunneling dung

beetles were excluded by covering the pat with a coarse

metal mesh cage (mesh size 1 3 1 cm), (4) pats from

which large tunneling dung beetles and earthworms were

excluded by a coarse metal mesh cage (mesh size 131

cm), combined with a dense cloth underneath the pat,

and (5) pats from which all decomposers were excluded

by a fine metal mesh cage (mesh size 1 3 1 mm).

To record dung decomposition, dung pats along with

their mesh cages were weighed every 10 days for a time

period of two months (six weighings in total). Changes

in dung mass will then reflect both desiccation and

actual dung removal and/or respiratory loss of mass by

pat-dwelling species (see, e.g., Wall and Strong 1987,

Slade et al. 2007, Rosenlew and Roslin 2008). As our

primary interest was in the relative contributions of

different invertebrates to decomposition rates, we used

treatment 5 (see above) as a point of comparison: in this

treatment, all changes in mass will reflect desiccation

and microbial decomposition, with no contribution by

macroscopic invertebrates. The participating citizen

scientists submitted mass data through a web service

after each weighing (see Hardwick et al. 2011). As

precipitation may add variation to pat mass, partici-

pants were asked to report the incidence of rain during

the course of the experiment.

Overall, we note that our general approach of

estimating functional contributions of decomposer taxa

as differences in mass among dung pats to which they

had vs. lacked access relies on the explicit assumption

that relevant taxa were present across sites. For pat-

dwelling dung beetles in the genus Aphodius, this

assumption was verified through direct sampling of five

bait pats (see above and Appendix B). For earthworms

and dor beetles, the sampling techniques employed will

suffice to establish neither their presence nor their

abundance. Nonetheless, earthworms are nearly ubiqui-

tous in the pasture soils of Finland (Nieminen et al.

2011), and a high incidence of dor beetles has been well

documented by a large number of Finnish entomologists

(as another example of citizen science, see Roslin and

Heliovaara [2007]). We note that slight variation in the

exact fauna present on individual sampling sites may

NOTES2646 Ecology, Vol. 94, No. 11

still add some variation to the data, but that finding

clear differences among treatments despite this variation

will offer strong evidence for true taxon-specific effects.

That the exclusion devices did not, on their own,

influence the physical desiccation rate of dung pats

was verified by comparing mass loss across treatments in

the absence of all invertebrates. This separate experi-

ment is described in Appendix C.

Statistical methods

To quantify the decomposition rate of dung in

different experimental treatments, we modeled dung

pat mass as a function of experimental treatment, time,

latitude (a continuous variable included to estimate the

effect of the northward cooling climate on decomposi-

tion rate), the occurrence of rain (a binary variable

indicating whether it had rained during or a day before

weighing), as well as biologically motivated two-way

interactions between experimental treatment and rain,

between experimental treatment and time, and between

time and rain. (The interaction treatment3 latitude was

removed from the final model as being far from

significant: F4, 768 ¼ 0.67, P ¼ 0.61.) The repeated-

measures model with pat identity specified as the subject

was fitted with PROC MIXED in SAS for Windows

(version 9.2; SAS Institute 2008). To account for the

nonindependence of consecutive measurements, we

explicitly modeled the covariance structure between

measurements. To allow the strength of dependence to

vary with the specific pair of measurements being

referenced, we assumed a heterogeneous Toeplitz error

structure (Moser 2004). This flexible model structure

will not only allow us to obtain separate variance

estimates for each week, but also to estimate a

correlation parameter for each time interval that

measurements can be separated by. For additional

statistical analyses testing for effects of local variation

in community composition, see Appendix B.

