library(metafor) burn<-read.csv(file="Additional File 7. Data used in meta-analyses.csv") attach(burn) str(burn) ############ #1. VascRichS ############ # Select the subset of data you want to include (& means AND, | means OR) VascRichS<-subset(burn,(Taxon=="Vasc"|Taxon=="NativNNativ"|Taxon=="NativV"|Taxon=="VascHVascW")&(Outcome=="RichS"|Outcome=="RichG/S")) # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="VascRichS.png", type="cairo", units="in", width=10, height=11, res=1000) forest(rmaData,slab=VascRichS$Citation_site,cex=0.75,main="Taxon Richness of Plants") dev.off() # funnel plot using inverse square root of sample size plot(VascRichS$SMD_g,1/sqrt(VascRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=0.3967,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=VascRichS) #failsafe N=331 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data VascRichSHigh<-subset(VascRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=VascRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Plants") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=VascRichS,xlab="Forest type",ylab="Effect size") # Broadleaf VascRichSB<-subset(burn,(Taxon=="Vasc"|Taxon=="NativNNativ"|Taxon=="NativV"|Taxon=="VascHVascW")&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=VascRichSB$Citation_site,cex=0.75,main="Taxon Richness of Plants in Broadleaf Forest") # Coniferous VascRichSC<-subset(burn,(Taxon=="Vasc"|Taxon=="NativNNativ"|Taxon=="NativV"|Taxon=="VascHVascW")&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=VascRichSC$Citation_site,cex=0.75,main="Taxon Richness of Plants in Coniferous Forest") # Mixed VascRichSM<-subset(burn,(Taxon=="Vasc"|Taxon=="NativNNativ"|Taxon=="NativV"|Taxon=="VascHVascW")&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=VascRichSM$Citation_site,cex=0.75,main="Taxon Richness of Plants in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time since burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) ############ #2. NNativRichS ############ # Select the subset of data you want to include (& means AND, | means OR) NNativRichS<-subset(burn,Taxon=="NNativ"&(Outcome=="RichS"|Outcome=="RichG/S")) # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=NNativRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="NNativRichS.png", type="cairo", units="in", width=8, height=6, res=1000) forest(rmaData,slab= NNativRichS $Citation_site,cex=0.75,main="Taxon Richness of Non-Native Plants") dev.off() # funnel plot using inverse square root of sample size plot(NNativRichS$SMD_g,1/sqrt(NNativRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=0.3862,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=NNativRichS) #failsafe N=23 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data NNativRichSHigh<-subset(NNativRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=NNativRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=NNativRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Non-native Plants") # Subgroup analyses on forest type # Cannot run forest models - all are data from coniferous forests rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=NNativRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=NNativRichS,xlab="Forest type",ylab="Effect size") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=NNativRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time since burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=NNativRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=NNativRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=NNativRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) ############ #3. VascHRichS ############ # Select the subset of data you want to include (& means AND, | means OR) VascHRichS<-subset(burn,Taxon=="VascH"&(Outcome=="RichS"|Outcome=="RichG/S")) # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="VascHRichS.png", type="cairo", units="in", width=8, height=6, res=1000) forest(rmaData,slab=VascHRichS$Citation_site,cex=0.75,main="Taxon Richness of Herbaceous Plants") dev.off() # funnel plot using inverse square root of sample size plot(VascHRichS$SMD_g,1/sqrt(VascHRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=0.3574,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=VascHRichS) #failsafe N=62 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data VascHRichSHigh<-subset(VascHRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=VascHRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Herbaceous Plants") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=VascHRichS,xlab="Forest type",ylab="Effect size") # Broadleaf VascHRichSB<-subset(burn,Taxon=="VascH"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=VascHRichSB$Citation_site,cex=0.75,main="Taxon Richness of Herbaceous Plants in Broadleaf Forest") # Coniferous VascHRichSC<-subset(burn,Taxon=="VascH"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=VascHRichSC$Citation_site,cex=0.75,main="Taxon Richness of Herbaceous Plants in Coniferous Forest") # Mixed VascHRichSM<-subset(burn,Taxon=="VascH"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=VascHRichSM$Citation_site,cex=0.75,main="Taxon Richness of Herbaceous Plants in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time since burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) library(plotrix) size<-rescale(1/VascHRichS$SMD_Vg,c(1,5)) plot(SMD_g~Time_since_burn_y,data=VascHRichS,xlab="Time since burn (years)",ylab="Effect size",cex=size) abline(a=0.7582,b=-0.1296) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascHRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) plot(VascHRichS$SMD_g~factor(VascHRichS$CZ)) #spurious significance, only one Cf point ############ #4. VascWRichS ############ # Select the subset of data you want to include (& means AND, | means OR) VascWRichS<-subset(burn,Taxon=="VascW"&(Outcome=="RichS"|Outcome=="RichG/S")) # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="VascWRichS.png", type="cairo", units="in", width=9, height=6, res=1000) forest(rmaData,slab=VascWRichS$Citation_site,cex=0.75,main="Taxon Richness of Woody Plants") dev.off() # funnel plot using inverse square root of sample size plot(VascWRichS$SMD_g,1/sqrt(VascWRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=-0.2530,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=VascWRichS) #failsafe N=112 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data VascWRichSHigh<-subset(VascWRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=VascWRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Woody Plants") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=VascWRichS,xlab="Forest type",ylab="Effect size") # Broadleaf VascWRichSB<-subset(burn,Taxon=="VascW"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=VascWRichSB$Citation_site,cex=0.75,main="Taxon Richness of Woody Plants in Broadleaf Forest") # Coniferous VascWRichSC<-subset(burn,Taxon=="VascW"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=VascWRichSC$Citation_site,cex=0.75,main="Taxon Richness of Woody Plants in Coniferous Forest") # Mixed VascWRichSM<-subset(burn,Taxon=="VascW"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=VascWRichSM$Citation_site,cex=0.75,main="Taxon Richness of Woody Plants in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time sinxze burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascWRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) ############ #5. TreeRichS ############ # Select the subset of data you want to include (& means AND, | means OR) TreeRichS<-subset(burn,Taxon=="Tree"&Outcome=="RichS") # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="TreeRichS.png", type="cairo", units="in", width=9, height=6, res=1000) forest(rmaData,slab=TreeRichS$Citation_site,cex=0.75,main="Taxon Richness of Trees") dev.off() # funnel plot using inverse square root of sample size plot(TreeRichS$SMD_g,1/sqrt(TreeRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=-1.0347,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=TreeRichS) #failsafe N=184 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data TreeRichSHigh<-subset(TreeRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=TreeRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Trees") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=TreeRichS,xlab="Forest type",ylab="Effect size") # Broadleaf TreeRichSB<-subset(burn,Taxon=="Tree"&Outcome=="RichS"&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=TreeRichSB$Citation_site,cex=0.75,main="Taxon Richness of Trees in Broadleaf Forest") # Coniferous TreeRichSC<-subset(burn,Taxon=="Tree"&Outcome=="RichS"&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=TreeRichSC$Citation_site,cex=0.75,main="Taxon Richness of Trees in Coniferous Forest") # Mixed TreeRichSM<-subset(burn,Taxon=="Tree"&Outcome=="RichS"&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=TreeRichSM$Citation_site,cex=0.75,main="Taxon Richness of Trees in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time sinxze burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=TreeRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) ############ #6. VascDivS ############ # Select the subset of data you want to include (& means AND, | means OR) VascDivS<-subset(burn,Taxon=="Vasc"&Outcome=="DivS") # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="VascDivS.png", type="cairo", units="in", width=8, height=6, res=1000) forest(rmaData,slab=VascDivS$Citation_site,cex=0.75,main="Diversity of PLants") dev.off() # funnel plot using inverse square root of sample size plot(VascDivS$SMD_g,1/sqrt(VascDivS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=-0.0646,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=VascDivS) #failsafe N=0 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data VascDivSHigh<-subset(VascDivS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=VascDivSHigh$Citation_site,cex=0.75,main="Diversity of Plants") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=VascDivS,xlab="Forest type",ylab="Effect size") # Broadleaf VascDivSB<-subset(burn,Taxon=="Vasc"&Outcome=="DivS"&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=VascDivSB$Citation_site,cex=0.75,main="Diversity of Plants in Broadleaf Forest") # Coniferous VascDivSC<-subset(burn,Taxon=="Vasc"&Outcome=="DivS"&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=VascDivSC$Citation_site,cex=0.75,main="Diversity of Plants in Coniferous Forest") # Mixed VascDivSM<-subset(burn,Taxon=="Vasc"&Outcome=="DivS"&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=VascDivSM$Citation_site,cex=0.75,main="Diversity of Plants in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time sinxze burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=VascDivS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) ############ #7. FungRichS ############ # Select the subset of data you want to include (& means AND, | means OR) FungRichS<-subset(burn,Taxon=="Fung"&(Outcome=="RichS"|Outcome=="RichG/S")) # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="FungRichS.png", type="cairo", units="in", width=7, height=5, res=1000) forest(rmaData,slab=FungRichS$Citation_site,cex=0.75,main="Taxon Richness of Fungi") dev.off() # funnel plot using inverse square root of sample size plot(FungRichS$SMD_g,1/sqrt(FungRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=-1.1628,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=FungRichS) #failsafe N=16 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data FungRichSHigh<-subset(FungRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=FungRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Fungi") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=FungRichS,xlab="Forest