Combined
effects
of
global
change
pressures
on
animal-mediated
pollination
Juan
P.
Gonza
´
lez-Varo
1
,
Jacobus
C.
Biesmeijer
2
,
Riccardo
Bommarco
3
,
Simon
G.
Potts
4
,
Oliver
Schweiger
5
,
Henrik
G.
Smith
6
,
Ingolf
Steffan-Dewenter
7
,
Hajnalka
Szentgyo¨
rgyi
8
,
Michał
Woyciechowski
8
,
and
Montserrat
Vila
`
1
1
Estacio´
n
Biolo´
gica
de
Don
˜
ana
(EBD-CSIC),
Avda.
Ame´
rico
Vespucio,
s/n,
Isla
de
la
Cartuja,
41092
Sevilla,
Spain
2
Naturalis
Biodiversity
Center,
PO
Box
9517,
2300
RA
Leiden,
The
Netherlands
3
Swedish
University
of
Agricultural
Sciences,
Department
of
Ecology,
SE-75007
Uppsala,
Sweden
4
School
of
Agriculture,
Policy,
and
Development,
University
of
Reading,
Reading,
RG6
6AR,
UK
5
UFZ,
Helmholtz
Centre
for
Environmental
Research
(UFZ),
Department
of
Community
Ecology,
Theodor-Lieser-Str.
4,
06120
Halle,
Germany
6
Department
of
Biology
&
Centre
of
Environmental
and
Climate
Research,
Lund
University,
S-223
62
Lund,
Sweden
7
Department
of
Animal
Ecology
and
Tropical
Biology,
Biocentre,
University
of
Wu
¨
rzburg,
Am
Hubland,
97074
Wu
¨
rzburg,
Germany
8
Institute
of
Environmental
Sciences,
Jagiellonian
University,
Gronostajowa
7,
30-387
Krakow,
Poland
Pollination
is
an
essential
process
in
the
sexual
repro-
duction
of
seed
plants
and
a
key
ecosystem
service
to
human
welfare.
Animal
pollinators
decline
as
a
conse-
quence
of
five
major
global
change
pressures:
climate
change,
landscape
alteration,
agricultural
intensifica-
tion,
non-native
species,
and
spread
of
pathogens.
These
pressures,
which
differ
in
their
biotic
or
abiotic
nature
and
their
spatiotemporal
scales,
can
interact
in
nonad-
ditive
ways
(synergistically
or
antagonistically),
but
are
rarely
considered
together
in
studies
of
pollinator
and/or
pollination
decline.
Management
actions
aimed
at
buff-
ering
the
impacts
of
a
particular
pressure
could
thereby
prove
ineffective
if
another
pressure
is
present.
Here,
we
focus
on
empirical
evidence
of
the
combined
effects
of
global
change
pressures
on
pollination,
highlighting
gaps
in
current
knowledge
and
future
research
needs.
Animal-mediated
pollination
under
global
change
Pollination
is
an
essential
process
in
the
sexual
reproduction
of
angiosperm
species,
more
than
260
000
of
which
(88%)
rely
on
animals
for
pollen
transfer
[1].
In
turn,
approximate-
ly
300
000
animal
species
are
attracted
to
visit
angiosperm
flowers
by
pollen
and
nectar
rewards
[2].
Besides
the
critical
role
of
this
mutualism
for
the
maintenance
of
biodiversity,
animal-mediated
pollination
also
provides
a
key
ecosystem
service
to
society.
Approximately
70%
of
the
major
crop
species
worldwide
are
at
least
partly
reliant
on
animal
pollination
(mainly
by
insects)
for
yield
production,
account-
ing
for
35%
of
global
food
production
[3].
Pollinator
declines
have
been
attributed
to
different
glob-
al
change
pressures
[4–7].
Climate
change
[8,9],
landscape
alteration
[10,11],
species
invasions
[12,13],
agricultural
intensification
[14–16],
and
spread
of
pathogens
[17]
have
been
identified
as
the
main
causes
of
declines
in
pollinator
abundances
and
extinctions,
with
the
latter
causing
shifts
in
pollinator
community
composition
[18],
disruption
of
plant–
pollinator
interactions
[19],
and
loss
or
destabilisation
of
pollination
services
to
wild
[19]
and
crop
plants
[10,20,21].
Terrestrial
ecosystems
are
currently
impacted
by
multi-
ple
pressures
and,
thus,
knowledge
of
the
interactive
effects
between
them
is
essential
for
both
biodiversity
conservation
and
the
maintenance
of
the
ecosystem
services
provided
by
pollinators
[22].
Indeed,
the
effects
of
one
pressure
can
be
amplified
or
buffered
by
the
effects
of
another
pressure
[22,23].
The
management
implication
of
such
interactive
effects
is
that
action
plans
aimed
at
buffering
the
effects
of
a
particular
pressure
can
become
ineffective
if
another
pres-
sure
is
present,
potentially
resulting
in
a
waste
of
resources
devoted
to
mitigation
(e.g.,
[24]).
In
this
paper,
we
focus
on
the
empirical
evidence
of
combined
effects
of
multiple
global
change
pressures
on
animal-mediated
pollination,
and
discuss
both
the
conse-
quences
for
pollination
services
and
the
potential
implica-
tions
for
management.
We
draw
attention
to
the
spatiotemporal
scales
of
impact,
the
experimental
approaches
used
to
study
them,
the
gaps
in
current
knowl-
edge,
and
future
research
needs.
From
single
effects
of
global
change
pressures
to
interactions
between
them
Single
global
change
pressures
have
characteristic
spatio-
temporal
scales
of
action
and
generate
impacts
at
different
rates
and
at
different
levels
of
ecological
organisation,
from
individuals
to
ecosystems
(Boxes
1
and
2).
It
is
important
to
consider
the
contrasting
biotic
or
abiotic
nature
of
these
pressures
to
understand
their
interactive
impacts
on
ani-
mal-mediated
pollination;
environmental
pressures
can
shape
the
distribution
of
species,
but
the
presence
of
Opinion
Corresponding
author:
Gonza
´
lez-Varo,
J.P.
(juanpe@ebd.csic.es).
species
cannot
(normally)
change
the
magnitude
of
the
environmental
changes.
A
given
pressure
can
impact
animal-mediated
pollina-
tion
directly
by
disrupting
the
occurrence,
abundance
and
phenology
of
partner
species.
However,
a
pressure
can
also
impact
pollination
indirectly,
by
interacting
with
other
pressures,
either
additively
or
nonadditively.
Nonadditive
effects
occur
if
the
effect
of
a
given
pressure
is
amplified
or
buffered
when
it
occurs
in
combination
with
another
pres-
sure.
Many
interactive
effects
are
indirect
effects
in
which
a
pressure
modifies
the
magnitude
(quantity)
of
another
pressure
(i.e.,
interaction
chain
effects)
and/or
its
impact
per
capita
(quality;
i.e.,
interaction
modification
effects
[22]).
Indirect
effects
are
expected
to
be
particularly
Box
1.
The
global
change
pressures
and
their
spatiotemporal
scales
of
action
Climate
change
Climate
change,
mostly
warming,
typically
occurs
at
broad
spatial
and
temporal
scales.
However,
increased
climatic
variability
can
result
in
climatically
anomalous
seasons
and/or
years
at
a
regional
scale,
whereas
anomalous
weather
events
can
occur
locally
during
a
short
time
period.
Climate
change
entails
changes
in
community
composition
through
shifts
in
the
geographical
range
and/or
phenology
of
species.
Landscape
alteration
Landscape
alteration
comprises
the
degradation
(including
diffuse
pollution),
destruction,
and
fragmentation
of
natural
habitats,
result-
ing
in
associated
changes
in
landscape
configuration,
habitat
diversity,
and
community
composition.
Although
landscape
altera-
tion
occurs
at
local
and
landscape
scales,
shared
environment
(e.g.,
orography)
and
policies
can
lead
to
similar
alteration
regimes
at
broader
spatiotemporal
scales.
Agricultural
intensification
Intensive
agriculture
is
characterised
by
an
increase
in
input
of
pesticides
and
fertilisers,
farm
size,
monocultures,
and
simplified
crop
rotations.
Agricultural
intensification
and
landscape
alteration
are
usually
difficult
to
separate
because
the
highest
levels
of
intensification
generally
occur
in
the
most
altered
landscapes.
Thus,
agricultural
intensification
shares
similar
spatiotemporal
scales
with
landscape
alteration,
but
also
comprises
processes
(e.g.,
ploughing
and
herbicide
and/or
pesticide
application)
that
occur
at
the
narrowest
scales
(plots
and
days).
Invasive
species
The
effects
of
biological
invasions
on
animal-mediated
pollination
have
usually
been
addressed
considering
non-native
plants
and
non-
native
pollinators
(but
see
[12]
for
a
study
considering
an
invasive
predator).
Whereas
non-native
plants
can
require
long
lag-times
until
significant
representation
in
the
wild,
non-native
managed
pollinators
(mainly
honeybees
Apis
mellifera
and
bumblebees
Bombus
spp.)
can
achieve
huge
abundances
at
short
spatiotemporal
scales
after
the
introduction
of
hives.
Pathogens
The
huge
increase
during
the
past
decades
in
the
trade
of
managed
pollinators
has
promoted
pathogen
transmission
to
wild
pollinators,
and
vice
versa.
Pathogen
transmission
occurs
at
landscape
scales
and
during
the
first
weeks
after
the
release
of
managed
hives.
However,
large-scale
trends
in
the
use
of
managed
pollinators
can
lead
to
widespread
transmission
across
a
region.
Box
2.
Main
documented
impacts
of
single
global
change
pressures
on
animal-mediated
pollination
Climate
change
Climate
change
is
predicted
to
cause
spatial
and
temporal
mis-
matches
between
pollinators
and
their
food
plants
owing
to
differential
shifts
in
the
distribution
ranges
and
phenology
of
interacting
species,
respectively
([8,9,23],
but
see
[61]).
Mismatches
can
cause
pollen
limitation
to
plants
and
gaps
in
food
supply
to
pollinators,
and
both
processes
are
expected
to
be
particularly
detrimental
in
specialist
species
[8].
Overall,
the
more
generalist
the
relations
(i.e.,
multiple
pollinator
species
for
a
plant
or
broad
diet
in
pollinators),
the
more
resilient
the
interactions
are
under
climate
change.
Consequently,
non-random
novel
communities
overrepre-
sented
by
generalist
species
are
expected
after
the
spatial
and
phenological
shifts
of
species
distributions
imposed
by
climate
change
[8,23].
Landscape
alteration
Landscape
alteration
generally
involves
the
reduction
of
floral
and
nesting
resources,
isolation
of
populations,
and
shifts
of
biotic
interactions.
Landscape
alteration
results
in
significant
reductions
in
species
richness
and
abundance
of
pollinators,
particularly
of
habitat-
and
food-specialist
insect
taxa
that
locate
their
nests
above
ground
[11,62].
As
a
consequence,
habitat
fragmentation
produces
strong
negative
effects
on
plant
pollination
and
fecundity
[63].
The
compat-
ibility
system
of
plants,
which
reflects
the
degree
of
dependence
on
pollinator
mutualism,
explains
the
differences
among
species
in
their
response
to
fragmentation
[63].
Agricultural
intensification
Agricultural
intensification
is
thought
to
be
a
major
driver
of
loss
of
pollinators
[14,15,24]
and,
thus,
of
pollination
services
[21].
Mechan-
ical
and
chemical
(herbicides)
practices
result
in
the
loss
of
field
margins
and
weeds
that
provide
nest
sites
and
forage
resources
for
pollinators.
Pesticides
can
directly
affect
the
fitness
of
pollinators,
leading
to
declines,
particularly
of
wild
species
[6,27,64].
Increased
use
of
inorganic
fertilisers
might
also
result
in
pollinator
losses
via
homogenisation
of
floral
communities
[65].
Invasive
species
Many
non-native
plants
are
ornamental
entomophilous
plants
with
floral
displays
attractive
to
native
pollinators.
These
non-native
plants
integrate
well
within
local
pollination
networks,
receiving
on
average
more
pollinator
visits
than
coexisting
native
plant
species
and,
thus,
acting
as
super-generalists
[30,34].
Competition
with
native
plants
seems
to
prevail
over
facilitation
[66],
although
the
sign
and
magnitude
of
such
effects
are
likely
density
dependent.
Non-native
pollinators
can
change
the
composition
of
local
pollinator
assem-
blages
as
a
result
of
their
high
abundance
[38]
and
their
direct
competition
with
native
pollinators.
Furthermore,
non-native
pollina-
tors
can
disrupt
pollination
patterns
of
native
plants
[66].
Managed
honeybees
can
reduce
both
fecundity
and
progeny
performance
through
pollen
limitation
and
inbreeding
depression
[37–39].