RESULTS

Citizen scientists in Finland successfully recorded the

mass of a total of 733 individual dung pats across 73

farms during six individual weighing occasions (ntot ¼4180 measurements, given the accidental destruction of

a few pats during the course of the experiment). Overall,

the mass of experimental cow pats decreased with time

(Fig. 2a; Table 1), with specific rates of change differing

significantly among treatments (Table 1). By the end of

the experiment, the exposed pats (treatment 1) were the

lightest, while pats protected from all decomposers

(treatment 5) were the heaviest (Fig. 2a). For each

additional group of decomposers excluded, the amount

of dung remaining by the end of the experiment

significantly increased (Fig. 2a), but specific differences

FIG. 1. Experimental design. (a) Map of Finland with the 73 target farms indicated by red circles. (b) At each farm fiveexperimental treatments were implemented: (1) cow pats exposed to all local decomposers, (2) pats from which the burying activityof dor beetles (genus Geotropes) was blocked, (3) pats from which dor beetles were completely excluded, (4) pats from which bothdor beetles and earthworms were excluded, and (5) pats from which all local decomposers were excluded.

November 2013 2647NOTES

between treatments with and without Aphodius beetles

(treatments 4 and 5), and between the treatment

preventing dor beetles from burying the dung vs. the

fully exposed treatment (treatments 1 and 2, respective-

ly) proved nonsignificant (P ¼ 0.35 and P ¼ 0.16,

respectively). With respect to individual taxa, the largest

difference in pat mass (85 6 20 g [mean 6 SE]) occurred

between treatments with and without the activity of dor

beetles (treatments 1 and 3). Treatment 2, in which dor

beetles were stopped from tunneling but not from

entering, fell half-way between these values, whereas

the exclusion of earthworms and/or dung-dwelling

Aphodius beetles had less of an effect (Fig. 2a).

Beyond the biotic effects of invertebrates, abiotic

climate had a major impact on dung pat mass loss.

Overall, the rate of decomposition slowed significantly

with increasing latitude, and by the end of the

experiment, an additional 16 6 3 g of dung remained

for each 100 km north (Fig. 2b). Pats were, on average,

97 6 10 g heavier after rain (Table 1), with pats

underlaid with dense cloth to exclude earthworms

(treatment 4, Fig. 1b) gaining most mass from rain

water and exposed pats gaining the least (Fig. 2c).

FIG. 2. The development of dung mass as a function of (a) treatment and time, (b) latitude, and (c) treatment and incidence ofrain. (a) Shown are estimates of dung pat mass (means 6 SE) as derived from a generalized linear mixed-effects model (see Table 1).Different treatments (see Fig. 1) are indicated with different colors. Treatments lines identified by different lowercase letters differedsignificantly in final pat mass (P , 0.05). For day 60, at the far right, treatment-specific least-squares means m (6SE) are given innumeric form, as used in the Discussion. (b) To illustrate the relative impact of latitude and treatment on final pat mass, the twomost divergent treatments are shown: treatment 5 (excluding all decomposers; black squares and line) and treatment 1 (including alldecomposers; green circles and line). (c) Here bars indicate final pat masses (mean 6 SE): after rain (striped bars) and in the absenceof rain (solid bars).

TABLE 1. Generalized linear mixed-effects model of dung massas a function of experimental treatment, time, latitude, theoccurrence of rain, and biologically motivated two-wayinteractions, showing Type 3 F ratios for fixed effects, withassociated P values.

Factor df F P

Treatment 4, 743 19.79 ,0.0001Time 5, 1628 237.77 ,0.0001Latitude 1, 773 33.10 ,0.0001Rain 1, 2767 573.01 ,0.0001Treatment 3 rain 4, 2728 5.09 0.0004Treatment 3 time 20, 2258 1.68 0.03Time 3 rain 5, 2122 3.67 0.003

NOTES2648 Ecology, Vol. 94, No. 11

DISCUSSION

In this study we adopted the principles of citizenscience to quantify the specific contribution of individual

invertebrate taxa to the nationwide ecosystem service ofdung decomposition. This was only made possible

through the participation of young volunteers, who

together performed a well-controlled experiment asreplicated across all parts of Finland. To pinpoint the

biological insights gained from this concerted effort, we

will convert our estimates of effect sizes (from Fig. 2a) toan explicit assessment of the relative contribution of

climate and of different invertebrate groups to the

nationwide decomposition of cattle dung (Fig. 3).Of all Finnish cattle dung deposited in pastures, an

average fraction of f will disappear during the summer

due to desiccation and microbial activity alone: f ¼(morig� m5)/morig. Here, morig is the original fresh mass

of the dung pat (average 1165 g; see Materials and

methods), and m5 is the mean mass of pats in treatment 5by the end of the experiment (i.e., the dung pats from

which all invertebrates were excluded; Fig. 1b).

