Advantages of volunteer-based biodiversity monitoring in Europe.

Abstract

Without robust and unbiased systems for monitoring, changes in natural systems will remain enigmatic for policy makers, leaving them without a clear idea of the consequences of any environmental policies they might adopt. Generally, biodiversity-monitoring activities are not integrated or evaluated across any large geographic region. The EuMon project conducted the first large-scale evaluation of monitoring practices in Europe through an on-line questionnaire and is reporting on the results of this survey. In September 2007 the EuMon project had documented 395 monitoring schemes for species, which represents a total annual cost of about 4 million euro, involving more than 46,000 persons devoting over 148,000 person-days/year to biodiversity-monitoring activities. Here we focused on the analysis of variations of monitoring practices across a set of taxonomic groups (birds, amphibians and reptiles, mammals, butterflies, plants, and other insects) and across 5 European countries (France, Germany, Hungary, Lithuania, and Poland). Our results suggest that the overall sampling effort of a scheme is linked with the proportion of volunteers involved in that scheme. Because precision is a function of the number of monitored sites and the number of sites is maximized by volunteer involvement, our results do not support the common belief that volunteer-based schemes are too noisy to be informative. Just the opposite, we believe volunteer-based schemes provide relatively reliable data, with state-of-the-art survey designs or data-analysis methods, and consequently can yield unbiased results. Quality of data collected by volunteers is more likely determined by survey design, analytical methodology, and communication skills within the schemes rather than by volunteer involvement per se.

Full text

Page 1

Contributed Paper
Advantages of Volunteer-Based Biodiversity
Monitoring in Europe
DIRK S. SCHMELLER,a,b,cPIERRE-YVES HENRY,dROMAIN JULLIARD,dBERND GRUBER,e
JEAN CLOBERT,aFRANK DZIOCK,b,fSZABOLCS LENGYEL,gPIOTR NOWICKI,h
ESZTER D´ ERI,iEDUARDAS BUDRYS,jTIIU KULL,kKADRI TALI,kBIANCA BAUCH,b
JOSEF SETTELE,lCHRIS VAN SWAAY,mANDREJ KOBLER,nVALERIJA BABIJ,o
EVA PAPASTERGIADOU,pAND KLAUS HENLEb
aStation d’Ecologie Exp´ erimentale du CNRS ` a Moulis, 09200 Saint Girons, France
bUFZ – Helmholtz-Centre for Environmental Research, Department of Conservation Biology, Permoserstrasse 15, 04318 Leipzig,
Germany
dUMR 5173 & UMR 7179 MNHN-CNRS-UPMC, Mus´ eum National d’Histoire Naturelle, CP 51, 55 rue Buffon, Paris, France
eUFZ – Helmholtz-Centre for Environmental Research, Department of Computational Landscape Ecology, Permoserstrasse 15,
04318 Leipzig, Germany
fTechnische Universit¨ at Berlin, Department of Biodiversity Dynamics, Rothenburgstrasse 12, 12165 Berlin, Germany
gDepartment of Ecology, University of Debrecen, 4032 Debrecen, Egyetem t´ er 1, Hungary
hInstitute of Environmental Sciences, Jagiellonian University, Gronostajowa 7, 30-387 Krak´ ow, Poland
iDepartment of Evolutionary Zoology, University of Debrecen, 4032 Debrecen, Egyetem t´ er 1, Hungary
jInstitute of Ecology of Vilnius University, Akademijos 2, 08412 Vilnius, Lithuania
kEstonian University of Life Sciences, Institute of Agricultural and Environmental Sciences, Riia 181, 51014 Tartu, Estonia
lUFZ – Helmholtz-Centre for Environmental Research, Department of Community Ecology, Theodor-Lieser-Str. 4, 06120 Halle,
Germany
mDe Vlinderstichting – Dutch Butterfly Conservation, P.O. Box 506, 6700 AM Wageningen, The Netherlands
nDepartment of Forest Inventory and Spatial Information Systems, Slovenian Forestry Institute, Vecna pot 2, 1000 Ljubljana,
Slovenia
oScientific Research Centre of Slovenian Academy of Sciences and Arts Jovan Hadˇ zi Institute of Biology, Novi trg 2, 1000
Ljubljana, Slovenia
pDepartment of Biology Plant Ecology & Ecosystems Management, 26500 Patras, Greece
Abstract: Without robust and unbiased systems for monitoring, changes in natural systems will remain
enigmatic for policy makers, leaving them without a clear idea of the consequences of any environmental
policies they might adopt. Generally, biodiversity-monitoring activities are not integrated or evaluated across
any large geographic region. The EuMon project conducted the first large-scale evaluation of monitoring
practicesinEuropethroughanon-linequestionnaireandisreportingontheresultsofthissurvey.InSeptember
2007 the EuMon project had documented 395 monitoring schemes for species, which represents a total annual
cost of about €4 million, involving more than 46,000 persons devoting over 148,000 person-days/year to
biodiversity-monitoring activities. Here we focused on the analysis of variations of monitoring practices
across a set of taxonomic groups (birds, amphibians and reptiles, mammals, butterflies, plants, and other
insects) and across 5 European countries (France, Germany, Hungary, Lithuania, and Poland). Our results
suggest that the overall sampling effort of a scheme is linked with the proportion of volunteers involved in that
scheme. Because precision is a function of the number of monitored sites and the number of sites is maximized
by volunteer involvement, our results do not support the common belief that volunteer-based schemes are too
noisy to be informative. Just the opposite, we believe volunteer-based schemes provide relatively reliable data,
cAddress correspondence to Dirk S. Schmeller, email ds@die-schmellers.de
Paper submitted November 12, 2007; revised manuscript accepted July 28, 2008.
307
Conservation Biology, Volume 23, No. 2, 307–316
C ?2008 Society for Conservation Biology
DOI: 10.1111/j.1523-1739.2008.01125.x

