Climate Change, Disease, and
Amphibian Declines
by Jason R. Rohr
University of South Florida
Department of Biology, SCA 110
4202 E. Fowler Ave.
Tampa, FL 33620
jasonrohr@gmail.com
Climate Change, Amphibian
Declines, and Bd
Also evidence that Bd-related
declines are linked to climate
change (Pounds et al. 2006,
Bosch et al. 2006)
Outline for Talk
Does global climate change affect worldwide
amphibian declines associated with chytrid
fungal infections?
Enigmatic Amphibian Declines
Genus Atelopus
71 of 113 spp. presumed extinct,
many of which were ostensibly
caused by chytridiomycosis
from La Marca et al. 2005. Biotropica
Climate, Bd, and Conservation
Planning
• If we understand the climatic factors that
accelerate Bd spread, increase host
susceptibility, or elevate pathogen
virulence, we can identify present and
future geographic locations that might have
amphibians at risk of Bd-related declines
• Hence, we can better target areas that
warrant monitoring and remediation
Tenuous Links Between Climate
and Amphibian Declines
0.20
0.25
• Most of the evidence supporting climate
Log Beer/20
0.20
0.18
change as a factor in Bd-related
LogaBanana/25
0.15 from
0.16
amphibian extinctions comes
positive, but temporally confounded,
multi0.10
0.14
decade correlation between air Amphibian
extinctions
0.05
0.12
temperature and extinctions in the toad
Temp.
0.00
0.10
genus Atelopus
anomalies/10
1970
1980
1990
Year
Rohr et al. 2008 PNAS
2000
Need to Conduct Detrended
Analyses?
• If there is a true relationship between
climate and Bd-related extinctions,
fluctuations around temporal trends in
temperature and extinctions should also
positively correlate
• There would many fewer non-causal
explanations for this correlation than the
multidecadal relationship between
declines and temperature
Objectives
Use the Atelopus database to simultaneously test various climate-related
hypotheses for amphibian declines,
controlling for multidecadal correlations
and the intrinsic spatiotemporal spread
of Bd
Ultimate Hypothesis: ENSO
Drives Amphibian Declines
ENSO: El Niño-Southern
Oscillation
• Commonly referred to as simply El Niño is a
global coupled ocean-atmosphere phenomenon
– The Pacific ocean signatures, El Niño (warm and
wet) and La Niña (cool and dry) are important
temperature fluctuations in surface waters of the
tropical Eastern Pacific Ocean
– The atmospheric signature, the Southern Oscillation
(SO) reflects the monthly or seasonal fluctuations in
the air pressure
• Effects of El Niño in South America are direct
and stronger than in North America
Proximal Hypotheses for
Enigmatic/Bd-related Declines
•
Spatiotemporal spread hypothesis: declines are
caused by the introduction and spread of Bd,
independent of climate (Bell et al. 2004, Lips et al. 2006)
•
Climate-based hypotheses:
–
Chytrid-thermal-optimum hypothesis: Increased cloud cover,
due to warmer oceanic temperatures, leads to temperature
convergence on the optimum temperature for growth of Bd
(Pounds et al. 2006, Bosch et al. 2006)
–
Mean-climate hypothesis: changes in mean temp. and/or
moisture conditions affect the distributions of amphibians
(Whitfield et al. 2007, Buckley & Jetz 2007)
–
Climate-variability hypothesis: temporal variability in temp.
cause suboptimal host immunity facilitating declines (Raffel, Rohr,
et al. 2006)
Climate-Variability Hypothesis
Ectotherms:
*seasonal changes in body temperature*
Climate Variability Hypothesis
• Hypothesis: unpredictable temperature
shifts, which are increasing with GCC,
benefit pathogens more than hosts.
– faster metabolisms of parasites should allow them
to acclimate more quickly to unpredictable
temperature shifts, especially for ectothermic hosts
– parasites have fewer cells and processes to adjust
and generally withstand greater temperature
extremes than hosts (Portner 2002)
– owing to their shorter generation times, parasites
should evolve more quickly than hosts to changes
in climate
Climate Variability Hypothesis
• The categorically faster metabolisms,
smaller size, and greater reproductive
capabilities of parasites than hosts
provides a general hypothesis for how
global climate change will affect
disease risk– unpredictable climate
variability should increase disease.
Is there Spatiotemporal Spread of Atelopus
Extinctions?
Difference in year of decline
25
20
15
10
5
0
0
Rohr et al. 2008 PNAS
10
20
Distance
30
Atelopus Extinctions Through
Time
0.18
Best fit curve
Proportion Extinct
0.15
0.12
0.09
0.06
0.03
0.00
1979
Rohr et al. 2008 PNAS
1984
1989
Year
1994
1999
0.3
Partial residual (Last year observed)
0.2
How we controlled for
the likely epidemic
spread of the pathogen
0.1
0.0
-0.1
-0.2
-0.3
1980
1985
1990
1995
2000
Ultimate Hypothesis: ENSO
Drives Amphibian Declines
0.3
El Niño Years?
