osvaldo sala. projected precipitation change 1970-99 vs 2071-99 us national climate assessment 2014
TRANSCRIPT
The Effect of Climate Change on Arid and Semi-Arid Ecosystems:
Directional Changes in Precipitation Amount and
Variability
Osvaldo Sala
Projected Precipitation Change
1970-99 vs 2071-99
US National Climate Assessment 2014
Projected Changes in Precipitation
Precipitation Variability is projected to increase
IPCC. 2013. Climate Change 2013: The physical science basis. Contribution of working group I to the fifth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New York, NY, USA.
Grassland <-> Shrubland
Hypothesis 2 Directional changes in water availability that favor grasses over shrubs or shrubs over grasses are reinforced through time.
Jornada LTER VI Project
Our scientific approach
Observations Experiments
Data Mining Modelling
Observation Multiyear Precipitation trend
+
Peters et al (2011)
Consequences of Multiyear Precipitation trend
+
Peters et al (2011)
Experimentation
2014
ambient
+ 80%
- 80%
Solar panel
Battery
Pump
Intermediary tank
(55 gal.)
Float switchInterception
plot
Irrigation plot
Filter
ARMSautomated rainfall
manipulation system
(Plots trenched to 40-60 cm) Gherardi and Sala, Ecosphere: 2013
• Assess ecosystem sensitivity to precipitation
• Not to replicate climate change scenarios
Experimental Design Objectives
CCImpact = ʄ (Δ Climate, Ecosystem Sensitivity)
PPTImp= ʄ (Δ PPT, Ecosystem Sensitivity to PPT)
Climate Change Impact
Total Aboveground Net Primary Production
+ 80%
Ambient
- 80 %
Grass Aboveground Net Primary Production
+ 80%
Ambient
- 80 %
Shrub Aboveground Net Primary Production
Ambient
+ 80 %
- 80 %
Direct and Indirect Effects of Precipitation
Plant-Species Diversity H’
+ 80%
Ambient
- 80 %
There was an effect of time on ecosystem response variables to long-term changes in PPT
The effect of time varied for different response variables
Asymmetry Hypothesis – The absolute magnitude of the effect was different for increasing or decreasing PPT
Conclusions
Effects of Enhanced Precipitation Variability
10 reps * 5 treat = 50 plots
(2.5 x 2.5 m)
Trenched 60 cm deep
20 rainout shelters
20 irrigated plots
10 control plots
Methods
Methods
Effect of PPT Variability
Gherardi & Sala PNAS 2015
The Mechanism
Gherardi & Sala PNAS 2015
How do we explain these responses?
Sala, Gherardi, Peters Climatic Change 2015
Modelling
Gherardi & Sala PNAS 2015
Effect of PPT Variance Increases through Time
Demise of grasses under high PPT variability favors shrubs
Further explore the existence of thresholds ◦ Cumulative endogenous ◦ Stochastic exogenous◦ Interaction between endogenous and exogenous
Mechanisms for indirect effects
Future Studies
Thank youLaureano Gherardi, Lara Reichmann Courtney Currier Kelsey DuffyOwen McKennaJosh Haussler
Pulse and Press
Collins et al 2011
Smith MD et al (2009)
Press Conceptual Model
Hypothesis 1
a) Ecosystem response variables are proportional to water availability
increased water
ambient
decreased water
Time
Eco
syst
em
resp
onse
vari
able
Response variable = b0 + b1*PPT
0 +
+
Hypothesis 1
b) Ecosystem response variables are proportional to changes in water and to the time that the ecosystem has been exposed to the new condition
increased water
ambient
decreased water
TimeEco
syst
em
resp
onse
vari
able
Response variable = b0 + b1*PPT + b2*Time
+
+
-
0
Multiyear PPT trend
0 +
+
Peters et al (2011)
Consequences of Multiyear PPT trend
0 +
+
Peters et al (2011)
Hypothesis 1
c) Acclimation / exhausting of resources
in-creased water
am-bient
TimeEco
syst
em
resp
onse
vari
able
Response variable = b0 + b1*PPT + b2*Time + b3*Time*PPT
0 +
+
-
Hypothesis 2 The effect of time is asymmetric for reduced
water and increased water
increased water
ambient
decreased water
TimeEco
syst
em
resp
onse
vari
able
0 +
+
-
Hypothesis 3 The effect of time varies for different response variables
response variable A
response variable B
TimeEco
syst
em
resp
onse
vari
able
0 +
+
-
Solar panel
Battery
Pump
Intermediary tank
(55 gal.)
Float switchInterception
plot
Irrigation plot
Filter
ARMSautomated rainfall
manipulation system
(Plots trenched to 40-60 cm) Gherardi and Sala, Ecosphere: 2013
Total ANPP
Shrub ANPP
Grass ANPP
Spp Richness
Diversity
PPT
Conclusions Rejected H1a. There was an effect of
time on ecosystem response variables to long-term changes in PPT, due to legacies in the ecosystem response.
