merge.data.PARACOU.R 7.73 KiB
### MERGE paracou DATA
### Edited by FH
rm(list = ls()); source("./R/format.function.R"); library(reshape)
#########################
## READ DATA
####################
### read individuals tree data
data.paracou <- read.table("./data/raw/DataParacou/20130717_paracou_1984_2012.csv",header=TRUE,stringsAsFactors=FALSE,sep = ";", na.strings = "NULL")
#barplot(apply(!is.na(data.paracou[,paste("circ_",1984:2012,sep="")]),MARGIN=2,FUN=sum),las=3)
# select good columns
data.paracou <- data.paracou[,c("foret","parcelle","carre","arbre","vernaculaire","idtaxon",
                                "x","y","circ_2001","code_2001","circ_2005","code_2005",
                                "circ_2009","code_2009","campagne_mort","type_mort")]
colnames(data.paracou) <- c("forest","plot","subplot","tree","vernacular","taxonid","x","y","circum2001","code2001","circum2005","code2005","circum2009","code2009","yeardied","typedeath")
for(k in 7:14) { 
	data.paracou[,k] <- gsub(",",".",data.paracou[,k]); data.paracou[,k] <- as.numeric(data.paracou[,k]) } ## Replace all , in decimals with .
data.paracou$treeid <- apply(data.paracou[,1:4],1,paste,collapse="."); ## Create a tree id
data.paracou <- data.paracou[,c(ncol(data.paracou),1:(ncol(data.paracou)-1))]	
## REMOVE ALL TREES WITH X OR Y >250 m or NA ASK GHISLAIN!!!
data.paracou <- subset(data.paracou,subset=(!is.na(data.paracou[["x"]])) & data.paracou[["x"]]<251 &  data.paracou[["y"]]<251)
## plot each 
pdf("./figs/plots.paracou.pdf")
lapply(unique(data.paracou[["plot"]]),FUN=fun.circles.plot,data.paracou[['x']],data.paracou[['y']],data.paracou[["plot"]],data.paracou[["circum2009"]],inches=0.2)
dev.off()
pdf("./figs/plots.paracou.pdf")
by(data.paracou[["circum2009"]],INDICES=data.paracou[["plot"]],FUN=function(x) hist(x,breaks=50))
dev.off()
####################
#### PLOT 16 17 18 ARE STRANGE ASK GHSILAIN
#####################
data.paracou <- subset(data.paracou,subset=! data.paracou[["plot"]] %in% 16:18)
## keep only tree alive in 2001
data.paracou <- subset(data.paracou,subset=!(as.numeric(data.paracou[["yeardied"]])<=2001 & !is.na(data.paracou[["yeardied"]])))
### read species names
species.clean <- read.csv("./data/raw/DataParacou/20130717_paracou_taxonomie.csv",stringsAsFactors=FALSE, header = T, sep = ";")
## SPECIES CODE COME FROM idTaxon in paracou_taxonomie and taxonid in paracou_1984_2012 to match the traits data we need to use the "Genus species"
## we better work not work with vernacular because this doesn't match necesseraly the Genus species taxonomie
species.clean$sp <- species.clean[["idTaxon"]]
## data.paracou <- merge(data.paracou, as.data.frame(species.clean[!duplicated(species.clean[["sp"]]),c("Genre","Espece","sp")]), by = "sp", sort = FALSE)
######################################
## MASSAGE TRAIT DATA
############################
##########################################
## FORMAT INDIVIDUAL TREE DATA
#############
data.paracou2 <- data.paracou[rep(1:nrow(data.paracou),each=2),c(1:10,(ncol(data.paracou)-2):ncol(data.paracou))]
rownames(data.paracou2) <- 1:nrow(data.paracou2); data.paracou2 <- as.data.frame(data.paracou2)
data.paracou2$yr1 <- rep(c(2001,2001+4),nrow(data.paracou)); data.paracou2$yr2 <- rep(c(2005,2005+4),nrow(data.paracou))
data.paracou2$year <- rep(c(4,4),nrow(data.paracou))
data.paracou2$dbh1 <- c(rbind(data.paracou$circum2001/pi,data.paracou$circum2005/pi))
data.paracou2$dbh2 <- c(rbind(data.paracou$circum2005/pi,data.paracou$circum2009/pi))
data.paracou2$code1 <- c(as.numeric(rbind(data.paracou$code2001,data.paracou$code2005)))
data.paracou2$code2 <- c(as.numeric(rbind(data.paracou$code2005,data.paracou$code2009)))
data.paracou2$dead <- rep(0,nrow(data.paracou)*2)
data.paracou2$dead[c(as.numeric(data.paracou[["yeardied"]]) %in% 2002:2005 & (!is.na(data.paracou[["yeardied"]])),
                     as.numeric(data.paracou[["yeardied"]]) %in% 2006:2009 & (!is.na(data.paracou[["yeardied"]])))] <- 1
7172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140