Page 2

308
Volunteer-Based Biodiversity Monitoring
with state-of-the-art survey designs or data-analysis methods, and consequently can yield unbiased results.
Quality of data collected by volunteers is more likely determined by survey design, analytical methodology,
and communication skills within the schemes rather than by volunteer involvement per se.
Keywords: biodiversity monitoring, costs, data collection, sampling effort, volunteer involvement
Ventajas del Monitoreo de Biodiversidad Basado en Voluntarios en Europa
Resumen: Sin sistemas de monitoreo robustos y objetivos, los cambios en los sistemas naturales seguir´ an
siendo un enigma para los pol´ ıticos, ya que no tendr´ an una idea clara de las consecuencias de las pol´ ıticas
ambientales que puedan adoptar. Generalmente, las actividades de monitoreo de la biodiversidad no est´ an
integradas o evaluadas en ninguna regi´ on geogr´ afica extensa. El proyecto EuMon llev´ o a cabo la primera
evaluaci´ on a gran escala de las pr´ acticas de monitoreo en Europa por medio de un cuestionario en l´ ınea
y presenta los resultados de este muestreo. En septiembre 2007, el proyecto EuMon hab´ ıa documentado
395 esquemas de monitoreo de especies, lo que representa un costo anual total de casi €4 millones, y la
participaci´ on de m´ as de 46,000 personas que dedican m´ as de 148,000 personas d´ ıas/a˜ no en actividades
de monitoreo de la biodiversidad. Aqu´ ı, nos centramos en el an´ alisis de variaciones en las pr´ acticas de
monitoreo de un conjunto de grupos taxon´ omicos (aves, anfibios y reptiles, mam´ ıferos, mariposas, plantas y
otros insectos) en cinco pa´ ıses europeos (Francia, Alemania, Hungr´ ıa, Lituania y Polonia). Nuestros resultados
sugieren que el esfuerzo de muestreo total de un esquema est´ a vinculado con la proporci´ on de voluntarios
que participan en el ese esquema. Debido a que la precisi´ on es una funci´ on del n´ umero de sitios monitoreados
y el n´ umero de sitios es maximizado por la participaci´ on de voluntarios, nuestros resultados no soportan la
creencia popular de que los esquemas basados en voluntarios tienen demasiado ruido para ser informativos.
Al contrario, consideramos que los esquemas basados en voluntarios proporcionan datos confiables, con
dise˜ nos de muestreo o m´ etodos de an´ alisis de datos de vanguardia, y consecuentemente pueden proporcionar
resultados objetivos. La calidad de los datos recolectados por voluntarios probablemente est´ a determinada
por el dise˜ no del muestreo, la metodolog´ ıa anal´ ıtica y las habilidades de comunicaci´ on de los esquemas y no
por la participaci´ on de voluntarios per se.
Palabras Clave: costos, esfuerzo de muestreo, monitoreo de biodiversidad, participaci´ on de voluntarios,
recolecci´ on de datos
Introduction
International conventions, such as the Convention on
Biodiversity (CBD) and the European nature directives,
compel national governments to implement and identify
existing biodiversity-monitoring schemes (92/43/EEC,
conservation of natural habitats and of wild fauna and
flora [Habitats Directive]; 79/409/EEC, conservation of
wild birds [Birds Directive], see also Evans 2005). With-
out robust and unbiased systems for monitoring changes
in natural systems, policy makers will not realize the ex-
tent of change, and they will have no clear way to eval-
uate the consequences of environmental policies they
might adopt (Balmford et al. 2003, 2005). A good recent
example of how policy can be informed by biodiversity
monitoring and analysis was shown for European bird
species (Donald et al. 2007).
Biodiversity monitoring entails 3 activities: collecting
monitoring data, such as occurrence or abundance in-
dices in a target area over an appropriate timescale; an-
alyzing the spatial and temporal patterns of biodiversity
components, including diagnosing the causes of change
as robustly as possible; and deriving results oriented
toward policy makers, such as status and trend assess-
ments or management evaluations. Biodiversity monitor-
ing faces at least 2 practical difficulties: the need to main-
tain a sustained effort of monitoring across years to en-
sure the collection of relevant monitoring time series and
the need to obtain precise monitoring data that allow the
detection of significant changes across space and time
in biodiversity. These needs come into conflict with the
usually limited amount of available financial and human
resources. A trade-off solution is to involve volunteers
in monitoring activities. In Europe and North America
many monitoring organizations rely on such volunteer
programs (van Swaay et al. 1997; Link & Sauer 1998; Van-
denbosch 2003; Thomas 2005; Gregory et al. 2005). The
involvementofthepublicisimperativetoreducethecost
of biodiversity monitoring and has the added benefit of
enhancing citizen participation in science practices and
thusenvironmentalawareness(e.g.,Belletal.2008).Nev-
ertheless, volunteer-based monitoring is often perceived
as simplistic, prone to higher biases than professional
monitoring schemes (Engel & Voshell 2002; Genet & Sar-
gent 2003), and as a trade-off between precision and cost
(Brashares & Sam 2005).
Here we report on the results of the first large-
scale evaluation of monitoring practices in Europe. We
compared monitoring practices and potential benefits re-
sulting from volunteer involvement in different species
Conservation Biology
Volume 23, No. 2, 2009

