-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathdata_diogo.R
executable file
·153 lines (132 loc) · 6.2 KB
/
data_diogo.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
#Analysis of Amy Downing's experimental data to incorporate PD as a predictor of
#ecosystem functioning. Analysis based on linear mixed effect models.
#PhD chapter of Diogo Provete. Last modified: JAN 2014.
#(c) Diogo B. Provete | [email protected]
library(lme4)
library(bbmle)
library(MuMIn)
## Data input
###########
dados <- read.csv("data_diogo.csv", header = T, sep = ";")
dados=dados[,-15]
head(dados)
attach(dados)
dados$PD2<-dados$PD/1000
source("p.values.lmer.R")
#---Exploratory Plots
plot(prod~as.factor(PD2), xlab="Faith's PD", ylab="Productivity")
plot(prod~as.factor(div), xlab="Species Richness (Treatments)", ylab="Productivity")
boxplot(prod~as.factor(div)+time)
plot(prod~PD2, xlab="Faith's PD", ylab="Productivity")
boxplot(prod~as.factor(time), xlab="Time", ylab="Productivity")
## models
###########
#--Productivity
##Establishing the best random effects
model1 <- lmer(prod~PD2 + time + factor(div) + (1+time+PD2|tank), verbose=T, data=dados)
model2 <- lmer(prod~PD2 + time + factor(div) + (1+time|tank),data=dados)
model3 <- lmer(prod~PD2 + time + factor(div) + (1|tank),data=dados)#best model
model31 <- lmer(prod~PD2 + time + (1|tank)+(1|time),data=dados)#best model
model32 <- lmer(prod~PD2 + time + factor(div) + (1|tank/time),data=dados)#best model
model4 <- lmer(prod~PD2 + time + factor(div) + (1+PD2|tank), data=dados)
anova(model1, model2, model3, model4)
(ICtab(model1, model2, model3, model4,model31,
type = c("AICc"), weights = TRUE, delta = TRUE, sort = TRUE, nobs=504))
##Establishing the best fixed effects model
ic<-function(dados){
mymodel2 <- lmer(prod~PD* time + (1|tank)+(1|time),data=dados)
mymodel3 <- lmer(prod ~ PD + time + (1|tank)+(1|time),data=dados)
#mymodel04<-lmer(prod ~ PD2*div + (1|tank),data=dados)#best model
#mymodel24<-lmer(prod ~ PD2*factor(div) + time+ (1|tank),data=dados)
#mymodel4<-lmer(prod ~ PD2 + factor(div) + (1|tank),data=dados)
#mymodel5 <- lmer(prod ~ factor(div) + time + (1|tank),data=dados)
mymodel6<-lmer(prod ~ comp + (1|tank)+(1|time),data=dados)
#mymodel7 <- lmer(prod ~ factor(div) + (1|tank),data=dados)
mymodel8<-lmer(prod ~ time + (1|tank)+(1|time),data=dados)
mymodel9<-lmer(prod ~ 1 + (1|tank)+(1|time),data=dados)#"null" model only with intercept
modelos<-list(mymodel2=mymodel2, mymodel6=mymodel6, mymodel8=mymodel8, mymodel3=mymodel3)
tab1<-do.call(rbind, lapply(modelos, r.squaredGLMM))
#Model selection table for the models
tab2<-ICtab(mymodel2, mymodel3, mymodel6, mymodel8,mymodel9,
type = c("AICc"), weights = TRUE, delta = TRUE, sort = TRUE,nobs=84)
return(list(R2=tab1, AICTab=tab2))
}
lapply(dados1,ic)
#Calculating conditional pseudo R2 for each model
#r.squaredGLMM(mymodel04)
#r.squaredGLMM(mymodel4)
r.squaredGLMM(mymodel2)
#r.squaredGLMM(mymodel7)
r.squaredGLMM(mymodel6)
#r.squaredGLMM(mymodel5)
r.squaredGLMM(mymodel8)
r.squaredGLMM(mymodel3)
#Calculating the % Deviance Explained (~ R2) by each model
100*(deviance(mymodel9) - deviance(mymodel2))/deviance(mymodel9)
100*(deviance(mymodel9) - deviance(mymodel3))/deviance(mymodel9)
