Creative citizen science illuminates complex ecological responses to climate change


COMMENTARY

Creative citizen science illuminates complex
ecological responses to climate change
Abraham J. Miller-Rushinga,1, Amanda S. Gallinatb,c, and Richard B. Primackd

Climate change is causing the timing of key behaviors
(i.e., phenology) to shift differently across trophic
levels and among some interacting organisms (e.g.,
plants and pollinators, predators and prey), suggest-
ing that interactions among species are being dis-
rupted (1, 2). Studying the phenology of interactions,
however, is difficult, which has limited researchers’
ability to zero in on changes in specific interactions or
on the consequences of mismatches. In PNAS, Hassall
et al. (3) use a combination of citizen science tech-
niques to investigate the effects of climate change
on dozens of specific interactions. They focus on a Bates-
ian mimicry complex involving stinging bees and wasps,
stingless syrphid flies (also known as hoverflies) that
mimic their appearance, and avian predators. The meth-
ods used by Hassall et al. (3) continue an upsurge of
innovations in climate change ecology research, in
which the role of citizen science is expanding to pro-
vide new approaches to complex challenges.

Citizen Science Has a Long History in Ecology
and Climate Change
Citizen science (i.e., the participation of nonspe-
cialists in scientific research) has yielded important
contributions throughout the history of ecology (4).
Our understanding of ecological and phenological
responses to climate change, in particular, has
benefited from long-term observations collected by
individuals and networks, like those of naturalists
Henry David Thoreau and Robert Marsham (5, 6)
and historical phenology and weather observation
networks (7), and the ongoing insect monitoring
schemes used by Hassall et al. (3). It is this type
of citizen science—enthusiastic volunteers collect-
ing and submitting field observations—that many
people imagine when they think of citizen science in
ecology. In fact, the amount of data being gener-
ated from these types of observations is staggering;
citizen scientists have recorded hundreds of mil-
lions of observations, yielding much of the existing
data describing how flowering, bird migration, and

insect flight times are changing (8–10). These observations,
as many as there are, account for only a small portion
of the growing breadth of projects and ways to engage
citizen scientists (11). For example, citizen scientists are
categorizing, describing, and transcribing digital im-
ages and museum specimens (12); community mem-
bers are working jointly with researchers on all stages of
some research projects (13); and volunteers are de-
veloping new technologies for open research (14).

Why Studying Interactions or Mismatches Is
Hard
The methodology of Hassall et al. (3) gives us a way
to approach notoriously difficult-to-study questions
about species interactions and temporal mismatches.
Ecologists have speculated that phenological mis-
matches may be occurring throughout nature, disrupt-
ing ecological relationships; however, this speculation
has been accompanied by surprisingly little evidence

Fig. 1. The online game Hassall et al. (3) created to identify hymenopteran
model (Right) and syrphid mimic (Left) pairs, in action.

aAcadia National Park, National Park Service, Bar Harbor, ME 04609; bDepartment of Biology, Utah State University, Logan, UT 84322; cEcology
Center, Utah State University, Logan, UT 84322; and dDepartment of Biology, Boston University, Boston, MA 02215
Author contributions: A.J.M.-R., A.S.G., and R.B.P. wrote the paper.
The authors declare no conflict of interest.
Published under the PNAS license.
See companion article 10.1073/pnas.1813367115.
1To whom correspondence should be addressed. Email: abe_miller-rushing@nps.gov.

www.pnas.org/cgi/doi/10.1073/pnas.1820266116 PNAS Latest Articles | 1 of 3

C
O

M
M

E
N
T
A
R
Y

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 

http://crossmark.crossref.org/dialog/?doi=10.1073/pnas.1820266116&domain=pdf&date_stamp=2019-01-04
https://www.pnas.org/site/aboutpnas/licenses.xhtml
mailto:abe_miller-rushing@nps.gov
https://www.pnas.org/cgi/doi/10.1073/pnas.1820266116


(15). Research is limited by the difficulty of examining interactions
in the field and the possibility that mismatches may be uncommon
or quickly corrected by evolutionary processes (15), and many
interactions are hidden or would take too much time to observe.
In cases of mimicry, like those studied by Hassall et al. (3), iden-
tifying model-mimic pairs can require complex measurements and
comparisons of morphology. As a result, testing the influence of
shifts in phenology on the fitness of organisms is rare (but see refs.
16 and 17). In the case of model species, mimics, and predators,
measuring the fitness consequences of phenological change can
require experiments to test predator preference or behavior, but
such experiments are difficult to carry out and might have un-
expected artifacts depending on their design.

