Hannah Zoller, July 2025

In order to reproduce the results presented in the manuscript "A global map of Earth system interactions" (Hannah Zoller, Steven J. Lade, Juan C. Rocha, C. Kendra Gotangco Gonzales, Ingo Fetzer, and Nitin Chaudhary), the scripts need to be executed for both the reanalysis data set CRU TS3.21 and the other six climate models in the following order:


1a. Cellwise_interactions.R 

This script takes as input
- LPJmL model outputs provided by Steven J. Lade et al. (https://doi.org/10.5281/zenodo.4738009, fpc.nc, mrunoff_surf.nc, co2_1841-2018.dat)

This script produces 
- cell-wise interaction data for the 
* effects of land use change: df_interaction_cellwise.Rdata 
* effects of climate change: CC_df_interaction_cellwise.Rdata
- the file raster.template.Rdata to be used in Top_down_approach.R
- figures showing world maps of cell-wise interaction strength, as presented in the manuscript

1b. LPJmL_based_classifications.R

This script takes as input
- LPJmL model outputs provided by Steven J. Lade et al. (https://doi.org/10.5281/zenodo.4738009, (fpc.nc, mrunoff_surf.nc, co2_1841-2018.dat)
- the file temp_1984_2013_cru3.23.avg.nc

This script produces data frames containing a global classification by 
- main plant function type (df.NoLU.late.Rdata)
- plant functional type (df.NoLU.late.fine.Rdata)
- biome(Ostberg.NoLU.late.Rdata)

2a. Bottom_up_approach.R

This script takes as input
- the world continents shapefile world_continents.shp (https://doi.org/10.5281/zenodo.4738009)
- the cell-wise interaction data df_interaction_cellwise and CC_df_interaction_cellwise.Rdata (as being produced by Cellwise_interactions.R)

This script produces
- the files metrics_separate_BU_alpha[1-5].Rdata
- the files metrics_BU_alpha[1-5].Rdata 
- the figures presented in the manuscript section "The bottom-up approach"

2b. Top_down_approach.R

This script takes as input
- the world continents shapefile world_continents.shp (https://doi.org/10.5281/zenodo.4738009)
- the Köppen-Geiger climate zones shapefile (https://koeppen-geiger.vu-wien.ac.at/present.htm)
- the realms shapefile (https://ecoregions.appspot.com/)
- the classifications df.NoLU.late (main PFTs), df.NoLU.late.fine (PFTs), and df.Ostberg.NoLU.late (biomes) (as being produced by LPJmL_classifications.R)
- the cell-wise interaction data df_interaction_cellwise and CC_df_interaction_cellwise.Rdata (as being produced by Cellwise_interactions.R)
- the bottom-up cluster validity indices 
   * metrics_separate_BU_alpha1-5.Rdata
   * metrics_BU_alpha1-5.Rdata 
   (as being produced by Bottom_up_approach.R)

This script produces the figures presented in the manuscript section "The top-down approach"


Station_coverage.R

This script takes as input the station count data (stn) provided along with the CRU TS3.21 dataset. The data is publicly available under https://data.ceda.ac.uk/badc/cru/data/cru_ts/cru_ts_3.21/station (accessible with a free CEDA account). 

This script produces maps displaying the average station cover underlying the CRU TS3.21 dataset for the variables of temperature and diurnal temperature range combined (tmpdtr), total precipitation (pre), rainday counts (wet), and cloud cover (cld)