Global Climate Change Program • WWF-UK & WWF-US
Dragonfly MigrationContinental-scale Movement with a Climate-
driven Pulse
John H. Matthews ° 2 August 2008
Do Dragonflies Migrate South in Fall?
with Jay Banner, Tom Juenger, Larry Mack, & Len Wassenaar
What Is Migration?Johnson, Kennedy, & Dingle: 1960 to 1996
Consensus: “Persistent, undistracted motion”
Need not be two-way movement or to/from specific localities
Often associated with special behaviors, physiological states (Dingle)
Large-scale movement rarely described in small-bodied organisms, especially insects
Movement over large scales presents strong conservation challenges
>5 x 106 individuals passing single localities in a day
Telemetry & gross (not net) movement: up to 100 km in a day (Wikelski et al. 2006)
Allegations of Migration
One of ~15 odonate species believed to migrate long distances
More Than a century of anecdotal reports
Oil platforms: individuals >100km from shore crossing Gulf of Mexico
Larvae are completely aquatic, specializing in small standing wetlands
Hydrogen isotopeprocessing
Strontium isotopeprocessing
Microsatellite DNAEstimates north-south movement
Are Adults from Coastal Areas?
Reproductive Strategy,Biogeography
Studying Movement by Individuals and
Across Generations
Individual-based Methods : : Population-based Methods
Isotopic Ratios & MovementPrecipitation Hydrogen isotope ratios (δD) Have a Strong Latitudinal Component
Marine and near-coastal water strontium ratios (87Sr/86Sr) are uniform and distinct
Runoff, underlying geology 87Sr/86Sr ratios
Precipitation δD values
— Negligible Fractionation —
— fractionated —new δD value, locked
in wing tissue
Ambient 87Sr/86Sr, locked in wing tissue
August–October 2005:16,000-km Sampling Route
19° latitude
47° latitude
How Many Degrees of Latitude Are They Flying? northwa
d
Southward
Net
latit
ude
(°)
15
20
25
30
35
40
45
50
55
60
65
0 20 40 60 80 100 120 140 160 18015
20
25
30
35
40
45
50
55
60
65
0 20 40 60 80 100 120 140 160 180Individuals, sorted by collection latitude & est. origin latitude late August mid October
Collection latitude
95% Prediction Intervals for Migration Origination Latitude
Where Did They Come From?
Most are from this range
Intra-generational Movement
northwardd
movement
southward movement
Net north-south movement (km)
Freq
uenc
y
Sμ = 909 km
Globalμ = 683 km
Nμ = 558 km
no
mov
emen
t
How Are Swarms Organized?
0.70800
0.70850
0.70900
0.70950
0.71000
0.71050
-100 -80 -60 -40 -20 0 20
Calculated natal precip dD value
522937545118444649169264023322NJ marsh
above marine ratio
below marine ratio
~ carbonate ratios in Ontario & Quebec
Profiling one Flight Group in southern New Jersey
+Δ Distance flown before capture
What is microsatellite DNA?
. . . TTCAAAATCATCATCATCATCATCATCATCATCATACCAGCC . . .
Forward flanking region Reverse flanking region
1 repan arbitrary number of repeats
primer designed to fit this locus
PCR amplifies this region; with two chromosomes with this locus (one from each parent), an individual might inherit two different versions of this locus
5’-—>3’
Inter-generational Movement180 adults
12 microsatellite loci developed, 9 used
Global Fst: 0.04
Mantel: no relationship
Bayesian clustering methods (Pritchard et al. 2000; Structure 2.2)
Optimal clustering is low (1 to 2 “populations”) Adults Form a Single Population Across 27°
Latitude in Eastern North America
Testing genetic hypotheses:Microsatellite resultsclustering Type (130
genotypes)
Number of clusters Overall Fst P Value
Adult collect site 10 0.032 <0.001
Males v. females 2 0.007 0.003
Atlantic coast, Gulf coast, Lake Huron, Ontario 3 0.009 <0.001
Collected above/below 30° N Lat 2 0.003 0.02
Individual based Bayesian clustering optimal: 2 0.052 NA
Migration Conclusions
First conclusive proof of large-scale odonate migration
Individuals traveling up to 2800 km net north-south distance
One of the largest-scale insect migrations ever described
Spring northern movement must be explored in more detail
New suite of inexpensive & synergistic techniques to describe large-scale movement by insects
Should have a wide variety of conservation applications
Are Anax junius Populations Spatially Structured?