data.paracou2$sp <- data.paracou[["taxonid"]] ## remove tree dead in 2005 in second census (2005-2009) data.paracou <- subset(data.paracou2,subset=!(data.paracou2[['yr1']] ==2005 & (as.numeric(data.paracou[["yeardied"]]) %in% 2002:2005 & (!is.na(data.paracou[["yeardied"]]))))) ## change unit and names of variables to be the same in all data for the tree data.paracou$G <- 10*(data.paracou$dbh2-data.paracou$dbh1)/data.paracou$year ## diameter growth in mm per year data.paracou$G[data.paracou$code1>0] <- NA ## indivs with code indicating problem in dbh measurment at dbh1 data.paracou$G[data.paracou$code2>0] <- NA ## indivs with code indicating problem in dbh measurment at dbh2 data.paracou[which(data.paracou$G < -50),] ## THERE SEEMS TO BE SOME PROBLEMS WITH THE DBH DATA ## much less issue data.paracou$D <- data.paracou[["dbh1"]]; data.paracou$D[data.paracou$D == 0] <- NA ;## diameter in cm data.paracou$plot <- data.paracou$plot#apply(data.paracou[,c("forest","plot","subplot")],1,paste,collapse=".") ## plot code data.paracou$htot <- rep(NA,length(data.paracou[["G"]])) ## height of tree in m - MISSING data.paracou$obs.id <- 1:nrow(data.paracou) ### delete recruit in 2001 or 2005 for first census data.paracou <- subset(data.paracou,subset=!is.na(data.paracou$D)) ## minimum circumfer 30 delete all tree with a dbh <30/pi, data.paracou <- subset(data.paracou,subset= data.paracou[["D"]]>(30/pi)) ###################### ## ECOREGION ################### ## paracou has only 1 eco-region YES NO ECOREGION ###################### ## PERCENT DEAD ################### ## variable percent dead ## compute numer of dead per plot to remove plot with disturbance ## THERE ARE LOTS OF NAs - DID YOU WANT TO REMOVE THEM OR TREAT THEM AS ALIVE perc.dead <- tapply(data.paracou[["dead"]],INDEX=data.paracou[["plot"]],FUN=function.perc.dead2) data.paracou <- merge(data.paracou,data.frame(plot=names(perc.dead),perc.dead=perc.dead), by = "plot", sort=FALSE) ########################################################### ### PLOT SELECTION FOR THE ANALYSIS ################### ## Nothing to remove #vec.abio.var.names <- c("MAT","MAP") ## MISSING NEED OTHER BASED ON TOPOGRAPHY ASK BRUNO vec.basic.var <- c("obs.id","treeid","sp","plot","D","G","dead","year","htot","x","y","perc.dead") data.tree <- subset(data.paracou,select=c(vec.basic.var)) ############################################## ## COMPUTE MATRIX OF COMPETITION INDEX WITH SUM OF BA PER SPECIES IN EACH PLOT in m^2/ha without the target species ########################### ## NEED TO COMPUTE BASED ON RADIUS AROUND TARGET TREE ## TODO NEED TO ADD BUFFER ZONE data.tree.s <- subset(data.tree,subset=data.tree[["plot"]] ==1) BA.SP.FUN.XY(obs.id,xy.table,diam,sp,Rlim){ data.BA.SP <- BA.SP.FUN(id.tree=as.vector(data.paracou[["treeid"]]), diam=as.vector(data.paracou[["D"]]), sp=as.vector(data.paracou[["sp"]]), id.plot=as.vector(data.paracou[["plot"]]), weights=1/(pi*(0.5*data.paracou$D/100)^2), weight.full.plot=NA) ## change NA and <0 data for 0 data.BA.SP[is.na(data.BA.SP)] <- 0; data.BA.SP[,-1][data.BA.SP[,-1]<0] <- 0 ### CHECK IF sp and sp name for column are the same if(sum(!(names(data.BA.SP)[-1] %in% unique(data.paracou[["sp"]]))) >0) stop("competition index sp name not the same as in data.tree") #### compute BA tot for all competitors
141142143144145146147148149150151152153
BATOT.COMPET <- apply(data.BA.SP[,-1],1,sum,na.rm=TRUE) data.BA.SP$BATOT.COMPET <- BATOT.COMPET; rm(BATOT.COMPET) ### create data frame names(data.BA.SP) <- c("tree.id",names(data.BA.SP)[-1]) data.BA.sp <- merge(data.frame(tree.id=data.paracou[["tree.id"]],ecocode=data.paracou[["ecocode"]]),data.BA.SP,by="tree.id",sort=FALSE) ## test if(sum(!data.BA.sp[["tree.id"]] == data.tree[["tree.id"]]) >0) stop("competition index not in the same order than data.tree") ## save everything as a list list.paracou <- list(data.tree=data.tree,data.BA.SP=data.BA.sp,data.traits=data.traits) save(list.spain,file="./data/process/list.paracou.Rdata")