SNAP Data QA/QC

Introduction

The samples_setup_CRU.R script prepares extracted CRU 3.x regional temperature and precipitation data for direct use.

Motivation

This script transforms and organizes output data from a parent script into a format that is easier to work with in certain frameworks.

Details

The data are extracted based on regional shapefiles and the full Alaska-Canada 2-km extent for SNAP’s CRU 3.x data set. Data manipulation is performed to organize the inputs into more usefully formatted outputs.

Capabilities

This script uses parallel processing via the base parallel package and the function mclapply. It uses as many CPU cores as are available since it is parallelized across regions, which exceed the number of cores. It becomes serial beyond the number of cores simultaneously in use.

Limitations

This code expects CRU 3.x data have been processed by the parent script. The code does not make use of Rmpi or similar options so it cannot take advantage of multi-node cluster parallel processing.

R Code

Setup

Setup is minimal. Set working directory. List CRU 3.x files.

setwd("/workspace/UA/mfleonawicz/projects/SNAPQAQC/data/regional/samples")

library(parallel)
library(data.table)
library(reshape2)

cru <- "32"
files <- list.files(pattern = paste0("^CRU", cru, ".*.regions_samples.RData$"))

models <- paste0("CRU", cru)

Load and compile data

dlist <- vector("list", length(files))
for (i in 1:length(files)) {
    load(files[i])
    m <- do.call(rbind, samples.out)
    n <- nrow(m)
    dlist[[i]] <- data.frame(Var = rep(c("Temperature", "Precipitation"), each = n/(2 * 
        length(samples.out))), m, stringsAsFactors = F)
    print(i)
}
d <- as.data.frame(rbindlist(dlist))
rm(samples.out, n, m, i, files, models, dlist)
names(d)[3:ncol(d)] <- gsub("X", "", names(d)[3:ncol(d)])

Organize data frame and support objects

f <- function(i, n, index, cru, multiplier = c(1, 1), decimals = NULL) {
    name.tmp <- samples.names[i]
    rsd <- subset(d, Location == name.tmp)
    rsd <- melt(rsd, id.vars = names(rsd)[1:2], measure.vars = names(rsd)[-c(1:2)], 
        variable.name = "Time", value.name = "Vals_Probs")
    rsd$vals.ind <- rep(rep(c(T, F), each = n), length = nrow(rsd))
    rsd$dcastIsDumb <- rep(1:n, length = nrow(rsd))
    rsd <- dcast(rsd, Var + Location + Time + dcastIsDumb ~ vals.ind, value.var = "Vals_Probs")
    names(rsd)[6:5] <- c("Val", "Prob")
    rsd$Year <- substr(as.character(rsd$Time), 1, 4)
    rsd$Month <- substr(as.character(rsd$Time), 6, 8)
    rsd$Month <- factor(rsd$Month, levels = month.abb)
    rsd$Decade <- paste0(substr(rsd$Year, 1, 3), 0)
    rownames(rsd) <- NULL
    rsd <- rsd[c(1:2, 6, 5, 7:9)]
    grp <- rep(names(region.names.out), times = sapply(region.names.out, length))[i]
    dir.create(outDir <- paste0("../../final/region_files_CRU", cru, "/samples/", 
        grp, "/", name.tmp), recursive = T, showWarnings = F)
    if (!is.null(decimals)) 
        rsd$Val <- round(rsd$Val, decimals[1]) * multiplier[1]
    if (!is.null(decimals)) 
        rsd$Prob <- round(rsd$Prob, decimals[2]) * multiplier[2]
    if (i == 1) 
        x <- subset(rsd, Var == "Precipitation") else x <- NULL
    gc()
    rsd.cru <- unlist(subset(rsd, Var == "Precipitation")[, index])
    names(rsd.cru) <- NULL
    save(rsd.cru, file = paste0(outDir, "/", "precipitation.RData"))
    gc()
    rsd.cru <- unlist(subset(rsd, Var == "Temperature")[, index])
    names(rsd.cru) <- NULL
    save(rsd.cru, file = paste0(outDir, "/", "temperature.RData"))
    gc()
    print(i)
    x
}

Save output files

if (file.exists("../../final/meta.RData")) load("../../final/meta.RData")
samples.columns.cru <- 3:4
samples.multipliers.cru <- c(10, 1e+16)  # see function arguments 'multiplier' and 'decimals'

out <- mclapply(1:length(samples.names), f, n = n.samples, index = samples.columns.cru, 
    cru = cru, multiplier = samples.multipliers.cru, decimals = c(1, 16), mc.cores = 32)
cru.samples.df <- out[[1]]
cru.samples.df[, samples.columns.cru] <- NA
cru.samples.df$Var <- NA
cru.samples.df$Location <- NA
if (as.numeric(cru) == 31) cru31.samples.df <- cru.samples.df else if (as.numeric(cru) == 
    32) cru32.samples.df <- cru.samples.df

rm(cru, cru.samples.df, d, f, out)
save.image("../../final/meta.RData")