Some
short-tongue
bumblebees
(e.g.,
Bombus
terrestris)
can
bite
a
hole
in
the
corolla
of
long-tube
flowers
[20,31],
which
can
also
be
used
by
subsequent
visitors,
leading
to
illegitimate
visits
and
reduction
of
plant
fitness.
Pathogens
Pathogen
transmission
from
managed
to
wild
pollinators,
and
vice
versa,
has
resulted
in
widespread
pollinator
declines
[6,17].
The
best-
known
cases
are
the
transmissions
(mainly
the
Varroa
mite
and
intestinal
protozoans
Nosema
spp.
and
Crithidia
spp.)
between
native
Asiatic
honeybee
(Apis
cerana)
and
non-native
and
managed
European
honeybees
(Apis
mellifera;
[6,67]),
and
between
managed
(mainly
originating
from
Europe)
and
wild
bumblebees
[40,41,44,58].
Although
some
of
these
pathogens
can
be
transferred
between
phylogenetically
more
distant
species
(e.g.,
bumblebees
and
honey-
bee),
there
is
a
lack
of
studies
on
infection
of
wild
pollinators
not
belonging
to
the
Apis
or
Bombus
genera.
Transfer
of
RNA
viruses
has
also
been
found
among
honeybees,
bumblebees,
and
other
non-Apis
taxa
of
wild
bees
[43].
2
common
in
interactions
between
environmental
(climate
change,
landscape
alteration,
or
agricultural
intensifica-
tion)
and
biotic
pressures
(such
as
invasion
of
non-native
species,
or
spread
of
pathogens),
because
the
former
can
potentially
affect
both
the
abundance
and
the
per
capita
impact
of
the
latter
(Box
3,
Figure
I).
Only
a
few
empirical
studies
have
explicitly
explored
the
interactive
effects
of
multiple
global
change
pressures
on
pollinators
and
pollination
(Table
1).
In
the
following
sec-
tions,
we
summarise
studies
focusing
on
the
paired
com-
binations
of
these
pressures,
aiming
to
identify
gaps
in
current
knowledge
and
priorities
for
future
research.
Landscape
alteration
and
agricultural
intensification
The
interactive
effect
between
agricultural
intensification
and
landscape
alteration
is
the
most
commonly
assessed
interaction,
largely
motivated
by
evaluations
of
the
context
dependent
effects
of
agri-environmental
schemes
on
biodiversity
(Table
1).
A
common
experimental
approach
is
a
factorial
design
with
two
levels
within
each
factor.
Levels
for
landscape
alteration
are
usually
‘simple
versus
complex’,
‘homogeneous
versus
heterogeneous’
[25]
or
‘close
to
versus
far
from’
edges
of
semi-natural
habitats
[26],
whereas
levels
for
agricultural
intensification
usually
are
‘conventional
versus
organic’
farming,
although
some
studies
have
compared
farms
with
and
without
pesticide
application
[27].
In
general,
the
negative
effects
of
agricultural
intensifi-
cation
on
pollinator
species
richness
and
abundance
are
stronger
in
simple
(i.e.,
low
cover
of
semi-natural
habitats)
than
in
complex
landscapes,
indicating
synergistic
effects
between
both
pressures
[24].
A
plausible
explanation
is
that
farms
in
complex
landscapes
are
more
likely
to
have
higher
pollinator
diversity
than
are
those
in
simple
land-
scapes.
Therefore,
the
effects
of
organic
farming
are
weak
in
the
former
and
stronger
in
the
latter
[24].
Although
most
Box
3.
Complex
interactive
effects
in
real-world
ecosystems
In
the
real
world,
animal-mediated
pollination
is
impacted
by
more
than
one
pair
of
global
change
pressures,
so
that
multi-pressure
complex
interactive
effects
are
probably
the
norm
rather
than
the
exception
[23,52].
Figure
I
represents
the
possible
combined
negative
effects
of
three
pressures
on
native
bumblebees
and
their
pollination
services:
landscape
alteration,
invasion
by
a
non-native
bumblebee,
and
spread
of
non-native
pathogens
(based
on
[17,20,31,40,41,58]).
Landscape
alteration
might
impact
native
bumblebees
directly
by
reducing
floral
and
nesting
resources.
Indirect
impacts
include:
(i)
‘interaction
chain
effects’
favouring
the
abundance
of
the
non-native
bumblebee;
and
(ii)
‘interaction
modification
effects’
increasing
its
per
capita
impact
through
resource
limitation,
which
additionally
would
increase
the
prob-
ability
of
pathogen
spillover
[22].
