Biological Conservation 253 (2021) 108912

Contents lists available at ScienceDirect

Biological Conservation
journal homepage: www.elsevier.com/locate/biocon

Short communication

Conservation birding: A quantitative conceptual framework for prioritizing

citizen science observations
Corey T. Callaghan a, b, *, James E.M. Watson c, d, Mitchell B. Lyons a, William K. Cornwell a,
Richard A. Fuller c
a
Centre for Ecosystem Science, School of Biological, Earth and Environmental Sciences, UNSW Sydney, Sydney, 2052, New South Wales, Australia
b
Community Ecology & Conservation Research Group, Faculty of Environmental Sciences, Czech Univ. of Life Sciences Prague, Prague, Czech Republic
c
School of Biological Sciences, The University of Queensland, St. Lucia, 4072, Queensland, Australia
d
Wildlife Conservation Society, Global Conservation Program, Bronx, 10460, NY, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Despite impressive growth in global biodiversity data, knowledge about the occurrence of species in many parts
Biodiversity data of the world remains incomplete because of major gaps in the underlying data. This can lead to ill-informed
Big data conservation decisions. The collective effort of citizen scientists can generate a great deal of data quickly, but
Citizen science
how might we prioritize the powerful — but finite — effort? We argue that instead of simply filling empty spots
Community science
Spatial and temporal gaps
on the map based solely on where biodiversity information is incomplete, near-term threats to the integrity or
Biodiversity sampling persistence of biodiversity assemblages could also be incorporated to prioritize citizen science sampling. Here we
Species richness develop a quantitative framework illustrating how citizen science sampling and initiatives can be prioritized
when simultaneously considering both the completeness of biodiversity sampling and the risk of habitat con­
version. We illustrate this framework for birds using global citizen science data from the eBird platform and a
global model of the risk of habitat conversion. We find that regions in Africa and southeast Asia would rank as the
highest priorities for new and expanded citizen science initiatives. Our framework provides a mechanism to
quantify where new biodiversity data are most urgently needed, ultimately helping to improve environmental
decision-making. We anticipate this framework can be used in the future at a suite of relevant planning scales,
ranging from local to regional to global.

1. Introduction et al., 2019; La Sorte and Somveille, 2020). In such regions, biodiversity
data are inadequate to support conservation decision-making. Our un­
Biodiversity loss is accelerating globally (Butchart et al., 2010; derstanding of biodiversity is typically biased taxonomically (e.g., more
IPBES, 2019), and targeted monitoring is crucial for implementing observations for charismatic fauna and flora), temporally (more data in
effective conservation (Yoccoz et al., 2001; Groves et al., 2002). Yet for recent decades coinciding with the rise of online citizen science pro­
monitoring to be effectively integrated into a conservation strategy, a grams), and spatially (e.g., more observations in more populated re­
clear understanding of the distribution and abundance of species needs gions). Spatially, our understanding of biodiversity is biased towards
to be established. At a global scale, there has been a tremendous increase sites in two categories: first, accessible areas (i.e., sites that are acces­
in biodiversity monitoring, resulting in large open-access biodiversity sible to either professional or citizen scientists), and second, “hotspots”
databases such as the Global Biodiversity Information Facility (GBIF) — sites already known to support high levels of biodiversity. In many
which now hosts more than 1.5 billion biodiversity records. cases those hotspots are already protected areas (Boakes et al., 2010;
Despite the impressive volume of global biodiversity data, estimates Martin et al., 2012). These biases leave many parts of the world poorly
of biodiversity in many parts of the world remain at best imprecise, and sampled, and specifically, we are often unsure of total species richness
at worst non-existent because of a lack of underlying data (Stork, 1993; and the number and identity of threatened species in these parts of the
Boakes et al., 2010; Scheffers et al., 2012; Essl et al., 2013; Cornwell world. This fundamental lack of understanding of which species exist

* Corresponding author at: Centre for Ecosystem Science, School of Biological, Earth and Environmental Sciences, UNSW Sydney, Sydney, 2052, New South Wales,
Australia.
E-mail address: [email protected] (C.T. Callaghan).