100*(deviance(mymodel9) - deviance(mymodel4))/deviance(mymodel9)
100*(deviance(mymodel9) - deviance(mymodel5))/deviance(mymodel9)
100*(deviance(mymodel9) - deviance(mymodel6))/deviance(mymodel9)
100*(deviance(mymodel9) - deviance(mymodel7))/deviance(mymodel9)
100*(deviance(mymodel9) - deviance(mymodel8))/deviance(mymodel9)
## FIGURE
############
boxplot(prod ~ factor(time), ylab = "Productivity") ## productivity decreases with time regardless of richness
#--Models for Decomposition
##Establishing the best random effects
m1 <- lmer(dc~PD2 + time + factor(div) + (1+time+PD2|tank), data=dados)
m2 <- lmer(dc~PD2 + time + factor(div) + (1+time|tank),data=dados)
m3 <- lmer(dc~PD2 + time + factor(div) + (1|tank),data=dados)#best model
m4 <- lmer(dc~PD2 + time + factor(div) + (1+PD2|tank), data=dados)
anova(m1, m2, m3, m4)
(ICtab(m1, m2, m3, m4, type = c("AICc"), weights = TRUE, delta = TRUE, sort = TRUE, nobs=504))
##Establishing the best fixed effects model
my2 <- lmer(dc~PD2 + time + factor(div) + (1|tank),data=dados)
my3 <- lmer(dc ~ PD2 + time + (1|tank),data=dados)
my4<-lmer(dc ~ PD2 + factor(div) + (1|tank),data=dados)
my40<-lmer(dc ~ PD2*factor(div) + (1|tank),data=dados)
my5 <- lmer(dc ~ factor(div) + time + (1|tank),data=dados)
my6<-lmer(dc ~ PD2 + (1|tank),data=dados)
my7 <- lmer(dc ~ factor(div) + (1|tank),data=dados)
my8<-lmer(dc ~ time + (1|tank),data=dados)#best model
my9<-lmer(dc ~ 1 + (1|tank),data=dados)#"null" model only with intercept
#Model selection table for the models
(ICtab(my2, my3,my4, my40,my5, my6, my7,my8,my9,
type = c("AICc"), weights = TRUE, delta = TRUE, sort = TRUE,nobs=504))
r.squaredGLMM(my8)[[2]]
r.squaredGLMM(my3)[[2]]
r.squaredGLMM(my5)[[2]]
r.squaredGLMM(my2)[[2]]
r.squaredGLMM(my6)[[2]]
r.squaredGLMM(my7)[[2]]
r.squaredGLMM(my4)[[2]]
r.squaredGLMM(my40)[[2]]
#--Models for Respiration
##Establishing the best random effects
mo1 <- lmer(resp~PD2 + time + factor(div) + (1+time+PD2|tank), data=dados)
mo2 <- lmer(resp~PD2 + time + factor(div) + (1+time|tank),data=dados)
mo3 <- lmer(resp~PD2 + time + factor(div) + (1|tank),data=dados)#best model
mo4 <- lmer(resp~PD2 + time + factor(div) + (1+PD2|tank), data=dados)
(ICtab(mo1, mo2, mo3, mo4, type = c("AICc"), weights = TRUE, delta = TRUE, sort = TRUE, nobs=504))
##Establishing the best fixed effects model
mm2 <- lmer(resp~PD2 + time + factor(div) + (1|tank),data=dados)
mm3 <- lmer(resp ~ PD2 + time + (1|tank),data=dados)
mm4<-lmer(resp ~ PD2 + factor(div) + (1|tank),data=dados)
mm14<-lmer(resp ~ PD2 * factor(div) + (1|tank),data=dados)
mm5 <- lmer(resp ~ factor(div) + time + (1|tank),data=dados)
mm6<-lmer(resp ~ PD2 + (1|tank),data=dados)
mm7 <- lmer(resp ~ factor(div) + (1|tank),data=dados)
mm8<-lmer(resp ~ time + (1|tank),data=dados)#best model
mm9<-lmer(resp ~ 1 + (1|tank),data=dados)#"null" model only with intercept
#Model selection table for the models
(ICtab(mm2, mm3,mm4, mm14,mm5, mm6, mm7,mm8,mm9,
type = c("AICc"), weights = TRUE, delta = TRUE, sort = TRUE,nobs=504))
#Calculating conditional pseudo R2 for each model
r.squaredGLMM(mm8)[[2]]
r.squaredGLMM(mm3)[[2]]
r.squaredGLMM(mm5)[[2]]
r.squaredGLMM(mm2)[[2]]
r.squaredGLMM(mm6)[[2]]
r.squaredGLMM(mm7)[[2]]
r.squaredGLMM(mm4)[[2]]
r.squaredGLMM(mm14)[[2]]