Page 3

Schmeller et al.
309
groups and across several eastern and western European
countries.
Methods
The EuMon project (Schmeller et al. 2006) conducted
a survey on biodiversity-monitoring practices across Eu-
rope from 2005 to 2007. We examined the data gathered
up to September 2007. Information on monitoring pro-
grams and practices was and still is collected through
an Internet questionnaire. The questionnaire had 8 ques-
tions on basic features of biodiversity monitoring and
33 questions specific to species-monitoring methods and
design (see Supporting Information). For reaching rep-
resentatives of stakeholder groups involved in monitor-
ing activities (governmental and nongovernmental bod-
ies), we distributed announcements of the questionnaire
through emails, letters, and at conferences to over 1600
individuals and several national and international mailing
lists. We asked respondents to provide data on-line, but
also accepted written responses, which were inputted
to the database by us. Responses were checked for com-
pleteness, and missing details were sought from the co-
ordinators of monitoring projects. After validation data
were accepted and made publicly available in a database.
Complete information was not available for all schemes;
hence, the sample sizes and degrees of freedom differ
slightly among tests and comparisons.
Determination of Biases in Taxonomic and Geographical
Coverage
Despite the fact that the EuMon survey is the first large-
scale survey of its kind, it may suffer from biases in taxo-
Table 1. Variables describing biodiversity-monitoring practices in our study.
VariableAbbreviation Description
Annual frequency ann. freq.between-year frequency of monitoring;
1, every year; 2, every second year, etc.
person-days/monitoring scheme
person-days/monitoring scheme
Work effort
Work effort per species divided
by the number of species monitored
No. of professionals
No. of samples
No. of sites
No. of species
No. of visits
No. of volunteers
Persons
Proportion of volunteers
Total costs
person-days
person-daysspec
Nprof
Nsamples
Nsites
Nspecies
Nvisits
Nvolunteers
Npersons
% vol
costs
no. professionals in a scheme
no. samples collected/visit
no. sites visited
no. species monitored simultaneously in a scheme
no. visits per year
no. volunteers in a scheme
sum of volunteer and professionals in one scheme
proportion of volunteers in a scheme
sum of costs of equipment and salaries of professionals, as given
by the average salary per country in 2005 (World Bank 2006)
sum of costs of equipment and salaries of professionals, as given
by the average salary per country in 2005 (World Bank 2006)
divided by no. of species monitored in one scheme
start year of monitoring scheme
Total costs per speciescostsspec
Yearyear
nomic and geographic coverage. We assessed the geo-
graphical bias by searching for country-related moni-
toring entries in Zoological Records. We used the fol-
lowing search string: biodiversity AND monitoring AND
species AND country. The taxonomic bias was assessed
by searching the Zoological Records for animal species
only and Google Scholar for references to monitoring.
The search query we used was monitoring AND species
group AND europ. AND biodiversity. We computed the
bias as logit(observed) – logit(expected), where the ob-
served values were the values from our database and the
expected values the records from Zoological Records or
Google Scholar that met our search criteria. Neverthe-
less, the 2 databases may suffer from the same type of
biases as our survey—differential inclinations of monitor-
ingschemestopublishtheirresults.Therefore,thebiases
in our data would not differ much from usual publication
biases.
Variables and Statistical Analysis
Intotalwecollecteddataon14variables(Table1),which
characterized monitoring practices and resource needs
of biodiversity monitoring in Europe. We examined dif-
ferences of monitoring schemes among species groups
that were sufficiently covered by our survey (birds, n =
149; amphibians and reptiles, n = 53; mammals, n = 90;
plants, n = 58; butterflies n = 37; other insects, n = 31).
The bird group comprised all bird-monitoring schemes,
regardless of whether they targeted passerines, water
birds, or any other subgroup. In addition, the mammal
groupcomprisedspeciesasdifferentasbats,rodents,and
large carnivores. In plants we considered only monitor-
ing schemes of vascular plants. The other-insects group
included monitoring schemes of dragonflies, beetles, and
Conservation Biology
Volume 23, No. 2, 2009