https://doi.org/10.1016/j.biocon.2020.108912
Received 1 July 2020; Received in revised form 23 November 2020; Accepted 2 December 2020
Available online 13 December 2020
0006-3207/© 2020 Elsevier Ltd. All rights reserved.
C.T. Callaghan et al. Biological Conservation 253 (2021) 108912

where is highlighted by modern day ‘rediscoveries’ (e.g., Nguyen et al., biodiversity of that patch based on already-submitted citizen science
2019; Scheffers et al., 2011). observations; referred to as ‘survey completeness’ (i.e., a value, such as
An incomplete understanding of biodiversity estimates can lead to percentage probability, representing our confidence of how much we
rash development decisions, allowing development to proceed unin­ know about the biodiversity in that patch; see details in the following
formed by an understanding of biodiversity distributions in that given section). Assume Patch A has 5000 citizen science observations and that
region. For example, the southern subspecies of Black-throated Finch survey completeness has been estimated at 60%. In contrast, assume
Poephila cincta cincta is a critically endangered bird in eastern Australia, Patch B has only 2000 citizen science observations with an estimated
and development activities — ranging from mining to urbanisation to survey completeness of 20%. Most practitioners and biologists would
agriculture — throughout its range are ongoing at least in part because intuitively encourage citizen scientists to submit samples from Patch B
there is a lack of fundamental understanding of the distribution of as opposed to Patch A because we know less about the biodiversity of
biodiversity in much of the region (Ramesh et al. 2017; Reside et al., that area. But what if we knew that Patch A has a 90% chance of being
2019). Unfortunately, however, the current pace of funding for profes­ cleared to make way for a development, while Patch B has only a 5%
sional conservation and ecology is not compatible with the increasing chance of being developed? From a conservation perspective, it is
need for robust biodiversity monitoring (Bakker et al., 2010). arguably more beneficial to allocate new citizen science sampling effort
So how do we combat the pervasive lack of complete species’ in­ to Patch A providing conservationists with the necessary data to argue
ventories? Citizen science — scientific research conducted in whole or in against inappropriate development, or design appropriate conservation
part by people for whom science is not their profession — has clear mitigation and offset strategies. In other words, because of the low risk
potential to deliver comprehensive monitoring of biodiversity in both of development, Patch B can ‘wait’ until we more fully understand the
space and time (Tulloch et al., 2013a; Danielsen et al., 2014; Chandler biodiversity of Patch A.
et al., 2017a, 2017b; McKinley et al., 2017). Indeed, data from citizen
science projects have shifted our ability to understand spatiotemporal 3. Illustrating the framework
dynamics of many species’ populations (e.g., Schuster et al., 2019).
Nevertheless, these data are still far from complete: there are still many We can generalize this example by defining two axes: (1) survey
spatial and temporal biases in these datasets (Boakes et al., 2010; La completeness — i.e., how well-sampled the biodiversity is in an area,
Sorte and Somveille, 2020), reflecting a general bias in global under­ and (2) risk of habitat conversion — i.e., development pressure in an
standing of biodiversity towards areas with high human population area. As illustrated in our conceptual example, as risk of habitat con­
density and hotspots of both ecotourism and scientific activity (Boakes version in an area increases, so should the importance of improving our
et al., 2010; Cornwell et al., 2019). Many researchers and/or practi­ understanding of the biodiversity present in that area, especially where
tioners have suggested that to ensure more robust estimates of biodi­ survey completeness is low (Fig. 1).
versity, citizen scientists could serve to ‘fill the gaps’ in biodiversity To illustrate an example of how this conceptual framework could be
datasets in regions that are not well surveyed (e.g., Tulloch et al., 2013a; used, we use a global citizen science project — eBird — and a global
