The seafloor-character map, GLMs, and the Marine Geospatial Ecology Tool (MGET; <http://mgel.env.duke.edu>)
in ArcGIS were used to develop predictive-probability maps of occurrence per 12 sq m of seafloor for selected invertebrate
taxa. The 12 sq m area was calculated by multiplying the average width of the video images by the average distance
covered during the 10-second samples. The seafloor-character map provided the habitat-class data.
Taxa Model Formulas for GLM:
Brittle Star
taxa~Depth+I(Depth^2)+ Class+Block+Depth:Block, # Depth+ Depth^2+Class+Block+Depth:Block tryCatch(taxa33.glm <- glm(taxa~Depth+I(Depth^2)+
Class+Block+Depth:Block, family=binomial(link="logit"), data=taxa.df),
warning = function(x) { output.df[33,"flag"] <<- 1 }, finally = taxa33.glm <- glm(taxa~Depth+I(Depth^2)+ Class+Block+Depth:Block, family=binomial(link="logit"), data=taxa.df)
)
summary(taxa33.glm)
MODEL SUMMARY:
==============
Call:
glm(formula = brit_in ~ Depth + I(Depth^2) + factor(Class) + factor(Block) + Depth:factor(Block), family = binomial(link="logit"), data = na.omit(d))
Deviance Residuals:
Min 1Q Median 3Q Max
-1.5551 -0.6983 -0.4912 -0.1107 2.7408
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) -1.6441219 0.6640992 -2.476 0.01330 *
Depth 0.1144813 0.0234621 4.879 1.06e-06 ***
I(Depth^2) -0.0013578 0.0002068 -6.567 5.13e-11 ***
factor(Class)2 -0.9254904 0.3303862 -2.801 0.00509 **
factor(Class)3 -0.8756778 0.5539213 -1.581 0.11391
factor(Block)2 -1.5455724 0.5615242 -2.752 0.00591 **
factor(Block)3 -4.4970380 0.7311817 -6.150 7.73e-10 ***
Depth:factor(Block)2 0.0061871 0.0092539 0.669 0.50376
Depth:factor(Block)3 0.0807144 0.0147237 5.482 4.21e-08 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for binomial family taken to be 1)
Null deviance: 1009.10 on 922 degrees of freedom
Residual deviance: 841.58 on 914 degrees of freedom
AIC: 859.58
Number of Fisher Scoring iterations: 5
Analysis of Deviance Table
Model: binomial, link: logit
Response: brit_in
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev
NULL 922 1009.10
Depth 1 2.14 921 1006.96
I(Depth^2) 1 85.43 920 921.53
factor(Class) 2 8.31 918 913.22
factor(Block) 2 43.91 916 869.30
Depth:factor(Block) 2 27.72 914 841.58
resampled to 10 meters
Krigsman, L.M., M.M. Yoklavich, E.J. Dick, and G.R. Cochrane (2012) Models and maps: predicting the distribution of corals and other benthic macro-invertebrates in shelf habitats. Ecosphere 3:1-16.