Page 4

310
Volunteer-Based Biodiversity Monitoring
grasshoppers, but not butterflies. Country-specific differ-
ences were examined for France (n = 91), Germany (n
= 37), Hungary (n = 30), Lithuania (n = 32), and Poland
(n = 101).
We used principal component analyses with 14 vari-
ables to describe differences in monitoring practices in
different species groups and countries. For determina-
tion of the most discriminating variables among species
groups and countries, we used a fully saturated discrim-
inate analysis with stepwise exclusion of statistically in-
significantvariables.Wecheckedvisuallythatallpairwise
relationships between the variables were linear (Support-
ing Information). For correlations, we used the nonpara-
metric Spearman rank correlation to obtain most robust
results. The statistical analyses were done with the R sta-
tistical package (R Development Core Team 2008) and
the package ADE4 (Dray & Dufour 2007).
Results
The EuMon survey compiled descriptive data for 395
species-monitoring schemes led by 227 organizations
in 28 European countries (data extracted September
2007). The total annual cost of these monitoring schemes
was approximately €4 million, and they engaged more
than 46,000 people, who devoted over 148,000 person-
days per year to monitoring activities. Only 13.3%
of the participants in the monitoring programs were
professionals.
Biases
Despite a large sample we found some biases in our
survey. Schemes from Hungary, Lithuania, Poland, and
Slovenia were overrepresented, whereas schemes from
Greece,theUnitedKingdom,Italy,andSpainwereunder-
represented (Supporting Information). The taxonomic
bias following from Google Scholar and from Zoological
Record entries was consistent in a few taxonomic groups
only (Supporting Information). Whereas amphibian and
reptile monitoring schemes were overrepresented in the
EuMon database, schemes monitoring fungi were the
mostunderrepresentedfollowed,respectively,bylichens
and fishes. Schemes monitoring vascular plants were
slightly overrepresented according to Google Scholar.
For the remaining analyzed taxonomic groups, their rela-
tiverepresentationintheEuMondatabase,theZoological
Records, andGoogleScholarweresimilar(SupportingIn-
formation).
Monitoring in Different Species Groups
The different variables not only showed some differences
between the different species groups analyzed but also
had some common features (Fig. 1). For plant monitor-
ing only, the starting year was negatively related to the
number of species and sites and the proportion of vol-
unteers, whereas starting year hardly contributed to the
principal components in the other species groups. For
all species groups, number of sites was positively related
to variables describing the work effort per scheme, with
varying degrees in the different species groups. Other
variables of sampling effort were related to work-effort
variables, especially to the proportion of volunteers (Fig.
1). For monitoring of all species groups, except butter-
flies, the proportion of volunteers was positively associ-
ated with the number of sites and species. In butterfly
monitoring, the proportion of volunteers was positively
associated with the number of sites and visits. The pro-
portion of volunteers correlated well with the number of
sites in schemes monitoring herpetofauna (R48= 0.39;
p = 0.006), birds (R142= 0.28; p < 0.001), butterflies
(R36= 0.38; p = 0.024), and mammals (R85= 0.23; p =
0.031).
The variable costs contributed strongly to the princi-
pal components in all species groups, but related differ-
ently to the other variables (Fig. 1). For birds, costs were
negatively associated with proportion of volunteers. For
herpetofauna, monitoring costs were positively associ-
ated with number of species and the between-year fre-
quency of sampling, and negatively related to the pro-
portion of volunteers. For schemes monitoring mammals
and the group other insects, costs were positively related
to number of sites and species. For butterflies the annual
frequency and number of species were positively associ-
ated with the total costs. For plants costs were positively
associatedwiththeperspeciesvariables,person-daysand
costs per species.
The main differences in monitoring practices in the
different species groups resulted from differences in 3
variables, the starting year of a scheme, the number
of sites monitored, and the proportion of volunteers
per scheme (Fig. 2a). Across different species groups,
the variables overlapped largely (Table 2). Nevertheless,
costs by species were the highest for mammal, amphib-
ian, and reptile species, whereas the overall costs of a
monitoring scheme were highest for butterfly monitor-
ing.Obviously,bird-monitoringschemeswereamongthe
longest running schemes, but usually only 23 sites were
monitored, whereas most butterfly-monitoring schemes
surveyed more than 50 sites (Table 2).
Monitoring in Different Countries
The principal component analysis revealed some differ-
ences in monitoring practices in France, Germany, Hun-
gary, Poland, and Lithuania (Fig. 3). In France variables
of the sampling effort, except of number of sites, con-
tributed weakly to any principal component. In German
monitoring schemes, proportion of volunteers was posi-
tively associated with several variables of sampling effort
Conservation Biology
Volume 23, No. 2, 2009

Page 5

Schmeller et al.
311
Amphibians/Reptiles
Year
Costsspec
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Nsamples
Ann.Freq
Nsites
Persondays
Costs
Persondaysspec
Eigenvalues
First: 0.29
Second: 0.22
Birds
Year
Nspecies
Nvolunteers
Persondays
Nprof
Persondaysspec
Costsspec
Npersons
Nsites
%vol
Nvisits
Ann.Freq
Nsamples
Costs
Eigenvalues
First: 0.27
Second: 0.17
Other Insects
Year
Nspecies
Ann.Freq
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Nsites
Nsamples
Persondays
Costs
Persondaysspec
Costsspec
Eigenvalues
First: 0.32
Second: 0.18
Plants
Year
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Ann.Freq
Nsites
Nsamples
Persondays
CostsPersondaysspec
Costsspec
Eigenvalues
First: 0.27
Second: 0.19
Butterflies
Year
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Ann.Freq
Nsites
Persondaysspec
Nsamples
Persondays
Costs
Costsspec
Eigenvalues
First: 0.28
Second: 0.21
Mammals
Year
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Ann.Freq
Nsites
Nsamples
Persondays
Costs
Persondaysspec
Costsspec
Eigenvalues
First: 0.29
Second: 0.17
Figure 1. Results of the principal component analysis of monitoring practices in different species groups.
Abbreviations are defined in Table 1 (x-axis = axis 1; y-axis = axis 2).
Conservation Biology
Volume 23, No. 2, 2009

