4  Download precipitation gridded products

4.1 About

The following code downloads, crops, and calculates climatologies, as well as monthly and yearly time series for 11 precipitation products. Be aware that downloading the full datasets can take several hours. Once the data are cropped and processed into climatologies, and monthly and yearly time series, the final dataset is approximately 250 MB. For convenience, a compressed, cropped version is available. Please contact vasquezvicente@ufl.edu for access.

4.2 Libraries

We utilize the terra package for cropping, stacking and calculating rasters (Hijmans 2024). The ncdf4 package is used to read NC files(Pierce 2024).

library(terra)
library(ncdf4)
library(dplyr)
library(tidyr)

4.3 Requirements

We have gather all the permanent links directly to each of the repositories and are available in the csv. We utilize the curl method for downloads but you might need to use another method depending on your machine.

dir.create("../data_reanalysis", showWarnings = FALSE)
extentMap  <- terra::ext(c(-80.2,-79.4, 8.8, 9.5))
options(download.file.method = "curl")
options(timeout = 1080000)

links<-read.csv("../tables/reanalysis_links.csv", header = TRUE, sep = ",") 

4.4 Functions

Some of the datasets follow the same file organization and format, thus created functions for each of the cases.

checkdir<-function(dataset1) {
  if (dir.exists(paste0("../data_reanalysis/",dataset1)) == FALSE) {
        dir.create(paste0("../data_reanalysis/", dataset1), showWarnings = FALSE)
        print(paste0("created directory", "./data_reanalysis/",dataset1))
      }else{print(paste0("directory", "./data_reanalysis/",dataset1, " already exists"))}
}
download_chelsa<-function(dataset){
  for (i in 1:nrow(dataset)) {
  destfile <- paste0("../data_reanalysis/", dataset$source[i], "/", dataset$filename[i])
  if (file.exists(destfile) == FALSE) {
    print(paste0("downloading ", dataset$filename[i]))
    download.file(dataset$link[i], destfile = destfile, method = "curl")
    ras <- terra::rast(destfile)
    ras <- terra::crop(ras, extentMap)
    writeRaster(ras, filename = destfile, overwrite = TRUE)
  }else{
    print(paste0("file ", dataset$filename[i], " already exists"))
    }
  }
}
download_chelsa2<-function(dataset){
  for (i in 1:nrow(dataset)) {
  destfile <- paste0("../data_reanalysis/", dataset$source[i], "/", dataset$filename[i])
  if (file.exists(destfile) == FALSE) {
    print(paste0("downloading ", dataset$filename[i]))
    download.file(dataset$link[i], destfile = destfile, method = "curl")
    ras <- terra::rast(destfile)
    ras <- terra::crop(ras, extentMap)
    ras <- terra::app(ras, fun = function(x) { x * 0.1 })
    writeRaster(ras, filename = destfile, overwrite = TRUE)
  }else{
    print(paste0("file ", dataset$filename[i], " already exists"))
    }
  }
}
download_chp<-function(dataset){
  for (i in 1:nrow(dataset)) {
    destfile <- paste0("../data_reanalysis/", dataset$source[i], "/", dataset$filename[i])
    destfile2<- gsub(".gz","",destfile)
    if (file.exists(destfile2) == FALSE) {
      download.file(dataset$link[i], destfile = destfile, method = "curl")
      R.utils::gunzip(destfile, remove = TRUE)
      destfile=gsub(".gz","",destfile)
      ras <- terra::rast(destfile)
      ras <- terra::crop(ras, extentMap)
      writeRaster(ras, filename = destfile, overwrite = TRUE)
    }else {
      print(paste0(destfile2, " already exists"))
    }
  }
}
download_terra <- function(dataset) {
  for (i in 1:nrow(dataset)) {
    destfile <- paste0("../data_reanalysis/", dataset$source[i], "/", dataset$filename[i])
    if (!file.exists(paste0(destfile, "_01.tif"))) {
      download.file(dataset$link[i], destfile = destfile, method = "curl")
      ras <- terra::rast(destfile)
      ras <- terra::crop(ras, extentMap)
      Sys.sleep(1)
      for (t in 1:nlyr(ras)) {
        writeRaster(ras[[t]], filename = sprintf("%s_%02d.tif", destfile, t), overwrite = TRUE)
      }
      unlink(destfile)
    } else {
      print(paste0(destfile, " already exists"))
    }
  }
}
download_pbcor<-function(dataset){
  for (i in 1:nrow(dataset)) {
    destfile <- paste0("../data_reanalysis/", dataset$source[i], "/", dataset$filename[i])
    if (file.exists(destfile) == FALSE) {
      download.file(dataset$link[i], destfile = destfile, method = "libcurl")
      unzip(destfile, exdir = dirname(destfile))
      data<-list.files(dirname(destfile),pattern = ".nc")
      raschelsa<- terra::rast(paste0(dirname(destfile),"/",data[1]))
      raschelsa<- terra::crop(raschelsa, extentMap)
      for (t in 1:nlyr(raschelsa)) {
        print("writting")
        writeRaster(raschelsa[[t]], filename = paste0("../data_reanalysis/","_",t, ".tif"), overwrite = TRUE)
      }
      raschp<- terra::rast(paste0(dirname(destfile),"/",data[2]))
      raschp<- terra::crop(raschp, extentMap)
      for (t in 1:nlyr(raschp)) {
        print("writting")
        writeRaster(raschp[[t]], filename = paste0(destfile,"_",t, ".tif"), overwrite = TRUE)
      }
      rasworl<- terra::rast(paste0(dirname(destfile),"/",data[3]))
      rasworl<- terra::crop(rasworl, extentMap)
      for (t in 1:nlyr(rasworl)) {
        print("writting")
        writeRaster(rasworl[[t]], filename = paste0(destfile,"_",t, ".tif"), overwrite = TRUE)
      }
      unlink(paste0(dirname(destfile),"/",data[1]))
      unlink(paste0(dirname(destfile),"/",data[2]))
      unlink(paste0(dirname(destfile),"/",data[3]))
      unlink(paste0(dirname(destfile),"/","readme.txt"))
      unlink(destfile)
    }else {
      print(paste0(destfile, " already exists"))
    }
  }
}

