library(metafor)

# 1.Species richness

richness<-read.csv('C:/Users/ssa/Desktop/WKH/Diversity/richness.csv')
dat<-richness
yi<-dat$d
vi<-dat$var_d
overall<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML")
summary(overall)
profile(overall, sigma2=1)
profile(overall, sigma2=2)


#Forest plot
overall<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", slab=paste(Author, Pub_year))
par(mai=c(0.82,3,0.82,3))
forest(overall, cex=0.7, xlab="Effect size")

#Funnel plot and Fail-safe number
par(mai=c(0.82,2,0.82,2))
plot(dat$d,dat$N_C,ylim=c(0,70),ylab="Sample size",xlim=c(-1,12), xlab="Effect size", cex=1.2, cex.lab=1.2)
abline(v=0.81,lty=2)
fsn(yi,vi, data=dat)

#Study characteristics
mods1<- rma.mv(yi, vi, random = ~ 1 | Study/Id, mods=~Pub_type+Study_year+Country, data=dat)
summary(mods1)

#Influence of effect modifiers
##Comparator, taxa and their interaction
mods3<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Comparator+Taxa+Comparator*Taxa)
summary(mods3)
##Owner of the forest
owner<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Owner)
summary(owner)
##management intensity
manintensity<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Sur_man)
summary(manintensity)
##Age of the comparator forest
age<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Age_C)
summary(age)


#Sensitivity analysis without study id no 7
richness7<-read.csv('C:/Users/ssa/Desktop/WKH/Diversity/richnessWO7.csv')
dat<-richness7
yi<-dat$d
vi<-dat$var_d
overall7<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat)
summary(overall7)
mods7<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, mods=~Comparator+Taxa+Comparator*Taxa)
summary(mods7)

#Sensitivity analysis without studies with corrected sample sizes
sensitivity<-read.csv('C:/Users/ssa/Desktop/WKH/Diversity/sensitivity.csv')
dat<-sensitivity
yi<-dat$d
vi<-dat$var_d
ressen<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat)
summary(ressen)


#2.Individual abundance
abundance<-read.csv('C:/Users/ssa/Desktop/WKH/Abundance/Abundance.csv')
dat<-abundance
yi<-dat$d
vi<-dat$var_d
overall<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat)
summary(overall)
profile(overall, sigma2=1)
profile(overall, sigma2=2)

#Forest plot
overall<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", slab=paste(Author, Pub_year))
par(mai=c(0.82,3,0.82,3))
forest(overall, cex=0.7, xlab="Effect size")

#Funnel plot and Fail-safe number
par(mai=c(0.82,2,0.82,2))
plot(dat$d,dat$N_C,ylim=c(0,50),ylab="Sample size",xlim=c(-2,10), xlab="Effect size", cex=1.2, cex.lab=1.2)
abline(v=1.91,lty=2)
fsn(yi,vi, data=dat)

#Study characteristics
studych<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Pub_type+Country+Study_year)
summary(studych)

#Influence of effect modifiers
mods2<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Comparator+Taxa)
summary(mods2)

#Sensitivity analysis
abundancesen<-read.csv('C:/Users/ssa/Desktop/WKH/Abundance/Abundance_sensitivity.csv')
dat<-abundancesen
yi<-dat$d
vi<-dat$var_d
sensitivity<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat)
summary(sensitivity)


#3.Deadwood
deadwood<-read.csv('C:/Users/ssa/Desktop/WKH/Deadwood/Deadwood.csv')
dat<-deadwood
yi<-dat$d
vi<-dat$var_d
overall<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", slab=paste(Author, Pub_year))
summary(overall)
profile(overall, sigma2=1)
profile(overall, sigma2=2)
overallr<-rma(yi, vi, data=dat, method="REML")
###Note: Both rma.mv and rma models produce same results as Study level variance is estimated to be zero in the rma.mv model.

#Forest plot
par(mai=c(0.82,3,0.82,3))
forest(overall, cex=0.7, xlab="Effect size")

#Funnel plots and Fail-safe number
par(mai=c(0.82,2,0.82,2))
plot(dat$d,dat$N_I,ylim=c(0,639),ylab="Sample size",xlim=c(-3,8), xlab="Effect size", cex=1.2, cex.lab=1.2)
abline(v=0.625,lty=2)
fsn(yi,vi, data=dat)

#Study characteristics
study<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Country+Pub_type+Study_year)
summary(study)

#Influence of effect modifiers
##Comparator
comparator<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Comparator)
summary(comparator)
##Age of the production forest
age<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Age_C, subset=(Comparator=="production"))
summary(age)


#Sensitivity analysis without studies with imputed SDs
sdcorrected<-read.csv('C:/Users/ssa/Desktop/WKH/Deadwood/sensitivity.csv')
dat<-sdcorrected
yi<-dat$d
vi<-dat$var_d
sdres<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML")
summary(sdres)

#Influence of study characteristics without studies with imputed SDs
study<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Country+Pub_type+Study_year)

#Influence of effect modifiers without studies with imputed SDs
##Comparator
comparator<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Comparator)

##Age of the production forest
age<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Age_C, subset=(Comparator=="production"))

##Funnel plots and Fail-safe number without studies with imputed SDs
par(mai=c(0.82,2,0.82,2))
plot(dat$d,dat$N_I,ylim=c(0,639),ylab="Sample size",xlim=c(-3,8), xlab="Effect size", cex=1.2, cex.lab=1.2)
abline(v=0.48,lty=2)
fsn(yi,vi, data=dat)

#Sensitivity analysis without high risk studies
deadwoods<-read.csv('C:/Users/ssa/Desktop/WKH/Deadwood/Deadwood_sensitivity.csv')
dat<-deadwoods
yi<-dat$d
vi<-dat$var_d
overalls<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML")
summary(overalls)

#Study characteristics
studys<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Country+Pub_type+Study_year)
summary(studys)

#Influence of effect modifiers
##Comparator
comparators<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Comparator)
summary(comparators)

##Age of the comparator forest
age<- rma.mv(yi, vi, random = ~ 1 | Study/Id, data=dat, method="REML", mods=~Age_C)
summary(agepro)