Grassland ecosystems and climate change: Insights from experiments

Alan K. KnappAlan K. KnappDepartment of BiologyDepartment of Biology

Graduate Degree Program in EcologyGraduate Degree Program in Ecology

Year

1860 1880 1900 1920 1940 1960 1980 2000

Glo

bal p

opul

atio

n (m

illio

ns)

0

1000

2000

3000

4000

5000

6000

Car

bon

emis

sion

s (G

t)

0

2

4

6

8

10

Atm

osph

eric

CO

2 (p

pm)

280

300

320

340

360

380

Carbon

Population

CO2

Year

1860 1880 1900 1920 1940 1960 1980 2000

Tota

l rea

ctiv

e N

(Tg

N y

-1)

0

20

40

60

80

100

120

140

160

180

Nitrogen

Year

1860 1880 1900 1920 1940 1960 1980 2000

Glo

bal t

empe

ratu

re a

nom

aly

(C)

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

Temperature

We have moved beyond questions such as “Is climate change happening?” and “What is the cause?”…The uncertainty is with the consequences of climate change…

…we are changing Earth more rapidly than we are understanding it.

-Vitousek et al. 1997

What do we know?We know that climate change is chronic and directional,

but also that variability will be increased

Rainfall regimes

CO2 effects

Temperature effects

Soil moisture+

+ -

-

Global changes in rainfall, temperature and CO2 can be linked in most ecosystems through their indirect effects…

OrganismWater relations

Ecosystem Process & Function

Sala et al. 1988

NA Grasslands (most) are water limited systems and precipitation amount

predicts spatial patterns of ANPP well…

Rainfall vs. Production across the Central US

What do we need to know?How will climate change affect earth’s ecosystems?

Can we forecast the future?

"I would not say that the future is necessarily less predictable than the past. I think the past was not predictable when it started."

~Former Defense Secretary Donald Rumsfeld

Resources

Biota

A. Disturbance

Resources

Biota

B. Global Change

Chronic resourcealteration

Discretedisturbance

Moving from a “disturbance centric” world

How we “think” - A resource perspective…

- Smith, Knapp and Collins - Ecology 2009

Mesic

Grasslands of North AmericaThe Flint Hills of Kansas –

The Konza

Prairie

• Tallgrass prairie is the most mesic of Great Plains grasslands – sufficient precipitation to support forest

• Flint Hills are located at the dry edge of the original range of tallgrass prairie – Mean annual precipitation = 835 mm

KPBS

C4 grass dominance

Diverse C3 forbs

A dynamic & tractable system

Chronic (water) resource addition

What is the degree of water limitation in this “mesic” grassland?If water is limiting this system, what are the consequences of a chronic change in resource abundance (soil moisture)?

Long-term Irrigation Study

• Treatments initiated in 1991

• Supplemental water added during the growing season to replicate 140 m transects (paired with control transects)

• Designed to meet plant water demand and minimize intra-annual variability in soil water deficits

Summary of the first eight years…

• Water availability limited ANPP 6 out of 8 years

• Irrigation increased ANPP by ~25% (physiological response)

Tota

l AN

PP

(g/m

2 )

300

400

500

600

700

800

900

Year1992 1994 1996 1998

PP

T (mm

)

0

200

400

600

800

1000

1200

Control

Irrigated

*

* *

*

*

*

Annual ppt + irrigation (mm)600 800 1000 1200 1400

AN

PP

(g/m

2 )

300

400

500

600

700

800

r2 = 0.81

Annual ppt + irrigation (mm)600 800 1000 1200 1400

AN

PP

(g/m

2 )300

400

500

600

700

800

900

r2=0.52

1991-1998irrigated

The value of long-term experiments…(+15 years)

• Mean increase in ANPP for the next nine years was 70% (vs. 25%)

• An 8 year lagged response…

65%

1999 – 2007 irrigated

Control

Physiological response

Community re-ordering

Community change

Res

pons

e

Time

Resource alteration

Physiological response

Community re-ordering

Community change

Res

pons

e

Time

Resource alteration

Smith, Knapp and Collins 2009

Irrigation water added (mm)0 100 200 300 400 500

Stim

ulat

ion

in A

NP

P (g

/m2 )

0

100

200

300

400

500

r2 = 0.66

= 1999-2003

= 1991-1998

1999 - 2007 What would change the relationship between ANPP and precipitation?

8 yrs

9 yrs

Difference in Absolute cover between Irrigation and Control transects

Date1990 1992 1994 1996 1998 2000 2002 2004 2006 2008

Abs

olut

e C

over

(%)

-20

-10

0

10

20

30

40

Panicum virgatum

Δ ANPP = 355.2 g/m2

Δ ANPP = 145.2 g/m2

Precipitation (mm)

0 200 400 600 800 1000 1200 1400 1600

AN

PP

(g/m

2 )

0

200

400

600

800

1000

Sala et al. 1988 relationshipControl plotsControl relationship17-yr mean control plots

Initial 8-yr mean, irrigated lowlandsNext 9-yr mean, irrigated lowlands

Proportion of total rainfall in the US from large (>5 cm) rainfall events

Prop

ortio

n (%

) of r

ainf

all i

n “e

xtre

me

even

ts”

Abundant evidence that rainfall regimes are changing

b.