4.5 Download regular climatologies

CHELSA 1.2 timeseries and climatologies, as well as the version 2.1 (Dirk N. Karger et al. 2018). CHPclim v1.0 climatology(C. Funk et al. 2015). CHIRPS v2.0 climatology and timeseries(Chris Funk et al. 2015). TerraClimate climatology and timeseries (Abatzoglou et al. 2018)

#CHELSA 1.2 climatology
dataset_chelsa1 <- links %>% filter(source == "CHELSA 1.2" & type == "climatology") 
checkdir(dataset_chelsa1$source[1])
download_chelsa(dataset_chelsa1)

#CHELSA 2.1 climatology
dataset_chelsa2 <- links %>% filter(source == "CHELSA 2.1" & type == "climatology")
checkdir(dataset_chelsa2$source[1])
download_chelsa2(dataset_chelsa2)

#CHIRPS 2.0 timeseries
dataset_chirps <- links %>% filter(source == "CHIRPS 2.0" & type == "timeseries")
checkdir(dataset_chirps$source[1])
download_chp(dataset_chirps)

#CHPclim climatology
dataset_chpclim <- links %>% filter(source == "CHPclim" & type == "climatology")
checkdir(dataset_chpclim$source[1])
download_chp(dataset_chpclim)

#TerraClimate climatology
dataset_terra <- links %>% filter(source == "TerraClimate" & type == "climatology") 
checkdir(dataset_terra$source[1])
download_terra(dataset_terra)

Calculate CHIRPS V2 Climatology

chirps<- list.files("../data_reanalysis/CHIRPS 2.0", pattern=".tif", full.names = TRUE)
month <- numeric(length(chirps))
year <- numeric(length(chirps))
for (i in 1:length(chirps)) {
    file_parts <- unlist(strsplit(basename(chirps[i]), "\\."))
    month[i] <- as.numeric(file_parts[4])
    year[i] <- as.numeric(file_parts[3])
  }
chirps <- data.frame(chirps, month, year)
if (file.exists("../data_reanalysis/CHIRPS 2.0/climatology")==FALSE){
  dir.create("../data_reanalysis/CHIRPS 2.0/climatology", showWarnings = FALSE)
  for (m in unique(chirps$month)){
    print(paste0("Processing ", m))
    ras <- terra::rast(chirps[chirps$month == m, ]$chirps)
    print(paste0("Stacked ", m))
    ras <- terra::app(ras, mean)
    print(paste0("Calculated ", m))
    output_file <- paste0("../data_reanalysis/CHIRPS 2.0/climatology/CHIRPS_2.0_", m, ".tif")
    terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
  }
}else{
  print("climatology already proccessed")
}

