#load packages ==== library(ggplot2) library(ggfortify) library(dplyr) library(plyr) library(lme4) library(data.table) library(gmodels) library(ape) library(permuco) library(emmeans) library(lmPerm) library(multcomp) library(Ecdat) library(hexbin) library(sp) library(geometry) library(MASS) library(car) library(ClassDiscovery) library(PCDimension) library(factoextra) library(rela) library(psych) library(rela) library(qqplotr) library(mvnormtest) library(devtools) library(factoextra) library(ClassDiscovery) library(PCDimension) library(ggpubr) library(ggrepel) library(boot) library(grid) library(gridExtra) library(gtable) library(patchwork) library(svglite) library(cowplot) library(vioplot) library(resample) library(coin) library(mosaic) library(tidyverse) library(rcompanion) library(ICC) library(multilevel) library(purrr) library(ggbiplot) library(lmerTest) library(Rmisc) library(GGally) library(lmtest) ######################################################################## #####Coral cover analysis##### ######################################################################## coral<-coral_2019 #check distribution using residuals==== lm.residual<- lm(CC~Site, data=coral) sresid <- studres(lm.residual) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) #ANOVA ANOVA_coral<-aov(CC~Site+Year+Year*Site, data=coral) summary(ANOVA_coral) #Boxplot coral_cover<-ggplot(coral, aes(x=Site, y=CC))+ geom_boxplot(width=0.4, position=position_dodge(0.65))+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), axis.text.y=element_text(size= 12, colour="Black"), axis.text.x = element_text(angle=90, vjust=0.5, size=12, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=14))+ guides(fill=guide_legend(title=NULL))+ labs(x="Site", y="Coral cover (%)")+ scale_x_discrete(labels=c("HIGH", "LOW"))+ facet_wrap(~Year) ggsave(file="coral_cover.png") #calculate mean coral mean_coral<- aggregate(coral$CC, by=list(coral$Year,coral$Site), FUN=mean) ######################################################################## #####Territory mapping analysis##### ######################################################################## #load data territories<-chaet_territories_area #Calculate area for each territory #SAM_T1 SAM_T1_Hull_X<-c(18.2,3.6,0.1,4.2,15.7,19.7) SAM_T1_Hull_Y<-c(0.4,3.7,9,11.3,12.8,9.1) SAM_T1<-polyarea(SAM_T7_Hull_X, SAM_T7_Hull_Y) #PK_T1 PK_T1_Hull_X<-c(8.1,8.1,1.3,0.2,0,0,0.5,6.8) PK_T1_Hull_Y<- c(5.8,2.4,3.1,3.4,6.3,8,10.2,9.4) PK_T1<-polyarea(PK_T8_Hull_X, PK_T8_Hull_Y) #SAM_T2 SAM_T2_Hull_X<-c(22.3,16.1,15,10,3.9,5.7,16,18.3,22.1) SAM_T2_Hull_Y<-c(6.8,5.4,5.3,5.3,6.4,8.8,12,11.2,8.1) SAM_T2<-polyarea(SAM_T5_Hull_X,SAM_T5_Hull_Y) #PK_T2 PK_T2_Hull_X<-c(13.6,6,0,2.5,9.6) PK_T2_Hull_Y<-c(3.2,0,0.8,11.9,10.1) PK_T2<-polyarea(PK_T4_Hull_X, PK_T4_Hull_Y) #SAM_T3 SAM_T3_Hull_X<-c(27,18.7,4,4,6.8,12.2,15.1,18.2,22.4) SAM_T3_Hull_Y<-c(3.6,2.8,10,11,13,15.5,15.5,14.3,12.6) SAM_T3<-polyarea(SAM_T6_Hull_X, SAM_T6_Hull_Y) #PK_T3 PK_T3_Hull_X<-c(15,14.5,8.5,4.4,0.4,1.2,5.1,11.3) PK_T3_Hull_Y<-c(1.2,0.9,0.6,0.7,4.7,6.3,7.2,8) PK_T3<-polyarea(PK_T1_Hull_X, PK_T1_Hull_Y) #SAM_T4 SAM_T4_Hull_X<-c(16.1,8.9,1.2,0.4,10,21) SAM_T4_Hull_Y<-c(0.5,1,3,8.3,13.4,8.4) SAM_T4<-polyarea(SAM_T4_Hull_X, SAM_T4_Hull_Y) #PK_T4 PK_T4_Hull_X<-c(16.6,16.7,6,2.1,4.3,12.2) PK_T4_Hull_Y<-c(2.4,2,2.8,4.2,4.9,4.3) PK_T4<-polyarea(PK_T2_Hull_X, PK_T2_Hull_Y) #PK_T5 PK_T5_Hull_X<-c(16.1,15.3,13.9,11.4,7.6,7.1,7.9) PK_T5_Hull_Y<-c(3.9,0.4,0,2.6,7.9,12.4,17.1) PK_T5<-polyarea(PK_T3_Hull_X, PK_T3_Hull_Y) #SAM_T5 SAM_T5_Hull_X<-c(23.8,28.1,22.7,17.3,7.2,3.6,5,15.3) SAM_T5_Hull_Y<-c(6.8,0.9,0,0.3,3.5,6.9,21.2,18) SAM_T5<-polyarea(SAM_T3_Hull_X, SAM_T3_Hull_Y) Site_name<-c("PK","PK","PK","PK","PK","SAM","SAM","SAM","SAM","SAM") Pair_ID<-c("PK_T1", "PK_T2", "PK_T3", "PK_T4", "PK_T5", "SAM_T1", "SAM_T2", "SAM_T3", "SAM_T4", "SAM_T5") Area<-c(PK_T1, PK_T2, PK_T3, PK_T4, PK_T5, SAM_T1, SAM_T2, SAM_T3, SAM_T4, SAM_T5) territory_size<-cbind(Site_name, Pair_ID, Area) territory_size<-as.data.frame(territory_size) territory_size$Area<-as.numeric(territory_size$Area) ##################################################################### #Analyse territory_PK<- territory_size%>% filter(Site_name=="PK") mean(territory_PK$Area) sd(territory_PK$Area)/sqrt(length(territory_PK$Area)) territory_SAM<- territory_size%>% filter(Site_name=="SAM") mean(territory_SAM$Area) sd(territory_SAM$Area)/sqrt(length(territory_SAM$Area)) lm.residual_territory<- lm(Area~Site_name, data=territory_size) sresid <- studres(lm.residual_territory) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) territory_anova<-aov(Area~Site_name, data=territory_size) summary(territory_anova) ##################################################################### #plot find_hull<- function(df) df[chull(df$X, df$Y),] hulls<- ddply(territories, "Territory", find_hull) territory_size<-ggplot(territories, aes(x= X, y=Y, group=1, fill=Site))+ geom_point(size=2, col="Black")+ geom_polygon(data=hulls, alpha=0.5)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), axis.title.x = element_text(colour = "Black", face="bold", size=14), axis.title.y = element_text(colour = "Black", face="bold", size=14), strip.background = element_rect( fill="White"), legend.title = element_text(size = 16), legend.text = element_text(size = 12), strip.text.x = element_text(face="bold", size=12))+ facet_wrap(~Label, nrow=2)+ labs(x="X Transect (m)", y= "Y Transect (m)", fill="Coral cover")+ scale_fill_manual(values =c("deepskyblue2", "firebrick4"),labels=c("HIGH", "LOW")) ggsave("territory_size.tiff", units="in", width=9, height=6) ####################################################################### #Behavioural analyses==== ###################################################################### #Mean times for each pair chaet_behaviour_mean<-aggregate(cbind(Aggression, Feeding, Searching, Travelling, Sociability, Exploration)~Pair_ID+Site_name, data=chaet_behaviour, FUN="mean") ###################################################################### #MANOVA#==== MANOVA_prop<-manova(cbind(Aggression,Exploration,Feeding, Sociability)~Site_name, data=chaet_behaviour_mean) summary(MANOVA_prop) print(MANOVA_prop) summary.aov(MANOVA_prop, p.adjust.method="bonferroni") ###################################################################### #Boxplots for each trait using proportions ###################################################################### #Create dataframe chaet_boxplot$Behaviour <- factor(chaet_boxplot$Behaviour, levels=c("Exploration", "Feeding", "Sociability", "Aggression")) p1 <- ggplot(chaet_boxplot, aes(x=Behaviour, y=Time_s, fill=Site_name))+ geom_boxplot(width=0.3, position=position_dodge(0.45), alpha=0.8)+ geom_point(position=position_dodge(0.45), size=3, alpha=0.6)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.text.y=element_text(size= 14, colour="Black"), axis.text.x = element_text(vjust=0.5, size=14, colour="Black"), axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), strip.background = element_rect( fill="White"), legend.position="none")+ annotate(geom = 'segment', x= Inf, xend = Inf, y = -Inf, yend = Inf)+ labs(x="Behaviour category", y="Time (seconds)", fill="Coral cover")+ scale_fill_manual(values =c("deepskyblue2", "firebrick4"),labels=c("HIGH", "LOW")) p2 <- ggplot(chaet_boxplot, aes(x=Behaviour, y=Time_s_Agg, fill=Site_name))+ geom_boxplot(width=0.3, position=position_dodge(0.45), alpha=0.8)+ geom_point(position=position_dodge(0.45), size=3, alpha=0.6)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.text.y=element_text(size= 14, colour="Black"), axis.text.x = element_text(vjust=0.5, size=14, colour="Black"), axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), strip.background = element_rect( fill="White"), legend.position="none")+ annotate(geom = 'segment', x= Inf, xend = Inf, y = -Inf, yend = Inf)+ labs(x="Behaviour category", y="Proportion of time invested", breaks =c(0,0.01,0.02,0.03,0.04,0.05), fill="Coral cover")+ scale_x_discrete(labels=c("Exploration", "Feeding", "Sociability", "Aggression"))+ scale_fill_manual(values =c("deepskyblue2", "firebrick4"),labels=c("HIGH", "LOW"))+ theme(panel.background = element_rect(fill = NA)) g1 <- ggplot_gtable(ggplot_build(p1)) g2 <- ggplot_gtable(ggplot_build(p2)) pp <- c(subset(g1$layout, name == "panel", se = t:r)) g <- gtable_add_grob(g1, g2$grobs[[which(g2$layout$name == "panel")]], pp$t, pp$l, pp$b, pp$l) ia <- which(g2$layout$name == "axis-l") ga <- g2$grobs[[ia]] ax <- ga$children[[2]] ax$widths <- rev(ax$widths) ax$grobs <- rev(ax$grobs) ax$grobs[[1]]$x <- ax$grobs[[1]]$x - unit(1, "npc") + unit(0.15, "cm") g <- gtable_add_cols(g, g2$widths[g2$layout[ia, ]$l], length(g$widths) - 1) g <- gtable_add_grob(g, ax, pp$t, length(g$widths) - 1, pp$b) boxplot<-grid.draw(as_ggplot(g)) ggsave("overall_plot.tiff") #Check to see if aggression is a consequence of abundance #sum for each pair prob<-encounter_data prob_A<- aggregate(Prop_A~Pair_ID+Site_name, prob,FUN="mean") number<-aggregate(Total~Pair_ID+Site_name,prob,FUN="mean") #Test residuals lm.residual<- lm(Prop_A~Site_name, data=prob) sresid <- studres(lm.residual) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) prob_glm<-glm(Prop_A~Site_name, family=quasibinomial(), data=prob_A) summary(prob_glm) plot(prob_glm) #number of encounters lm.residual.n<- lm(Total~Site_name, data=number) sresid <- studres(lm.residual.n) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) encounter_glm<-glm(Total~Site_name, data=number, family=quasipoisson()) summary(encounter_glm) plot(encounter_glm) ##################################################################### #comparing feeding and foraging behaviours across sites==== ###################################################################### #Between sites feeding lm.residual.F<- lm(Feeding~Site_name, data=chaet_behaviour_mean) sresid <- studres(lm.residual.F) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) ANOVA_F<-aov(Feeding~Site_name, data=chaet_behaviour_mean) summary(ANOVA_F) #Between sites searching lm.residual.FS<- lm(Searching~Site_name, data=chaet_behaviour_mean) sresid <- studres(lm.residual.FS) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) ANOVA_FS<-aov(Searching~Site_name, data=chaet_behaviour_mean) summary(ANOVA_FS) #Individual sites #PK behav_FFS_PK<- chaet_FFS%>% filter(Site_name=="High") lm.residual.FFS.PK<- lm(time_seconds~Behaviour, data=behav_FFS_PK) sresid <- studres(lm.residual.FFS.PK) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) ANOVA_FFS.PK<-aov(time_seconds~Behaviour, data=behav_FFS_PK) summary(ANOVA_FFS.PK) ###################################################################### #SAM behav_FFS_SAM<- chaet_FFS%>% filter(Site_name=="Low") lm.residual.FFS.SAM<- lm(time_seconds~Behaviour, data=behav_FFS_SAM) sresid <- studres(lm.residual.FFS.SAM) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) ANOVA_FFS.SAM<-aov(time_seconds~Behaviour, data=behav_FFS_SAM) summary(ANOVA_FFS.SAM) ###################################################################### #plot FFS_plot<- ggplot(chaet_FFS, aes(x=Site_name, y=time_seconds, fill=Behaviour))+ geom_boxplot(width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), axis.text.y=element_text(size= 12, colour="Black"), axis.text.x = element_text(angle=90, vjust=0.5, size=12, colour="Black"), strip.background = element_rect( fill="White"), legend.text=element_text(size=14), strip.text.x = element_text(face="bold", size=9.5))+ guides(fill=guide_legend(title=NULL, size=60))+ labs(x="Resource availability", y="Time (seconds)")+ scale_x_discrete(labels=c("HIGH", "LOW"))+ scale_fill_manual(values=c("#808080", "#E0E0E0"), labels = c("Feeding", "Searching")) ggsave("FFS_plot.png", units="in", width=9, height=5, dpi=300) ###################################################################### #Principal component analysis==== ###################################################################### behav_PCA <- chaet_behaviour_mean[, 3:7] head(behav_PCA) ###################################################################### #compute PCA behav.pca.site <- prcomp(behav_PCA, scale = TRUE) summary(behav_PCA) #get eigenvalues ev <- behav.pca.site$sdev^2 ev Eigenvalue.plot<- function(ev) { n = length(ev) bsm = data.frame(j=seq(1:n), p=0) bsm$p[1] = 1/n for (i in 2:n) bsm$p[i] = bsm$p[i-1] + (1/(n + 1 - i)) bsm$p = 100*bsm$p/n op = par(mfrow=c(2,1),omi=c(0.1,0.3,0.1,0.1), mar=c(1, 1, 1, 1)) barplot(ev, col="Grey", las=2) abline(h=mean(ev), col="red") legend("topright", "Mean eigenvalue", lwd=1, col=2, bty="n") barplot(t(cbind(100*ev/sum(ev), bsm$p[n:1])), beside=TRUE, col=c("Grey",2), las=2) legend("topright", c("% variation", "Broken stick model"), pch=15, col=c("Grey",2), bty="n") par(op) } ###################################################################### #Visualise PCA ###################################################################### fviz_eig(behav.pca.site) #PCA loadings behav.pca.site$rotation loadings<-behav.pca.site$rotation loadings<-format(round(loadings, 3), nsmall = 3) loadings_plot<-loadings[, 1:2] loadings_plot<-ggtexttable(loadings_plot, theme=ttheme(base_style="blank")) loadings_plot<-loadings_plot%>% tab_add_hline(at.row = c(1, 2), row.side = "top", linewidth = 3, linetype = 1) %>% tab_add_hline(at.row = c(6), row.side = "bottom", linewidth = 3, linetype = 1) %>% tab_add_vline(at.column = c(1), column.side = "left", linewidth = 3, linetype = 1) %>% tab_add_vline(at.column = c(3), column.side = "right", linewidth = 3, linetype = 1) %>% tab_add_vline(at.column = 2:tab_ncol(loadings_plot), column.side = "left", from.row = 2, linetype = 2) #Join individuals and variables in biplot fviz_pca_biplot(behav.pca.site, repel = TRUE,col.var = "Black", col.ind = "#696969") #Add in site as group set.seed(1) groups<-as.factor(behav.PCA$Site_name[1:20]) biplot<-fviz_pca_biplot(behav.pca.site, repel=T, habillage=chaet_behaviour_mean$Site_name, palette =c("deepskyblue2", "firebrick4"), col.var = "Black", alpha.var=0.6, addEllipses=TRUE, labelsize=7, expand=20, ellipse.type = "confidence", ellipse.alpha=0.3, geom=c("point", "text"))+ theme(axis.line = element_line(colour = "Black"), panel.grid.major = element_blank(), panel.grid.minor = element_blank(), legend.position= "none", axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), axis.text.y=element_text(size= 12, colour="Black"), axis.text.x = element_text( vjust=0.5, size=12, colour="Black"), title=NULL)+ ggtitle(NULL)+ labs(x="PC1(62.4%)", y="PC2 (16.7%)") biplot<-fviz_pca_biplot(behav.pca.site, repel=T, habillage=chaet_behaviour_mean$Site_name, palette =c("deepskyblue2", "firebrick4"), col.var = "Black", alpha.var=0.9, addEllipses=TRUE, labelsize=7, expand=20, ellipse.type = "confidence", ellipse.alpha=0.3, geom=c("point", "text"))+ theme(axis.line = element_line(colour = "Black"), panel.grid.major = element_blank(), panel.grid.minor = element_blank(), legend.position= "none", axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), axis.text.y=element_text(size= 12, colour="Black"), axis.text.x = element_text( vjust=0.5, size=12, colour="Black"), title=NULL)+ ggtitle(NULL)+ labs(x="PC1(62.4%)", y="PC2 (16.7%)") biplot_figure<-biplot+inset_element(loadings_plot, 0, 0.3, 1.8, 0.1, on_top=FALSE) ggsave(file="biplot_figure.tiff") ###################################################################### #Checking assumptions ###################################################################### #Sphericity (bartlett test) assumptions <- paf(as.matrix(behav_PCA), eigcrit = 1, convcrit = .001) bartlettTest <- cortest.bartlett(behav.pers.site.pca, n = 20) bartlettTest #Sampling accuracy print(assumptions$KMO) ###################################################################### ##### Abundance/ Density ##### ###################################################################### #Abundance/density abundance<-fish_abundance abundance_lun<- abundance%>% filter(Species=="C.lunulatus") abundance_lun$Site<-as.factor(abundance_lun$Site) abundance_lun$Depth<-as.factor(abundance_lun$Depth) mean_abundance<- aggregate(abundance_lun$Abundance, by=list(abundance_lun$Site), FUN=mean) sd_abundance<- aggregate(abundance_lun$Abundance, by=list(abundance_lun$Site), FUN=sd) mean_depth<-aggregate(abundance_lun$Abundance, by=list(abundance_lun$Site, abundance_lun$Depth), FUN=mean) sd_depth<- aggregate(abundance_lun$Abundance, by=list(abundance_lun$Site, abundance_lun$Depth), FUN=sd) lm.residual_abundance_lun<- lm(Abundance~Site, data=abundance_lun) sresid <- studres(lm.residual_abundance_lun) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) glm_abundance<-glm(Abundance~Site, data=abundance_lun, family = gaussian()) summary(glm_abundance) plot(glm_abundance) anova(glm_abundance) #plot sites abundance_plot<-ggplot(abundance_lun, aes(x=Site, y=Abundance))+ geom_boxplot(width=0.5, position=position_dodge(0.65))+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), axis.text.y=element_text(size= 12, colour="Black"), axis.text.x = element_text(angle=90, vjust=0.5, size=12, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ guides(fill=guide_legend(title=NULL))+ labs(x="Site", y="Density") ggsave(file="abundance_plot.png") #plot depth ggplot(abundance_lun, aes(x=Site, y=Abundance, fill=Depth))+ geom_boxplot(width=0.5, position=position_dodge(0.65))+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold"), axis.title.y = element_text(colour = "Black", face="bold"), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(angle=90, vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ guides(fill=guide_legend(title=NULL))+ scale_fill_brewer(palette="Greys")+ labs(x="Site", y="Abundance") #Species analysis with each site separately abundance_PK<- abundance%>% filter(Site=="HIGH") lm.residual_abundance_pk<- lm(Abundance~Species, data=abundance_PK) sresid <- studres(lm.residual_abundance_pk) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) glm_species<-glm(Abundance~Species, data=abundance_PK, family=gaussian()) summary(glm_species) plot(glm_species) #Compare with depth as interaction glm_species_depth<-glm(Abundance~Species*Depth, data=abundance_PK, family = gaussian()) summary(glm_species_depth) plot(glm_species_depth) anova(glm_species_depth) #Compare with depth as additional factor glm_species_depth_2<-glm(Abundance~Species+Depth, data=abundance_PK, family = gaussian()) summary(glm_species_depth_2) plot(glm_species_depth_2) anova(glm_species_depth_2) #Compare with likelihood ratio test# lrtest(glm_species,glm_species_depth,glm_species_depth_2) #Use interaction term model in results. abundance_SAM<- abundance%>% filter(Site=="LOW") lm.residual_abundance_SAM<- lm(Abundance~Species, data=abundance_SAM) sresid <- studres(lm.residual_abundance_SAM) hist(sresid, freq=FALSE) xfit<-seq(min(sresid),max(sresid),length=40) yfit<-dnorm(xfit) lines(xfit, yfit) glm_species_sam<-glm(Abundance~Species, data=abundance_SAM, family=gaussian()) summary(glm_species_sam) plot(glm_species_sam) #Compare with depth as interaction glm_species_depth<-glm(Abundance~Species*Depth, data=abundance_SAM, family = gaussian()) summary(glm_species_depth) plot(glm_species_depth) anova(glm_species_depth) #Compare with depth as additional factor glm_species_depth_2<-glm(Abundance~Species+Depth, data=abundance_SAM, family = gaussian()) summary(glm_species_depth_2) plot(glm_species_depth_2) anova(glm_species_depth_2) #Compare with likelihood ratio test# lrtest(glm_species,glm_species_depth,glm_species_depth_2) #Use interaction term model in results. abundance.spp<-ggplot(abundance, aes(x=Site, y=Abundance, fill=Species))+ geom_boxplot(width=0.5, position=position_dodge(0.65))+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=16), axis.title.y = element_text(colour = "Black", face="bold", size=16), axis.text.y=element_text(size= 12, colour="Black"), axis.text.x = element_text(angle=90, vjust=0.5, size=12, colour="Black"), strip.background = element_rect( fill="White"), legend.text=element_text(size=12), legend.key = element_rect(fill = "transparent"), strip.text.x = element_text(face="bold", size=9.5))+ guides(fill=guide_legend(title=NULL))+ labs(x="Resource availability", y="Density")+ scale_fill_manual(values=c("White", "Dark Grey"))+ facet_wrap(~Depth) ggsave(file="abundance.spp.png") ###################################################################### #Individual level analyses==== ###################################################################### #Pair ID MANOVA #PK behav.pk<- chaet_behaviour%>% filter(Site_name=="PK") MANOVA_PK<- manova(cbind(Aggression,Exploration,Feeding,Sociability, Searching, Travelling)~Pair_ID, data=behav.pk) summary.aov(MANOVA_PK, p.adjust.method="bonferroni") #SAM behav.sam<- chaet_behaviour%>% filter(Site_name=="SAM") MANOVA_SAM<- manova(cbind(Aggression,Exploration,Feeding,Sociability, Searching, Travelling)~Pair_ID, data=behav.sam) summary.aov(MANOVA_SAM, p.adjust.method="bonferroni") #boxplot feeding ggplot(behav.sam, aes(x=Pair_ID, y=Feeding, fill=Site_name))+ geom_boxplot(width=0.3, position=position_dodge(0.45))+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=15), axis.title.y = element_text(colour = "Black", face="bold", size=15), axis.text.y=element_text(size= 15, colour="Black"), axis.text.x = element_text(angle=45, vjust=0.5, size=15, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=15))+ labs(x="Pair", y="Feeding")+ scale_y_continuous("Proportion", breaks =c(0,0.25,0.5,0.75,1))+ scale_fill_manual(values =c("grey39", "grey"),labels=c("HIGH", "LOW")) #Among individual variation data frame# summary.aov(MANOVA_PK) Site<-c("High", "Low","High", "Low","High", "Low","High", "Low","High", "Low","High", "Low","High", "Low","High", "Low","High", "Low","High", "Low") Behaviour<-c("Aggression", "Aggression", "Feeding", "Feeding", "Searching", "Searching", "Travelling", "Travelling", "Sociability", "Sociability","Aggression", "Aggression", "Feeding", "Feeding", "Searching", "Searching", "Travelling", "Travelling", "Sociability", "Sociability") Total<-c(1104.78, 65.041, 13740, 16931.1, 12876.9, 14351.4, 3390.2,8060.6,22624,25681) Variance<-c(413.52, 37.505, 6517.9, 11703.2, 7236.2, 5958.8, 1849.7, 3984.0, 9964,10434) Residual<-Total-Variance Variance<-c(413.52, 37.505, 6517.9, 11703.2, 7236.2, 5958.8, 1849.7, 3984.0, 9964,10434, Residual) Type<-c("Among","Among","Among","Among","Among","Among","Among","Among","Among","Among", "Residual","Residual","Residual","Residual","Residual","Residual","Residual","Residual","Residual","Residual") Proportion<-Variance/Total Variance<-as.data.frame(cbind(Site, Behaviour, Variance, Type, Proportion)) #Plots of variation #Actual variations variation ggplot(Variance, aes(x=Site, y=Proportion, fill=Type))+ geom_bar(stat="Identity", width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold"), axis.title.y = element_text(colour = "Black", face="bold"), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(angle=45, vjust=0.5, size=10, colour="Black"))+ labs(x="Behaviour", y= "Data points")+ guides(fill=guide_legend(title=NULL))+ scale_fill_manual(values=c("Black", "Grey"))+ facet_wrap(~Behaviour) Variance$Behaviour<-factor(Behaviour, levels=c("Aggression","Feeding","Searching", "Travelling", "Sociability")) #Proportion of variation Variance$Proportion<-as.numeric(Variance$Proportion) variance_prop<-ggplot(Variance, aes(x=Site, y=Proportion, fill=Type))+ geom_bar(stat="Identity", width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), strip.background = element_rect( fill="White"), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), legend.text = element_text(size = 10, face="bold"), panel.spacing=unit(0.7, "cm"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"))+ labs(x=NULL, y= "Proportion")+ scale_y_continuous(breaks=seq(0, 1, 0.25))+ guides(fill=guide_legend(title=NULL))+ scale_fill_manual(values=c("Black", "Grey"))+ facet_wrap(~Behaviour, nrow=1, scales="free_x")+ theme(strip.text = element_text(size = 10, face="bold")) #Individual behaviours #Aggression Variance_agg<- Variance%>% filter(Behaviour=="Aggression") agg_variance<-ggplot(Variance_agg, aes(x=Site, y=Proportion, fill=Type))+ geom_bar(stat="Identity", width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), plot.title=element_text(face="bold"), strip.background = element_rect( fill="White"), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), legend.text = element_text(size = 10, face="bold"), panel.spacing=unit(0.7, "cm"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"))+ labs(x=NULL, y= "Proportion")+ ggtitle("Aggression")+ guides(fill=guide_legend(title=NULL))+ scale_y_continuous(breaks=seq(0, 1, 0.25))+ scale_fill_manual(values=c("Black", "Grey")) #Feeding Variance_feed<- Variance%>% filter(Behaviour=="Feeding") feed_variance<-ggplot(Variance_feed, aes(x=Site, y=Proportion, fill=Type))+ geom_bar(stat="Identity", width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), strip.background = element_rect( fill="White"), plot.title=element_text(face="bold"), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), legend.text = element_text(size = 10, face="bold"), panel.spacing=unit(0.7, "cm"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"))+ labs(x=NULL, y= "Proportion")+ ggtitle("Feeding")+ guides(fill=guide_legend(title=NULL))+ scale_y_continuous(breaks=seq(0, 1, 0.25))+ scale_fill_manual(values=c("Black", "Grey")) #Searching Variance_search<- Variance%>% filter(Behaviour=="Searching") search_variance<-ggplot(Variance_search, aes(x=Site, y=Proportion, fill=Type))+ geom_bar(stat="Identity", width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), plot.title=element_text(face="bold"), strip.background = element_rect( fill="White"), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), legend.text = element_text(size = 10, face="bold"), panel.spacing=unit(0.7, "cm"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"))+ labs(x=NULL, y= "Proportion")+ ggtitle("Searching")+ guides(fill=guide_legend(title=NULL))+ scale_y_continuous(breaks=seq(0, 1, 0.25))+ scale_fill_manual(values=c("Black", "Grey")) #Travelling Variance_travel<- Variance%>% filter(Behaviour=="Travelling") travel_variance<-ggplot(Variance_travel, aes(x=Site, y=Proportion, fill=Type))+ geom_bar(stat="Identity", width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), plot.title=element_text(face="bold"), strip.background = element_rect( fill="White"), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), legend.text = element_text(size = 10, face="bold"), panel.spacing=unit(0.7, "cm"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"))+ labs(x=NULL, y= "Proportion")+ ggtitle("Travelling")+ guides(fill=guide_legend(title=NULL))+ scale_y_continuous(breaks=seq(0, 1, 0.25))+ scale_fill_manual(values=c("Black", "Grey")) #Sociability Variance_soc<- Variance%>% filter(Behaviour=="Sociability") sociability_variance<-ggplot(Variance_soc, aes(x=Site, y=Proportion, fill=Type))+ geom_bar(stat="Identity", width=0.7)+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), strip.background = element_rect( fill="White"), plot.title=element_text(face="bold"), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), legend.text = element_text(size = 10, face="bold"), panel.spacing=unit(0.7, "cm"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"))+ labs(x=NULL, y= "Proportion")+ ggtitle("Sociability")+ guides(fill=guide_legend(title=NULL))+ scale_y_continuous(breaks=seq(0, 1, 0.25))+ scale_fill_manual(values=c("Black", "Grey")) variance_plot<-ggarrange(agg_variance, feed_variance, search_variance, travel_variance, sociability_variance, nrow=5) #bootstrap means for each behaviour at each site behav_mean_PK<- chaet_behaviour_mean%>% filter(Site_name=="PK") behav_mean_SAM<- chaet_behaviour_mean%>% filter(Site_name=="SAM") #bootstrap populations for each behaviour boot.mean.full = function(x,B,binwidth=NULL) { n = length(x) boot.samples = matrix( sample(x,size=n*B,replace=TRUE), B, n) boot.statistics = apply(boot.samples,1,mean) se = sd(boot.statistics) require(ggplot2) if ( is.null(binwidth) ) binwidth = diff(range(boot.statistics))/30 p = ggplot(data.frame(x=boot.statistics),aes(x=x)) + geom_histogram(aes(y=..density..),binwidth=binwidth, fill="grey55")+ geom_density(color="Black")+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=20), axis.title.y = element_text(colour = "Black", face="bold", size=20), axis.text.y=element_text(size= 15, colour="Black"), axis.text.x = element_text(angle=45, vjust=0.5, size=15, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ labs(x="Aggression",y="Density")+ geom_vline(xintercept = mean(boot.statistics), linetype="dashed") plot(p) svg(filename="plot.svg") plot(p) dev.off() interval = mean(x) + c(-1,1)*2*se print( interval ) return( list(boot.statistics = boot.statistics, interval=interval, se=se, plot=p) ) } #just mean boot.mean = function(x,B,binwidth=NULL) { n = length(x) boot.samples = matrix( sample(x,size=n*B,replace=TRUE), B, n) boot.statistics = apply(boot.samples,1,mean) } #Aggression set.seed(1234) out.pk = with(behav_mean_PK, boot.mean(Aggression, B = 1000)) quantile(out.pk,c(0.025, 0.975)) set.seed(1234) out.sam = with(behav_mean_SAM, boot.mean(Aggression, B = 1000)) quantile(out.sam,c(0.025, 0.975)) PK<-data.frame(out.pk) names(PK)[1] <- "BOOT" SAM<-data.frame(out.sam) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' Agg <-rbind(PK, SAM) Aggression<-ggplot(Agg, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.5, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ geom_vline(xintercept=c(8.67,19.15), linetype="solid",color = "black", size=1)+ geom_vline(xintercept=c(0.19,3.26), linetype="dashed", color = "black", size=1)+ scale_fill_manual(values=c("Black", "Grey45"))+ labs(x="Aggression",y="Density") #Sociability set.seed(1234) out.pk = with(behav_mean_PK, boot.mean(Sociability, B = 1000)) quantile(out.pk,c(0.025, 0.975)) set.seed(1234) out.sam = with(behav_mean_SAM, boot.mean(Sociability, B = 1000)) quantile(out.sam,c(0.025, 0.975)) PK<-data.frame(out.pk) names(PK)[1] <- "BOOT" SAM<-data.frame(out.sam) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' Soc <-rbind(PK, SAM) Sociability<-ggplot(Soc, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.5, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(525.35,576.06), linetype="solid",color = "black", size=1)+ geom_vline(xintercept=c(467.41,522.42), linetype="dashed", color = "black", size=1)+ labs(x="Sociability",y=NULL) #Searching set.seed(1234) out.pk = with(behav_mean_PK, boot.mean(Searching, B = 1000)) quantile(out.pk,c(0.025, 0.975)) set.seed(1234) out.sam = with(behav_mean_SAM, boot.mean(Searching, B = 1000)) quantile(out.sam,c(0.025, 0.975)) PK<-data.frame(out.pk) names(PK)[1] <- "BOOT" SAM<-data.frame(out.sam) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' Search <-rbind(PK, SAM) Searching<-ggplot(Search, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.5, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ geom_vline(xintercept=c(249.68,294.17), linetype="solid",color = "black", size=1)+ geom_vline(xintercept=c(303.21,344.61), linetype="dashed", color = "black", size=1)+ scale_fill_manual(values=c("Black", "Grey45"))+ labs(x="Searching",y=NULL) #Travelling set.seed(1234) out.pk = with(behav_mean_PK, boot.mean(Travelling, B = 1000)) quantile(out.pk,c(0.025, 0.975)) set.seed(1234) out.sam = with(behav_mean_SAM, boot.mean(Travelling, B = 1000)) quantile(out.sam,c(0.025, 0.975)) PK<-data.frame(out.pk) names(PK)[1] <- "BOOT" SAM<-data.frame(out.sam) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' Travel <-rbind(PK, SAM) Travelling<-ggplot(Travel, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.5, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(19.20,40.63), linetype="solid",color = "black", size=1)+ geom_vline(xintercept=c(50.71,86.02), linetype="dashed", color = "black", size=1)+ labs(x="Travelling",y=NULL) #Feeding set.seed(1234) out.pk = with(behav_mean_PK, boot.mean(Feeding, B = 1000)) quantile(out.pk,c(0.025, 0.975)) set.seed(1234) out.sam = with(behav_mean_SAM, boot.mean(Feeding, B = 1000)) quantile(out.sam,c(0.025, 0.975)) PK<-data.frame(out.pk) names(PK)[1] <- "BOOT" SAM<-data.frame(out.sam) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' Feed <-rbind(PK, SAM) Feeding<-ggplot(Feed, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.5, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ geom_vline(xintercept=c(258.54,300.33), linetype="solid",color = "black", size=1)+ geom_vline(xintercept=c(166.50,224.42), linetype="dashed", color = "black", size=1)+ scale_fill_manual(values=c("Black", "Grey45"))+ labs(x="Feeding",y=NULL) #Join plots# bootstrap_plot<-ggarrange(Aggression, Feeding, Searching, Travelling, Sociability, nrow=1) ggsave("bootstrap_plot.tiff", width=12) #Repeatability and bootstrapped distributions #PK repeat_PK<- chaet_behaviour%>% filter(Site_name=="PK") MANOVA_PK<- manova(cbind(Aggression,Exploration,Feeding,Sociability, Searching, Travelling)~Pair_ID, data=repeat_PK) summary.aov(MANOVA_PK) #SAM repeat_SAM<- chaet_behaviour%>% filter(Site_name=="SAM") MANOVA_SAM<- manova(cbind(Aggression,Exploration,Feeding,Sociability, Searching, Travelling)~Pair_ID, data=repeat_SAM) summary.aov(MANOVA_SAM) #Sociability set.seed(1234) out.mean = with(repeat_PK, boot.mean(Sociability, B = 1000)) PK_soc<-ICCest(Pair_ID, Sociability, data=repeat_PK, CI.type="Smith") PK_soc_boot<-boot.icc(repeat_PK$Sociability, repeat_PK$Pair_ID, 1000, aov.est=TRUE) summary(PK_soc_boot) quantile(PK_soc_boot,c(0.025, 0.975)) as.vector(PK_soc_boot) t.test(PK_soc_boot,conf.level = 0.95) #bootstrap among individual variation PK_soc_aov<- aov(Sociability~Pair_ID, data=repeat_PK) summary(PK_soc_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(PK_soc_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Sociability~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.PK <- data.frame(out.mean, between) data.PK$MS=100*(between/(out.mean^2)) mean(data.PK$MS) quantile(data.PK$MS,c(0.025, 0.975)) #SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Sociability, B = 1000)) ICCest(Pair_ID, Sociability, data=repeat_SAM, CI.type="Smith") SAM_soc_boot<-boot.icc(repeat_SAM$Sociability, repeat_SAM$Pair_ID, 1000, aov.est=TRUE) summary(SAM_soc_boot) quantile(SAM_soc_boot,c(.025,0.975)) as.vector(SAM_soc_boot) t.test(SAM_soc_boot,conf.level = 0.95) #bootstrap among individual variation SAM_soc_aov<- aov(Aggression~Pair_ID, data=repeat_SAM) summary(SAM_soc_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(SAM_soc_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Sociability~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) #combine vectors out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.SAM <- data.frame(out.mean, between) data.SAM$MS=100*(between/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Soc.MS <-rbind(PK, SAM) #Feeding set.seed(1234) out.mean = with(repeat_PK, boot.mean(Feeding, B = 1000)) ICCest(Pair_ID, Feeding, data=repeat_PK, CI.type="Smith") PK_feed_boot<-boot.icc(repeat_PK$Feeding, repeat_PK$Pair_ID, 1000, aov.est=TRUE) summary(PK_feed_boot) quantile(PK_feed_boot,c(0.025, 0.975)) as.vector(PK_feed_boot) t.test(PK_feed_boot,conf.level = 0.95) #bootstrap among individual variation PK_F_aov<- aov(Feeding~Pair_ID, data=repeat_PK) summary(PK_F_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(PK_F_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Feeding~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.PK <- data.frame(out.mean, between) data.PK$MS=(between/(out.mean^2)*100) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #Feeding SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Feeding, B = 1000)) ICCest(Pair_ID, Feeding, data=repeat_SAM, CI.type="Smith") #GLM equivalent= 0.200 SAM_feed_boot<-boot.icc(repeat_SAM$Feeding, repeat_SAM$Pair_ID, 1000, aov.est=TRUE) summary(SAM_feed_boot) quantile(SAM_feed_boot,c(0.025, 0.975)) as.vector(SAM_feed_boot) t.test(SAM_feed_boot,conf.level = 0.95) #bootstrap among individual variation SAM_F_aov<- aov(Feeding~Pair_ID, data=repeat_SAM) summary(SAM_F_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(SAM_F_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Feeding~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) #combine vectors out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.SAM <- data.frame(out.mean, between) data.SAM$MS=100*(between/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Feed.MS <-rbind(PK, SAM) #Searching set.seed(12334) out.mean = with(repeat_PK, boot.mean(Searching, B = 1000)) ICCest(Pair_ID, Searching, data=repeat_PK, CI.type="Smith") #GLM equivalent=0.054 PK_search_boot<-boot.icc(repeat_PK$Searching, repeat_PK$Pair_ID, 1000, aov.est=TRUE) summary(PK_search_boot) quantile(PK_search_boot,c(0.025, 0.975)) as.vector(PK_search_boot) t.test(PK_search_boot,conf.level = 0.95) #bootstrap among individual variation PK_FS_aov<- aov(Searching~Pair_ID, data=repeat_PK) summary(PK_FS_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(PK_FS_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Searching~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) data.PK <- data.frame(out.mean, between) data.PK$MS=100*(between/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #Searching SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Searching, B = 1000)) ICCest(Pair_ID, Searching, data=repeat_SAM, CI.type="Smith") SAM_search_boot<-boot.icc(repeat_SAM$Searching, repeat_SAM$Pair_ID, 1000, aov.est=TRUE) summary(SAM_search_boot) quantile(SAM_search_boot,c(0.025, 0.975)) as.vector(SAM_search_boot) t.test(SAM_search_boot,conf.level = 0.95) #bootstrap among individual variation SAM_FS_aov<- aov(Searching~Pair_ID, data=repeat_SAM) summary(SAM_FS_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(SAM_FS_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Searching~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) #combine vectors out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.SAM <- data.frame(out.mean, between) data.SAM$MS=100*(between/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS, c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Search.MS <-rbind(PK, SAM) #Travelling set.seed(1234) out.mean = with(repeat_PK, boot.mean(Travelling, B = 1000)) ICCest(Pair_ID, Travelling, data=repeat_PK, CI.type="Smith") #GLM equivalent=0.039 PK_travel_boot<-boot.icc(repeat_PK$Travelling, repeat_PK$Pair_ID, 1000, aov.est=TRUE) summary(PK_travel_boot) quantile(PK_travel_boot,c(0.025, 0.975)) as.vector(PK_travel_boot) t.test(PK_travel_boot,conf.level = 0.95) #bootstrap among individual variation PK_S_aov<- aov(Travelling~Pair_ID, data=repeat_PK) summary(PK_S_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(PK_S_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Travelling~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) data.PK <- data.frame(out.mean, between) data.PK$MS=100*(between/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #Travelling SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Travelling, B = 1000)) ICCest(Pair_ID, Travelling, data=repeat_SAM, CI.type="Smith") SAM_travel_boot<-boot.icc(repeat_SAM$Travelling, repeat_SAM$Pair_ID, 1000, aov.est=TRUE) summary(SAM_travel_boot) quantile(SAM_travel_boot,c(0.025, 0.975)) as.vector(SAM_travel_boot) t.test(SAM_travel_boot,conf.level = 0.95) #bootstrap among individual variation SAM_S_aov<- aov(Travelling~Pair_ID, data=repeat_SAM) summary(SAM_S_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(SAM_S_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Travelling~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) #combine vectors out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.SAM <- data.frame(out.mean, between) data.SAM$MS=100*(between/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Travel.MS <-rbind(PK, SAM) #Aggression set.seed(1234) #bootstrap population out.mean = with(repeat_PK, boot.mean(Aggression, B = 1000)) ICCest(Pair_ID, Aggression, data=repeat_PK, CI.type="Smith") PK_agg_boot<-boot.icc(repeat_PK$Aggression, repeat_PK$Pair_ID, 1000, aov.est=TRUE) summary(PK_agg_boot) quantile(PK_agg_boot,c(0.025, 0.975)) as.vector(PK_agg_boot) t.test(PK_agg_boot,conf.level = 0.95) #bootstrap among individual variation PK_Agg_aov<- aov(Aggression~Pair_ID, data=repeat_PK) summary(PK_Agg_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(PK_Agg_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Aggression~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) #combine vectors out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.PK <- data.frame(out.mean, between) data.PK$MS=100*(between/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #Aggression SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Aggression, B = 1000)) ICCest(Pair_ID, Aggression, data=repeat_SAM, CI.type="Smith") SAM_agg_boot<-boot.icc(repeat_SAM$Aggression, repeat_SAM$Pair_ID, 1000, aov.est=TRUE) summary(SAM_agg_boot) quantile(SAM_agg_boot,c(0.025, 0.975), na.rm=TRUE) as.vector(SAM_agg_boot) t.test(SAM_agg_boot,conf.level = 0.95) #bootstrap among individual variation SAM_agg_aov<- aov(Aggression~Pair_ID, data=repeat_SAM) summary(SAM_agg_aov) get_MS <- function(model) { anova(model)$`Mean Sq`[1] } get_MS(SAM_agg_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Aggression~Pair_ID, data = .))) %>% mutate(between_group_ms = map_dbl(models, get_MS)) #Mean and confidence interval between<-as.vector(permuted$between_group_ms) t.test(between,conf.level = 0.95) #combine vectors out.mean<-as.vector(out.mean) between<-as.vector(permuted$between_group_ms) data.SAM <- data.frame(out.mean, between) data.SAM$MS=100*(between/(out.mean^2)) data.clean <- data.SAM[-which(data$out.mean == "0" ), ] t.test(data.SAM$MS,conf.level = 0.95) quantile(data.clean$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Agg.MS <-rbind(PK, SAM) #Differences between sites- plot ICC distributions==== #Sociability PK<-data.frame(PK_soc_boot) names(PK)[1] <- "BOOT" SAM<-data.frame(SAM_soc_boot) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' Soc.ICC <-rbind(PK, SAM) Sociability.ICC<-ggplot(Soc.ICC, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.6, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("deepskyblue2", "firebrick4"))+ geom_vline(xintercept=c(-0.08,0.45), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(-0.11,0.43), linetype="dashed", color = "black", size=0.5)+ labs(x="Sociability",y="Density", fill="Coral cover") #Feeding PK<-data.frame(PK_feed_boot) names(PK)[1] <- "BOOT" SAM<-data.frame(SAM_feed_boot) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' feed.ICC <-rbind(PK, SAM) Feeding.ICC<-ggplot(feed.ICC, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.6, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("deepskyblue2", "firebrick4"))+ geom_vline(xintercept=c(-0.08,0.45), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(-0.02,0.61), linetype="dashed",color = "black", size=0.5)+ labs(x="Feeding",y="Density", fill="Coral cover") #Searching PK<-data.frame(PK_search_boot) names(PK)[1] <- "BOOT" SAM<-data.frame(SAM_search_boot) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' search.ICC <-rbind(PK, SAM) Searching.ICC<-ggplot(search.ICC, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.6, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("deepskyblue2", "firebrick4"))+ geom_vline(xintercept=c(-0.07,0.50), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(-0.15,0.48), linetype="dashed",color = "black", size=0.5)+ labs(x="Searching",y="Density", fill="Coral cover") #Travelling PK<-data.frame(PK_travel_boot) names(PK)[1] <- "BOOT" SAM<-data.frame(SAM_travel_boot) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' travel.ICC <-rbind(PK, SAM) Travelling.ICC<-ggplot(travel.ICC, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.6, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("deepskyblue2", "firebrick4"))+ geom_vline(xintercept=c(-0.03,0.59), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(-0.08,0.46), linetype="dashed",color = "black", size=0.5)+ labs(x="Travelling",y="Density", fill="Coral cover") #Aggression PK<-data.frame(PK_agg_boot) names(PK)[1] <- "BOOT" SAM<-data.frame(SAM_agg_boot) names(SAM)[1] <- "BOOT" PK$Site <- 'HIGH' SAM$Site <- 'LOW' Agg.ICC <-rbind(PK, SAM) Aggression.ICC<-ggplot(Agg.ICC, aes(BOOT, fill = Site)) + geom_histogram(alpha = 0.6, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text( vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("deepskyblue2", "firebrick4"))+ geom_vline(xintercept=c(-0.10,0.40), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(-0.05,0.59), linetype="dashed",color = "black", size=0.5)+ labs(x="Aggression",y="Density", fill="Coral cover") #combine plots# ICC_plot<-ggarrange(Aggression.ICC, Feeding.ICC, Searching.ICC, Travelling.ICC, Sociability.ICC, nrow=1) #Add summary table #Aggression R<-c("0.108", "0.195") Ia<-c("421.735", "4300.551") Iw<-c("410.463", "4497.196") Table_Aggression<-rbind(R,Ia,Iw) colnames(Table_Aggression) <- c("High", "Low") rownames(Table_Aggression) <- c("R", "Ia", "Iw") #Feeding R<-c("0.164", "0.298") Ia<-c("8.987", "16.964") Iw<-c("9.091", "17.272") Table_Feeding<-rbind(R,Ia,Iw) colnames(Table_Feeding) <- c("High", "Low") rownames(Table_Feeding) <- c("R", "Ia", "Iw") #Searching R<-c("0.211", "0.104") Ia<-c("8.067", "7.638") Iw<-c("8.133", "7.537") Table_Searching<-rbind(R,Ia,Iw) colnames(Table_Searching) <- c("High", "Low") rownames(Table_Searching) <- c("R", "Ia", "Iw") #Travel R<-c("0.227", "0.162") Ia<-c("194.054", "100.117") Iw<-c("197.323", "99.463") Table_Travel<-rbind(R,Ia,Iw) colnames(Table_Travel) <- c("High", "Low") rownames(Table_Travel) <- c("R", "Ia", "Iw") #Sociability R<-c("0.112", "0.099") Ia<-c("3.952", "5.915)") Iw<-c("3.946", "5.821") Table_Sociability<-rbind(R,Ia,Iw) colnames(Table_Sociability) <- c("High", "Low") rownames(Table_Sociability) <- c("R", "Ia", "Iw") #Finalise tables Table_Aggression<-ggtexttable(Table_Aggression, theme=ttheme(base_style="blank")) Table_Aggression<-Table_Aggression%>% tab_add_vline(at.column = c(2), column.side = "right", linewidth = 3, linetype = 1) %>% tab_add_hline(at.row = c(1), row.side = "bottom", linewidth = 3, linetype = 1) Table_Feeding<-ggtexttable(Table_Feeding, theme=ttheme(base_style = "blank")) Table_Feeding<-Table_Feeding%>% tab_add_vline(at.column = c(2), column.side = "right", linewidth = 3, linetype = 1) %>% tab_add_hline(at.row = c(1), row.side = "bottom", linewidth = 3, linetype = 1) Table_Searching<-ggtexttable(Table_Searching, theme=ttheme(base_style = "blank")) Table_Searching<-Table_Searching%>% tab_add_vline(at.column = c(2), column.side = "right", linewidth = 3, linetype = 1) %>% tab_add_hline(at.row = c(1), row.side = "bottom", linewidth = 3, linetype = 1) Table_Travel<-ggtexttable(Table_Travel, theme=ttheme(base_style = "blank")) Table_Travel<-Table_Travel%>% tab_add_vline(at.column = c(2), column.side = "right", linewidth = 3, linetype = 1) %>% tab_add_hline(at.row = c(1), row.side = "bottom", linewidth = 3, linetype = 1) Table_Sociability<-ggtexttable(Table_Sociability, theme=ttheme(base_style = "blank")) Table_Sociability<-Table_Sociability%>% tab_add_vline(at.column = c(2), column.side = "right", linewidth = 3, linetype = 1) %>% tab_add_hline(at.row = c(1), row.side = "bottom", linewidth = 3, linetype = 1) Aggression_plot<-ggarrange(agg_variance, Aggression.ICC, Table_Aggression, nrow=1, widths=c(1,1,0.5)) Feeding_plot<-ggarrange(feed_variance, Feeding.ICC, Table_Feeding, nrow=1,widths=c(1,1,0.5)) Searching_Plot<-ggarrange(search_variance, Searching.ICC, Table_Searching, nrow=1,widths=c(1,1,0.5)) Travelling_plot<-ggarrange(travel_variance, Travelling.ICC, Table_Travel, nrow=1,widths=c(1,1,0.5)) Sociability_plot<-ggarrange(sociability_variance, Sociability.ICC, Table_Sociability, nrow=1,widths=c(1,1,0.5)) #combine plots and table repeatability_plot<-ggarrange(Aggression_plot, Feeding_plot, Searching_Plot, Travelling_plot, Sociability_plot, nrow=5) ggsave("repeatability_plot.tiff", height=9, width=9) #plot mean standardisation distributions# Sociability.MS<-ggplot(Soc.MS, aes(MS, fill = Site)) + geom_histogram(alpha = 0.8, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(3.87, 4.03), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(5.78, 6.05), linetype="dashed", color = "black", size=0.5)+ labs(x="Sociability",y=NULL) Feeding.MS<-ggplot(Feed.MS, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(2.92, 17.48), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(6.29, 34.75), linetype="dashed",color = "black", size=0.5)+ labs(x="Feeding",y=NULL) Searching.MS<-ggplot(Search.MS, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(2.54,15.48), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(2.70, 13.60), linetype="dashed",color = "black", size=0.5)+ labs(x="Searching",y=NULL) Travelling.MS<-ggplot(Travel.MS, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(37.15, 226.93), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(26.45, 171.79), linetype="dashed",color = "black", size=0.5)+ labs(x="Travelling",y=NULL) Aggression.MS<-ggplot(Agg.MS, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(102.04, 1019.61), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(217.19, 27788.38), linetype="dashed",color = "black", size=0.5)+ labs(x="Aggression",y="Density") #join plots MS_plot<-ggarrange(Aggression.MS, Feeding.MS, Searching.MS, Travelling.MS, Sociability.MS, nrow=1) ggsave("MS_plot.tiff", width=12) ##### Bootstrap residual MS ##### #Aggression #PK set.seed(1234) out.mean = with(repeat_PK, boot.mean(Aggression, B = 1000)) PK_Agg_aov<- aov(Aggression~Pair_ID, data=repeat_PK) summary(PK_Agg_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(PK_Agg_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Aggression~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.PK <- data.frame(out.mean, within) data.PK$MS=100*(within/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Aggression, B = 1000)) #bootstrap within individual variation SAM_Agg_aov<- aov(Aggression~Pair_ID, data=repeat_SAM) summary(SAM_Agg_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(SAM_Agg_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Aggression~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.SAM <- data.frame(out.mean, within) data.SAM$MS=100*(within/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) mean(data.SAM$MS) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Agg.MS.R <-rbind(PK, SAM) Aggression.MS.R<-ggplot(Agg.MS.R, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(165.05, 854.78), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(267.78, 84980.47), linetype="dashed",color = "black", size=0.5)+ labs(x="Aggression",y="Density") #Feeding #PK set.seed(1234) out.mean = with(repeat_PK, boot.mean(Feeding, B = 1000)) PK_Feed_aov<- aov(Feeding~Pair_ID, data=repeat_PK) summary(PK_Feed_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(PK_Feed_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Feeding~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.PK <- data.frame(out.mean, within) data.PK$MS=100*(within/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Feeding, B = 1000)) #bootstrap within individual variation SAM_Feed_aov<- aov(Feeding~Pair_ID, data=repeat_SAM) summary(SAM_Feed_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(SAM_Feed_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Feeding~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.SAM <- data.frame(out.mean, within) data.SAM$MS=100*(within/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Feed.MS.R <-rbind(PK, SAM) Activity_F.MS.R<-ggplot(Feed.MS.R, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(6.86, 11.26), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(11.61, 24.19), linetype="dashed",color = "black", size=0.5)+ labs(x="Feeding",y=NULL) svg(filename="Activity_F_MS_R.svg") plot(Activity_F.MS.R) dev.off() #Searching #PK set.seed(1234) out.mean = with(repeat_PK, boot.mean(Searching, B = 1000)) #bootstrap within individual variation PK_search_aov<- aov(Searching~Pair_ID, data=repeat_PK) summary(PK_search_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(PK_search_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Searching~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.PK <- data.frame(out.mean, within) data.PK$MS=100*(within/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Searching, B = 1000)) #bootstrap within individual variation SAM_search_aov<- aov(Searching~Pair_ID, data=repeat_SAM) summary(SAM_search_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(SAM_search_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Searching~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.SAM <- data.frame(out.mean, within) data.SAM$MS=100*(within/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' search.MS.R <-rbind(PK, SAM) FS.MS.R<-ggplot(search.MS.R, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(6.06, 10.33), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(7.47, 7.57), linetype="dashed",color = "black", size=0.5)+ labs(x="Searching",y="Density") #Travelling #PK set.seed(1234) out.mean = with(repeat_PK, boot.mean(Travelling, B = 1000)) #bootstrap within individual variation PK_travel_aov<- aov(Travelling~Pair_ID, data=repeat_PK) summary(PK_travel_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(PK_travel_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Travelling~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.PK <- data.frame(out.mean, within) data.PK$MS=100*(within/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Travelling, B = 1000)) #bootstrap within individual variation SAM_travel_aov<- aov(Travelling~Pair_ID, data=repeat_SAM) summary(SAM_travel_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(SAM_travel_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Travelling~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.SAM <- data.frame(out.mean, within) data.SAM$MS=100*(within/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' travel.MS.R <-rbind(PK, SAM) S.MS.R<-ggplot(travel.MS.R, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), legend.position="none", axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(91.56, 457.18), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(56.29, 161.75), linetype="dashed",color = "black", size=0.5)+ labs(x="Travelling",y="Density") #Sociability #PK set.seed(1234) out.mean = with(repeat_PK, boot.mean(Sociability, B = 1000)) #bootstrap within individual variation PK_soc_aov<- aov(Sociability~Pair_ID, data=repeat_PK) summary(PK_soc_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(PK_soc_aov) permuted <- repeat_PK %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Sociability~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.PK <- data.frame(out.mean, within) data.PK$MS=100*(within/(out.mean^2)) t.test(data.PK$MS,conf.level = 0.95) quantile(data.PK$MS,c(0.025, 0.975)) #SAM set.seed(1234) out.mean = with(repeat_SAM, boot.mean(Sociability, B = 1000)) #bootstrap within individual variation SAM_soc_aov<- aov(Sociability~Pair_ID, data=repeat_SAM) summary(SAM_soc_aov) get_WS <- function(model) { anova(model)$`Mean Sq`[2] } get_WS(SAM_soc_aov) permuted <- repeat_SAM %>% modelr::permute(1000, Pair_ID) %>% mutate(models = map(perm, ~ aov(Sociability~Pair_ID, data = .))) %>% mutate(within_group_ms = map_dbl(models, get_WS)) #Mean and confidence interval within<-as.vector(permuted$within_group_ms) t.test(within,conf.level = 0.95) out.mean<-as.vector(out.mean) within<-as.vector(permuted$within_group_ms) data.SAM <- data.frame(out.mean, within) data.SAM$MS=100*(within/(out.mean^2)) t.test(data.SAM$MS,conf.level = 0.95) quantile(data.SAM$MS,c(0.025, 0.975)) PK<-data.frame(data.PK) SAM<-data.frame(data.SAM) PK$Site <- 'HIGH' SAM$Site <- 'LOW' Soc.MS.R <-rbind(PK, SAM) Sociability.MS.R<-ggplot(Soc.MS.R, aes(MS, fill = Site)) + geom_histogram(alpha = 0.7, aes(y = ..density..), position = 'identity')+ theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), panel.background = element_blank(), axis.line = element_line(colour = "Black"), axis.title.x = element_text(colour = "Black", face="bold", size=12), axis.title.y = element_text(colour = "Black", face="bold", size=12), axis.text.y=element_text(size= 10, colour="Black"), axis.text.x = element_text(vjust=0.5, size=10, colour="Black"), strip.background = element_rect( fill="White"), strip.text.x = element_text(face="bold", size=9.5))+ scale_fill_manual(values=c("Black", "Grey45"))+ geom_vline(xintercept=c(3.08, 4.64), linetype="solid",color = "black", size=0.5)+ geom_vline(xintercept=c(4.46, 6.92), linetype="dashed",color = "black", size=0.5)+ labs(x="Sociability",y=NULL) #combine plots# Residual_plot<-ggarrange(Aggression.MS.R, Activity_F.MS.R, FS.MS.R, S.MS.R, Sociability.MS.R, nrow=1) ggsave("Residual_plot.tiff", width=12)