Chandler et al., 2017a, 2017b; McKinley et al., 2017). Exactly how these model of risk of habitat conversion. We used eBird to illustrate our
gaps are filled, however, is still up for debate. Previous studies, for framework because it is a globally relevant, successful, citizen science
example, have used species distribution models to target the places with project and the data are openly available. Further, survey completeness
the highest probability of finding target species (e.g., Udyawer et al., has been previously quantified using eBird data (La Sorte and Somveille,
2020); designed prioritization approaches for bird surveys with the 2020). As a result of relying on eBird to illustrate our framework, our
highest benefits to specific project goals (e.g., Tulloch et al., 2013b); analysis is restricted to birds, but we highlight two key points: (1) birds
used Value of Information from experts to highlight the critical knowl­ are often considered an effective surrogate to represent biodiversity
edge gaps to be filled (e.g., Nicol et al., 2018); and quantified where data more broadly (Gregory et al., 2003; Larsen et al., 2012); and (2) our
are most important for informing conservation decisions using various framework can easily be adapted to other taxa such as amphibians,
mathematical models (e.g., Kujala et al., 2018). mammals, plants, and insects for which there are citizen science datasets
Crucially, to be of maximum conservation utility, this strategy of available.
gap-filling could be further targeted towards areas with known or pro­
jected threats to the integrity or persistence of an assemblage; particu­ 3.1. Survey completeness
larly important given the lack of adequate resources available for
biodiversity monitoring. Here we produce a conceptual quantitative eBird is arguably one of the most successful biodiversity citizen sci­
framework that can be used to prioritize citizen science observations ence projects in the world. In May 2019 alone, eBird averaged 7.5 ob­
based on both the biodiversity survey completeness of a region and the servations per second. With >800 million observations globally of
level of threats in a region. This framework is intended for citizen sci­ >99% of the world’s known avifauna, there is clearly the potential to use
ence practitioners, environmental managers hoping to leverage citizen these data to fundamentally enhance our understanding of conservation
science data, and those involved in various citizen science projects. We need at a global scale, and these data are continuously available to
provide an example of how this framework could be operationalized, decision-makers and planners around the world through an open-access
conducting a global analysis of where new citizen science observations portal. We used a recently published assessment of eBird survey
of birds are most likely to have the greatest return on understanding completeness (La Sorte and Somveille, 2020), based on data submitted
biodiversity in areas at risk of loss through development. We call this between 2012 and 2018. La Sorte and Somveille (2020) estimated sur­
“conservation birding”. vey completeness by modelling the relationship between the number of
species and sampling effort to develop a species accumulation curve
2. A conceptual example describing the relationship between the accumulated number of species
and survey effort (for full details see La Sorte and Somveille, 2020 and
Choosing where and how to implement effective conservation stra­ Lobo et al., 2018). We used the equal-area hexagonal cells (49,811 km2)
tegies fundamentally requires information about which species occur from La Sorte and Somveille (2020) to derive an average completeness
where, combined with information about threats to the integrity or score — rounded to the nearest integer between 0 and 100 — across all
persistence of an assemblage. Consider two different, equally sized, months. We restricted our analysis to those hexagonal cells that are
habitat patches equidistant from a large city in which reside dedicated terrestrial (Fig. S1).
birdwatchers who could choose to allocate their surveying to either
patch. For both patches, we can estimate how well we understand the