Calculate CHIRPS V2 yearly timeseries

if (file.exists("../data_reanalysis/CHIRPS 2.0/yearly")==FALSE){
  dir.create("../data_reanalysis/CHIRPS 2.0/yearly")
  for (y in unique(chirps$year)){
  print(paste0("Processing ", y))
  ras <- terra::rast(chirps[chirps$year == y, ]$chirps)
  print(paste0("Stacked ", y))
  ras <- terra::app(ras, sum)
  print(paste0("Calculated ", y))
  output_file <- paste0("../data_reanalysis/CHIRPS 2.0/yearly/CHIRPS_2.0_", y, ".tif")
  terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
}
}else{
  print("yearly timseries already proccessed")
}

4.6 Download non-regular precipitation products

Worldclim 2.1 comes as a zip file and requires a different download method “libcurl”. It also include some metadata files (Fick and Hijmans 2017).

dataset_worlclim <- links %>% filter(source == "WorldClim 2.1" & type == "climatology")
checkdir(dataset_worlclim$source[1])
destfile<- gsub(".zip","_01.tif",paste0("../data_reanalysis/", dataset_worlclim$source[1], "/", dataset_worlclim$filename[1]))
if (file.exists(destfile)==FALSE){
  download.file(dataset_worlclim$link[1], destfile = paste0("../data_reanalysis/", dataset_worlclim$source[1], "/",  dataset_worlclim$filename[1]), method = "libcurl")
  
  unzip(paste0("data_reanalysis/", dataset_worlclim$source[1], "/", dataset_worlclim$filename[1]), exdir = paste0("data_reanalysis/",      dataset_worlclim$source[1]))
  data<-list.files(paste0("data_reanalysis/", dataset_worlclim$source[1], "/"),pattern = ".tif",full.names = TRUE)
  for (t in 1:length(data)) {
    ras<- terra::rast(data[t])
    ras<- terra::crop(ras, extentMap)
    writeRaster(ras, filename = data[t], overwrite = TRUE)
  }
  unlink(paste0("data_reanalysis/", dataset_worlclim$source[1], "/", dataset_worlclim$filename[1]))
  unlink(paste0("data_reanalysis/", dataset_worlclim$source[1], "/readme.txt"))
}else{
  print("Files already exist")
}

PBCOR correction factors

This correction factors are applied to the already downloaded products CHELSA 1.2, CHP and Worldlclim. We resample the higuest resolution products to match the PBCOR resolution (Beck et al. 2020).

dataset_pbcor <- links %>% filter(source == "PBCOR" & type == "correction factors")
checkdir(dataset_pbcor$source[1])
process_data<- function(source,nc_file, orig_raster,outputfolder) {
  stacked<- terra::rast(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/",nc_file))
  cropped<- terra::crop(stacked, extentMap)
  orig<- list.files(paste0("../data_reanalysis/",orig_raster),pattern = ".tif",full.names = TRUE)
  rasorig<- lapply(orig,raster)
  dir.create(paste0("../data_reanalysis/",source,"/",outputfolder), showWarnings = FALSE)
    for (t in 1:nlyr(cropped)) {
      print("writting")
      chelsa_resampled<- terra::projectRaster(rasorig[[t]],cropped[[t]],method = "ngb")
      correctedfactor<- cropped[[t]]*chelsa_resampled
      writeRaster(correctedfactor, filename = paste0("../data_reanalysis/",source,"/",outputfolder,"/",sub("//.nc$","",nc_file),"_",t, ".tif"), overwrite = TRUE)
    }
}
if (dir.exists("../data_reanalysis/PBCOR/CHELSA 1.2 corrected") == FALSE) {
  download.file(dataset_pbcor$link[1], destfile = paste0("../data_reanalysis/", dataset_pbcor$source[1], "/", dataset_pbcor$filename[1]), method = "libcurl")
  unzip(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/", dataset_pbcor$filename[1]), exdir = paste0("../data_reanalysis/", dataset_pbcor$source[1]))
data<-list.files(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/"),pattern = ".nc")  
process_data("PBCOR",data[1],"CHELSA 1.2","CHELSA 1.2 corrected")
process_data("PBCOR",data[2],"CHPclim","CHPclim corrected")
process_data("PBCOR",data[3],"WorldClim 2.1","WorldClim corrected")
unlink(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/", dataset_pbcor$filename[1]))
unlink(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/", data[1]))
unlink(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/", data[2]))
unlink(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/", data[3]))
unlink(paste0("../data_reanalysis/", dataset_pbcor$source[1], "/readme.txt"))  
}else{
  print("File already exists")
}

