Code for "Global analysis suggests nitrogen deposition as an underestimated driver of vegetation greening"

Github: https://github.com/JonasTrepel/GlobalChangeEcosystemFunctioning ## Repository Structure: --- ### Data Acquisition and PreparationAll datasets used in this study are publicly available. Atmospheric nitrogen deposition estimates are taken from Rubin et al. 2023 (https://doi.org/10.5194/acp-23-7091-2023). Climate variables are derived from the ERA5 monthly reanalysis dataset (https://developers.google.com/earth-engine/datasets/catalog/ECMWF_ERA5_MONTHLY). Human land-modification data are obtained from Theobald et al. (2020) (https://zenodo.org/records/3963013). Vegetation data (EVI) are sourced from the MODIS MOD13A1 product (https://developers.google.com/earth-engine/datasets/catalog/MODIS_061_MOD13A1). The alternative nitrogen-deposition product from Zhu et al. is available at https://doi.org/10.6084/m9.figshare.26778574. Fire-occurrence data are taken from the MCD64A1 Collection 6.1 product (https://lpdaac.usgs.gov/products/mcd64a1v061/). Temporal N-deposition datasets can be downloaded for the United States from https://nadp.slh.wisc.edu/committees/tdep/#tdep-maps and for Europe from https://thredds.met.no/thredds/catalog/data/EMEP/2025_Reporting/catalog.html. Protected-area boundaries are obtained from the World Database on Protected Areas (https://www.protectedplanet.net/en/search-areas?geo_type=site). Since we are not the owner of the data (and files exceed storage capacity of most providers), we provide all analysis ready datasets at OSF to ensure scripts in the analysis and viz folders can be reproduced: https://osf.io/zn8kg/files/osfstorage #### **R/rgee/: Remote Sensing Data from Google Earth Engine (via `rgee`)**- `R/rgee/initialize_rgee.R` – Set up `rgee` for accessing GEE.- `R/rgee/rgee_annual_ _download.R` – Download annual composites of RS variables (e.g., EVI, temperature). #### **R/prep/: Data preparation**- `R/prep/combine_hmi_and_get_change.R` - calculate difference in human land modification between 2015-2000- `R/prep/calculate_europe_total_n_depo.R` - Aggregate annual N deposition totals for Europe.- `R/prep/get_pa_shapes.R` – Combine and prepare WDPA shapefiles.- `R/prep/rasterize_biomes_and_pas.R` – Rasterize PA boundaries and biomes for easy extraction later on.- `R/prep/create_grid.R` - create global 5x5 km grid- `R/prep/extract_covs.R` - extract covariates- `R/prep/find_controls_for_pas.R` - find suitable control sites for protected areas- `R/prep/extract_time_series_values.R` - extract time series values to calculate changes in EVI and climate - `R/prep/calculate_trends.R` - calculate trends in time series - `R/prep/combine_data_for_analysis.R` - Combine final datasets to be used in subsequent analysis ### