A developmental version of C-REML is on another github repository and can be assessed by the command devtools::install_github("kohleth/spcreml") in R.
The folder simStudy contains the script simstudy.R which runs the simulation study. It compares geoR, asreml, and C-REML. This means you need to have all three packages installed on R. Note that asreml is a proprietary package.
Depending on the processing power of your computer, you might want to adjust the parameter Nlist and Nsim accordingly. Nlist controls the sample size of the generated dataset. Nsim controls the number of replication in the study.
The folder pH_analysis contains the datasets and script for the analysis of the pH dataset.
The script analysis.R is the main script which contains the analysis code. If you run this on a normal computer with <20 cores, it will take a long time to fit the model. Therefore we have provided the fitted model fit1.rda so you can just load it onto R load("fit1.rda") when necessary.
The prediction grid in this script is done on a much lower resolution than the 50X50m said in the paper, this is so that the file size of the covariate map (covar.tif) can be greatly reduced (from >1 GB to <1 MB).