Of the absolute amount of dung decomposed byinvertebrates, the relative fractions decomposed by

individual invertebrate groups i can then be calculatedas di ¼ (mk � mi )/(m5 � m1), where mk and mi are the

average amounts of dung remaining by the end of the

experiment in the absence vs. presence of group i,

respectively; m5 is (as above) the average amount of

dung remaining in the absence of all invertebrate

decomposers, and m1 is the average amount of dung

remaining when all decomposers are included.

Given the treatment-specific means (reported in Fig.

2a), we estimate that an average fraction f of 68% [(morig

� m5)/morig] of the original dung disappears due to

evaporation and microbial activity. Another 13%

disappears due to invertebrate activity, whereas 18%

remains by the end of the season. Nonetheless, these

fractions will differ considerably among different parts

of the country. Given the observed gradient in

decomposition rate with latitude, we estimate that by

the end of the experiment, 10% of the original dung pat

will remain in southernmost Finland, as compared to

26% in the northernmost parts of the country (Fig. 3).

This proportion amounts to a net difference of a full 186

g per pat between the South coast and northernmost

Lapland.

The striking difference in decomposition rate between

southern and northern sites corresponds to a rather

limited difference in temperatures: the summertime

temperatures of southern Finland are on average 3.68C

higher than in the north (Finnish Meteorological

Institute 2012). Nonetheless, it results in an effect as

big as the summed contribution of all invertebrate

FIG. 3. The scale and players of an ecosystem service. Left-hand cube: Finnish cattle produce an annual total of ;4 billion kgof dung. This volume corresponds to a cube with an edge of 160 m (the U.S. Statue of Liberty is superimposed for comparison).Middle sub-cubes: Focusing on the dung deposited in pastures, we estimate the relative fractions (top) disappearing due toevaporation of water and microbial activity alone, (center) decomposing through the action of invertebrates, and (bottom)remaining after two months. The lighter box refers to the situation in northernmost Finland, the darker cube to southernmostFinland (for the derivation of estimates, see Discussion). Right-hand sub-cubes: Of the dung decomposed by invertebrates, (center)dor beetles account for 61%, (bottom) earthworms for 28%, and (top) dung-dwelling Aphodius species for 11%.

November 2013 2649NOTES

decomposers present at any one farm (163g; for

graphical comparisons, see Figs. 2b and 3).

Of the dung decomposed by invertebrates, dor beetles

account for 61% [(m3� m1)/(m5� m1)], earthworms for

28% [(m4 � m3)/(m5 � m1)] and Aphodius dung dwellers

for 11% [(m5 � m4)/(m5 � m1)]; Fig. 3). Overall, each

invertebrate group then seems to add its individual share

to the national ecosystem service, but consistent with a

priori predictions, the dor beetles proved the most

efficient decomposers. When this group was excluded,

about 1.5 times more dung remained by the end of the

season than when all decomposers had access to the

dung, as compared to 1.8 times when all decomposers

were excluded (Fig. 2a). Of the functional efficiency of

dor beetles, about one-third was due to dung being

buried directly under the pat, and about two-thirds to

the beetles actively consuming dung in the pat (see the

difference between treatments 2 and 3; Fig. 2a).

