ESA talk on pollination declines, ecosystem functions, response effect traits and

1. Ignasi Bartomeus
nacho.bartomeus@gmail.com
@ibartomeus #ESA2013
Pollinator ecological
traits mediate the loss
of pollination services
with agricultural
intensification

2. 76% of crops are animal
dependent (Klein et al 2007)
Pollination is a key function

3. complement
species (14,
or “samplin
other mecha
evenness m
complement
dominant sp
the most eff
date, the f
portance of
crop pollina
results (22),
on pollinatio
unknown, a
insect loss o
evaluated
pollinated cr
We teste
from the ass
effectively p
crops, and t
placed by i
of honey be
(1) for mos
honey bee v
deposition o
consequently
insect and h
proves fruit
insects prom
honey bees
negatively in
wild-insect
Wild Pollinators Enhance Fruit Set of
Crops Regardless of Honey Bee
Abundance
Lucas A. Garibaldi,1
* Ingolf Steffan-Dewenter,2
Rachael Winfree,3
Marcelo A.
Aizen,4
Riccardo Bommarco,5
Saul A. Cunningham,6
Claire Kremen,7
Luísa G.
Carvalheiro,8,9
Lawrence D. Harder,10
Ohad Afik,11
Ignasi Bartomeus,12
Faye
Benjamin,3
Virginie Boreux,13,14
Daniel Cariveau,3
Natacha P. Chacoff,15
Jan H.
Dudenhöffer,16
Breno M. Freitas,17
Jaboury Ghazoul,14
Sarah Greenleaf,7
Juliana Hipólito,18
Andrea Holzschuh,2
Brad Howlett,19
Rufus Isaacs,20
Steven
K. Javorek,21
Christina M. Kennedy,22
Kristin Krewenka,23
Smitha Krishnan,14
Yael Mandelik,11
Margaret M. Mayfield,24
Iris Motzke,13,23
Theodore Munyuli,25
Brian A. Nault,26
Mark Otieno,27
Jessica Petersen,26
Gideon Pisanty,11
Simon G.
Potts,27
Romina Rader,28
Taylor H. Ricketts,29
Maj Rundlöf,5,30
Colleen L.
Seymour,31
Christof Schüepp,32,33
Hajnalka Szentgyörgyi,34
Hisatomo Taki,35
Teja Tscharntke,23
Carlos H. Vergara,36
Blandina F. Viana,18
Thomas C.
Wanger,23
Catrin Westphal,23
Neal Williams,37
Alexandra M. Klein13
*To whom correspondence should be addressed. E-mail: lgaribaldi@unrn.edu.ar
Affiliations are listed at the end of the text
Diversity and abundance of wild-insect pollinators have declined in many
agricultural landscapes. Whether such declines reduce crop yields, or are mitigated
by managed pollinators such as honey bees, is unclear. Here we show universally
positive associations of fruit set with wild-insect visits to flowers in 41 crop systems
worldwide, and thus clearly demonstrate their agricultural value. In contrast, fruit
set increased significantly with visitation by honey bees in only 14% of the systems
surveyed. Overall, wild insects pollinated crops more effectively, because increase
Fig. 1. Locations of the 41 crop systems studied. Symbols indicate the percentage of total visitation rate to crop flowers
Garibaldi et al. 2013 Science

4. How bee declines will affect pollination function?

6. 40%
47%
13%
Winfree, Bartomeus, Cariveau 2011,AREES
Positive Neutral Negative

7. Not all species are equally vulnerable
Historical changes in northeastern US bee pollinators
related to shared ecological traits
Ignasi Bartomeusa,b,1
, John S. Ascherc,d
, Jason Gibbse
, Bryan N. Danforthe
, David L. Wagnerf
, Shannon M. Hedtkee
,
and Rachael Winfreea,g
a
Department of Entomology, Rutgers University, New Brunswick, NJ 08901; b
Department of Ecology, Swedish University of Agricultural Sciences, Uppsala
SE-75007, Sweden; c
Division of Invertebrate Zoology, American Museum of Natural History, New York, NY 10024-5192; d
Department of Biological Sciences,
Raffles Museum of Biodiversity Research, National University of Singapore, Singapore 117546; e
Department of Entomology, Cornell University, Ithaca,
NY 14853; f
Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT 06269-3043; and g
Department of Ecology, Evolution,
and Natural Resources, Rutgers University, New Brunswick, NJ 08901