4.7 Download CHELSA EarthEnv daily timeseries

The daily timeseries can take hours to download. We take a daily timeseries and calculated it into a monthly timeseries that is used to get a climatology (Dirk Nikolaus Karger et al. 2021).

dataset_earthenv <- links %>% filter(source == "CHELSA EarthEnv" & type == "timeseries")
checkdir(dataset_earthenv$source[1])

for (i in 1:nrow(dataset_earthenv)) {
    if (file.exists(paste0("../data_reanalysis/", dataset_earthenv$source[1],"/",sub("//.zip$", "", dataset_earthenv$filename[i]))) == FALSE) {
    download.file(dataset_earthenv$link[i], destfile = paste0("data_reanalysis/", dataset_earthenv$source[1], "/", dataset_earthenv$filename[i]), method = "curl")
    unzip(paste0("../data_reanalysis/", dataset_earthenv$source[1], "/", dataset_earthenv$filename[i]), exdir = paste0("../data_reanalysis/", dataset_earthenv$source[1]))
    file.rename(paste0("../data_reanalysis/", dataset_earthenv$source[1],"/source"),
                paste0("../data_reanalysis/", dataset_earthenv$source[1],"/",gsub("//.zip$", "", dataset_earthenv$filename[i])))
    unlink(paste0("../data_reanalysis/", dataset_earthenv$source[1], "/", dataset_earthenv$filename[1]))
    data<-list.files(paste0("../data_reanalysis/", dataset_earthenv$source[1],"/",gsub("//.zip$", "", dataset_earthenv$filename[i])),pattern = ".tif",full.names = TRUE)
    for (t in 1:length(data)) {
      ras<- terra::rast(data[t])
      ras<- terra::crop(ras, extentMap)
      ras<-terra::app(ras, fun=function(x){x/100})
      writeRaster(ras, filename = data[t], overwrite = TRUE)
    }
  }else {
     print("already exists")
  }
}

CHELSA EarthEnv from daily to monthly.

envearth<- list.files("../data_reanalysis/CHELSA EarthEnv",full.names = TRUE, pattern='.zip')
if (dir.exists("../data_reanalysis/CHELSA EarthEnv/monthly")==FALSE){
dir.create("../data_reanalysis/CHELSA EarthEnv/monthly", showWarnings = FALSE)
for (subdir in 2:length(envearth)){
  print(paste0("processing ",envearth[subdir]))
  data<-list.files(envearth[subdir],pattern = ".tif",full.names = TRUE)
  ras<- terra::rast(data)
  ras<- terra::app(ras,sum)
  writeRaster(ras, filename = paste0("../data_reanalysis/CHELSA EarthEnv/monthly/",gsub(".zip", "", strsplit(envearth[[subdir]], "/")[[1]][4]),".tif"), overwrite = TRUE)
}}else{
  print("already proccessed into monthly timeseries")
}

CHELSA EarthEnv from monthly to yearly

February of 2016 is not usable, thus we exclude year 2016.

plot(terra::rast("../data_reanalysis/CHELSA EarthEnv/monthly/CHELSA_preccor_land_02_2016.tif"))

files <- list.files("../data_reanalysis/CHELSA EarthEnv/monthly", pattern = ".tif", full.names = TRUE)
data <- data.frame(files)
years <- vector("character", length(files))
month <- vector("character", length(files))
for (i in seq_along(files)) {
  years[i] <- gsub(".tif", "", strsplit(files[i], "_")[[1]][6])
  month[i] <-strsplit(files[i], "_")[[1]][5]
}
data$year <- years
data$month <- month
years<- 2003:2015
if (dir.exists("../data_reanalysis/CHELSA EarthEnv/yearly")==FALSE){
  dir.create("../data_reanalysis/CHELSA EarthEnv/yearly", showWarnings = FALSE)
  for (year in years){
    print(paste0("processing ",year))
    ras<- terra::rast(data[data$year==year,]$files)
    ras<- terra::app(ras,sum)
    writeRaster(ras, filename = paste0("/..data_reanalysis/CHELSA EarthEnv/yearly/","CHELSA_EarthEnv_",year,".tif"), overwrite = TRUE)
  }
}else{
  print("Yearly data already processed")
}

