Kriging: (multi-core) Kriging

In ErikKusch/KrigR: Downloading, Aggregating, and Kriging of ECMWF CDS-Data

(multi-core) Kriging

Description

This function statistically downscales input data using covariate data and the kriging methodology. Use optional arguments such as Dir, Keep_Temporary, KrigingEquation, nmax and Cores for ease of use, substitution of non-default covariates, localisation of kriging, and parallel processing.

Usage

Kriging(
  Data,
  Covariates_training,
  Covariates_target,
  Equation = NULL,
  Cores = parallel::detectCores(),
  nmax = Inf,
  Dir = getwd(),
  FileName,
  FileExtension = ".nc",
  Compression = 9,
  Keep_Temporary = FALSE,
  verbose = TRUE
)

Arguments

Data	SpatRaster which is to be downscaled.
Covariates_training	SpatRaster containing covariates at training resolution.
Covariates_target	SpatRaster containing covariates at target resolution.
Equation	Formula or character string specifying which covariates to use and how. Layer names in Covariates_training and Covariates_target need to match parameters in this formula. Do not include ". ~", just supply the righthand side of this formula like so: "Covariate1+Covariate2" or "Covariate1*Covariate2", etc.
Cores	Numeric. How many cores to use. Parallel processing is carried out when Cores is bigger than 1. Default is detecting all cores of your machine.
nmax	NUmeric. Controls local kriging. Number of nearest observations to be used kriging of each observation. Default is to use all available (Inf). You can specify as a number (numeric).
Dir	Character/Directory Pointer. Directory specifying where to place final kriged product. Default is current working directory.
FileName	Character. A file name for the produced files.
FileExtension	Character. A file extension for the produced file. Supported values are ".nc" (default) and ".tif" (better support for metadata).
Compression	Integer between 1 to 9. Applied to final .nc file that the function writes to hard drive. Same as compression argument in terra::writeCDF(). Ignored if FileExtension = ".tif".
Keep_Temporary	Logical, whether to delete individual kriging products of layers in Data after processing. Default is TRUE. These temporary files are stored in a newly created directory in Dir which is pre-pended with "TEMP-" and is deleted if Keep_Temporary = FALSE upon completion.
verbose	Optional, logical. Whether to report progress of data download (if queried) in the console or not.

Value

A list object containing SpatRasters reporting (1) the downscaled data as well as (2) the standard deviation for downscaling. Also produces two files of specified extension in the specified directory which are the two data contents of the aforementioned list. A temporary directory is populated with individual files during the execution of this function which is deleted upon completion if Keep_Temporary = FALSE and all layers in the Data raster object were kriged successfully.

The produced SpatRasters contains metadata/attributes as a named vector that can be retrieved with terra::metags(...):

• Citation - A string which to use for in-line citation of the data product obtained with Kriging.

• KrigRCall.X - Arguments passed to the Kriging function that produced the file.

ATTENTION: If data is loaded again from disk at a later point with a different function, take note that citation and KrigR-call metadata will not be loaded properly from a .nc when loading data through a different function. Kriging() handles these .nc specific issues when loading .nc files created previously with Kriging() from disk.

See Also

CovariateSetup, Plot.Kriged.

Examples

## Not run: 
## Kriging using pre-fab data with a rectangular extent and a fives layers of data with parallel processing
### Loading data
CDS_rast <- terra::rast(system.file("extdata", "CentralNorway.nc", package = "KrigR"))
Cov_train <- terra::rast(system.file("extdata", "Covariates_Train.nc", package = "KrigR"))
Cov_target <- terra::rast(system.file("extdata", "Covariates_Target.nc", package = "KrigR"))
names(Cov_train) <- names(Cov_target) <- "GMTED2010"

### kriging itself
ExtentKrig <- Kriging(
  Data = CDS_rast,
  Covariates_training = Cov_train,
  Covariates_target = Cov_target,
  Equation = "GMTED2010",
  Cores = 2,
  FileName = "KrigTest1",
  FileExtension = ".nc",
  Keep_Temporary = TRUE,
  nmax = 40,
  verbose = TRUE
)
Plot.Kriged(Krigs = ExtentKrig)
## Kriging using full KrigR pipeline with shapefile data
### Shapefile loading
data("Jotunheimen_poly")
### CDS data download
Qsoil_rast <- CDownloadS(
  Variable = "Volumetric soil water layer 1", # can also specify as "volumetric_soil_water_layer_1"
  # time-window, default set to range of dataset-type
  DateStart = "1995-01-01 00:00",
  DateStop = "1995-01-03 23:00",
  TZone = "CET",
  # temporal aggregation
  TResolution = "day",
  # spatial
  Extent = Jotunheimen_poly,
  # file storing
  FileName = "KrigTest2_Raw",
  # API credentials
  API_User = API_User,
  API_Key = API_Key
)

### Covariate preparations
Covariates_ls <- CovariateSetup(
  Training = Qsoil_rast,
  Target = 0.03,
  Covariates = c("tksat", "tkdry", "csol", "k_s", "lambda", "psi", "theta_s"),
  Source = "Drive",
  Extent = Jotunheimen_poly,
  Keep_Global = TRUE
)

### kriging itself
ShapeKrig <- Kriging(
  Data = Qsoil_rast,
  Covariates_training = Covariates_ls[[1]],
  Covariates_target = Covariates_ls[[2]],
  Equation = "tksat + tkdry + csol + k_s + lambda + psi + theta_s",
  Cores = 1,
  FileName = "KrigTest2",
  FileExtension = ".nc",
  Keep_Temporary = FALSE,
  nmax = 40,
  verbose = TRUE
)
Plot.Kriged(Krigs = ShapeKrig, SF = Jotunheimen_poly)

## End(Not run)

ErikKusch/KrigR documentation built on Feb. 17, 2025, 2:09 p.m.