Cascading
effects
on
plant
pollination
are
expected
if
the
non-native
bumblebee
is
less
efficient
than
the
native
pollinator
or
if
it
visits
flowers
illegitimately,
for
example
by
nectar
robbing
[31].
Nectar
robbing
is
expected
to
be
more
frequent
in
the
commercially
traded
bumblebee
Bombus
terrestris
owing
to
its
shorter
tongue
length
compared
with
other
bumblebee
species
[20,31].
Loss of floral and
nesng resources
Compeon
Resource
limitaon
Infecon
(in capvity)
Propagule pressure/
invaded habitat
Pathogen
spillover
Infecon
(in the wild)
Legimate visits
Legimate visits
Pollinaon
services
Nectar robbing
Nectar robbing
Non-nave
pollinator
Nave
pollinator
TRENDS in Ecology & Evolution
L
oss of floral and
nes
ng resources
Compe
on
R
esour
ce
limitaon
P
ropagule pressure/
invaded
h
abita
t
Path
spill
t
e vis
i
ts
Nectar robbing
Nectar robbing
Nave
po
ll
inator
Non-nave
Land-use
alteraon
Figure
I.
Scheme
showing
possible
synergistic
effects
between
landscape
alteration,
invasion
by
a
non-native
pollinator,
and
pathogen
spread
impacting
native
pollinators
and
their
pollination
services.
Black
arrows
represent
direct
effects,
whereas
red
arrows
represent
(indirect)
interactive
effects
by
which
a
pressure
(landscape
alteration
or
pathogens)
change
the
per
capita
impact
of
the
non-native
pollinator
on
the
native
pollinator
[22].
Positive
or
negative
signs
in
the
arrows
denote
an
increase
or
a
decrease,
respectively,
in
the
variable
of
study,
whereas
the
text
close
to
each
arrow
denotes
the
mechanism(s)
responsible
for
its
effects.
The
shaded
ellipse
denotes
a
higher
probability
of
pathogen
spillover
due
to
flower
resource
limitation
in
altered
landscapes.
The
pollination
services
provided
by
both
pollinators
will
depend
on
whether
they
perform
legitimate
visits
or
nectar
robbing.
Photo
reproduced
with
permission
from
A.
Montero-Castan
˜
o
(top),
H.
Szentgyorgyi
(right),
and
J.P.
Gonza´
lez-Varo
(bottom
and
left).
Opinion
3
studies
have
only
focused
on
pollinators,
changes
in
their
abundance
and
composition
are
expected
to
have
conse-
quences
on
the
magnitude
and
stability
of
the
pollination
service
needed
for
fruit
and
seed
production
of
wild
plants
[19,28]
and
agricultural
crops
[10,14,21].
The
implication
for
management
of
this
relatively
well-studied
interactive
effect
is
that
certain
policy
actions
aimed
at
buffering
the
negative
effects
of
agricultural
intensification
can
be
more
efficient
in
moderate
to
highly
altered
landscapes
com-
pared
with
little
altered
landscapes
[24].
Landscape
alteration
and
non-native
species
Several
studies
have
considered
the
effect
of
landscape
type
in
combination
with
the
occurrence
of
non-native
pollinators
or
plants
(Table
1).
In
these
studies,
the
degree
of
landscape
alteration
has
been
accounted
for
either
cat-
egorically;
for
example,
‘continuous
versus
fragmented’
or
‘disturbed
versus
undisturbed’
[29,30],
or
continuously
along
a
gradient
of
landscape
naturalness
[31,32].
Invasion
is
assessed
at
the
local
plot
level
(‘invaded
versus
non-
invaded’).
In
general,
both
non-native
plants
and
pollina-
tors
are
disproportionally
more
abundant
in
highly
altered
landscapes,
such
as
in
disturbed
habitats,
or
in
small
remnant
patches
of
semi-natural
habitat
[29,31,33,34].
Thus,
it
is
difficult
to
disentangle
the
causal
effects
of
landscape
alteration
from
those
of
invasion.
With
regard
to
non-native
pollinators,
only
a
few
case
studies
exist.
Ishii
et
al.
[31]
studied
the
distribution
of
non-
native
(Bombus
terrestris)
and
native
(Bombus
spp.)
bum-
blebees
as
well
as
their
foraging
behaviour
visiting
flowers
along
transitions
from
open
farmland
habitats
to
forests
in
central
Hokkaido,
Japan.