2
C.T. Callaghan et al. Biological Conservation 253 (2021) 108912

Fig. 1. Conceptual framework illustrating the basis for

prioritizing citizen science sampling based on existing
sample completeness (x-axis) and the risk of habitat con­
version (y-axis). The priority of a citizen science observa­
tion becomes increasingly important moving from the
bottom right-hand corner of the space to the top left-hand
corner. The red line represents the 1:1 slope, and the dis­
tance of a site below or above this line indexes its sampling
priority from a conservation perspective. (For interpretation
of the references to color in this figure legend, the reader is
referred to the web version of this article.)

3.2. Risk of land conversion Fig. 1), leaving us with each hexagonal cell assigned a sampling priority
score ranging from 0 (lowest sampling priority) to 1 (highest sampling
We used a recently published global model of the risk of habitat priority).
conversion in 30 km2 planning units (Allan et al., 2019). This map used
spatially explicit data on future land-use scenarios from the Land Use 4. Results & discussion
Harmonisation Dataset v2 (Hurtt et al., 2020) and projections under
socioeconomic pathway 3 (RCP7.0; AIM), a business-as-usual pessi­ Our analyses revealed that large swathes of terrestrial Earth (53% of
mistic scenario where land use change is poorly regulated. The hexagonal cells) had virtually no recorded eBird data (see La Sorte and
harmonised land-use data contains 12 state layers for the years 2015 and Somveille, 2020). Many of these same areas are at risk of conversion to
2030, and we used the difference between these to estimate the pro­ more intensive human uses (Fig. 2), with 16% of terrestrial Earth being
portion of intact habitat projected to be converted to human uses by at high risk of conversion (Fig. S2). Our results indicate strong potential
2030 in each 30 km2 planning unit. If a planning unit had >50% to prioritize where citizen scientists could contribute future biodiversity
probability of conversion to human uses, it was defined as being ‘at risk’ sampling (e.g., Figs. 1, 2a), where those cells with high risk of habitat
of conversion (see Allan et al., 2019). This thresholding approach rec­ conversion receive the greatest allocation of effort. While our current
ognizes the considerable uncertainty in land-use projections given that illustration is limited to eBird only, we highlight that different regions
our aim was to derive broad estimates of the risk of habitat conversion have different bird monitoring schemes that may not have data available
for a given region, rather than calculating exact amounts of habitat lost. in eBird, and our work is currently only applicable to birds. More work is
We were left with a binomial representation of whether a planning unit necessary to generalize these findings to other taxa. But importantly, our
was associated with the potential for conversion to human uses. For our framework can easily be adapted to focus on or incorporate other taxa.
analysis, we spatially aggregated the values (i.e., at risk = 1, or not at Regardless, our framework offers one way to differentiate the conser­
risk = 0) of risk of habitat conversion (at 30 km2 planning units) to each vation value of additional citizen science sampling efforts: biodiversity
hexagonal cell by taking the mean value of the binomial planning units conservation strategies could aim to incorporate information about
within a hexagonal cell. threats to the integrity or persistence of an assemblage.
Our results showed that some of the highest priority regions in the
3.3. Assigning sampling priority world for biodiversity sampling — and thus developing citizen science
infrastructure — are in Africa, parts of central and southeast Asia,
We calculated a sampling priority score for each hexagonal cell in the Mongolia, parts of the Middle East, and parts of Brazil (Fig. 2b; inter­
two-dimensional space given by our two axes (survey completeness and active version here). Across the world there was high variability in the
risk of habitat conversion; Fig. 1). This was done by plotting sampling priority, with different cells in our analysis in markedly
completeness vs. risk for each hexagonal cell and calculating the different regions of our conceptual space (e.g., Fig. 2a), and this was true
Euclidean distance to the 1:1 slope line, whereby the furthest distance even within specific regions (e.g., Australia). Areas with a high risk of
was in the top left and top right corners. Based on our conceptual figure, habitat conversion and low average completeness (e.g., regions in
any value that fell above the 1:1 slope line was assigned a positive value Guinea, Congo, Brazil, Central African Republic) are where biodiversity
(i.e., highly important regions), whereas any value that fell below the data are most urgently needed. Conversely, some regions have relatively
1:1 slope line was assigned a negative value (i.e., less important re­ low sampling priority either because the risk of habitat conversion is
gions). These values were then scaled from 0 to 1 after adding the ab­ low, or their biodiversity is relatively well known (e.g., relatively less-
solute value of the minimum value (i.e., those in the bottom right of diverse regions such as the boreal in Canada, or part of Europe). These

3
C.T. Callaghan et al. Biological Conservation 253 (2021) 108912

Fig. 2. a) Real-world data placed into our conceptual space, from eBird average completeness (x-axis) and the risk of habitat conversion (y-axis). As illustrated in
Fig. 1, cells in the top left of our conceptual space have the highest sampling priority and cells in the bottom right have the lowest sampling priority. b) The sampling
priority of a cell mapped across the world (see here for an interactive version).

results highlight the differences among regions in their current under­ need to be considered, the increase in global ecotourism holds much
standing of biodiversity relative to the current projections in risk of potential to increase biodiversity sampling in high priority regions (see
habitat conversion until the year 2030. examples in Callaghan et al., 2020). Such international initiatives could
The ranking of sampling priority regions, at a global scale, could be be coupled with local-scale investment and collaboration since this is a
used as a basis for incentivizing and expanding coverage by both in­ proven technique for biodiversity monitoring in traditionally data-poor
ternational and local citizen science programs and as a means to move regions (Ortega-Álvarez et al., 2012). For example, community-based
citizen science beyond the ‘western world’ through a combination of monitoring has successfully helped monitor tropical forest diversity
participatory, contributory, and field-based projects (Pocock et al., (Chandler et al., 2017a, 2017b) and large carnivores in remote montane
2019). While the potential downsides of excessive carbon footprints landscapes (Farhadinia et al., 2018). Further, some global citizen