[1] "Yearly data already processed"

Calculate climatology for CHELSA EarthEnv fromn 2003 to 2015

if (dir.exists("../data_reanalysis/CHELSA EarthEnv/climatology")==FALSE){
  dir.create("../data_reanalysis/CHELSA EarthEnv/climatology", showWarnings = FALSE)
  for (i in 1:length(unique(data$month))){
    print(paste0("processing ",unique(data$month)[i]))
    ras<- terra::rast(data[data$month==unique(data$month)[i],]$files)
    ras <- terra::app(ras, function(x) { mean(x, na.rm = TRUE) })
    writeRaster(ras, filename = paste0("../data_reanalysis/CHELSA EarthEnv/climatology/","CHELSA_EarthEnv_",unique(data$month)[i],".tif"), overwrite = TRUE)
  }
}else{
  print("Climatologies are already calculated")
}

[1] "Climatologies are already calculated"

4.8 Download CHELSA W5E5 daily timeseries

This is a daily timeseries and can take hour to complete the download (Dirk N. Karger et al. 2022).

dataset_chelsa_e5e5 <- links %>% filter(source == "CHELSA-W5E5v1.0" & type == "timeseries")
checkdir(dataset_chelsa_e5e5$source[1])
for (i in 1:nrow(dataset_chelsa_e5e5)) {
  if (file.exists(paste0("../data_reanalysis/", dataset_chelsa_e5e5$source[1],"/",sub("//.nc$", "", dataset_chelsa_e5e5$filename[i]),"_",1,".tif")) == FALSE) {
    download.file(dataset_chelsa_e5e5$link[i], destfile = paste0("../data_reanalysis/", dataset_chelsa_e5e5$source[1], "/", dataset_chelsa_e5e5$filename[i]), method = "curl")
    ras<-terra::rast(paste0("../data_reanalysis/", dataset_chelsa_e5e5$source[1], "/", dataset_chelsa_e5e5$filename[i]))
    ras<- terra::crop(ras, extentMap)
    ras<-terra::app(ras, fun=function(x){x*86400})
    for (t in 1:nlyr(ras)) {
      writeRaster(ras[[t]], filename = paste0("../data_reanalysis/", dataset_chelsa_e5e5$source[1],"/",sub("//.nc$", "", dataset_chelsa_e5e5$filename[i]),"_",t,".tif"), overwrite = TRUE)
    }
    unlink(paste0("../data_reanalysis/", dataset_chelsa_e5e5$source[1], "/", dataset_chelsa_e5e5$filename[i]))
  }else {
    print("already exists")
  }
}

CHELSA W5E5 to monthly timeseries

chelsae5 <- list.files("../data_reanalysis/CHELSA-W5E5v1.0", pattern=".tif", full.names = TRUE)
day <- numeric(length(chelsae5))
month <- numeric(length(chelsae5))
year <- numeric(length(chelsae5))
for (i in 1:length(chelsae5)) {
  file_parts <- unlist(strsplit(basename(chelsae5[i]), "_"))
  day[i] <- as.numeric(gsub(".tif", "", file_parts[8]))
  month[i] <- as.numeric(substring(file_parts[7], 5, 6))
  year[i] <- as.numeric(substring(file_parts[7], 1, 4))
}
chelsae5 <- data.frame(chelsae5, day, month, year)

if (file.exists("../data_reanalysis/CHELSA-W5E5v1.0/monthly")==FALSE){
dir.create("../data_reanalysis/CHELSA-W5E5v1.0/monthly", showWarnings = FALSE)
for (y in unique(chelsae5$year)) {
  for (m in unique(chelsae5$month)) {
    print(paste0("Processing ", y, "_", m))
    chelsae5[chelsae5$year == y & chelsae5$month == m, ]$chelsae5
    ras <- terra::rast(chelsae5[chelsae5$year == y & chelsae5$month == m, ]$chelsae5)
    print(paste0("Stacked ", y, "_", m) )
    ras <- terra::app(ras, sum)
    print(paste0("Calculated ", y, "_", m))
    output_file <- paste0("../data_reanalysis/CHELSA-W5E5v1.0/monthly/CHELSA-W5E5v1.0_", y, "_", m, ".tif")
    
    terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
  }
}
}else{
  print("Monthly timeseries already processed")
}