The current observations of the impact of dor beetles

offer large-scale proof for previous findings obtained at

smaller spatial scales: that dung beetle species of high

body mass may be disproportionately important for

sustaining an overall ecosystem service in both temper-

ate (Rosenlew and Roslin 2008) and tropical regions

(Larsen et al. 2005, Slade et al. 2007, 2011). The fate of a

major ecosystem service may then hinge disproportion-

ately on the fate of individual, large-sized species—a

finding with profound consequences for conservation

priorities.

The calculations and inference outlined above show

how important biological insights can be gained from an

experiment implemented at a national scale. In partic-

ular, they illustrate how large-scale experiments built on

public involvement may serve as a bridge between

hypotheses derived from one scale to patterns realized at

another. In the case of dung decomposition, experiments

conducted at a local scale (typically smaller than a

pasture; Finn 2001, Slade et al. 2007, Rosenlew and

Roslin 2008) have been used to derive predictions

regarding the relative roles of different dung decompos-

ers. Our experiment now shows how specific hypotheses

formulated at a small scale (Rosenlew and Roslin 2008)

can be verified at a large one (our present study), and

how local influences thus quantified can be gauged

against large-scale variation in climate. Overall, these

approaches will usefully complement each other in

accounting for large-scale patterns in dung decomposi-

tion across the pastures of Finland.

To date, numerous citizen science projects have

gathered data on species distribution, abundance, and

phenology over large areas, from different taxa and

habitats (see Dickinson et al. 2010 for a review). Our

study serves to identify another use for citizen science, as

recently proposed by multiple authors (including Dick-

inson et al. 2010): that citizen scientists may implement

well-controlled experiments across large numbers of

sites and vast areas. At least two prior studies have

already achieved this in terms of specific manipulations

(Jones et al. 1998, Hames et al. 2002), and our study

illustrates how this crucial idea may be further expanded

to more complex designs.

As a field of particular relevance for a new brand of

experimental citizen science, we offer experiments aimed

at quantifying vital ecosystem services. As illustrated by

our experiment on dung decomposition rates, such joint

ventures between scientists and citizens may not only

offer unique data sets to the scientists, but also give the

citizens direct insights into the services from which they

benefit. Agricultural ecosystem services, including nat-

ural pest control, pollination (see Kremen et al. 2011),

soil aeration and nutrient recycling (Kremen et al. 2002,

Tscharntke et al. 2005, Macfadyen et al. 2009), may be

identified as promising targets for further projects in

citizen science. These services are of direct relevance to

potential citizen volunteers—and given their scope and

scale, they are hard to target by individual research

groups. While we urgently need accurate predictions of

how declining biodiversity will affect agroecosystem

services worldwide (Naeem and Wright 2003,

Tscharntke et al. 2005, Macfadyen et al. 2009, Carval-

heiro et al. 2011), our current assumptions regarding key

processes are typically derived from small-scale experi-

ments focusing on limited types of systems (Rustad et al.

2001, Naeem and Wright 2003). Here, citizen science

might offer a way to expand novel experimental designs

to new levels and systems.

But if the volunteer-based approach is as powerful as

we claim, why then have so few previous studies turned

to volunteers to implement ecological experiments? The

scarcity of such studies can perhaps be traced to

concerns regarding the quality of the data collected by

citizen scientists (Kremen et al. 2011). Data from

experiments built on volunteer-based sampling may

indeed include more noise than data from small-scale

experiments implemented by a small number of profes-

sional biologists (Penrose and Call 1995, Engel and

Voshell 2002, Nerbonne and Vondracek 2003). Howev-

er, recent studies suggest that quality of the data in

volunteer-based projects is more likely to be determined

by study design, analytical methodology, and commu-

nication skills, rather than by the volunteer-based

method per se (Engel and Voshell 2002, Lovell et al.

2009, Schmeller et al. 2009). Here, higher sampling

effort, as achieved by larger numbers of volunteers and

more visited sites, may actually result in more precise

and unbiased results than traditional approaches—since

precision is also a function of the number of samples, as

maximized by volunteer involvement (Hochachka et al.

2000, Schmeller et al. 2009).