4
C.T. Callaghan et al. Biological Conservation 253 (2021) 108912

science projects (e.g., eBird and iNaturalist) have proved successful on a prioritize citizen science observations (e.g., development pressure at
global scale by providing local-scale community curation with infra­ regional or state scales) at more localized scales; (2) extending these
structure and initiatives (e.g., websites in the relevant languages). As an analyses to biodiversity more broadly, for example through the use of
example, Taiwan, despite its relatively small geographic area, is GBIF data; and (3) developing predictive frameworks which are
currently ranked number eight in terms of eBird submissions (htt continuously updated as citizen science observations are submitted (e.
ps://ebird.org/home) due in large part to a local portal which pro­ g., Callaghan et al., 2019a).
vides relevant news to Taiwanese birders (https://ebird.org/taiwan/h A robust understanding of biodiversity is essential to protect against
ome). As another example, iNaturalist has truly global reach due in development of natural areas and effectively plan conservation strate­
part to their concerted effort to highlight global biodiversity through gies. Biodiversity monitoring is likely to rely on citizen science data, at
initiatives such as the iNaturalist World Tour (https://inaturalist.github. least in part, in the future (Danielsen et al., 2014; Chandler et al., 2017a,
io/internationals_all.html). Importantly, there is not a ‘one-size-fits-all’ 2017b; McKinley et al., 2017). We argue that citizen science projects
recipe to engage with local citizen scientists, due to many different — could focus on the highest priority observations, by encouraging or
and complicated — issues; such as access and literacy surrounding incentivizing participants to sample in the most meaningful way (e.g.,
smartphones (e.g., Leibenberg et al., 2017; Pejovic and Skarlatidou, Callaghan et al., 2019b). Biodiversity monitoring will continue to rely
2020). To truly reach a global vision for citizen science (Pocock et al., on a diverse set of end-users and contributors (Bayraktarov et al., 2019),
2018), many different types of projects will be necessary (Pocock et al., ideally working together to further our understanding of biodiversity,
2019), and for example different recruitment strategies are likely and what threatens it, in space and time. The capacity to prioritize
necessary to engage with different types of volunteers (Requier et al., where biodiversity data are most urgently needed will provide the
2020). Nevertheless, our results suggest that similar efforts could be fundamental data to improve environmental decision-making.
spearheaded in areas such as Africa or southeast Asia (Pocock et al.,
2018, 2019) — where it might be difficult to acquire local funding — to CRediT authorship contribution statement
increase the biodiversity sampling in these parts of the world in the face
of potential threats. Corey T. Callaghan: Conceptualization, Methodology, Investiga­
Although we currently provide an illustration of our quantitative tion, Writing - Original draft preparation, Visualization. James E. M.
conceptual framework at a global scale, the applicability of the frame­ Watson: Methodology, Data Curation, Writing – Review & Editing.
work will likely be more powerful at a local scale. There is currently Mitchell B. Lyons: Visualization, Methodology, Investigation, Writing –
strong correlation between our sampling priority scores and the Review & Editing. William K. Cornwell: Investigation, Writing – Re­
completeness scores by La Sorte and Somveille (2020), likely as a result view & Editing. Richard A. Fuller: Conceptualization, Investigation,
of the coarseness of our analysis (i.e., 50,000 km2 cells). This is because Visualization, Writing – Review & Editing.
of the heterogeneity of threats and completeness that occurs at a spatial
scale less than the size of the specific cells used here in our case study.
Data availability
We highlight, however, that our conceptual framework can still shift the