They
found
that
B.
terrestris
occupied
deforested
areas,
where
they
had
displaced
native
Bombus
spp.
to
forest
habitats,
which
seem
to
act
as
barriers
against
the
expansion
of
B.
terrestris.
Given
that
B.
terrestris
is
a
short-tongued
pollinator,
the
consequence
for
many
plants
is
an
increase
in
nectar
robbing
(i.e.,
illegitimate
visits
performed
by
biting
a
hole
in
the
corolla
tube)
on
long-tube
flowers
in
the
most
deforested
landscapes.
In
another
study,
Dick
et
al.
[29]
examined
the
mating
patterns
and
pollen
dispersal
distances
of
Dinizia
excelsa
trees
in
fragmented
versus
continuous
rainforests
in
Brazil.
Where-
as
flower
visits
in
continuous
forests
were
performed
almost
exclusively
by
native
pollinators,
introduced
honeybees
were
the
main
flower
visitor
on
remnant
D.
excelsa
trees
located
in
pastures.
For
this
self-incompatible
tree,
honey-
bees
provide
genetic
rescue
by
promoting
long-distance
mating
events
that
connect
continuous
and
fragmented
populations.
Managed
honeybees
can
also
reduce
pollen
limitation
of
self-compatible
plants
through
high
flower
visitation
rates
[35].
Nevertheless,
because
they
tend
to
forage
on
many
flowers
of
the
same
individual
plant
[36],
honeybees
usually
promote
geitonogamous
crossings,
par-
ticularly
in
those
that
bear
large
numbers
of
flowers
[35,37],
which
can
reduce
fecundity
[38]
and
plant
progeny
perfor-
mance
through
self-incompatibility
and/or
inbreeding
de-
pression
[39].
With
regard
to
non-native
plants,
Williams
et
al.
[33]
studied
bee
visits
to
non-native
and
native
plants
in
transformed
and
semi-natural
habitats
of
California
and
New
Jersey
(USA).
They
found
a
positive
interaction
be-
tween
landscape
alteration
and
non-native
plants
on
bee–
plant
interactions;
bee
visits
(species
richness
and
abun-
dance)
to
non-native
plants
were
greater
in
transformed
than
in
semi-natural
habitats.
However,
bees
foraged
on
different
flower
species
according
to
their
local
abundance,
denoting
a
lack
of
preference
for
non-native
flowers.
Additionally,
non-native
pollinators
and
plants
can
form
‘invader
complexes’;
that
is,
groups
of
introduced
species
interacting
more
with
each
other
than
expected
by
chance,
which
might
have
positive
feedbacks
facilitating
the
invasion
of
undisturbed
habitats
[34].
For
example,
in
temperate
forests
of
the
southern
Andes,
non-native
visi-
tors,
mainly
Apis
mellifera
and
Bombus
ruderatus,
visited
flowers
of
non-native
plants
more
frequently
in
disturbed
than
in
undisturbed
habitats;
however,
there
was
no
in-
teraction
between
habitat
disturbance
and
plant
origin
(native
or
non-native)
[34].
As
in
the
case
of
Williams
et
al.
[33],
this
result
could
be
explained
by
the
greater
abundance
of
non-native
plants
in
disturbed
habitats.
Table
1.
Summary
of
studies
that
have
simultaneously
addressed
the
effects
of
two
global
change
pressures
on
animal-mediated
pollination
a,b
Global
change
pressures
Climate
change
Landscape
alteration
Non-native
species
Agricultural
intensification
Landscape
alteration
Positive
C:[19]
Negative
I:
[52]
Non-native
species
Negative
R:
[23]
Positive
I:
[30,32–34]
C:
[29,31]
Agricultural
intensification
Negative
I:
[53]
Positive
M:
[24]
I:
[25–27,60,68]
C:
[14,69]
Positive
C:
[40,41]
Spread
of
pathogens
Positive
R:
[23,56]
C:
[57]
–
Positive
C:
[40,41,70]
Positive
I:
[48–51]
C:
[40,41]
a
‘Positive’
and
‘Negative’
denote
the
type
of
combined
effect
between
pairs
of
pressures
on
diverse
response
variables
related
to
pollinators
(assemblages,
species,
populations,
and
individual
fitness)
and/or
pollination-associated
processes
(visitation
rates,
pollen
limitation,
mating
patterns,
and
fecundity).
b
I,
studies
that
explicitly
tested
for
interactive
effects
between
pressures;
C,
studies
that
assessed
simultaneously
the
effects
of
two
pressures
but
not
the
interaction;
R,
review
studies;
M,
meta-analytical
studies.