Year

8.0

10.0

12.0

r2=0.704

c.

6.0

7.0

8.0

r2=0.638

a.

12.0

14.0

16.0

r2=0.534

Rai

n In

tens

ity (m

m/e

vent

)

1940 1960 20001970 19801950 1990

Year

Oklahoma

Question:

Given increased variability and frequency of extreme events, to what extent will altered precipitation patterns affect ecological processes in “mesic” grasslands?

12 rainfall manipulation plots (RaMPs) + non-sheltered controls

Collect, store, and reapply natural rainfall on intact prairie plots

Treatments include ambient and altered rainfall patterns

The Rainfall Manipulation Plot (RaMP) Experiment

Address the impact of changes in size and timing of growing season rainfall

Target altered rainfall pattern: 50% longer inter-rainfall periods, larger individual rain events

Current

Predicted

Pre

cipi

tatio

n am

ount

Day of year

No change in total precipitation quantity

Current

Predicted

Pre

cipi

tatio

n am

ount

Day of yearFewer events

Greater storm size

Increased dry periods

Ambient ppt timing: • “typical” seasonal

pattern

Altered ppt timing:• repeated deficits• more extreme wetting

and drying cycles

Key driver of responses: Soil moisture dynamicsPrecipitation (m

m)0

20406080100

Soil

wat

er c

onte

nt 0

-30

cm (%

)10

20

30

40

50

DateJun Jul Aug Sep Oct

10

20

30

40

50

020406080100

Ambient Altered

Cha

nge

in S

WC

(%)

per

sam

plin

g in

terv

al

0.0

2.5

5.0

7.5

10.0Mean SWC = 24.7 ± 1.4%

P = 0.01

Mean SWC = 29.5 ± 1.2%A - Ambient rainfall treatment

B - Altered rainfall treatment

Precipitation (mm

)

0 10 20 30 40 50

Dep

th (c

m)

0-10

20-30

40-50

60-70

80-90

Root mass (% of total)

Average soil water content in top 30 cm:- reduced by 12%

Variability in soil moisture:- increased by 27%

Year1998 2000 2002 2004 2006 2008

AN

PP

(g m

-2)

300

400

500

600

700

800

900

1000 ambient timingaltered timing

*

* *

*

pre-treatment

*

ANPP responses to altered rainfall timing

13-22% reduction in years with significant responses

Across all years and all rainfall treatments

CV of soil water content (0-15 cm)20 30 40 50 60

AN

PP

(g m

-2)

400

500

600

700

800

900

1000

r2 = 0.57

Ambient timingAltered timing

Ojima and Lackett 2002

Tall-grass

Mid-grass

Short-grass

Precipitationgradient

A regional assessment

Jana Heisler-White et al. 2009. Contingent productivity responses to more extreme rainfall regimes across a grassland biome.Global Change Biology 15:2894-2904.

SGS

Experimental alteration of rainfall patterns across the Great Plains of North America

Frequent, small events

Fewer, infrequent large events

VS.

A = 12 eventsB = 6 eventsC = 4 eventsTotal = 191 mm

A = 12 eventsB = 6 eventsC = 4 eventsTotal = 450 mm

KNZ

Semi-arid Mesic

Jana Heisler

Mesic

GrasslandKNZ 2006

# of events

Ambient 12 events 6 events 4 events

Prod

uctio

n (g

/m2 )

100

200

300

400

500

600

700a

bc c

A shift from 12 to 4 events results in a 15% decrease

in

productivity.

More “extreme” rainfall regime

Shift to more extreme (large) rainfall events reduced productivity

Semi-arid Grassland

# of rain events

Ambient 12 events 6 events 4 events

Pro

duct

ion

(g/m

2 )

20

40

60

80

100

120

140

a

a

bb

A shift from 12 to 4 events resulted in a 30% increase in productivity.

More “extreme” rainfall regime

Shift to more extreme (large) rainfall events increased productivity

Why do arid vs. mesic grasslands respond differently?

Two aspects of more extreme rainfall regimes:

Current

Predicted

Pre

cipi

tatio

n am

ount

Day of yearFewer events

Greater storm size

Increased dry periods

1. Larger rain events2. Longer dry periods between events

When soils are usually moist, longer dry periods are most important…

When soil are already dry, larger rain events cause the greater response…

Where do we need to go with future field experiments?We need more long-term experiments where we “push” ecosystems…We need comparative analyses of ecosystem responses to climate change at large spatial scales…

Acknowledgements:

John Briggs, Scott Collins, John Blair, Patrick O’Neal, Melinda Smith, James Koelliker, Jana Heisler-White

Colorado State UniversityColorado State University

Kansas State UniversityKansas State University

KonzaKonza Prairie Biological StationPrairie Biological Station

USDA Managed Ecosystems ProgramUSDA Managed Ecosystems Program

NSF NSF –– LTER & Ecosystem Studies ProgramLTER & Ecosystem Studies Program

Department of Energy Program in Ecosystem ResearchDepartment of Energy Program in Ecosystem Research