sampling priority of some cells more or less, depending on the level of
Data to illustrate our framework were extracted from Allan et al.
threat — even at the large spatial scale used. Nevertheless, there still
2019 and La Sorte and Somveille 2020 as described above. Code and
exists variability among these scores (see Fig. 2a) highlighting regions
aggregated data to reproduce Figures 1 and 2 in this paper can be found
with relatively greater sampling priority (Fig. 2b). In New Zealand, for
here: https://doi.org/10.5281/zenodo.4316543.
example, most cells are unsurprisingly well-sampled, but after ac­
counting for the level of threat, grid cells are more variable in terms of
their sampling priority (see interactive figure here). Citizen science Declaration of competing interest
practitioners and conservationists could operationalize the conceptual
framework we present here at more localized scales; for example, within The authors declare no conflict of interest.
regional or state management units, or even within more localized areas
such as cities, nature reserves, or wetlands where practitioners may be
Acknowledgements
worried about threats to the habitat. At a local scale, it may also be more
likely that practitioners will be able to successfully incentivize and
We thank the >400,000 volunteers who have contributed data to
interact with local citizen scientists, promoting uptake of a prioritization
eBird, and the countless regional volunteers who manage the data
scheme for citizen science sampling (Callaghan et al., 2019a). In addi­
quality of the project. We also thank four anonymous reviewers who
tion, data on development pressures/ecological threats are likely more
provided suggestions which helped us improve this manuscript.
reliable and robust at smaller spatial scales. Importantly, other metrics
could also be integrated into our framework in the future such as the cost
of conservation interventions and mechanisms, or the accessibility of a Appendix A. Supplementary data
site. The accessibility of a site/region is important to consider because
professional monitoring will continue to be necessary in areas that are Supplementary data to this article can be found online at https://doi.
difficult to get to by citizen scientists (e.g., remote parts of Australia), org/10.1016/j.biocon.2020.108912.
and some regions of the world will continue to be difficult to monitor
with citizen science due to socioeconomic limitations (e.g., technology References
or limited education). Importantly, we expect that the correlation be­
Allan, J.R., Watson, J.E.M., Di Marco, M., O’Bryan, C.J., Possingham, H.P., Atkinson, S.
tween sampling priority and average completeness will be lessened at
C., Venter, O., 2019. Hospots f human impact on threatened terrestrial vertebrates.
smaller spatial scales, but this needs to be formally tested. PLoS Biol. 17 (12), e3000598.
The approach presented here is somewhat static: dependent on a Bakker, V.J., Baum, J.K., Brodie, J.F., Salomon, A.K., Dickson, B.G., Gibbs, H.K.,
snapshot of our understanding of survey completeness and risk of Jensen, O.P., McIntyre, P.B., 2010. The changing landscape of conservation science
funding in the United States. Conserv. Lett. 3 (6), 435–444.
habitat conversion. But citizen science observations are dynamic, Bayraktarov, E., Ehmke, G., O’Connor, J., Burns, E.L., Nguyen, H.A., McRae, L.,
constantly updated and are continuously being amassed (Callaghan Possingham, H.P., Lindenmayer, D.B., 2019. Do big unstructured biodiversity data
et al., 2019b). Similarly, the threat of habitat development can also be mean more knowledge? Frontiers in Ecology and the Environment 6.
Boakes, E.H., McGowan, P.J., Fuller, R.A., Chang-qing, D., Clark, N.E., O’Connor, K.,
dynamic, shifting with political regimes for instance. We see the next Mace, G.M., 2010. Distorted views of biodiversity: spatial and temporal bias in
steps in this research space being: (1) aggregating data necessary to species occurrence data. PLoS Biol. 8 (6), e1000385.