Page 6

312
Volunteer-Based Biodiversity Monitoring
Countries
Year
Nspeciesspecies
Nvolunteers
Npersons
persons
Nprof prof
%vol
Nvisits
Nvisits
Ann.Freq
Nsamples
Nsamples
Nsites
HU HU
Persondays
Costs
Costs
Persondaysspec
Costsspec
DE DE
FR FR
LT LT
PL PL
1.0
-1.0
0.5
0.0
-0.5
1.00.5 0.0-0.5-1.0
Species group
Year
Nspecies
Nvolunteers
Mammals Mammals Mammals Mammals
Nprof
Npersons
Nsamples
%vol
Nvisitsvisitsvisits visits
Ann.Freq
Nsites
Persondays
Costs
spec
spec spec
Costsspec
Birds Birds Birds Birds
Butterflies Butterflies Butterflies Butterflies
Other Insects Other Insects Other Insects Other Insects
Plants Plants Plants Plants
1.0
-1.0
0.5
0.0
-0.5
1.00.5 0.0-0.5
-0.5
-1.0
volunteers
spec
Persondaysspec
Amphibians/Reptiles
Figure 2. Canonical analysis results of the
discriminant analysis, indicating the contributions of
the different variables to monitoring schemes in the
(a) different species groups and for (b) different
countries. The labels of the species groups and
countries are located at the group’s mean.
Abbreviations are defined in Table 1 (x-axis is axis 1;
y-axis is axis 2).
(number of sites, number of samples, and number of
species). In Hungary, Germany, and Lithuania, the vari-
able year was negatively associated with proportion of
volunteers in a scheme.
Generally, costs increased with increasing work force
of a scheme, regardless of the country. In Hungary, costs
increased with increasing sampling effort and proportion
of volunteers. In Germany and Lithuania costs increased
with number of sites and person-days, whereas in France
costs were positively related to number of professionals
in a scheme. Across all schemes number of sites (R376=
0.48; p < 0.001), number of visits (R376 = 0.15; p =
0.004), and number of samples (R366= 0.28; p < 0.001)
were all significantly related to the sampling effort of a
monitoring scheme (number of people).
The variables person-days, proportion of volunteers,
starting year, and number of sites explained most of the
differences in monitoring practices in the different coun-
tries (Fig. 2b). Although in France monitoring was started
in 1993 for 50% of the schemes, the majority of monitor-
ing schemes in Lithuania started after 2005 (Table 2).
Furthermore, most of the schemes in Poland and Lithua-
nia were expert schemes, with no involvement of volun-
teers. Germany had the highest number of person-days
of all 5 countries. Poland had the lowest number of sites
per scheme (Table 2).
Three variables—total costs, number of samples, and
proportion of volunteers—contributed most to discrimi-
nation of monitoring practices in the different countries
(Fig.2b).Inaddition,numberofprofessionalschemesdif-
fered geographically. In Poland 63.7% (n = 101) of mon-
itoring schemes were run by professionals only, whereas
this proportion was 33.3% (n = 30) in Hungary, 28.3%
(n = 93) in France, and 23.7% (n = 38) in Germany.
Discussion
Our results show various differences among monitoring
practices in different species groups and different coun-
tries. Generally, our results support an increased sam-
pling effort with increasing involvement of volunteers in
biodiversity monitoring. Hence, our results do not sup-
port the common belief that volunteer-based schemes
yield noisy, imprecise results.
Geographic Bias
Despitelargeeffortsinseekingaddressesandsendingout
requeststofilloutouron-linequestionnaire,theresponse
rate was low and varied greatly among countries, which
led to a geographic bias in our survey. That led to under-
representation of Nordic countries in particular. Our sur-
vey, therefore, may be representative of only 15% of the
landmass of Europe (9,938,000 km2). Combined, France,
Germany, Hungary, Poland, and Lithuania are 1,502,773
km2. These countries contain 6 of the 11 European bio-
geographic regions (55%). The Mediterranean, however,
was underrepresented, and a large proportion of these
5 countries lie within the Atlantic and Continental bio-
geographic region only. Despite these biases, our survey
covered wide geographic area, with at least one scheme
from each country in Europe, and it sufficiently covered
postcommunistic (Poland, Hungary, Lithuania) and west-
ern EU countries (Germany, France) and thus different
monitoring histories (H. Kobierska et al., unpublished
data).
Taxonomic Bias
We focused on species groups for which we could ob-
tain a significant number of descriptions of monitoring
Conservation Biology
Volume 23, No. 2, 2009