With Tom Juenger & Sandra Boles
How related are migrants & residents?Is there population structure within/between categories?
Scales of Genetic Organization Across the Landscape
In a set of eggs
Within a single pond
Throughout a region
• Females, males mate with multiple individuals
• Other odonate females store sperm• May only be half-sibs
• Males guard shorelines, females visit ponds on a more-transient basis
• Hundreds visit a single pond/day• Kinship coefficients for 19 collection
sites: ~0.05 (Ritland `96)
• Regional patterns are weak descriptors of structure (R2<0.10)
• Migration over 100s of km• Mating/egglaying en route
Spatio-temporal Patterns~450 individuals, spanning adult and larval stages
Mantel shows no spatial pattern
Larval heterozygosity declines slightly with latitude
Structure supports a low number of populations across the full range
Consensus number is one to two populations
At two populations, less than 10% of individuals are mixed between populations
The combination of category, life stage, latitude, and longitudeexplains about 22% of the population membership data
Life-stage may be conflated with temporal segregation of mating, dispersal patterns
Biogeography conclusions
Two complex suites of behavior are maintained over large spatial scales
Novel form of migration, more typical of marine than terrestrial species
adults do not leave reproductive state during the movement process
reproduction coupled with migration
Why Don’t All Anax junius
Migrate South?determinants of life-history
trajectoryWith Camille Parmesan,
Morgan Kelly, & Tom Juenger
Genetically, residents & migrants form a single group with two different phenologies and adult behaviors.
•How does an individual know to follow one or another path?•Why are there two paths? Why not one? three? or seven?
Could life-history path be a form of phenotypic plasticity?
Phenotypic plasticity: an external cue signals to an organism to follow a particular developmental path. Two obvious cues:
Larval growth rate can be regulated by temperature (Trottier 1970); common in insects
Photoperiod (daylength) had “some”influence on developmental rate on a congeneric (Corbet 1956); reliable over large scales
| hold constant
| vary by treatment
Experimental design
Eggs from a single female (unknown # males)
Constant photoperiod
Increasing photoperiod
springtreatment
Decreasing photoperiodFall
treatment
Migrantsfast
Residentsslow
Individuals raised in cups in a growth chamber
Experimental Results
0.300
0.400
0.500
0.600
0.700
0.800
0.900
1.000
10/7/06 10/12/06 10/17/06 10/22/06 10/27/06 11/1/06 11/6/06 11/11/06
condecinc
Duration: 60 days N: 34 individuals
Constant DecreasingIncreasing
Photoperiod Treatment
Lf/Lf-25
1.5 —
3.0 —
3.5 —
Comparison of the last 26 days of growth-rate by treatment
2.5 —
2.0 —
bars show 95% CI
Constant DecreasingIncreasing
Photoperiod Treatment
Lf/Lf-20
1.5 —
2.0 —
2.5 —
Comparison of the last 21 days of growth-rate by treatment
bars show 95% CI
a
b
a
p < 0.05
Constant DecreasingIncreasing
Photoperiod Treatment
Lf/Lf-10
1.2 —
1.4 —
1.6 —
1.8 —
Comparison of the last 11 days of growth-rate by treatment
bars show 95% CI
a
b
a
p < 0.05
Life-history path conclusions
Increasing treatment was significantly different than constant & decreasing treatments
Treatments differentiated late in the experiment
Weak family effects; clear direction for future work
Rain, Rain, Go Away: Climate Change Impacts in Southern
Ontario with Camille Parmesan
Emergence phenology and precipitation normals are linked over large scales
Evapotranspiration rates are inversely related to emergence rates
Austin, Texas, 20-year climate normal data
Aus
tin, T
exas
, em
erge
nce
coun
ts
Aus
tin, T
exas
, EV
T
Caledon, Ontario: 1967–68
J. Can Ento, Trottier, 1971
Generalized Caledon emergence, 1967–68
Residents Migrants
µ=July 4 µ=Sept 13
Hiatus: 40 days
Indi
vidu
als
Calendar day
Source: Trottier 1971
Generalized Caledon emergence, 2003–06
µ~July 22 µ~August 31
Indi
vidu
als
Calendar day
Emerginglater
Emergingearlier
No hiatus
What explains opposing shifts in emergence phenology?