New Approaches to Investigating Interactions
In recent decades, new technologies have paved the way for more
indirect approaches—for example, environmental DNA, satellite
tracking, and time-lapse videos or camera traps—to investigate
interactions among species. These techniques enhance our ability
to assess which species are in the same locations at the same
times and to document occurrences of specific interactions that
might get captured on camera. But they leave open questions
about the nature of interactions and the consequences of phe-
nological shifts on interacting species. The approach of Hassall
et al. (3) shows one way to fill these gaps; their unusual method-
ology allowed them to investigate dozens of interactions across
complex trophic relationships. Each step of their approach dem-
onstrates a unique value of citizen science that other researchers
might find useful in their own work.

First, rather than observing model-mimic relationships directly
in the field or using feeding tests with birds, Hassall et al. (3)
created an online game in which citizen scientists could rate the
visual similarities of co-occurring pairs of syrphids and hymenop-
terans based on size, shape, color, hairiness, and other characters
(Fig. 1). Over the course of a year, volunteers (recruited via Twitter)
submitted 30,300 ratings and identified 237 high-fidelity model-
mimic pairs out of 2,352 potential combinations, with results that
correlated strongly with pigeon experiments. This technique
demonstrates the value of citizen science to allow researchers to
assess pairwise combinations faster and in greater numbers than
can be achieved using traditional methods. Variations on this
method could be appropriate for other studies involving mor-
phological comparisons, such as in other model-mimic or cam-
ouflage studies, or studies of specimens collected during bioblitzes
or biodiversity inventories.

Second, Hassall et al. (3) analyzed over a million records of
syrphid fly, bee, and wasp abundance from two long-term citizen
science monitoring projects in the United Kingdom. This exten-
sive dataset showed that hymenoptera emerge earlier, on average,
than syrphid flies, and that the spring emergence of hymenopterans
is advancing faster than that of syrphids in response to warming
temperatures. This study adds to the growing body of research
using citizen science data to document shifts in species phenology
and interactions, including community-level interactions at single
locations (18), possible interactions across trophic levels (1), and
known interactions (2).

Third, Hassall et al. (3) developed a virtual reality game to sim-
ulate model-mimic-predator interactions in different phenological
scenarios. Instead of offering live hymenopterans and syrphids to
birds in complicated feeding experiments, or using data-intensive
and highly parameterized computational algorithms, 45 humans
took on the role of bird predators in a game. Participants were

recruited from the University of Leeds and surrounding areas and
were compensated £5 for their time. They gained points by feeding
on benign mimic syrphids and lost points when they ate stinging
model hymenopterans. Participants had to identify mimics against
complex backgrounds, a simulation of the challenges that bird
predators face in deciding whether or not to feed on an insect. The
results of the game showed that model species (i.e., bees and
wasps) perform best (i.e., had highest survival rates) when they are
active before the syrphid mimics. Mimics, on the other hand, per-
form best when the phenology of model species and mimics
overlaps and occurs randomly from the perspective of predators.

In PNAS, Hassall et al. use a combination of
citizen science techniques to investigate the
effects of climate change on dozens of specific
interactions.

And predators do best when either the model bees and wasps
emerge first or when the syrphid mimics emerge first, with con-
sistency. This creative game-based approach worked because
researchers could be fairly confident that decisions made by hu-
mans in simulations were similar to those previously observed in
animal-based experiments and results from algorithms and neural
nets. This technique could be promising in other ecological
systems in which human behavior and decision making provides
an added value, such as in other predator-prey, competition, or
mutualism interactions.

Fourth, in the final phase of their study, Hassall et al. (3) re-
turned to the long-term citizen science data to analyze changes in
the phenology of the interacting species as the climate warms in
light of what they discovered about the fitness benefits and costs
of different phenological scenarios. They found that the phenol-
ogy of many model bee and wasp species is shifting so that they
appear before mimic syrphid species. These shifts create new
situations that benefit bees and wasps and bird predators, while
reducing the circumstances during which syrphids do best (over-
lapping phenology) and worst (mimics emerge first). These con-
clusions result from the synthesis of three citizen science tools,
each of which benefitted from different styles and values of par-
ticipant contributions, and each of which provided a different
piece of this complex story.

What Next?
The approaches of Hassall et al. (3) avoid labor-intensive field
research and sometimes impossible-to-get direct observations of
interactions in nature. We are confident that researchers will find
many more applications for these techniques. Most obviously,
similar methods might be applied to other Batesian mimicry sys-
tems such as those occurring in butterflies, in which there are models,
mimics, and bird predators (e.g., 19). This approach might also
be applied to other complex species interactions such as some
predator-prey interactions, or to human activities such as bushmeat
hunting or fishing strategies.