CHELSA W5E5 from monthly to yearly timeseries

if (file.exists("../data_reanalysis/CHELSA-W5E5v1.0/yearly")==FALSE){
  dir.create("../data_reanalysis/CHELSA-W5E5v1.0/yearly", showWarnings = FALSE)
  for (y in unique(chelsae5$year)){
    print(paste0("Processing ", y))
    ras <- terra::rast(chelsae5[chelsae5$year == y, ]$chelsae5)
    print(paste0("Stacked ", y))
    ras <- terra::app(ras, sum)
    print(paste0("Calculated ", y))
    output_file <- paste0("../data_reanalysis/CHELSA-W5E5v1.0/yearly/CHELSA-W5E5v1.0_", y, ".tif")
    terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
  }
}else{
  print("yearly timeseries have been processed")
}

CHELSA W5E5 climatology calculated

monthly_files<- list.files("../data_reanalysis/CHELSA-W5E5v1.0/monthly", pattern=".tif", full.names = TRUE)
month <- numeric(length(monthly_files))
year <- numeric(length(monthly_files))
for (i in 1:length(monthly_files)) {
    file_parts <- unlist(strsplit(basename(monthly_files[i]), "_"))
    month[i] <- as.numeric(substring(file_parts[3], 1, 2))
    year[i] <- as.numeric(file_parts[2])
  }
monthly_files <- data.frame(monthly_files, month, year)
if (file.exists("../data_reanalysis/CHELSA-W5E5v1.0/climatology")==FALSE){
  dir.create("../data_reanalysis/CHELSA-W5E5v1.0/climatology", showWarnings = FALSE)
  for (m in unique(monthly_files$month)){
    print(paste0("Processing ", m))
    ras <- terra::rast(monthly_files[monthly_files$month == m, ]$monthly_files)
    print(paste0("Stacked ", m))
    ras <- terra::app(ras, mean)
    print(paste0("Calculated ", m))
    output_file <- paste0("../data_reanalysis/CHELSA-W5E5v1.0/climatology/CHELSA-W5E5v1.0_", m, ".tif")
    terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
  }
}else{
  print("climatology already calculated")
}

4.9 Download CHELSA 1.2 monthly timeseries

dataset_chelsa12 <- links %>% filter(source == "CHELSA 1.2" & type == "timeseries")
checkdir(dataset_chelsa12$source[1])
dir.create(paste0("../data_reanalysis/", dataset_chelsa12$source[1],"/monthly"), showWarnings = FALSE)
for (i in 1:nrow(dataset_chelsa12)){
  if (file.exists(paste0("../data_reanalysis/", dataset_chelsa12$source[1],"/monthly/",dataset_chelsa12$filename[i])) == FALSE) {
    download.file(dataset_chelsa12$link[i], destfile = paste0("../data_reanalysis/", dataset_chelsa12$source[1], "/monthly/", dataset_chelsa12$filename[i]), method = "curl")
    ras<-terra::rast(paste0("../data_reanalysis/", dataset_chelsa12$source[1], "/monthly/", dataset_chelsa12$filename[i]))
    ras<- terra::crop(ras, extentMap)
    writeRaster(ras, filename = paste0("../data_reanalysis/", dataset_chelsa12$source[1], "/monthly/", dataset_chelsa12$filename[i]), overwrite = TRUE)
    }
  else {
    print("already exists")
  }
}

Calculate CHELSA 1.2 yearly timeseries

files<- list.files("../data_reanalysis/CHELSA 1.2/monthly",full.names = TRUE)
month<- numeric(length(files))
year<- numeric(length(files))
for (i in 1:length(files)) {
  file_parts <- unlist(strsplit(basename(files[i]), "_"))
  month[i] <- as.numeric(file_parts[4])
  year[i] <- as.numeric(file_parts[3])
}
chelsa12<-data.frame(files,month,year)
if (file.exists("../data_reanalysis/CHELSA 1.2/yearly")==FALSE){
dir.create("../data_reanalysis/CHELSA 1.2/yearly", showWarnings = FALSE)
for (y in unique(chelsa12$year)){
  print(paste0("Processing ", y))
  ras <- terra::rast(chelsa12[chelsa12$year == y, ]$files)
  print(paste0("Stacked ", y))
  ras <- terra::app(ras, sum)
  print(paste0("Calculated ", y))
  output_file <- paste0("../data_reanalysis/CHELSA 1.2/yearly/CHELSA_1.2_", y, ".tif")
  terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
}
}else {
   print("yearly timeseries already processed")
}