8. Bee species declining
are a non random
subset of all species
Bartomeus et al 2013 PNAS

9. Bee species declining
are a non random
subset of all species
Do they have
common traits?
Response traits

10. Bee species declining
are a non random
subset of all species
Not all bees are as
efficient crop
pollinators
Do they have
common traits?
Response traits

11. Bee species declining
are a non random
subset of all species
Not all bees are as
efficient crop
pollinators
Do they have
common traits?
Do they have
common traits?
Effect traits
Response traits

12. The Data

13. Systems:
Rachael Winfree
Faye Benjamin
Dan Cariveau

14. Species
Abundance
14
15 sites per crop

15. 15
Species
Abundance

16. Pollen deposition
Abundance X
Pollination Function
16

17. Sites Agricultural intensification:
Tina Harrison
ember 2011 12:49
?
?
?
?
?
?
?
?
?
?
300 to 3,000 m radius
a
b % Agriculture in1000 radius

18. Species Response traits: which bee traits make them
more sensitive to land use change?

19. Traits: -Body size
-Specialization (corrected for plant phylogeny)
-Phenology
-Nest site
-Tongue length
Species Response traits: which bee traits make them
more sensitive to land use change?
Are those traits also important for pollination efficiency?

20. Questions:
Can we identify response traits typical of the species
that decline most with increasing agricultural intensification?
(Can we identify effect traits typical of the species that
provide the most pollination?)
How would trait-mediated species loss order affect
function?

21. - traits x agricultural intensification
What we need identify response traits?

22. - species x site
- species x traits
- sites x agricultural intensification
- traits x agricultural intensification
ES42CH01-Winfree ARI 26 September 2011 12:49
?
?
?
?
?
?
?
?
?
?
300 to 3,000 m radius
a
b
Figure 3
Schematic showing the two study designs contrasted in this review. (a) Design focused on surrounding
landscape cover. Sampling is generally done within a fixed habitat type. In the most common design, sites
vary in the proportion of surrounding land cover composed of specific habitat types such as forest (dark gree
or agriculture ( yellow). The radius at which landscape cover is assessed varies across studies but is typically
between 300 and 3,000 m. Other designs, which we include in this category, vary either the linear distance
the nearest habitat patch or the area of the habitat patch. (b) Design focused on local land-use type. These
studies compare pollinator communities among different habitat types. The surrounding landscape cover
and the spatial extent of the habitat type where pollinators are sampled are generally not reported.
Figure 4)]. Bees and butterflies both show strong negative responses to land-use change in extrem
systems, but more mixed responses in moderate systems (Supplemental Tables 2 and 3). Extrem
land use causes a strong decrease in abundance and/or richness (e.g., Aizen & Feinsinger 199
Koh & Sodhi 2004, Kremen et al. 2002, Ockinger & Smith 2006), whereas studies in moderate
anthropogenic landscapes find more varied responses (e.g., Bartomeus et al. 2010, Bergman et
2008).
Study designs that make comparisons across habitat types, rather than across landscape gr
dients, find even fewer negative effects, and responses are predominantly positive for most ta
(Supplemental Table 4). For bees, the ratio of negative-to-positive responses decreases fro
8.2 for extreme landscape studies to 2.0 for moderate landscape studies, to 0.5 for across-habit
comparisons. For butterflies, the ratios decrease from 6.0 to 3.0 to 1.1, respectively (Suppl
mental Tables 2–4). The responses of syrphid flies and vertebrates are difficult to interpr
due to the limited number of landscape-scale studies that have been conducted (Supplement
Annu.
Rev.
Ecol.
Evol.
Syst.
2011.42:1-22.
Downloaded
from
www.annualreviews.org
by
67.139.62.82
on
11/18/11.
For
personal
use
only.
What we need identify response traits?

