merge.data.FUSHAN.R

### MERGE Fushan DATA
rm(list = ls())
source("./R/format.function.R")
library(reshape)

######################### READ DATA read individuals tree data
load("./data/raw/DataFushan/fushan.rdata")
data.fushan <- data.frame(fushan)
rm(fushan)

### read species names
species.clean <- read.csv("./data/raw/DataFushan/Splist_Fushan_En.csv", stringsAsFactors = FALSE)

###################################### MASSAGE TRAIT DATA Obtain maximum height per species from data.trait no sd
###################################### available as we have only one observation for species
data.trait <- read.table("./data/raw/DataFushan/fs_trait_Kunstler.txt", header = T, 
    sep = "\t")
colnames(data.trait) <- c("sp", "sla", "wd", "seedmassmg", "meanN", "maxheightm")
data.trait <- merge(data.trait, data.frame(sp = species.clean$sp, Latin = apply(species.clean[,c("genus","epithet")],1,paste,collapse="_"), 
	stringsAsFactors = F), by = "sp")
data.trait$Leaf.N.mean <- data.trait$meanN*10; data.trait$meanN <- NULL; ## conversion from percentage to mg/g
data.trait$Leaf.N.sd <- NA
data.trait$Seed.mass.mean <- data.trait$seedmassmg; data.trait$seedmassmg <- NULL
data.trait$Seed.mass.sd <- NA
data.trait$SLA.mean <- data.trait$sla/10; data.trait$sla <- NULL
data.trait$SLA.sd <- NA
data.trait$Wood.density.mean <- data.trait$wd; data.trait$wd <- NULL
data.trait$Wood.density.sd <- NA
data.trait$Max.height.mean <- log10(data.trait$maxheightm); data.trait$maxheightm <- NULL
data.trait$Max.height.sd <- NA

data.max.height <- data.frame(sp = data.trait$sp, Max.height = log10(data.trait$maxheightm))
data.fushan <- merge(data.fushan, data.max.height, by = "sp")

########################################## FORMAT INDIVIDUAL TREE DATA
data.fushan$year <- data.fushan$Interval  ## number of year between measurement - MISSING - FIVE YEARS
data.fushan$G <- 10*(data.fushan$dbh2 - data.fushan$dbh1)/data.fushan$Interval  ## diameter growth in mm per year
data.fushan$D <- data.fushan[["dbh1"]] ## diameter in cm
data.fushan$dead <- as.numeric(data.fushan$status2 == "dead")  ## dummy variable for dead tree 0 alive 1 dead
data.fushan$plot <- rep(NA,nrow(data.fushan))  ## plot code - MISSING - ONE BIG PLOT
data.fushan$htot <- rep(NA,nrow(data.fushan)) ## height of tree in m - MISSING
data.fushan$tree.id <- data.fushan$tag  ## tree unique id - use tag here as it is unique for each row
data.fushan$sp.name <- rep(NA,nrow(data.fushan)) 
v <- species.clean$sp
for (i in 1:length(unique(data.fushan$sp))) {
    sel.spp <- which(data.fushan$sp == unique(data.fushan$sp)[i])
    data.fushan$sp.name[sel.spp] <- paste(species.clean$family[sel.spp], species.clean$genus[sel.spp], 
        species.clean$epithet[sel.spp], sep = "_")
}
data.fushan$weights <- 1/(pi*(0.5*data.fushan$D/100)^2)
data.fushan$obs.id <- data.fushan$tag 

########################################## CHANGE COORDINATE SYSTEM DON'T KNOW!

###################### 1 ECOREGION ONLY? 

###################### PERCENT DEAD CANNOT DO THIS WITHOUT A PLOTID compute numer of dead per plot to
###################### remove plot with disturbance
perc.dead <- tapply(data.fushan[["dead"]], INDEX = data.fushan[["plot"]], FUN = function.perc.dead2)
data.fushan <- merge(data.fushan, data.frame(plot = names(perc.dead), perc.dead = perc.dead), 
    by = "plot", sort = FALSE)

########################################################### PLOT SELECTION FOR THE ANALYSIS Remove data with dead == 1
table(data.fushan$dead)
data.fushan <- data.fushan[data.fushan$dead == 0, ]

vec.abio.var.names <- c("MAT", "MAP")
vec.basic.var <- c("tree.id", "sp", "sp.name", "plot", "D", "G", "dead", "year", 
    "htot", "gx", "gy", "perc.dead")
data.tree <- subset(data.fushan, select = c(vec.basic.var, vec.abio.var.names))