4.10 Download CHELSA 2.1 monthly timeseries

dataset_chelsa21 <- links %>% filter(source == "CHELSA 2.1" & type == "timeseries")
checkdir(dataset_chelsa21$source[1])
dir.create(paste0("../data_reanalysis/", dataset_chelsa21$source[1],"/monthly"), showWarnings = FALSE)
for (i in 1:nrow(dataset_chelsa21)){
  if (file.exists(paste0("../data_reanalysis/", dataset_chelsa21$source[1],"/monthly/",dataset_chelsa21$filename[i])) == FALSE) {
    download.file(dataset_chelsa21$link[i], destfile = paste0("../data_reanalysis/", dataset_chelsa21$source[1], "/monthly/", dataset_chelsa21$filename[i]), method = "curl")
    ras<-terra::rast(paste0("../data_reanalysis/", dataset_chelsa21$source[1], "/monthly/", dataset_chelsa21$filename[i]))
    ras<- terra::crop(ras, extentMap)
    writeRaster(ras, filename = paste0("../data_reanalysis/", dataset_chelsa21$source[1], "/monthly/", dataset_chelsa21$filename[i]), overwrite = TRUE)
    }
  else {
    print("already exists")
  }
}

Calculate CHELSA 2.1 yearly timeseries

files<- list.files("../data_reanalysis/CHELSA 2.1/monthly",full.names = TRUE)
month<- numeric(length(files))
year<- numeric(length(files))
for (i in 1:length(files)) {
  file_parts <- unlist(strsplit(basename(files[i]), "_"))
  month[i] <- as.numeric(file_parts[3])
  year[i] <- as.numeric(file_parts[4])
}
chelsa21<-data.frame(files,month,year)
if (file.exists("../data_reanalysis/CHELSA 2.1/yearly")==FALSE){
dir.create("../data_reanalysis/CHELSA 2.1/yearly", showWarnings = FALSE)
for (y in unique(chelsa21$year)){
  print(paste0("Processing ", y))
  ras <- terra::rast(chelsa21[chelsa21$year == y, ]$files)
  print(paste0("Stacked ", y))
  ras <- terra::app(ras, sum)
  print(paste0("Calculated ", y))
  output_file <- paste0("../data_reanalysis/CHELSA 2.1/yearly/CHELSA_2.1_", y, ".tif")
  terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
}
}else {
   print("yearly timeseries already processed")
}

4.11 Download TerraClimate monthly timeseries

(Abatzoglou et al. 2018).

dataset_terra <- links %>% filter(source == "TerraClimate" & type == "timeseries")
checkdir(dataset_terra$source[1])
dir.create(paste0("../data_reanalysis/", dataset_terra$source[1],"/monthly"), showWarnings = FALSE)
for (i in 1:nrow(dataset_terra)){
  tif_files <- paste0("../data_reanalysis/", dataset_terra$source[1],"/monthly/",dataset_terra$filename[i],"_", sprintf("%02d", 1:12), ".tif")
  if (!all(file.exists(tif_files))) {
    download.file(dataset_terra$link[i], destfile = paste0("../data_reanalysis/", dataset_terra$source[1], "/monthly/", dataset_terra$filename[i]), method = "curl")
    ras<-terra::rast(paste0("../data_reanalysis/", dataset_terra$source[1], "/monthly/", dataset_terra$filename[i]))
    ras<- terra::crop(ras, extentMap)
    for (t in 1:nlyr(ras)) {
      t_formatted <- sprintf("%02d", t)
      writeRaster(ras[[t]], filename = paste0("../data_reanalysis/", dataset_terra$source[1],"/monthly/",dataset_terra$filename[i],"_",t_formatted,".tif"), overwrite = TRUE)
    }
    unlink(paste0("../data_reanalysis/", dataset_terra$source[1], "/monthly/", dataset_terra$filename[i]))
  } else {
    print("already exists")
  }
}