23. - species x site
- species x traits
- sites x agricultural intensification
- traits x agricultural intensification
ES42CH01-Winfree ARI 26 September 2011 12:49
?
?
?
?
?
?
?
?
?
?
300 to 3,000 m radius
a
b
Figure 3
Schematic showing the two study designs contrasted in this review. (a) Design focused on surrounding
landscape cover. Sampling is generally done within a fixed habitat type. In the most common design, sites
vary in the proportion of surrounding land cover composed of specific habitat types such as forest (dark gree
or agriculture ( yellow). The radius at which landscape cover is assessed varies across studies but is typically
between 300 and 3,000 m. Other designs, which we include in this category, vary either the linear distance
the nearest habitat patch or the area of the habitat patch. (b) Design focused on local land-use type. These
studies compare pollinator communities among different habitat types. The surrounding landscape cover
and the spatial extent of the habitat type where pollinators are sampled are generally not reported.
Figure 4)]. Bees and butterflies both show strong negative responses to land-use change in extrem
systems, but more mixed responses in moderate systems (Supplemental Tables 2 and 3). Extrem
land use causes a strong decrease in abundance and/or richness (e.g., Aizen & Feinsinger 199
Koh & Sodhi 2004, Kremen et al. 2002, Ockinger & Smith 2006), whereas studies in moderate
anthropogenic landscapes find more varied responses (e.g., Bartomeus et al. 2010, Bergman et
2008).
Study designs that make comparisons across habitat types, rather than across landscape gr
dients, find even fewer negative effects, and responses are predominantly positive for most ta
(Supplemental Table 4). For bees, the ratio of negative-to-positive responses decreases fro
8.2 for extreme landscape studies to 2.0 for moderate landscape studies, to 0.5 for across-habit
comparisons. For butterflies, the ratios decrease from 6.0 to 3.0 to 1.1, respectively (Suppl
mental Tables 2–4). The responses of syrphid flies and vertebrates are difficult to interpr
due to the limited number of landscape-scale studies that have been conducted (Supplement
Annu.
Rev.
Ecol.
Evol.
Syst.
2011.42:1-22.
Downloaded
from
www.annualreviews.org
by
67.139.62.82
on
11/18/11.
For
personal
use
only.
What we need:
The four corner problem: Permutation test
Legendre et al 2010 Ecology

24. Body size
Body size
Specialization
Results: Can we identify response traits typical of the
species that decline most with increasing agricultural
intensification?
r = “correlation with intensification”

25. Body size
Body size
Specialization
r = -0.34 p= 0.02
r = -0.22 p= 0.04
r = 0.17 p=0.04
Specialists are doing worst
Big are doing worst
Big are
doing better
Results: Can we identify response traits typical of the
species that decline most with increasing agricultural
intensification?

26. Simulation of order loss
How would trait-mediated species loss order affect function?
Function
% trait removed

27. Specialists first Big species first
%
Function
% trait removed
%
Function
% trait removed

28. Specialists first Big species first
%
Function
% trait removed
%
Function
% trait removed
**
ns

29. Specialists first Small species first
%
Function
% trait removed
%
Function
% trait removed
***
ns

30. Specialists first Big species first
%
Function
% trait removed
%
Function
% trait removed
ns
*

31. Pollen deposition
Abundance X
Pollination Function
31

32. all p < 0.05
Pollen
Deposition
% trait removed
Pollen
Deposition
Pollen
Deposition

33. Response traits vary in identity and strength
among crops.
The response-effect trait framework does not allow us
to generalize about how pollinator loss will affect
pollination function

34. Thank you - nacho.bartomeus@gmail.com
The response-effect trait framework does not allow us
to generalize about how pollinator loss will affect
pollination function
Response traits vary in identity and strength
among crops.
@ibartomeus
funding;

35. Trait mediated loss order vs
empirical species order loss
Traits mediated
Empirical loss order