Table 1 Differences in herbivore composition and herbivore function between exposure regimes.

From: Wave exposure reduces herbivory in post-disturbed reefs by filtering species composition, abundance and behaviour of key fish herbivores

Model

Random effects

Fixed effects

Goodness of fit

atoll

site

residual

 

coefficient

se

t

p

R2m

R2c

a. Species richness richness.transect1 ~exposure random= atoll/site

0.665

1.963

0.675

Intercept

13.666

0.930

14.683

<0.005

0.673

0.968

   

Exposure (Exposed)

−6

1.198

−5.004

0.001

  

b. Biomass log(kg.100 m−2) ~exposure random= atoll/site

0.298

0.532

0.203

Intercept

5.889

0.289

20.336

<0.005

0.323

0.932

   

Exposure (Exposed)

−0.851

0.329

−2.588

0.032

  

c. Bite rates sqrt(bites.min−1.m−2) ~ exposure random= atoll/site

0.383

0.352

0.154

Intercept

1.532

0.271

5.647

<0.005

0.515

0.608

   

Exposure (Exposed)

−0.796

0.222

−3.584

0.007

  
  1. Linear Mixed models results for a. Total rarefied species richness (richness. transect−1), b. Total biomass (kg. 100 m−2) and c. Total bite rates (bites. min−1.m−2) of the herbivore community. Response variables are modelled with exposure and atoll as fixed effects and sites within islands as random effects. The best models are presented according to a backward model selection procedure based on sequential Likelihood-Ratio tests (see methods). In addition to random and fixed effect coefficients and SE, goodness-of-fit (GOF) measures are provided for the model: marginal R2 (R2m), measuring variation explained by fixed effects only, and conditional R2 (R2c), measuring variation explained by both fixed and random effects. Bite rates is square root transformed and biomass is log transformed.
Back to article page