Calculate yearly timeseries of TerraClimate.

files<- list.files("../data_reanalysis/TerraClimate/monthly",pattern = ".tif",full.names = TRUE)    
year<- numeric(length(files))
month<- numeric(length(files))
for (i in 1:length(files)) {
  file_parts <- unlist(strsplit(basename(files[i]), "_"))
  year[i] <- as.numeric(substring(file_parts[3],1,4))
  month[i] <- as.numeric(gsub(".tif", "", file_parts[4]))   
}
if (file.exists("../data_reanalysis/TerraClimate/yearly")==FALSE){
  dir.create("../data_reanalysis/TerraClimate/yearly", showWarnings = FALSE)
  for (y in unique(year)){
    print(paste0("Processing ", y))
    ras <- terra::rast(files[year == y])
    print(paste0("Stacked ", y))
    ras <- terra::app(ras, sum)
    print(paste0("Calculated ", y))
    output_file <- paste0("../data_reanalysis/TerraClimate/yearly/TerraClimate_", y, ".tif")
    terra::writeRaster(ras, filename = output_file, overwrite = TRUE)
  }
}else{
  print("Yearly timeseries already processed")
}

Abatzoglou, John T., Solomon Z. Dobrowski, Sean A. Parks, and Katherine C. Hegewisch. 2018. “TerraClimate, a High-Resolution Global Dataset of Monthly Climate and Climatic Water Balance from 1958–2015.” Scientific Data 5 (1): 170191. https://doi.org/10.1038/sdata.2017.191.

Beck, Hylke E., Eric F. Wood, Tim R. McVicar, Mauricio Zambrano-Bigiarini, Camila Alvarez-Garreton, Oscar M. Baez-Villanueva, Justin Sheffield, and Dirk N. Karger. 2020. “Bias Correction of Global High-Resolution Precipitation Climatologies Using Streamflow Observations from 9372 Catchments.” Journal of Climate 33 (4): 1299–1315. https://doi.org/10.1175/JCLI-D-19-0332.1.

Fick, Stephen E., and Robert J. Hijmans. 2017. “WorldClim 2: New 1‐km Spatial Resolution Climate Surfaces for Global Land Areas.” International Journal of Climatology 37 (12): 4302–15. https://doi.org/10.1002/joc.5086.

Funk, Chris, Pete Peterson, Martin Landsfeld, Diego Pedreros, James Verdin, Shraddhanand Shukla, Gregory Husak, et al. 2015. “The Climate Hazards Infrared Precipitation with Stations—a New Environmental Record for Monitoring Extremes.” Scientific Data 2 (1): 150066. https://doi.org/10.1038/sdata.2015.66.

Funk, C., A. Verdin, J. Michaelsen, P. Peterson, D. Pedreros, and G. Husak. 2015. “A Global Satellite Assisted Precipitation Climatology.” Preprint. Data, Algorithms,; Models. https://doi.org/10.5194/essdd-8-401-2015.

Hijmans, Robert J. 2024. Terra: Spatial Data Analysis. https://github.com/rspatial/terra.

Karger, Dirk N., Olaf Conrad, Jürgen Böhner, Tobias Kawohl, Holger Kreft, Rodrigo W. Soria-Auza, Niklaus E. Zimmermann, H. Peter Linder, and Michael Kessler. 2018. “Data from: Climatologies at High Resolution for the Earth’s Land Surface Areas.” Dryad. https://doi.org/10.5061/DRYAD.KD1D4.

Karger, Dirk Nikolaus, Adam M. Wilson, Colin Mahony, Niklaus E. Zimmermann, and Walter Jetz. 2021. “Global Daily 1 Km Land Surface Precipitation Based on Cloud Cover-Informed Downscaling.” Scientific Data 8 (1): 307. https://doi.org/10.1038/s41597-021-01084-6.

Karger, Dirk N., Stefan Lange, Chantal Hari, Christopher P. O. Reyer, and Niklaus E. Zimmermann. 2022. “CHELSA-W5E5 V1.0: W5E5 V1.0 Downscaled with CHELSA V2.0.” ISIMIP Repository. https://doi.org/10.48364/ISIMIP.836809.1.

Pierce, David. 2024. Ncdf4: Interface to Unidata netCDF (Version 4 or Earlier) Format Data Files. https://CRAN.R-project.org/package=ncdf4.