merge.data.BCI.R

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

############# READ DATA read individuals tree data Requires careful formatting of 7 census
############ datasets The raw data is such that, once a tree dies in census X, then it no
########## longer exists in census X+1, X+2 etc...
data.bci1 <- read.table("./data/raw/DataBCI/census1/PlotsDataReport.txt", header = TRUE, 
    stringsAsFactors = FALSE, sep = "\t")
data.bci1$Date1 <- data.bci1$Date; data.bci1$Date <- NULL
data.bci1$DBH1 <- data.bci1$DBH; data.bci1$DBH <- NULL
big.bci <- NULL
for (k in 2:7) {
    new.directory <- paste("./data/raw/DataBCI/census", k, "/PlotsDataReport.txt", 
        sep = "")
    data.bci2 <- read.table(new.directory, header = TRUE, stringsAsFactors = FALSE, 
        sep = "\t"); 
    if (!is.null(big.bci)) {
        sub.bci <- merge(data.bci1[, c("Latin","Quadrat","Census","gx","gy","TreeID","Tag","Date1","DBH1")], data.frame(TreeID = data.bci2[["TreeID"]], 
            DBH2 = data.bci2[["DBH"]], Date2 = data.bci2[["Date"]], dead = as.numeric(data.bci2[["Status"]] == 
                "dead"),stringsAsFactors=F), sort = T, by = "TreeID") ## Uses the Date1 as the census number
        big.bci <- rbind(big.bci, sub.bci)
    }
    if (is.null(big.bci)) {
        big.bci <- merge(data.bci1[, c("Latin","Quadrat","Census","gx","gy","TreeID","Tag","Date1","DBH1")], data.frame(TreeID = data.bci2[["TreeID"]], 
            DBH2 = data.bci2[["DBH"]], Date2 = data.bci2[["Date"]], dead = as.numeric(data.bci2[["Status"]] == 
                "dead"),stringsAsFactors=F), sort = T, by = "TreeID")
    }
	data.bci1 <- data.bci2
	data.bci1$Date1 <- data.bci1$Date; data.bci1$Date <- NULL
	data.bci1$DBH1 <- data.bci1$DBH; data.bci1$DBH <- NULL
    cat("Census", k, "now included\n")
	print(summary(big.bci$DBH1)); print(summary(big.bci$DBH2))
}
rm(data.bci1, data.bci2, sub.bci)
big.bci <- big.bci[order(big.bci$TreeID), ]
data.bci <- big.bci
rm(big.bci)

### read species names
species.clean <- read.table("./data/raw/DataBCI/TaxonomyDataReport.txt", stringsAsFactors = FALSE, 
    header = T, sep = "\t")
species.clean$Latin_name <- paste(species.clean[["Genus"]], 
    species.clean[["species"]],sep=" ")

## ## Try to relate SpeciesID in species.clean species names in data.bci
## unique(data.bci$Latin) %in% species.clean$Latin_name

###################################### MASSAGE TRAIT DATA Use HEIGHT_AVG, LMALAM_AVD, SEED_DRY
data.trait <- read.csv("./data/raw/DataBCI/BCITRAITS_20101220.csv", stringsAsFactors = FALSE, 
    header = T)
data.trait$Latin <- apply(data.trait[, 1:2], 1, paste, collapse = " ")
data.trait <- data.trait[,c("GENUS.","SP.","Latin","SEED_DRY","LMALAM_AVD","LMALAM_SED","LMALAM_ND","SG100C_AVG","SG100C_SEM","SG100C_N","HEIGHT_AVG","HEIGHT_SEM","HEIGHT_N")]
data.trait$sp <- data.trait[["SP."]]
data.trait[["SP."]] <- NULL
data.trait$Leaf.N.mean <- NA
data.trait$Leaf.N.sd <- NA
data.trait$Seed.mass.mean <- data.trait$SEED_DRY*1000
data.trait$SEED_DRY <- NULL
data.trait$Seed.mass.sd <- NA 
data.trait$SLA.mean <- 1/data.trait$LMALAM_AVD
data.trait$SLA.mean <- data.trait$SLA.mean*1000 ## Conversion from g m^-2 to mm2 mg^-1
data.trait$SLA.sd <- 1/data.trait$LMALAM_SED
data.trait$SLA.sd <- data.trait$SLA.sd*1000 ## Conversion from g m^-2 to mm2 mg^-1
data.trait$SLA.sd <- data.trait$SLA.sd*sqrt(data.trait$LMALAM_ND) ## conversion of SEM in SD
data.trait$LMALAM_AVD <- data.trait$LMALAM_SED <- data.trait$LMALAM_ND <- NULL
data.trait$Wood.density.mean <- data.trait$SG100C_AVG; 
data.trait$Wood.density.sd <- data.trait$SG100C_SEM*sqrt(data.trait$SG100C_N) ## conversion of SEM in SD