Clearly, citizen science comes with limits on what can

be achieved. To allow strict standardization and efficient

communication of methods, our experimental treat-

NOTES2650 Ecology, Vol. 94, No. 11

ments were designed for maximal simplicity. As such,

they were based on removing species in a given sequence

(primarily based on body size). Hence, the resultant

design did not allow us to explore all interactions

between individual invertebrate groups, as might be

achieved when experimental communities are created by

adding rather that excluding taxa (e.g., Sheehan et al.

2006, O’Hea et al. 2010). Considering this limitation, we

believe that small-scale experiments conducted by

professional ecologists and large-scale experiments

based on citizen science should be seen as complemen-

tary rather than exclusive approaches to resolve major

questions in ecology. We also suggest that citizen science

may provide a rare chance to test for the scale-

dependency of patterns (e.g., Kunin 1998, Rahbek

2005, Azaele et al. 2012). With the help of citizen

scientists the same experiment may be performed

simultaneously at small and large scales, and the

resulting patterns directly compared. While our exper-

imental design did not exploit this option, it offers a

prime opportunity for studies to come.

Taken together, by using citizen scientists to imple-

ment an ecological experiment, we have been able to

usefully dissect the drivers of an ecosystem service on a

nationwide scale. Importantly, had this project been

implemented by professional scientists, the price would

have been manifold. The cost estimates presented in

Appendix A show how relying on local citizen scientists

present in the right place at the right time may

drastically reduce the expenses for a research project,

from an unrealistic level to one affordable by a finite

research grant. In our case, the savings amounted to

three-quarters of the budget. As implemented through

citizen science, the project also provided hands-on

insights into an ecosystem service for the young people

involved, and the same insights were widely publicized

through both local and national news. Thus, we believe

that some of the greatest prospects for citizen science

reside in the potential for expanding experiments in

ecology to new scales and scopes, and for allowing the

public to learn in the process. We therefore hope that

our approach may be used as a template for many

studies to come.

ACKNOWLEDGMENTS

We thank all the participating citizen scientists for conduct-ing the hard work forming the basis of this study. We are alsoindebted to the farmers allowing this experiment to becompleted on their pastures, and to the Finnish 4H Federationin general and Virpi Skippari and Anna-Kaisa Valaja inparticular for their tireless support and enthusiasm. Threeanonymous reviewers offered insightful comments that helpedus improve previous versions of the manuscript.

LITERATURE CITED

Azaele, S., S. J. Cornell, and W. E. Kunin. 2012. Downscalingspecies occupancy from coarse spatial scales. EcologicalApplications 22:1004–1014.

Beaubien, E., and A. Hamann. 2011. Spring flowering responseto climate change between 1936 and 2006 in Alberta, Canada.BioScience 61:514–524.

Carvalheiro, L. G., R. Veldtman, A. G. Shenkute, G. B. Tesfay,C. W. W. Pirk, J. S. Donaldson, and S. W. Nicolson. 2011.Natural and within-farmland biodiversity enhances cropproductivity. Ecology Letters 14:251–259.

Cooper, C. B., J. Dickinson, T. Phillips, and R. Bonney. 2007.Citizen science as a tool for conservation in residentialecosystems. Ecology and Society 12:11.

Devictor, V., R. J. Whittaker, and C. Beltrame. 2010. Beyondscarcity: citizen science programmes as useful tools forconservation biogeography. Diversity and Distributions 16:354–362.

Dickinson, J. L., B. Zuckerberg, and D. N. Bonter. 2010.Citizen science as an ecological research tool: challenges andbenefits. Annual Review of Ecology, Evolution, and System-atics 41:149–172.

Engel, S. R., and J. R. Voshell. 2002. Volunteer biologicalmonitoring: Can it accurately assess the ecological conditionof streams? American Entomologist 48:164–177.

Finn, J. A. 2001. Ephemeral resource patches as model systemsfor diversity–function experiments. Oikos 92:363–366.