Partial residual (Last year observed)
0.2
0.1
0.0
-0.1
La Niña Years?
-0.2
-0.3
1980
1985
1990
1995
2000
Ultimate Hypothesis: ENSO
Rohr and Raffel 2010 PNAS
Must Control for Intrinsic Dynamics to
Detect Extrinsic Factors!
• No significant ENSO signature if we don’t
control for probable epidemic spread
• Hence, the availability of susceptible
hosts appears the primary factor
influencing epidemic spread followed
secondarily by climate
But What is the Proximate
Explanation?
What is it about El Nino years
that is associated with amphibian
extinctions?
Proximal Hypotheses for
Enigmatic/Bd-related Declines
•
Spatiotemporal spread hypothesis: declines are
caused by the introduction and spread of Bd,
independent of climate (Bell et al. 2004, Lips et al. 2006)
•
Climate-based hypotheses:
–
Chytrid-thermal-optimum hypothesis: Increased cloud cover,
due to warmer oceanic temperatures, leads to temperature
convergence on the optimum temperature for growth of Bd
(Pounds et al. 2006, Bosch et al. 2006)
–
Mean-climate hypothesis: changes in mean temp. and/or
moisture conditions affect the distributions of amphibians
(Whitfield et al. 2007, Buckley & Jetz 2007)
–
Climate-variability hypothesis: temporal variability in temp.
cause suboptimal host immunity facilitating declines (Raffel, Rohr,
et al. 2006)
Regional Predictors
tested w/ and w/o a one year lag
1. Mean absolute value of monthly
differences (AVMD) in temp.
2. Cloud cover x temp. (Pounds et al. 2006)
3. Cloud cover (Pounds et al. 2006)
4. Temperature-dependent Bd growth
(Pounds et al. 2006)
5. Precip. x temp. (Whitfield et al. 2007)
6. Diurnal temp. range (DTR)
7. Frost freq.
8. Precip.
9. Temp.
10. Temp. max.
11. Temp. min.
12. Vapor press.
13. Wet day freq.
Results of Best Subset Model
Selection
t
Adjusted No. of
Model
Wet day
2
Ranking
R
effects Precip.
freq.
1
0.685
3
0.253
2
0.671
3
0.230
3
0.644
3
4
0.643
3
5
0.640
3
6
0.640
3
7
0.640
3
8
0.640
3
9
0.640
2
10
0.640
3
t-1
Diurnal
AVMD Cloud temp.
Temp. Temp. Vapor
temp. cover range Temp. max. min. Pres.
0.859
0.764
0.845
0.755
0.857 -0.154 0.692
0.804
0.788
0.212
0.807
0.738
0.177
0.807
0.693 0.161
0.807
0.649
0.157
0.806
-2.350 2.453
0.892
0.699
0.806
1.306 -1.286
results are similar using AIC
Can Monthly Temperature Variability Explain
Atelopus Extinctions?
Rohr and Raffel 2010 PNAS
Amphibian extinctions have
often occurred in warmer
years, at higher elevations,
and during cooler seasons.
Are monthly and daily
variability in temperature also
greater at these times and
locations?
Do Warmer Years Have Greater
Variability in Temperature?
0.35
10.5
0.30
10.4
10.3
0.20
10.2
0.15
10.1
0.10
10
0.05
0.00
-0.50
0.00
0.50
9.9
1.00
Annual temperature anomalies ( C)
Rohr and Raffel 2010 PNAS
DTR ( C)
AVMD ( C)
0.25
Do High Elevations Have Greater
Variability in Temp.?
0.9
15
13
0.8
12
11
0.7
10
9
0.6
8
0-199 m 2001001- >2400 m
(n=1269) 1000 m 2399 m (n=210)
(n=668) (n=185)
Elevation categories
DTR (°C)
AVMD (°C)
14
Do Cooler Months Have Greater
Variability in Temp.?
0.8
18
0.7
16
15
0.6
14
13
0.5
23.0
23.4
23.8
24.2
Monthly temperature (ºC)
Rohr and Raffel 2010 PNAS
12
24.6
DTR >2400 m (°C)
AVMD (°C)
17
Results of Path Analysis
0.15
Lag LYO (detrended)
0.25
DTR
0.05
-0.15
-0.35
-2 -1 0
1
2
3
El Niño 3.4
P<0.001
0.675
-0.05
-0.15
-0.25
-0.35 -0.15 0.05 0.25
DTR
2
R =0.633
DTR
P<0.001
0.694
P=0.044
-0.466
P=0.002
0.673
AVMD
anomalies
R2=0.567
0.10
0.05
0.00
-0.05
-0.10
Lag amphibian
extinctions
2
R =0.674
P<0.001
0.868
Lag LYO (detrended)
El Niño 3.4
AVMD
0.05
0.15
0.05
-0.05
-0.15
-0.25
-2 -1 0
1
2
El Niño 3.4
3
-0.10
0.00
AVMD
0.10
Rohr and Raffel 2010
PNAS
We Weren’t Convinced
Experimental Test
• Acclimated Cuban tree frogs to 15 or 25⁰ C
for four weeks
• Challenged with Bd at start of week five
• Quantified survival and pathogen loads
15⁰C
15⁰C
25⁰C
25⁰C
Does Temperature Variability
Increase Bd Loads on Frogs?