type",ylab="Effect size") # Broadleaf FungRichSB<-subset(burn,Taxon=="Fung"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=FungRichSB$Citation_site,cex=0.75,main="Taxon Richness of Fungi in Broadleaf Forest") # Coniferous FungRichSC<-subset(burn,Taxon=="Fung"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=FungRichSC$Citation_site,cex=0.75,main="Taxon Richness of Fungi in Coniferous Forest") # Mixed FungRichSM<-subset(burn,Taxon=="Fung"&(Outcome=="RichS"|Outcome=="RichG/S")&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=FungRichSM$Citation_site,cex=0.75,main="Taxon Richness of Fungi in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time since burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=FungRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) ############ #8. BirdRichS ############ # Select the subset of data you want to include (& means AND, | means OR) BirdRichS<-subset(burn,Taxon=="Bird"&Outcome=="RichS") # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="BirdRichS.png", type="cairo", units="in", width=7, height=4, res=1000) forest(rmaData,slab=BirdRichS$Citation_site,cex=0.75,main="Taxon Richness of Birds") dev.off() # funnel plot using inverse square root of sample size plot(BirdRichS$SMD_g,1/sqrt(BirdRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=-0.1693,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=BirdRichS) #Non-significant model so cannot run # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data BirdRichSHigh<-subset(BirdRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=BirdRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Birds") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=BirdRichS,xlab="Forest type",ylab="Effect size") # Broadleaf BirdRichSB<-subset(burn,Taxon=="Bird"&Outcome=="RichS"&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=BirdRichSB$Citation_site,cex=0.75,main="Taxon Richness of Birds in Broadleaf Forest") # Coniferous BirdRichSC<-subset(burn,Taxon=="Bird"&Outcome=="RichS"&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=BirdRichSC$Citation_site,cex=0.75,main="Taxon Richness of Birds in Coniferous Forest") # Mixed BirdRichSM<-subset(burn,Taxon=="Bird"&Outcome=="RichS"&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=BirdRichSM$Citation_site,cex=0.75,main="Taxon Richness of Birds in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time since burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BirdRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) ############ #9. BeetRichS - need to check groups of taxa are correct ############ # Select the subset of data you want to include (& means AND, | means OR) BeetRichS<-subset(burn,(Taxon=="BeetA"|Taxon=="BeetC"|Taxon=="BeetG"|Taxon=="BeetO")&Outcome=="RichS") # unmoderated model rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichS,random=~1|Site_ID) # meta-analysis results print(rmaData) # fores plot png(filename="BeetRichS.png", type="cairo", units="in", width=9, height=6, res=1000) forest(rmaData,slab=BeetRichS$Citation_site,cex=0.75,main="Taxon Richness of Beetles") dev.off() # funnel plot using inverse square root of sample size plot(BeetRichS$SMD_g,1/sqrt(BeetRichS$n),ylim=c(0.5,0),ylab="1/sqrt(n)",xlab="Effect size") abline(v=0.3980,lty=2) #failsafe number fsn(SMD_g,SMD_Vg,data=BeetRichS) #failsafe N=10 # Cook's distance plot x<-cooks.distance(rmaData) plot(x,type='o',pch=19,xlab="Study number",ylab="Cook's Distance") # sensitivity analysis for high validity data BeetRichSHigh<-subset(BeetRichS,Validity=="High validity") rmaDataHigh<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichSHigh,random=~1|Site_ID) print(rmaDataHigh) forest(rmaDataHigh,slab=BeetRichSHigh$Citation_site,cex=0.75,main="Taxon Richness of Beetles") # Subgroup analyses on forest type rmaData<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichS,mods=~Forest_type,random=~1|Site_ID) summary(rmaData) #not significant plot(SMD_g~Forest_type,data=BeetRichS,xlab="Forest type",ylab="Effect size") # Broadleaf BeetRichSB<-subset(burn,(Taxon=="BeetA"|Taxon=="BeetC"|Taxon=="BeetG"|Taxon=="BeetO")&Outcome=="RichS"&Forest_type=="Broadleaf") rmaDataB<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichSB,random=~1|Site_ID) print(rmaDataB) forest(rmaDataB,slab=BeetRichSB$Citation_site,cex=0.75,main="Taxon Richness of Beetles in Broadleaf Forest") # Coniferous BeetRichSC<-subset(burn,(Taxon=="BeetA"|Taxon=="BeetC"|Taxon=="BeetG"|Taxon=="BeetO")&Outcome=="RichS"&Forest_type=="Coniferous") rmaDataC<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichSC,random=~1|Site_ID) print(rmaDataC) forest(rmaDataC,slab=BeetRichSC$Citation_site,cex=0.75,main="Taxon Richness of Beetles in Coniferous Forest") # Mixed BeetRichSM<-subset(burn,(Taxon=="BeetA"|Taxon=="BeetC"|Taxon=="BeetG"|Taxon=="BeetO")&Outcome=="RichS"&Forest_type=="Mixed") rmaDataM<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichSM,random=~1|Site_ID) print(rmaDataM) forest(rmaDataM,slab=BeetRichSM$Citation_site,cex=0.75,main="Taxon Richness of Beetles in Mixed Forest") # Burn frequency meta-regression rmaDataBF<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichS,random=~1|Site_ID,mods=~Burn_freq_y) print(rmaDataBF) # Time since burn meta-regression rmaDataT<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichS,random=~1|Site_ID,mods=~Time_since_burn_y) print(rmaDataT) # Burn season subgroup analysis rmaDataBS<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichS,random=~1|Site_ID,mods=~factor(Burn_growdorm)) print(rmaDataBS) # Climate zone subgroup analysis rmaDataCZ<-rma.mv(yi=SMD_g,V=SMD_Vg,data=BeetRichS,random=~1|Site_ID,mods=~factor(CZ)) print(rmaDataCZ) #########################################################################################################