Finnish Meteorological Institute. 2012. Kesasaan tilastoja.www.ilmatieteenlaitos.fi/kesatilastot

Gittings, T., P. S. Giller, and G. Stakelum. 1994. Dungdecomposition in contrasting temperate pastures in relationto dung beetle and earthworm activity. Pedobiologia 38:455–474.

Hames, R. S., K. V. Rosenberg, J. D. Lowe, S. E. Barker, andA. A. Dhondt. 2002. Adverse effects of acid rain on thedistribution of the Wood Thrush Hylocichla mustelina inNorth America. Proceedings of the National Academy ofSciences USA 99:11235–11240.

Hardwick, B., T. Roslin, and R. Kaartinen. 2011. Kesan 2011lantakuoriaistutkimus. http://www.helsinki.fi/foodwebs/lantakuoriaiset/2011/index.htm

Hochachka, W. M., K. Martin, F. Doyle, and C. J. Krebs.2000. Monitoring vertebrate populations using observationaldata. Canadian Journal of Zoology 78:521–529.

Holter, P., and N. B. Hendriksen. 1988. Respiratory loss andbulk export of organic matter from cattle dung pats: a field-study. Holarctic Ecology 11:81–86.

Horgan, F. G. 2008. Dung beetle assemblages in forests andpastures of El Salvador: a functional comparison. Biodiver-sity and Conservation 17:2961–2978.

Jones, C. G., R. S. Ostfeld, M. P. Richard, E. M. Schauber, andJ. O. Wolff. 1998. Chain reactions linking acorns to gypsymoth outbreaks and Lyme disease risk. Science 279:1023–1026.

Kremen, C., K. S. Ullmann, and R. W. Thorp. 2011.Evaluating the quality of citizen-scientist data on pollinatorcommunities. Conservation Biology 25:607–617.

Kremen, C., N. M. Williams, and R. W. Thorp. 2002. Croppollination from native bees at risk from agriculturalintensification. Proceedings of the National Academy ofSciences USA 99:16812–16816.

Kunin, W. E. 1998. Extrapolating species abundance acrossspatial scales. Science 281:1513–1515.

Larsen, T. H., N. M. Williams, and C. Kremen. 2005.Extinction order and altered community structure rapidlydisrupt ecosystem functioning. Ecology Letters 8:538–547.

Losey, J. E., and M. Vaughan. 2006. The economic value ofecological services provided by insects. BioScience 56:311–323.

Lovell, S., M. Hamer, R. Slotow, and D. Herbert. 2009. Anassessment of the use of volunteers for terrestrial invertebratebiodiversity surveys. Biodiversity and Conservation 18:3295–3307.

November 2013 2651NOTES

Macfadyen, S., R. Gibson, A. Polaszek, R. J. Morris, P. G.Craze, R. Planque, W. O. C. Symondson, and J. Memmott.2009. Do differences in food web structure between organicand conventional farms affect the ecosystem service of pestcontrol? Ecology Letters 12:229–238.

Moser, E. B. 2004. Repeated measures modeling with PROCMIXED. Paper 188-29 in Proceedings of the SAS UsersGroup International 29th Conference, 9–12 May 2004,Montreal, Canada. http://www2.sas.com/proceedings/sigi29/188-29.pdf

Naeem, S., and J. P. Wright. 2003. Disentangling biodiversityeffects on ecosystem functioning: deriving solutions to aseemingly insurmountable problem. Ecology Letters 6:567–579.

Nerbonne, J. F., and B. Vondracek. 2003. Volunteer macroin-vertebrate monitoring: assessing training needs throughexamining error and bias in untrained volunteers. Journalof the North American Benthological Society 22:152–163.

Nichols, E., S. Spector, J. Louzada, T. Larsen, S. Amequita,M. E. Favila, and the Scarabaeinae Research Network. 2008.Ecological functions and ecosystem services provided byScarabaeinae dung beetles. Biological Conservation 141:1461–1474.