Page 7

Schmeller et al.
313
Table 2. Median and quartiles∗(in parentheses) of variables of monitoring practices by species group and country examined in our study.
Amphibians
and reptilesVariableBirds Butterflies Other insectsPlants Mammals
Year200019921999 200120002000
(1993–2002)
7 (1–21)
10 (0–100)
2 (1–5)
16 (3–103)
80.1 (50–98)
3 (1–5)
1 (1)
10 (2–120)
1 (1–5)
109 (16–456.6)
2056.6
(401.3–10401)
14.6 (3.1–94.3)
374.2
(39–1757.4)
(1984–2001)
12 (1–70)
11 (0–80)
4 (2–10)
23 (6–104)
71.4 (0–95)
2 (1–5)
1 (1)
23 (3–200)
1 (1–3)
150 (31–640)
2572.2
(811.2–14181.1)
17.4 (2.7–72)
331.9
(42.7–2027.2)
(1992–2003)
45.5 (2–148)
1.5 (0–91)
3 (2–9)
15.5 (3–150)
50 (0–93.5)
3 (1–20)
1 (1)
50 (5–500)
1 (1–5)
122.3 (45–550)
4665.8
(926.6–8162.2)
10.7 (2.3–45)
223.0
(39.2–867.6)
(1995–2003)
75.5 (32–400)
2.5 (0–15)
5 (2–9)
9.5 (3–40)
50 (0–71.8)
2 (1–6)
1 (1–2)
38 (7–100)
1 (1–10)
100 (35–251)
2791.8
(989.3–6550.8)
1.3 (0.4–3.5)
25.5
(9.8–108.7)
(1995–2003)
6 (1–71)
0 (0–2)
3 (2–4)
3.5 (2–23)
0 (0–36.7)
1 (1–2)
1 (1–2)
10.5 (3–150)
1 (1–2)
45 (8–100)
1886.9
(433.7–7620.4)
3.1 (0.7–11.1)
107.1
(21.6–449.8)
(1990–2003)
3 (1–12)
1 (0–10)
3 (1–10)
10 (3–21)
33.3 (0–76.9)
1 (1–2)
1 (1)
10 (2–55)
1 (1–4)
55 (15–150)
1215.9
(364.3–7886)
12 (4.5–38)
374.7
(72.8–1517.3)
Nspecies
Nvolunteers
Nprof
Npersons
% vol
Nvisits
Ann. freq.
Nsites
Nsamples
Person-days
Costs
Person-daysspec
Costsspec
France GermanyPolandLithuania Hungary
Year19932000 2000 20052001
(1985–2002)
6.5 (1–35)
5 (0–32)
2 (1–5)
12 (3–66)
66.7 (0–93.7)
3 (1–10)
1 (1)
22 (3–100)
1 (1–5)
68 (19–287)
2200.8
(735–9402)
13 (1.3–89)
373.6
(46.7–2012)
(1993–2003)
9 (1–70)
10 (0–30)
2 (1–3)
12 (3–31)
83.3 (0–96.5)
1 (1–5)
1 (1)
19 (4–235)
1 (1)
214 (60–412)
9494
(1912–20420)
46.3 (5–100)
412.1
(56.9–9524)
(1993–2003)
4 (1–20)
0 (0–3)
3 (1–8)
5 (2–20)
0 (0–50)
1 (1–2)
1 (1)
5 (1–50)
1 (1)
40 (14.7–150)
769.4
(246–2889)
7.5 (2–24)
139.6
(31–447)
(1994–2006)
5 (1–15)
0 (0)
3.5 (1–6)
4.5 (1.5–6)
0 (0)
1.5 (1–3)
1.5 (1–3)
15.5 (4–27)
1 (1–12)
16 (10–80)
381.8
(193–2028)
7 (2–20)
135.2
(43–510)
(1999–2003)
5 (1–35)
10 (0–15)
8.5 (4–12)
20 (10–60)
56.3 (0–75)
3 (2–6)
1 (1)
18.5 (8–96)
3.5 (1–15)
165 (108–313)
6178.7
(2000–14618)
25 (3.5–115)
815.3
(180–2873)
Nspecies
Nvolunteers
Nprof
Npersons
% vol
Nvisits
Ann. freq.
Nsites
Nsamples
Person-days
Costs
Person-daysspec
Costsspec
∗When the upper and lower quartile were the same, only one number is indicated. Variable abbreviations are defined in Table 1.
programs. Hence, the practices we could examine that
were included in the EuMon database may not be rep-
resentative of monitoring in other species groups, given
that we discovered significant differences among taxo-
nomic groups in several characteristics of monitoring
schemes. Monitoring of fishes, lichens, and fungi was
strongly underrepresented, whereas, for example, birds
and mammals showed small biases (relative to the num-
ber of publications listed in the Zoological Record).
Extrapolation of our conclusions to underrepresented
groups may not be warranted. Generally, however, the
databases on publication records may also be biased be-
cause in different countries different traditions of pub-
lishing exist (e.g., Gogolin et al. 2003). Overall, the Eu-
Mon database seemed biased toward more well-liked
species groups—or a large percentage of them are not
represented by publications accessible through Google
Scholar, as suggested by our database search. The bias
within each species group analyzed here might be con-
sidered low and the representativeness of our results for
these species groups high.
Monitoring Costs, Volunteer Involvement, and Sampling
Effort
The 395 documented monitoring schemes in the EuMon
database represent a total cost of about €4 million/year.
The 148,690 person-days spent per year to monitor Eu-
rope’s biodiversity adds up to roughly €13 million (av-
erage salary by country in 2005; The World Bank 2006).
Thus, if no volunteers were involved, the costs would be
increased 3-fold. Generally, costs in France and Germany
were much higher compared with Poland and Lithua-
nia because of considerably higher salary-related costs
of professionals and higher equipment costs in West-
ern compared with Eastern Europe. Hungary, despite a
Conservation Biology
Volume 23, No. 2, 2009