Asymmetry – The absolute magnitude of the effect was different for increasing or decreasing PPT, i.e. spp loss
with drought – no spp change with increased PPT
The effect of time varied for different response variables; may depend of the number of actors involved or the flow size relative to the pool size
Thank youLaureano GherardiLara ReichmannOwen McKennaJosh HausslerKelsey Duffy
Jose Anadon
NSF-Division of Environmental Biology
Jornada Basin LTER
Jornada Experimental Range - USDA
School of Life Sciences - ASU
AcknowledgmentsLara G. ReichmannR.C.A. GuchoOwen P.B.R. McKennaLaura YahdjianDeb PetersKelsey McGurrinJosh HausslerJohn Angel IIIShane & MiriamG. A. Gil
Funding sources
Contrasting productivity responses to interannual precipitation variability
Laureano A. Gherardi and Osvaldo E. SalaArizona State University, School of Life Sciences
Results of 6 years of precipitation manipulation at the Jornada Basin LTER
Projected Changes in Precipitation
Precipitation Variability is projected to increase
IPCC. 2013. Climate Change 2013: The physical science basis. Contribution of working group I to the fifth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New York, NY, USA.
Objective: To study the effect of inter-annual precipitation variability per se
Interannual Precipitation Variability
ANPP Mean
ANPP CV
Productivity
Stability
NSF-Division of Environmental Biology
Jornada Basin LTER
Jornada Experimental Range - USDA
School of Life Sciences - ASU
AcknowledgmentsLara G. ReichmannR.C.A. GuchoOwen P.B.R. McKennaLaura YahdjianDeb PetersKelsey McGurrinJosh HausslerJohn Angel IIIShane & MiriamG. A. Gil
Funding sources
ANPP, PPT, Space
ANPP = - 45.13 + 0.67*MAP r2= 0.76
Bai et al (2008)
ANPP=-34+0.60*MAPr2=0.94
Sala et al (1988)
ANPP=-30+0.47*MAP
McNaughton et al (1989)
Gre
at P
lain
s
Sou
th
Am
eric
a
Mon
goli
an
Pla
teau
Message A simple model accounts for a large fraction
of ANPP variability across space and for most grasslands of the world
0 500 1000 15000
200
400
600
800
1000
Annual Precipitation (mm)
Temporal ModelSpatial Model
Ab
ove
gro
un
d N
et P
rod
uct
ion
(g
/m2 /
yr)
Lauenroth and Sala 1992 Ecological Applications 2:397-403
-34 + 0.60*MAPr2 = 0.94, p < 0.001
56 + 0.13*MAPr2 = 0.39, p < 0.001
Spatial vs. temporal models of net primary production
Sala et al 2012, Philosophical Transactions of the Royal Society B
Spatial vs. Temporal models of net primary production
r2=0.39
Sala et al 2012, Philosophical Transactions of the Royal Society B
Message Time and Space cannot be exchanged for
the ANPP-MAP relationship
Spatial model does not work through time
Temporal model only accounts for a small fraction of the variability explained by spatial models and has shallower slope
Hypothesis
Differences between spatial and temporal models are explained by time lags in ecosystem response to changes in water availability
Legacies Time lags result from legacies of wet and
dry years
ANPP observed = F (PPTt, Legacy)
Legacies = ANPP observed – ANPP expected
◦ ANPP expected = F (PPTt)
Magnitude of Legacy= F (PPTt-1 – PPTt)
Global patterns of LegaciesMagnitude of Legacy= F (PPTt-1 – PPTt)
What is the shape of F?
How does this relationship change across a PPT gradient?
Legacy Symmetry Hypotheses
Sala et al 2012 PTRSB
Legacy Symmetry Hypotheses
Sala et al 2012 PTRSB
H 3.2
Knapp and Smith (2001)
Effect of previous-year PPT on ANPP across sites
Sala et al 2012, PTRSB
Effect of current year PPT on ANPP
Sala et al 2012, PTRSB
Effect of previous-year PPT on ANPP across sites
Sala et al 2012, PTRSB
Effect of previous-year ANPP on current-year ANPP across sites
Sala et al 2012, PTRSB
Effect of current- and previous-year PPT along a PPT gradient
Sala et al 2012, PTRSB
Experimental ApproachChihuahuan Desert Grassland
Jornada LTER
MAP 240 mm Dominant
species:◦ Bouteloua
eriopoda C4◦ Prosopis
glandulosa C3
Jornada Experimental RangeChihuahuan Desert Grassland
Experimental design
Fixed rainout shelters intercept different amounts of rain, depending on the number of shingles
irrigation
Water was added to the increased PPT treatments after each PPT event, year around
Total 132 plots
Legacy Magnitude
Reichmann, Sala, Peters, Ecology 2013
Legacy = -2.71 + 0.05 * ∆PPTR2 = 0.42
Rainout shelters in the Patagonian steppe
Yahdjian and Sala (2002)
Precipitation input (mm/year)
AN
PP
(g.