Air temperature?
Water temperature?
What drives water temperature in small standing wetlands?
What is the relationship between water temperature, the timing/amount of precipitation, and water volume?
Shallow-water temperatures are probably key since larvae are concentrated there
Testing drivers of shallow-zone water temperatures
From depth to volume:Some 120 sub-meter precision GIS sounding
waypoints, to generate a TIN to model volume through time
From July 2003 to August 2006:Water temps: 10 cm, 1 m depthsAir tempsWater depthsEmergence phenology
Deep-water temperatures drive shallow temperatures
Best-fit model also includes air temperatures and water volume (R2 = 0.9881)
R2 = 0.9774, p<2 x 10-6
Water temperature differences between 67 and 68, 04–06 also
explain small inter-annual phenological differences
Between May and October, water volume declines
As thermal mass decreases, differences between air and water temp also decrease
y = -0.0132x + 511.43
R2 = 0.1134
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
4/28/05 5/12/05 5/26/05 6/9/05 6/23/05 7/7/05 7/21/05 8/4/05 8/18/05 9/1/05 9/15/05 9/29/05
Ab
sVal(
Sh
allo
w t
em
p -
Air
tem
p)
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Vo
lum
e (
m^
3)
Abs(shallow-air) volume
Dragonfly emergence 2003–06
Indi
vidu
als
Calendar day
Much more rain in May since 1968
Much less rain in August since 1968
Emerginglater
Emergingearlier
More thermal mass, cooler water Less thermal mass,
warmer water, shorter hydroperiod
No hiatus
AcknowledgmentsAcknowledgments
• Field Assistance– Tony Alexander– Charles Britt– Nathan Burnett– John Crutchfield– Laurie Goodrich– Jeremy Harrison– Mike & Darleen McCormick– Marie Riddell– Alex Riddell, M.D.– Kieran Samuk
• Lab Assistance– Sandra Boles– Sung Chun– Abbie Green– Jamie Lee– Tierney Wayne– The Jansen lab
• Additional Support– Damon Broglie, GIS analyst– Anthony Cognato, Michigan State
Univ– Philip Corbet, Edinborough– Mike May, Rutgers University– Rob Plowes– The Parmesan lab
Th J l b
• Field Assistance– Tony Alexander– Charles Britt– Nathan Burnett– John Crutchfield– Laurie Goodrich– Jeremy Harrison– Mike & Darleen McCormick– Marie Riddell– Alex Riddell, M.D.– Kieran Samuk
• Lab Assistance– Sandra Boles– Sung Chun– Abbie Green– Jamie Lee– Tierney Wayne– The Jansen lab
• Additional Support– Damon Broglie, GIS analyst– Anthony Cognato, Michigan State
Univ– Philip Corbet, Edinborough– Mike May, Rutgers University– Rob Plowes– The Parmesan lab
The Juenger lab
• My Committee• Camille Parmesan, advisor
– John Abbott– Jay Banner, Geosciences – Thomas Juenger– Michael Singer
• Lab Mentors– Sandra Boles, Duke University– Larry Mack, Geosciences
• My Family & Friends• Kerry Watkins
Funding & SupportJohn AbbottJay BannerThomas JuengerCamille Parmesan
UT Brackenridge Field LaboratoryUT Environmental Science InstituteHawk MountainNational Science FoundationUT Section of Integrative BiologyUT Vice President of Research
Lorraine Stengl, M.D.The McLachlen family, Dripping SpringsThe Blattstein family, Tucson, AZ
More Acknowledgments