Null models and observed patterns of native and exotic diversity: Does native richness repel invasion? Rebecca L. Brown, 1,2 Jason D. Fridley, 1 and John.

Null models and observed patterns of native and exotic diversity: Does native

richness repel invasion?

Rebecca L. Brown,Rebecca L. Brown,1,21,2 Jason D. Fridley, Jason D. Fridley,11

and John F. Brunoand John F. Bruno11

11University of North Carolina-Chapel HillUniversity of North Carolina-Chapel Hill22Patrick Center for Environmental Research Patrick Center for Environmental Research

Diversity

Inva

sio

n

Prieur-Richard et al. 2000 Stachowitz et al. 2002 Dukes 2002 Tilman 1997 Hector et al. 2001 Knops et al. 1999

(mostly experimental)

Stohlgren and Chong 2002 Wiser et al. 1998 Bruno et al. 2002 Burger et al. 2001 Sax 2002Lonsdale 1999

(mostly observational)

Levine 2001Lavorel et al. 1999. Brown and Fridley 2003 Duncan 1996Stohlgren et al. 1998Brown and Peet 2003

Does diversity control invasion?

Diversity

Inva

sio

n+ -

Confounded or neutral relationship

Small scales:

Saturation, plant to plant competition

= negative relationship

Spatial scale effects

Larger scales:

Variation in other factors (disturbance, propagules, fertility)

= positive or no relationship

0 2 4 6 8 100

12

30 5 10 15 20

01

23

4

0 1 2 3 4 5

01

23

Native species richness

Exo

tic s

peci

es r

ichn

ess

100 m2 1 m2 0.1 m2 0.01 m2

0 20 40 60 80 100

010

2030

40Southern Appalachian Riparian Plant Communities

NS p = 0.02p < .001 p = 0.001

What is the scale-dependence of the native-exotic richness relationship in a randomly assembled community?

BUT – do these relationships imply biological mechanisms or could they be observed in randomly assembled

communities?

Randomly assembled communities: the null model

• Create simulated communities with native and exotic species sampled at multiple scales

• Randomize native and exotic species codes in real communities

5

Species Pool:

75% Native spp15% Exotic spp10% Blank

Richness varied randomly, 20-100 spp

Abundance varied randomly, 1 to 10,000 individuals

Quadrats:

6 scales

Species drawn from pool

Simulation of randomly assembled communities

800

100

50

20

10

Simulation method

Randomly pick # of sp in pool, assign abund. to each sp

Randomly pick 1 “individual” from pool

Add one individual of that sp to quadrat

Repeat until quadrat is filled

(5 to 800 spaces)