m-2.y
r-1)
PPT mm/year
50
70
90
110
130
20 60 100 140 180 220
without drought legacy
after 80% rainfall interception
after 55% rainfall interception
after 30% rainfall interception
Yahdjian and Sala (2006)
Drought legacies in the Patagonian Steppe
Conclusion Changes in precipitation result in legacies
Magnitude of Legacies is a function of difference in precipitation of current and previous year
Legacies in the Chihuahuan desert ecosystem are symmetrical
◦ │ Positive legacy │ = │ Negative legacy│
Corollary
Positive legacies would compensate negative legacies
Increased precipitation variability would not affect average productivity
Hypotheses for the Legacy Mechanisms Structural mechanism
◦ Meristem density constrains production response to a wet year after a dry year
◦ Meristem density enhances production after wet years
Biogeochemical mechanism◦ N limitation constrains production response to a wet
year after a dry year
◦ Abundant reactive N enhances production after wet years
Soil moisture carry-over
Structural mechanism
Reichmann, Sala, Peters, Ecology 2013
Structural mechanism
Reichmann, Sala, Peters, Ecology 2013
(Reichmann and Sala, Functional Ecology 2014)
Structural mechanisms
Structural mechanisms
(Reichmann and Sala, Functional Ecology 2014)
Biogeochemical mechanism
Reichmann, Sala, Peters, Ecology 2013
Biogeochemical mechanism
N mineralization effect
Reichmann et al Ecosphere 2012
N uptake and leaf N concentration
Reichmann et al Ecosphere 2012
N stocks
Reichmann et al Ecosphere 2012
Test of the soil-moisture carry-over hypothesis
• Tiller density determines magnitude of
legacies
• Biogeochemical mechanisms do not
determine legacies
• Soil water carry-over does not determine
legacies
Conclusions
Pulse and Press
Collins et al 2011
Present Future
+
Pre
cipi
tati
onTime
Hypothetical pattern
Observational
Short term manipulations
Most studies are
Central question
Can we predict press effects of directional changes in precipitation amount and variability based upon our understanding of pulse responses?
Smith MD et al (2009)
Proposal Hypothesis 2 a
Hypothesis 1
a) Ecosystem response variables are proportional to precipitation
Response variable = b0 + b1*PPT
Increased precipitation
Ambient
Decreased precipitation
0 +
+
Eco
syst
em
resp
onse
var
iabl
e
Time
b) Ecosystem response variables are proportional to changes in precipitation and to the time that the ecosystem has been exposed to the new condition
Hypothesis 1
Response variable = b0 + b1*PPT + b2*Time
Increased precipitation
Ambient
Decreased precipitation
0 +
+
Eco
syst
em
resp
onse
var
iabl
e
Time
c) Acclimation / exhausting of resources
Hypothesis 1
Response variable = b0 + b1*PPT + b2*Time + b3*Time*PPT
Increased precipitation
Ambient
Decreased precipitation
0 +
+
Eco
syst
em
resp
onse
var
iabl
e
Time
The effect of time varies for different response variables
Hypothesis 2
Increased precipitation
Ambient
Decreased precipitation
0 +
+
Eco
syst
em
resp
onse
var
iabl
e
Time
The effect of time is asymmetric for reduced and increased precipitation
Hypothesis 3
Increased precipitation
Ambient
Decreased precipitation
0 +
+
Eco
syst
em
resp
onse
var
iabl
e
Time
Central question
Can we predict press effects of directional changes in precipitation amount and variability based upon our understanding of pulse responses?
Individual Species Response + 80%
- 80%
Species Richness
+ 80%
Ambient
- 80 %
Mechanism
Linear and non-linear ANPP responses to
precipitation
Increased precipitation variance implies a higher frequency of extremely dry and wet years
1.Linear and non-linear ANPP responses to annual precipitation
Precipitation
An
nu
al A
NP
P (
g m
2 yr
-1)
Therefore, NULL precipitation variance effect on ANPP.
Linear ANPP responses to precipitation result in negative effects of dry years equal to positive effects of wet years.
Precipitation
An
nu
al A
NP
P (
g m
2 yr
-1)
Non-linear ANPP responses to precipitation result in different effects of dry and wet years.
Therefore, POSITIVE precipitation variance effect on ANPP.
Precipitation
An
nu
al A
NP
P (
g m
2 yr
-1)
Non-linear ANPP responses to precipitation result in different effects of dry and wet years.
Therefore, NEGATIVE precipitation variance effect on ANPP.
Precipitation
Interannual Precipitation Variability
ANPP Mean
ANPP CV
Productivity
Stability
Plant-functional types show different stability response to PPT variability
Ecosystem stability results from the aggregated response of plant types
Functional diversity increases with PPT variability
Changes in relative abundance support such effect
How do we explain these responses?
Concluding summary: effects on ANPP mean
1. Inter-annual precipitation variability itself has a negative effect on ANPP
2. Non-linear responses and changes in soil water distribution explain such effect
3. Aggregated plant-functional type responses determine overall ecosystem response
Concluding summary: effects on stability
1. Interannual precipitation variability has a positive effect on ANPP CV
2. Contrasting plant-functional type responses result in relative abundance change and increased diversity
3. Aggregated response of plant-types determines ecosystem stability