5
C.T. Callaghan et al. Biological Conservation 253 (2021) 108912

Butchart, S.H., Walpole, M., Collen, B., Van Strien, A., Scharlemann, J.P., Almond, R.E., Lobo, J.M., Hortal, J., Yela, J.L., Millán, A., Sánchez-Fernández, D., García-Roselló, E.,
Baillie, J.E., Bomhard, B., Brown, C., Bruno, J., et al., 2010. Global biodiversity: González-Dacosta, J., Heine, J., González-Vilas, L., Guisande, C., 2018. Knowbr: an
indicators of recent declines. Science 328 (5982), 1164–1168. application to map the geographical variation of survey effort and identify well-
Callaghan, C.T., Poore, A.G., Major, R.E., Rowley, J.J., Cornwell, W.K., 2019a. surveyed areas from biodiversity databases. Ecol. Indic. 91, 241–248.
Optimizing future biodiversity sampling by citizen scientists. Proc. R. Soc. B 286 Martin, L.J., Blossey, B., Ellis, E., 2012. Mapping where ecologists work: biases in the
(1912), 20191487. global distribution of terrestrial ecological observations. Front. Ecol. Environ. 10 (4),
Callaghan, C.T., Rowley, J.J., Cornwell, W.K., Poore, A.G., Major, R.E., 2019b. 195–201.
Improving big citizen science data: moving beyond haphazard sampling. PLoS Biol. McKinley, D.C., Miller-Rushing, A.J., Ballard, H.L., Bonney, R., Brown, H., Cook-
17 (6), e3000357. Patton, S.C., Evans, D.M., French, R.A., Parrish, J.K., Phillips, T.B., et al., 2017.
Callaghan, C.T., Poore, A.G.B., Mesaglio, T., Moles, A.T., Nakagawa, S., Roberts, C., Citizen science can improve conservation science, natural resource management,
et al., 2020. Three frontiers for the future of biodiversity research using citizen and environmental protection. Biol. Conserv. 208, 15–28.
science data. In press at BioScience. https://doi.org/10.1093/biosci/biaa131. Nguyen, A., Tran, V.B., Hoang, D.M., Nguyen, T.A.M., Nguyen, D.T., Tran, V.T., Long, B.,
Chandler, M., See., L., Andrianandrasana, H. T., Becker, D., Berardi, A., Bodmer, R., Meijaard, E., Holland, J., Wilting, A., Tilker, A., 2019. Camera-trap evidence that the
Brofeldt, S., et al. (2017a). The value and opportunities of community- and citizen- silver-backed chevrotain Tragulus versicolor remains in the wild in Vietnam. Nature
based approaches to tropical forest biodiversity monitoring. In A Sourcebook of Ecology & Evolution 3, 1650–1654.
Methods and Procedures for Monitoring Essential Biodiversity Variables in Tropical Nicol, S., Ward, K., Stratford, D., Joehnk, K.D., Chadès, I., 2018. Making the best use of
Forests With Remote Sensing (pp. 223–281). GOFC-GOLD Land Cover Project Office, experts’ estimates to prioritise monitoring and management actions: a freshwater
Wageningen University, The Netherlands. case study. J. Environ. Manag. 215, 294–304.
Chandler, M., See, L., Copas, K., Bonde, A.M., López, B.C., Danielsen, F., Legind, J.K., Ortega-Álvarez, R., Sánchez-González, L.A., Rodríguez-Contreras, V., Vargas-Canales, V.
Masinde, S., Miller-Rushing, A.J., Newman, G., et al., 2017b. Contribution of citizen M., Puebla-Olivares, F., Berlanga, H., 2012. Birding for and with people: integrating
science towards international biodiversity monitoring. Biol. Conserv. 213, 280–294. local participation in avian monitoring programs within high biodiversity areas in
Cornwell, W.K., Pearse, W.D., Dalrymple, R.L., Zanne, A.E., 2019. What we (don’t) know southern Mexico. Sustainability 4 (12), 1984–1998.
about global plant diversity. Ecography 42 (11), 1819–1831. Pejovic, V., Skarlatidou, A., 2020. Understanding interaction design challenges in mobile
Danielsen, F., Pirhofer-Walzl, K., Adrian, T.P., Kapijimpanga, D.R., Burgess, N.D., extreme citizen science. International Journal of Human-Computer Interaction 36
Jensen, P.M., et al., 2014. Linking public participation in scientific research to the (3), 251–270.
indicators and needs of international environmental agreements. Conserv. Lett. 7 Pocock, M.J.O., Chandler, M., Bonney, R., Thornhill, I., Albin, A., August, T.,
(1), 12–24. Bachman, S., et al., 2018. A vision for global biodiversity monitoring with citizen
Essl, F., Rabitsch, W., Dullinger, S., Moser, D., Milasowszky, N., 2013. How well do we science. Adv. Ecol. Res. 59, 169–223.
know species richness in a well-known continent? Temporal patterns of endemic and Pocock, M.J.O., Roy, H.E., August, T., Kuria, A., Barasa, F., Bett, J., Mwangi, G., Kairo, J.,