Moreover, the results of Hassall et al. (3) have intriguing im-
plications and raise questions about their generality and how they
play out in nature. Is selection driving these changes in phenology
or are the hymenopteran model species just more phenotypically
plastic than the mimic syrphids? The predictions generated from
this study (and other such studies) should also be tested with field
studies of the target species. Do syrphid mimics really do worse

2 of 3 | www.pnas.org/cgi/doi/10.1073/pnas.1820266116 Miller-Rushing et al.

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 

https://www.pnas.org/cgi/doi/10.1073/pnas.1820266116


when they emerge before hymenopteran models? And do birds
and hymenopterans do better when models emerge before mimics?

The approaches of Hassall et al. (3) add to the growing crea-
tivity of citizen science in ecology and climate change research.
New technologies, the collection and sharing of large datasets,
and greater online access to historical data and museum speci-
mens contribute to this growth, as does the willingness and in-
terest of so many people to participate in scientific projects.

We hope that ecologists continue to apply new, innovative
approaches to pressing current problems like understanding
ecological responses to climate change.

Acknowledgments
We thank Rick Bonney, Christopher Hassall, Tom Sherratt, and Jake Weltzin for
comments on earlier versions of this paper. The findings and conclusions in this
paper are those of the authors and do not necessarily represent the views of the
US Government.

1 Thackeray SJ, et al. (2016) Phenological sensitivity to climate across taxa and trophic levels. Nature 535:241–245.
2 Kharouba HM, et al. (2018) Global shifts in the phenological synchrony of species interactions over recent decades. Proc Natl Acad Sci USA 115:5211–5216.
3 Hassall C, Billington J, Sherratt TN (2019) Climate-induced phenological shifts in a Batesian mimicry complex. Proc Natl Acad Sci USA, 10.1073/pnas.1813367115.
4 Miller-Rushing A, Primack R, Bonney R (2012) The history of public participation in ecological research. Front Ecol Environ 10:285–290.
5 Sparks TH, Carey PD (1995) The responses of species to climate over two centuries: An analysis of the Marsham phenological record, 1736-1947. J Ecol
83:321–329.

6 Primack RB, Miller-Rushing AJ (2012) Uncovering, collecting, and analyzing records to investigate the ecological impacts of climate change: A template from
Thoreau’s Concord. Bioscience 62:170–181.

7 Schwartz MD, ed (2003) Phenology: An Integrative Environmental Science (Kluwer Academic Publishers, Dordrecht, The Netherlands), p 564.
8 Chandler M, et al. (2017) Contribution of citizen science towards international biodiversity monitoring. Biol Conserv 213:280–294.
9 Theobald EJ, et al. (2015) Global change and local solutions: tapping the unrealized potential of citizen science for biodiversity research. Biol Conserv
181:236–244.

10 Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37–42.
11 Shirk J, et al. (2012) Public participation in scientific research: A framework for deliberate design. Ecol Soc 17:29.
12 Ellwood ER, et al. (2015) Accelerating the digitization of biodiversity research specimens through online public participation. Bioscience 65:383–396.
13 Fernandez-Gimenez ME, Ballard HL, Sturtevant VE (2008) Adaptive management and social learning in collaborative and community-based monitoring: A study of

five community-based forestry organizations in the western USA. Ecol Soc 13:4.
14 Austen K (2015) Environmental science: Pollution patrol. Nature 517:136–138.
15 Renner SS, Zohner CM (2018) Climate change and phenological mismatch in trophic interactions among plants, insects, and vertebrates. Annu Rev Ecol Evol Syst

49:165–182.
16 Zimova M, Mills LS, Nowak JJ (2016) High fitness costs of climate change-induced camouflage mismatch. Ecol Lett 19:299–307.
17 Reed TE, Jenouvrier S, Visser ME (2013) Phenological mismatch strongly affects individual fitness but not population demography in a woodland passerine. J Anim

Ecol 82:131–144.
18 Ovaskainen O, et al. (2013) Community-level phenological response to climate change. Proc Natl Acad Sci USA 110:13434–13439.
19 Ries L, Mullen SP (2008) A rare model limits the distribution of its more common mimic:Atwist on frequency-dependent Batesian mimicry. Evolution

62:1798–1803.

Miller-Rushing et al. PNAS Latest Articles | 3 of 3

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1