Page 8

314
Volunteer-Based Biodiversity Monitoring
Lithuania
Year
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Ann.Freq
Nsites
Nsamples
Persondays
Costs
Persondaysspec
Costsspec
Eigenvalues
First: 0.26
Second: 0.2
Hungary
Year
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Ann.Freq
Nsites
Nsamples
Persondays
Costs
Persondaysspec
Costsspec
Eigenvalues
First: 0.29
Second: 0.22
Poland
Year
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Nsamples
Ann.Freq
Nsites
Persondays
Costs
Persondaysspec
Costsspec
Eigenvalues
First: 0.3
Second: 0.21
France
Year
Nspecies
Nvolunteers
Nprof
Costs
Persondaysspec
Costsspec
Npersons
%vol
Nvisits
Ann.Freq
Nsites
Nsamples
Persondays
Eigenvalues
First: 0.29
Second: 0.19
Germany
Year
Nspecies
Nvolunteers
Nprof
Npersons
%vol
Nvisits
Ann.Freq
Nsamples
Nsites
Persondays
Costs
Persondaysspec
Costsspec
Eigenvalues
First: 0.34
Second: 0.17
Figure 3. Results of the principal component analysis of species monitoring practices in different countries.
Abbreviations are defined in Table 1 (x-axis is axis 1; y-axis is axis 2).
Conservation Biology
Volume 23, No. 2, 2009

Page 9

Schmeller et al.
315
large proportion of volunteers, also had a large number
of professionals per scheme, which explains the consid-
erably higher costs in comparison with the other eastern
European countries Poland and Lithuania. One reason
could be the high biodiversity in the Pannonian region,
whichwouldexplainthelargeaveragenumberofspecies
monitored (the highest for the 5 countries assessed).
Hence, sampling effort per monitoring scheme in Hun-
gary seemed to be especially high, whereas the variable
person-days per species was comparable to France and
Germany. The case of Hungary may also give a good indi-
cation of costs and resource needs in other biodiversity-
rich regions, such as the Mediterranean biogeographic
region, which is so far insufficiently covered in the Eu-
Mon database.
Volunteer recruitment and the value of volunteers also
had a geographic component. For example, in Lithua-
nia and Poland, a majority of monitoring schemes were
professional, whereas this proportion was much smaller
in Hungary, France, and Germany. Our findings suggest
that it may be more difficult to attract volunteers, that
volunteers are not sought to participate in biodiversity
monitoring, or that the lack of volunteer participation
may be linked to unclear political and economic con-
ditions in some countries (H. Kobierska et al., unpub-
lished data). Furthermore, our findings may also suggest
that in countries with high salary costs (i.e., France and
Germany in our survey) biodiversity-monitoring obliga-
tions can only be implemented if volunteer involvement
is maintained or increased for some taxonomic groups
through targeted recruitment campaigns or if funds are
provided for the coordination of volunteer-based mon-
itoring programs. Nongovernmental organizations with
successful volunteer involvement could serve as partners
for such campaigns. For EU countries, the higher costs of
biodiversity monitoring are related to a high number of
endemicspecies,asindicatedinouranalysisforHungary,
and this should be accounted for in the development of
instrumentsforfiscaltransfers,astheyexist,forexample,
in Brazil (Young 2005; Ring 2008).
Data Quality
By definition the standard error of a species’ status es-
timate is negatively related to the total number of sam-
pling units monitored per year (i.e., sites, species) and
hence the overall sampling effort (e.g., Hochachka et al.
2000). We found a strong positive relationship between
number of observers and number of sites monitored (an
index for the sampling effort), number of visits per site
(an index for measurement precision), and number of
species monitored (a measure of the biological cover-
age). Hence, number of people (to a large, but varying
degree volunteers) involved in a monitoring scheme di-
rectly affected sampling effort of a scheme. Large vol-
unteer sampling efforts should counterbalance hypothe-
sized higher measurement errors in data collected by vol-
unteers (Hochachka et al. 2000). Therefore, all else being
equal, a higher sampling effort should yield a more pre-
ciseestimateofaspecies’statusandimprovethechances
of detecting a statistically verifiable status change (e.g.,
a change in species abundance). Hence, we believe that
with state-of-the-art survey designs or analysis methods,
volunteer-based schemes can provide relatively reliable
data and thus yield unbiased results. Quality of volunteer-
ing data is more likely determined by survey design, ana-
lyticalmethodology,andcommunicationskillswithinthe
schemes rather than by volunteer involvement per se.
Conclusion
Our results show that volunteer involvement is a good
trade-off solution and extremely important and valuable
for biodiversity monitoring because data from participa-
tory monitoring networks are not less informative, and
may be more informative, than those collected in pro-
fessional schemes. For effective biodiversity monitoring,
however, it is imperative to guarantee intra- as well as
inter nation comparability, data coherence, and harmo-
nization across large geographic regions, such as Europe
(e.g., Henry et al. 2008; Lengyel et al. 2008). Hence, re-
searchers should analyze the extent of inaccuracy of data
provided by volunteers compared with professionals and
should develop methods and protocols that help detect
and account for inaccuracies.
Acknowledgments
The EuMon project was funded within the 6th Frame-
work of the European Commission. We thank K. Zaun-
berger and G. Torta for their input from the political side,
A. Schmidt and S. Rattei for their excellent financial ad-
ministration, our colleagues for all the discussions, and B.
Sinervo and A. Chaine for language revision. We thank all
the contributors to our survey who spent their precious
timeansweringourquestions.Thedatasetweusedinthis
study is still growing as we add new data to the on line
database from monitoring coordinators across Europe.
Individuals, like you, may also add data. This database
is publicly available at http://eumon.ckff.si/monitoring/
search.php.
Supporting Information
The on line questionnaire of the EuMon survey of biodi-
versity monitoring (Appendix S1), pairwise relationships
between all variables describing monitoring practices
(Appendix S2), and relative geographic and taxonomic
bias of species monitoring schemes (Appendix S3) are
Conservation Biology
Volume 23, No. 2, 2009