When full, repeat process 100 times for each quadrat scale

20 – 100 spp1-10,000 individuals

Pool: 75% native 15% exotic 10% blank

Exo

tic R

ichn

ess

Native Richness

N = 800

Simulation of random quadrats

Exo

tic R

ichn

ess

Native Richness Native Richness

N = 800 N = 100

Simulation of random quadrats

Exo

tic R

ichn

ess

Native RichnessNative Richness Native Richness

N = 800 N = 100 N = 50

Simulation of random quadrats

Exo

tic R

ichn

ess

Native RichnessNative Richness

Native Richness

Exo

tic R

ichn

ess

N = 800 N = 100 N = 50

N = 20

Simulation of random quadrats

Exo

tic R

ichn

ess

Native Richness Native Richness

Exo

tic R

ichn

ess

Native Richness

N = 800 N = 100 N = 50N = 50

N = 20 N = 10

Simulation of random quadrats

Exo

tic R

ichn

ess

Native Richness Native Richness Native Richness

Exo

tic R

ichn

ess

N = 800 N = 100 N = 50

N = 20 N = 5N = 10

Simulation of random quadrats

0 50 100

Native species richness

05

01

00

Ex

oti

c s

pe

cie

s r

ich

ne

ss

Constraints on high native-exotic richness at smallest scales

100 individuals

Null relationship

Species pool: 75 Natives, 15 Exotics

0 50 100

Native species richness

05

01

00

Ex

oti

c s

pe

cie

s r

ich

ne

ss

Constraints on high native-exotic richness at smallest scales

100 individuals

50 individuals

Null relationship

Species pool: 75 Natives, 15 Exotics

0 50 100

Native species richness

05

01

00

Ex

oti

c s

pe

cie

s r

ich

ne

ss

Constraints on high native-exotic richness at smallest scales

100 individuals

50 individuals

Null relationship

Species pool: 75 Natives, 15 Exotics

10 individuals

0 50 100

Native species richness

05

01

00

Ex

oti

c s

pe

cie

s r

ich

ne

ss

100 individuals

50 individuals

Null relationship

Species pool: 75 Natives, 15 Exotics

10 individuals

0

1

2

3

4

5

6

0 1 2 3 4 5 6

Native Species

Inva

sive

Sp

ecie

s

Summary – Simulated Data

• In simulated randomly assembled communities, the relationship between native and exotic richness is positive at large scales, and negative at small scales

• Positive: because plots differ in total richness; slope is simply ratio of natives to exotics in the species pool

• Negative: due to constraints on total richness at very small scales

Next: real data

To test whether observed patterns of native and exotic species richness are different from pattern generated by random assembly (the null expectation):

– Randomize native and exotic species labels in the species pool

Permutation tests for observational data

Species pool

Sp A

Sp B

Sp C

Sp D

Nativity Label

Native

Exotic

Native

Native

Permuted Label

Exotic

Native

Native

Native

• Calculate correlation coefficient (r) or slope (s)Calculate correlation coefficient (r) or slope (s)

• Repeat 500x, compare null distribution to real valueRepeat 500x, compare null distribution to real value

0 2 4 6 8 100

12

30 5 10 15 20

01

23

4

0 1 2 3 4 5

01

23

Native species richness

Exo

tic s

peci

es r

ichn

ess

100 m2 1 m2 0.1 m2 0.01 m2

0 20 40 60 80 100

010

2030

40Actual results: Riparian plant communities

NS p = 0.02p < .001 p = 0.001

Permutation results: Riparian plant communities

0 2 4 6 8 100

12

30 5 10 15 20

01

23

4

0 1 2 3 4 5

01

23

r: NSs: NS

r: NSs: p < .05

r: p < .001s: NS

Native species richness

Exo

tic s

peci

es r

ichn

ess

100 m2 1 m2 0.1 m2 0.01 m2

0 20 40 60 80 100

010

2030

40

r: NSs: NS

Coastal plant communities, 24 sites at 500 m2

4 6 8 10 12 14 16

Site native richness

46

81

01

21

4

Site

exo

tic r

ichn

ess

Coastal plant communities, 24 sites at 500 m2

4 6 8 10 12 14 16

Site native richness

46

81

01

21

4

Site

exo

tic r

ichn

ess

0.0 0.2 0.4 0.6 0.8

Correlation coefficient (r)

Prob

ability

Observed r

Conclusions

• The native-exotic richness relationship is scale-dependent, BUT, this is the null expectation– With our null model – “competition” for space

among individuals, not species

• It is important to consider the null expectation when evaluating mechanistic explanations for patterns in data

Acknowledgements

Advising and DiscussionAdvising and Discussion: Bob Peet, Peter White, : Bob Peet, Peter White, Jim McNair, UNC Plant Ecology LabJim McNair, UNC Plant Ecology Lab

FundingFunding: National Science Foundation, UNC : National Science Foundation, UNC Graduate School, UNC Department of Biology, Graduate School, UNC Department of Biology, UNC Ecology Curriculum, Sigma Xi, The Nature UNC Ecology Curriculum, Sigma Xi, The Nature Conservancy, USDA National Forest Service, Conservancy, USDA National Forest Service, Patrick Center for Environmental ResearchPatrick Center for Environmental Research

Field crewsField crews