Opinion
4
Pathogens
and
non-native
species
The
interactive
effects
of
non-native
species
and
patho-
gen
transmission
have
been
examined
in
terms
of
path-
ogen
spillover
from
commercially
rea red
honeybees
and
bumblebees
to
wild
pollinators.
In
a
cas e
study
in
south-
ern
Ontario
(Canada),
three
bumblebee
pathogens
(two
microsporidia
and
a
tracheal
mite)
infected
native
bum-
blebees
via
shared
flowers
more
frequently
in
landscapes
with
greenhouses
than
in
those
lacking
them
[40].
A
follow-up
study
found
a
sharp
decline
in
infection
rates
by
the
microsporidian
Crithidia
bombi
in
wild
bumblebees
with
distance
from
greenhouses
with
commercial
bumblebee
hives
(Bombus
impatiens)
[41].
In
addition,
the
most
infected
wild
bumblebee
species
were
those
with
a
high
similarity
with
the
commercial
bumblebee
in
the
use
of
plant
species.
Pathogen
spillover
to
wild
bee
taxa
not
belonging
to
the
genera
Apis
and
Bombus
is
likely
to
also
be
important
[42,43],
and
there
is
a
huge
knowledge
gap
on
this
issue
[44].
Pathogens
and
landscape
alteration
or
agricultural
intensification
The
impact
of
pathogens
on
pollinators
is
expected
to
be
higher
in
altered
and
intensively
cultivated
landscapes,
where
pollinator
nutrition
and,
thus,
health
(immune
system
[45]),
relies
on
poor
flower
communities.
This
positive
interaction
is
supported
by
significant
correla-
tions
between
the
honeybee
colony
loss
suffered
by
the
states
in
the
USA
and
the
extent
of
their
main
land-use
types
[46].
Although
the
combined
effects
of
pathogen
spread
and
landscape
alteration
have
not
been
assessed
directly
so
far
(Table
1),
available
evidence
suggests
that
pathogen
spillover
from
commercial
pollinators
should
be
much
greater
(positive
interaction)
in
altered
landscapes,
where
floral
resources
are
usually
scarcer
and
mean
foraging
distances
are
larger
(e.g.,
[31,40,41,47])
(Box
3,
Figure
I).
Similarly,
a
positive
interactive
effect
is
also
expected
between
pathogen
spillover
and
agricul-
tural
intensification,
because
both
insecticides
and
pathogens
have
detrimental
effects
on
the
health
of
wild
pollinators.
Indeed,
Pettis
et
al.
[48]
recently
found
that
microsporidia
infections
(Nosema
sp.)
increased
signifi-
cantly
in
honeybees
exposed
to
a
widely
used
pesticide;
thus,
the
authors
demonstrated
experimentally
an
indi-
rect
positive
effect
of
pesticides
on
pathogen
spread.
Moreover,
several
tests
on
joint
effects
between
infection
by
a
microsporidian
(Nosema
ceranae)
and
exposure
to
a
neonicotinoid
insecticide
on
honeybee
performance,
show
that
several
fitness
parameters
decreased
only
by
the
combination
of
both
factors
[49–51].
These
results
pro-
vide
strong
evidence
of
synergistic
effects
between
path-
ogen
infection
and
pesticide
use.
As
noted,
agricultural
intensification
is
typically
associated
with
the
presence
of
managed,
and
often
non-native,
pollinators
used
to
provide
pollination
services
to
intensively
produced
crops
(e.g.,
[14,21,40]).
There
is,
however,
a
lack
of
studies
assessing
whether
native
pollinators
are
more
impacted
by
such
synergistic
effects
than
the
non-
natives.
Climate
change
and
landscape
alteration
or
agricultural
intensification
Landscape
alteration
and
climate
change
are
expected
to
affect
animal-mediated
pollination
synergistically,
causing
spatiotemporal
mismatches
between
interacting
species
[19]
(Box
2).
Only
one
study
has
experimentally
assessed
the
interactive
effect
of
climate
change
and
landscape
alter-
ation
on
animal-mediated
pollination
[52].