widespread species descriptions in the European fauna. Glob. Ecol. Biogeogr. 22 (1), Jimani, J., et al., 2019. Developing the global potential of citizen science: assessing
29–39. opportunities that benefit people, society and the environment in East Africa.
Farhadinia, M.S., Moll, R.J., Montgomery, R.A., Ashrafi, S., Johnson, P.J., Hunter, L.T.B., J. Appl. Ecol. 56 (2), 274–281.
Macdonald, D.W., 2018. Citizen science data facilitate monitoring of rare large Ramesh, V., Gopalakrishna, T., Barve, S., Melnick, D.J., 2017. IUCN greatly
carnivores in remote montane landscapes. Ecol. Indic. 94 (1), 283–291. underestimates threat levels of endemic birds in the Western Ghats. Biol. Conserv.
Gregory, R.D., Noble, D., Field, R., Marchant, J., Raven, M., Gibbons, D.W., 2003. Using 210, 205–221. https://doi.org/10.1016/j.biocon.2017.03.019.
birds as indicators of biodiversity. Ornis Hungarica 12-13, 11–24. Requier, F., Andersson, G.K.S., Oddi, F.J., Garibaldi, L.A., 2020. Citizen science in
Groves, C. R., Jensen, D. B., Valutis, L. L., Redford, K. H., Shaffer, M. L., Scott, J. M., developing countries: how to improve volunteer participation. Front. Ecol. Environ.
Baumgartner, J. V., Higgins, J. V., Beck M. W., and Anderson, M. G. (2002). Planning 18 (2), 101–108.
for biodiversity conservation: putting conservation science into practice: a seven- Reside, A.E., Cosgrove, A.J., Pointon, R., Trezise, J., Watson, J.E.M., Maron, M., 2019.
step framework for developing regional plans to conserve biological diversity, based How to send a finch extinct. Environ. Sci. Pol. 94, 163–173.
upon principles of conservation biology and ecology, is being used extensively by the Scheffers, B.R., Yong, D.L., Harris, J.B.C., Giam, X., Sodhi, N.S., 2011. The world’s
nature conservancy to identify priority areas for conservation. BioScience, 52(6): rediscovered species: back from the brink? PLoS One 6, e22531.
499–512. Scheffers, B.R., Joppa, L.N., Pimm, S.L., Laurance, W.F., 2012. What we know and don’t
Hurtt, G.C., Chini, L., Sahajpal, R., Frolking, S., Bodirsky, B.L., Calvin, K., Doelman, J.C., know about earth’s missing biodiversity. Trends Ecol. Evol. 27 (9), 501–510.
Fisk, J., Fujimori, S., Goldewijk, K.K., et al., 2020. Harmonization of global land-use Schuster, R., Wilson, S., Rodewald, A.D., Arcese, P., Fink, D., Auer, T., Bennett, J.R.,
change and management for the period 850-2100 (LUH2) for CMIP6. Geoscientifc 2019. Optimizing the conservation of migratory species over their full annual cycle.
Model Development Discussions. https://doi.org/10.5194/gmd-2019-360. Nat. Commun. 10, e1754.
IPBES, 2019. Global assessment report on biodiversity and ecosystem services. Available Stork, N.E., 1993. How many species are there? Biodivers. Conserv. 2 (3), 215–232.
at. https://ipbes.net/global-assessment. Accessed May 4th 2020. Tulloch, A.I., Possingham, H.P., Joseph, L.N., Szabo, J., Martin, T.G., 2013a. Realising
Kujala, H., Lahoz-Monfort, J.J., Elith, J., Moilanen, A., 2018. Not all data are equal: the full potential of citizen science monitoring programs. Biol. Conserv. 165,
influence of data type and amount in spatial conservation prioritisation. Methods 128–138.
Ecol. Evol. 9 (11), 2249–2261. Tulloch, A.I.T., Mustin, K., Possingham, H.P., Szabo, J.K., Wilson, K.A., 2013b. To boldly
La Sorte, F.A., Somveille, M., 2020. Survey completeness of a global citizen-science go where no volunteer has gone before: predicting volunteer activity to prioritize
database of bird occurrence. Ecography 43 (1), 34–43. surveys at the landscape scale. Divers. Distrib. 19, 465–480.
Larsen, F.W., Bladt, J., Balmford, A., Rahbek, C., 2012. Birds as biodiversity surrogates: Udyawer, V., Somaweera, R., Nitschke, C., d’Anastasi, B., Sanders, K., Webber, B.L.,
will supplementing birds with other taxa improve effectiveness? J. Appl. Ecol. 49 Hourston, M., Heupel, M.R., 2020. Prioritising search effort to locate previously
(2), 349–356. unknown populations of endangered marine reptiles. Global Ecology and
Leibenberg, L., Steventon, J., Brahman, N., Benadie, K., Minye, J., Langwane, H., Conservation e01013.
Xhukwe, Q., 2017. Smartphone icon user interface design for non-literate trackers Yoccoz, N.G., Nichols, J.D., Boulinier, T., 2001. Monitoring of biological diversity in
and its implications for an inclusive citizen science. Biol. Conserv. 208, 155–162. space and time. Trends Ecol. Evol. 16 (8), 446–453.