Add extra details and clean

clean comments

H1ResponseAnomaly.R

@@ -345,7 +345,8 @@ book.plot.field(
   inla.space1b$summary.random$s$mean  , 
   mesh = cmesh) 
 
-plot(inla.space1b$summary.linear.predictor$mean,butterfly.dt$n_change,col = alpha("grey", 0.4),xlab="Pred",ylab="obs")
+# Pred vs obs - similar to the first model 
+plot(inla.space1b$summary.fitted.values$mean[1:1504],butterfly.dt$n_change,col = alpha("grey", 0.4),xlab="Pred",ylab="obs")
 abline(0,1)
 
 # plots

H2CommunityAnalysis.R

@@ -29,6 +29,12 @@ inla.m4 <- inla(ICOV ~ 1 + rangescore + rangescore_2 + meanpopsize + sppnum,
 summary(inla.m4)
 bri.fixed.plot(inla.m4,together = F)
 
+# check obs vs pred - not brilliant but figure below shows better
+library(scales)
+plot(inla.m4$summary.linear.predictor$mean,stability$ICOV,col = alpha("grey", 0.4),xlab="Pred",ylab="obs")
+abline(0,1)
+
+
 # -------- Fit spatial barrier model ----------
 
 library(sf)
@@ -169,6 +175,9 @@ book.plot.field(
 
 epoints[,1] <- epoints[,1] + inla.space1c$summary.fixed$mean[1]
 
+# observed vs predicted
+plot(inla.space1c$summary.linear.predictor$mean,stability$ICOV,col = alpha("grey", 0.4),xlab="Pred",ylab="obs")
+abline(0,1)
 
 
 # -------- Plot spatial model marginal fit ----------

H3_H4Mechanisms.R

@@ -146,7 +146,6 @@ stackd <- inla.stack(
                     speciesname1= speciesstability$speciesname1 ,
                     speciesname3= speciesstability$speciesname1 ,
                     speciesname2= speciesstability$speciesname2 )) # spatial random effect
-  #iidx=1:nrow(df), # iid random effect
   ,
   tag= "data"
 )
@@ -160,7 +159,6 @@ inla.space1d<- inla(y~0+Intercept + rangepositionspring+ rangepositionspring_2+
                  f(speciesname3,rangepositionspring,model="iid",hyper=rand_prior) +f(speciesname4,rangepositionspring_2,model="iid",hyper=rand_prior)
                  , 
                  data = inla.stack.data(stackd),family="Gaussian",
-               #  control.inla =list(int.strategy ="eb",tolerance=1e-6,reordering ="metis",strategy="gaussian"),
                control.compute = list(config=TRUE,waic = TRUE,dic=TRUE),
                control.predictor = list(A = inla.stack.A(stackd),compute=TRUE),
                control.fixed = list(
@@ -254,7 +252,6 @@ simpop<-cbind.data.frame(simpop,estimP,estimlP,estimuP)
 #tiff("fit3_2.tif", res=600, compression = "lzw", height=6, width=6.5, units="in")
 ggplot(data=ptdt2,aes(meanpop,resid,color=country,group=country))+
   geom_point(size=4,alpha=0.7)+
-  # xlim(0,1)+
   scale_color_npg()+
   theme_cowplot() +
   geom_line(data=simpop,aes(simpop,estimP), size=2,alpha=0.9,inherit.aes = F) +
@@ -345,7 +342,6 @@ synchdat<-synchdat[complete.cases(synchdat),]
 # -------- Fit model ----------
 inla.m6<- inla(avgcor ~ 1 + rangescore+ rangescore_2 + nspp +varcor, 
                data=synchdat,
-               #  control.inla =list(int.strategy ="eb",tolerance=1e-6,reordering ="metis",strategy="gaussian"),
                control.compute = list(config=TRUE,waic = TRUE,dic=TRUE),
                control.fixed = list(
                  mean=0, prec= 0.25, # fixed effect priors