Pollinator
visits
and
seed
production
were
examined
in
experimental
patches
of
native
flowers.
Pots
with
wild
mustard
(Sinapsis
arvensis)
grown
with
‘normal’
and
‘advanced’
flowering
phenology
were
placed
both
‘close’
and
‘distant’
to
semi-
natural
grasslands.
Advanced
flowering
simulated
a
pheno-
logical
shift
in
flowering
due
to
global
warming,
and
distance
to
grasslands
represented
landscape
alteration.
A
negative
interaction
between
flowering
phenology
and
proximity
to
grasslands
was
found:
the
difference
in
the
number
of
flower
visits
by
wild
bees
to
‘distant’
(>500
m)
and
‘close’
(0
m)
flower
islands
was
higher
under
normal
than
under
ad-
vanced
flowering
phenology.
This
result
could
be
explained
by
more
similar
local
flower
abundance
between
close
and
distant
experimental
islands
in
the
advanced
phenology
scenario
as
compared
with
the
normal
one.
Recently,
Hoover
et
al.
[53]
examined
interactive
effects
between
warming,
increased
CO
2
and
nitrogen
(N)
deposi-
tion
in
laboratory
trials
on
several
plant
and
flower
traits
of
pumpkin
(Cucurbita
maxima)
as
well
as
on
domestic
bum-
blebee
(B.
terrestris)
foraging
preferences
and
longevity.
To
our
knowledge,
this
is
the
only
study
that
has
examined
interactive
effects
between
climate
change
and
agricultur-
al
intensification,
because
N
deposition
can
be
linked
to
agricultural
intensification
[54].
There
was
an
antagonistic
effect
between
warming
and
N
deposition
in
that
both
nectar
production
(by
plants)
and
nectar
consumption
(by
bumblebees)
in
the
N-enriched
treatment
were
higher
under
normal
than
under
elevated
temperatures.
Al-
though
such
experiments
are
valuable
because
they
pro-
vide
insights
into
the
mechanisms
underlying
plant
and
pollinator
responses,
they
often
represent
an
oversimplifi-
cation
of
the
real
world.
For
example,
the
studies
by
Parsche
et
al.
[52]
and
Hoover
et
al.
[53]
considered
climate
change
effects
(phenological
shifts
and
alterations
in
nec-
tar
composition,
respectively)
in
a
single
plant
species.
However,
climate
change
tends
to
impact
on
entire
com-
munities,
which
means
that
generalisations
based
on
mi-
crocosm
studies
should
be
made
with
caution
[23].
Climate
change
and
non-native
species
or
pathogens
There
is
a
lack
of
empirical
studies
testing
interactive
effects
of
climate
change
and
non-native
species
on
ani-
mal-mediated
pollination.
Schweiger
et
al.
[23]
compiled
literature
concerning
both
global
change
pressures
and
developed
hypotheses
about
their
possible
interactive
effects.
The
authors
hypothesised
that
atypical
flowering
phenology
of
non-native
species
might
buffer
(antagonistic
effect)
the
detrimental
effects
of
temporal
mismatches
between
interacting
species
caused
by
climate
change
[8].
In
temperate
regions,
many
non-native
plants
are
from
warmer
areas
and
exhibit
a
high
tolerance
to
a
wide
range
of
climatic
conditions;
therefore,
they
have
the
potential
to
fill
gaps
and
curtailments
in
food
supply
to
native
Opinion
5
![Figure I. Scheme showing possible synergistic effects between landscape alteration, invasion by a non-native pollinator, and pathogen spread impacting native pollinators and their pollination services. Black arrows represent direct effects, whereas red arrows represent (indirect) interactive effects by which a pressure (landscape alteration or pathogens) change the per capita impact of the non-native pollinator on the native pollinator [22]. Positive or negative signs in the arrows denote an increase or a decrease, respectively, in the variable of study, whereas the text close to each arrow denotes the mechanism(s) responsible for its effects. The shaded ellipse denotes a higher probability of pathogen spillover due to flower resource limitation in altered landscapes. The pollination services provided by both pollinators will depend on whether they perform legitimate visits or nectar robbing. Photo reproduced with permission from A. Montero-Castaño (top), H. Szentgyorgyi (right), and J.P. Gonzá lez-Varo (bottom and left).](/figures/figure-i-scheme-showing-possible-synergistic-effects-between-3r0cgdzt.webp)