Bd load:
Bd-induced mortality:
Temperature shifts increased Bd
loads and Bd-induced mortality
Raffel et al. in press Nature
Climate Change
Summary
• Availability of susceptible hosts appears to
be the primary factor influencing the spread
of Bd
• There is a strong ENSO signature to
extinctions after controlling for epidemic
spread
• Both field patterns of extinctions and
manipulative experiments support the
climate-variability hypothesis for amphibian
extinctions
Conclusions
Temperature, temperature
variability, and Central Pacific El
Niño events are increasing in
tropical and subtropical regions
because of climate change; thus,
global climate change might be
contributing to enigmatic amphibian
declines, by increasing disease risk
Conclusions
Elevated temperature variability might
represent a common, but under-appreciated,
link between climate change and both
disease and biodiversity losses and might
offer a general mechanism for why disease
would increase with GCC.
Parte 4: MODELO
CLIMÁTICO ECO
FISIOLÓGICO para
anfibios
Spea hammondi
Fisher and Shaffer 1996
Bufo boreas
Fisher and Shaffer 1996
Rana aurora (Bd induced?)
Fisher and Shaffer 1996
Preparación de modelos
• Molde de Látex a partir de un espécimen
de colección.
• Obtención de modelos de agar.
Experimento de campo
• 4 condiciones experimentales (factorial).
Día
SECO - SOL
(2 modelos)
HÚMEDO -SOL
(2 modelos)
SECO - SOMBRA
(2 modelos)
HÚMEDO - SOMBRA
(2 modelos)
Experimento de campo
• Modelos conectados a data loggers.
• Reemplazo de modelos cada 3-4 hr.
SECO
SOL
HÚMEDO
SOL
SECO
SOMBRA
HÚMEDO
SOMBRA
T operativa y Pérdida de agua
42
40
sombra húmedo
25
38
36
20
34
sol seco
15
32
30
10
28
5
09:00:00
26
17:00:00
01:00:00
Time
Del Puerto Canyon; 12 march 2012; nublado; Annaxyrus boreas model
09:00:00
Model mass (g)
Operative Temperature of the model (ºC)
30
T ambiental y Pérdida de agua
45
0.18
0.16
40
sol seco
0.14
Temperature (ºC)
35
0.12
30
0.10
sombra seco
0.08
25
sol húmedo
0.06
20
0.04
15
10
09:00:00
Tª sombra
11:00:00
sombra húmedo
13:00:00
Time
Los Baños Cistern; 5 march 2012; soleado; metamórficos
0.02
15:00:00
0.00
17:00:00
Relative Waterloss
Tª sol
Pseudacris sierra
Un sitio persistente
Un sitio Extinct
Del Puerto Canyon
Los Banos Well
4.00
Waterloss (g/hr)
4.00
Waterloss (g/hr)
3.50
3.00
2.50
3.50
3.00
2.50
Bufo
2.00
Bufo
1.50
1.50
Small
1.00
1.00
0.50
0.50
2.00
Small
0.00
0.00
Dry sun
Wet sun
Dry shade
Wet shade
Treatment
Dry night
Wet night
-0.50
Dry sun
Wet sun
Dry shade Wet shade
Treatment
Dry night
Wet night
Species Calibrations for CA Frogs
Hydric Loss Rates and Extinction, significant assoc:*,**,**
Low sensitivity to Bd:
• Spea hammondi
• Anaxyrus boreas***
• Anaxyrus punctatus
Immune to Bd:
• Pseudacris sierra***
• P. cadaverina*
• P. hypochondriaca
Very sensitive to Bd:
Rana sierrae***
Rana aurora
Bufo americanus
distancia recorrida en 10 minutos de la locomoción forzada
Preest and Pough, Funct Ecol 1989
• Víctor H. Luja lujastro@yahoo.com
• Octavio Jiménez octavio.jimenez.robles@gmail.com
• Eric Curiel ecuriell@ucsc.edu
Desarrollando un MODELO CLIMÁTICO ECO
FISIOLÓGICO para anfibios
• 1. Buscamos un modelo que explique mejor el límite de distribución de
las especies. Buscamos encontrar el conjunto de tasas de pérdida de
agua (y de temperatura) durante el día vs la noche que explique los
límites más secos (el umbral de desecación) debajo del cual la especie no
se encuentra mas. Este umbral que mejor ajusta está basado en las
ubicaciones de todas las poblaciones a partir de 1970.
• 2. Este umbral hídrico de pérdida de agua es conceptualmente similar al
h_r para las horas de restricción de actividad debido al calor.
• 3. Entonces, nosotros probaremos como se comporta este modelo para
predecir la distribución actual de las extinciones (por ejemplo cambios
en la pérdida hídrica del delta entre 1975–2010, debido al incremento
en la intensidad de la sequía).