Nieminen, M., E. Ketoja, J. Mikola, J. Terhivuo, T. Siren, andV. Nuutinen. 2011. Local land use effects and regionalenvironmental limits on earthworm communities in Finnisharable landscapes. Ecological Applications 21:3162–3177.

O’Hea, N. M., L. Kirwan, and J. A. Finn. 2010. Experimentalmixtures of dung fauna affect dung decomposition throughcomplex effects of species interactions. Oikos 119:1081–1088.

Penrose, D., and S. M. Call. 1995. Volunteer monitoring ofbenthic macroinvertebrates: regulatory biologists’ perspec-tives. Journal of the North American Benthological Society14:203–209.

Rahbek, C. 2005. The role of spatial scale and the perception oflarge-scale species-richness patterns. Ecology Letters 8:224–239.

Rosenlew, H., and T. Roslin. 2008. Habitat fragmentation andthe functional efficiency of temperate dung beetles. Oikos117:1659–1666.

Roslin, T., and K. Heliovaara. 2007. Suomen Lantakuoriaiset.Opas santiaisista lantiaisiin. Yliopistopaino, Helsinki, Fin-land.

Rustad, L. E., J. L. Campbell, G. M. Marion, R. J. Norby,M. J. Mitchell, A. E. Hartley, J. H. C. Cornelissen, J.Gurevitch, and GCTE-NEWS [Global Change and Terres-trial Ecosystems Network of Ecosystem Warming Studies].2001. A meta-analysis of the response of soil respiration, netnitrogen mineralization, and aboveground plant growth toexperimental ecosystem warming. Oecologia 126:543–562.

SAS Institute. 2008. SAS/STAT user’s guide, version 9.2. SASInstitute, Cary, North Carolina, USA.

Schmeller, D. S., et al. 2009. Advantages of volunteer-basedbiodiversity monitoring in Europe. Conservation Biology 23:307–316.

Sheehan, C., L. Kirwan, J. Connolly, and T. Bolger. 2006. Theeffects of earthworm functional group diversity on nitrogendynamics in soils. Soil Biology and Biochemistry 38:2629–2636.

Slade, E. M., D. J. Mann, and O. T. Lewis. 2011. Biodiversityand ecosystem function of tropical forest dung beetles undercontrasting logging regimes. Biological Conservation 144:166–174.

Slade, E. M., D. J. Mann, J. F. Villanueva, and O. T. Lewis.2007. Experimental evidence for the effects of dung beetlefunctional group richness and composition on ecosystemfunction in a tropical forest. Journal of Animal Ecology 76:1094–1104.

Tscharntke, T., A. M. Klein, A. Kruess, I. Steffan-Dewenter,and C. Thies. 2005. Landscape perspectives on agriculturalintensification and biodiversity—ecosystem service manage-ment. Ecology Letters 8:857–874.

Wall, D. H., et al. 2008. Global decomposition experimentshows soil animal impacts on decomposition are climate-dependent. Global Change Biology 14:2661–2677.

Wall, R., and L. Strong. 1987. Environmental consequences oftreating cattle with the antiparasitic drug ivermectin. Nature327:418–421.

Winkler, D. W., P. O. Dunn, and C. E. McCulloch. 2002.Predicting the effects of climate change on avian life-historytraits. Proceedings of the National Academy of Sciences USA99:13595–13599.

SUPPLEMENTAL MATERIAL

Appendix A

Detailed description of the project, including additional information on the study design and success rate, a budget estimate forthe project as implemented by citizen scientists vs. professional biologists, and the identification of nonmonetary benefits of usingcitizen scientists (Ecological Archives E094-242-A1).

Appendix B

Additional analyses of community structure vs. functional rates (Ecological Archives E094-242-A2).

Appendix C

Experiment addressing the physical impact of invertebrate exclusion devices on dung mass loss (Ecological ArchivesE094-242-A3).

NOTES2652 Ecology, Vol. 94, No. 11