Ecological Archives E095-282-A2

Karen M. Alofs, Donald A. Jackson. 2014. Meta-analysis suggests biotic resistance in freshwater environments is driven by consumption rather than competition. Ecology 95:3259–3270. http://dx.doi.org/10.1890/14-0060.1

Appendix B. Model selection and model averaging used for examining the factors determining biotic resistance in freshwater studies.

Table B1. Parameters fitted for predictive models of freshwater biotic resistance and their relative variable importance (summed Akaike weights) in model averaging.

Variable

relative
variable
importance

resistance mechanism (competitive or consumptive)

1.00

habitat (lotic, lentic, or reservoir)

1.00

experimental method (enclosure, exclosure, other)

0.96

latitude of study

0.49

measure of effect (establishment or performance)

0.32

length of study (days)

0.27

 

Table B2. Strongest predictive models of biotic resistance. D AICc is the difference in AICc between a given model and that with the lowest AICc observed. Akaike weight (wi) is the model likelihood, exp(-D AICc /2), over the sum of all model likelihoods; values near 1 indicate greater confidence. Models listed are those selected for model-averaging such that the cumulative sum of wis =0.9.

Model

df

AICc

ΔAICc

Wi

mechanism + habitat + method + latitude

9

283.04

0

0.24

mechanism + habitat + method

8

283.10

0.06

0.24

mechanism + habitat + method + latitude + measure

10

284.13

1.09

0.14

mechanism + habitat + method + measure

9

284.59

1.55

0.11

mechanism + habitat + method + length

9

285.99

1.95

0.09

mechanism + habitat + method + latitude + length

10

285.53

2.49

0.07

mechanism + habitat + length

7

286.83

3.79

0.04

mechanism + habitat + method + latitude +  measure + length

11

286.86

3.82

0.04

mechanism + habitat + method + measure + length

10

286.90

3.87

0.04

 

Table B3. Standardized coefficient estimates of effect size and their 95% confidence intervals from model averaging. Those estimates not overlapping 0 are in bold.

Variable

estimate

95% CI

mechanism: consumption (in contrast to competition)

0.553

0.338 – 0.768

habitat: lotic (in contrast to lentic)

-0.423

-0.608 – -0.237

habitat: reservoir (in to contrast to lentic)

-0.184

-0.367 – -0.002

method: exclosure (in contrast to enclosure)

-0.288

-0.493 – -0.083

method: other (in contrast to enclosure)

0.058

-0.126 – 0.241

latitude

0.156

-0.029 – 0.341

length

-0.067

-0.279 – 0.145

measure: performance (in contrast to establishment)

-0.111

-0.305 – 0.083


[Back to E095-282]