Page 10

316
Volunteer-Based Biodiversity Monitoring
available as part of the on line article. The author is
responsible for the content and functionality of these
materials. Queries (other than absence of the material)
should be directed to the corresponding author.
Literature Cited
Balmford, A., R. Green, and M. Jenkins. 2003. Measuring the changing
state of nature. Trends in Ecology & Evolution 18:326–330.
Balmford, A., P. Crane, A. Dobson, R. E. Green, and G. M. Mace. 2005.
The 2010 challenge: data availability, information needs and ex-
traterrestrial insights. Philosophical Transactions of the Royal Soci-
ety of London Series B Biological Sciences 360:221–228.
Bell, S., M. Marzano, J. Cent, H. Kobierska, D. Podjed, D. Vandzinskaite,
H. Reinert, A. Armaitiene, M. Grodzi´ nska-Jurczak, and R. Murˇ siˇ c.
2008.Whatcounts?Volunteersandtheirorganisationsintherecord-
ing and monitoring of biodiversity. Biodiversity and Conservation
13: in press.
Brashares, J. S., and M. K. Sam. 2005. How much is enough? Estimating
the minimum sampling required for effective monitoring of African
reserves. Biodiversity and Conservation 14:2709–2722.
Donald, P. F., F. J. Sanderson, I. J. Burfield, S. M. Bierman, R. D. Gregory,
and Z. Waliczky. 2007. International conservation policy delivers
benefits for birds in Europe. Science 317:810.
Dray, S., and A. B. Dufour. 2007. The ADE4 package: implementing
the duality diagram for ecologists. Journal of Statistical Software
22:1–20.
Engel, S. R., and J. R. Voshell Jr. 2002. Volunteer biological monitor-
ing: can it accurately assess the ecological condition of streams?
American Entomologist 48:164–177.
Evans, D. 2005. Natura. 2000 completing the EU’s network of sites to
conserve flora and fauna. PlantTalk 39:22–27.
Genet, K. S., and L. G. Sargent. 2003. Evaluation of methods and data
quality from a volunteer-based amphibian call survey. Wildlife Soci-
ety Bulletin 31:703–714.
Gogolin, I., P. Smeyers, Garcia Del Dujo, and D. Rusch-Feja. 2003. Eu-
ropean Social Science Citation Index: a chance for promoting Euro-
pean research? European Educational Research Journal 2:574–593.
Gregory, R. D., A. van Strien, P. Vorisek, A. W. G. Meyling, D. G. Noble,
R. P. B. Foppen, and D. W. Gibbons. 2005. Developing indicators
for European birds. Philosophical Transactions of the Royal Society
B—Biological Sciences 360:269–288.
Henry, P.-Y., S. Lengyel, P. Nowicki, R. Julliard, J. Clobert, T.ˇ Celik,
B. Gruber, D. S. Schmeller, V. Babij & K. Henle. 2008. Integrating
ongoing biodiversity monitoring: potential benefits and methods.
Biodiversity and Conservation 13: in press.
Hochachka, W. M., K. Martin, F. Doyle, and C. J. Krebs. 2000. Mon-
itoring vertebrate populations using observational data. Canadian
Journal of Zoology 78:521–529.
Lengyel, S., A. Kobler, L. Kutnar, E. Framstad, P.-Y. Henry, V. Babij, B.
Gruber, D. S. Schmeller, and K. Henle. 2008. A review and a frame-
work for the integration of biodiversity monitoring at the habitat
level. Biodiversity and Conservation 13: in press.
Link, W. A., and J. R. Sauer. 1998. Estimating population change from
count data: application to the North American breeding bird survey.
Ecological Applications 8:258–268.
RDevelopmentCoreTeam.2008.R:alanguageandenvironmentforsta-
tistical computing. R foundation for statistical computing. Vienna.
Available from http://www.R-project.org (accessed June 2008).
Ring, I. 2008. Biodiversity governance: adjusting local costs and global
benefits. Pages 107–126 in press in T. Sikor, editor. Public and
private natural resource governance: a false dichotomy? Earthscan,
London.
Schmeller, D. S., B. Gruber, B. Bauch, and K. Henle. 2006. EuMon–
Arten- und Lebensraum-monitoring in Europa. Naturschutz und
Landschaftsplanung 39:384–385.
Swaay, C. A. M. v., D. Maes, and D. Plate. 1997. Monitoring butterflies
in the Netherlands and Flanders: the first results. Journal of Insect
Conservation 1:81–87.
Thomas, J. A. 2005. Monitoring change in the abundance and distribu-
tion of insects using butterflies and other indicator groups. Philo-
sophical Transactions of the Royal Society of London Series B—
Biological Sciences 360:339–357.
Vandenbosch, R. 2003. Fluctuations of Vanessa cardui butterfly abun-
dance with El Nino and Pacific Decadal Oscillation climatic vari-
ables. Global Change Biology 9:785–790.
World Bank. 2006. Key development data and statistics. World
development indicators. World Bank, New York. Available
from http://go.worldbank.org/QKRICC4WI0 (accessed December
2007).
Young, C. E. F. 2005. Financial mechanisms for conservation in Brazil.
Conservation Biology 19:756–761.
Conservation Biology
Volume 23, No. 2, 2009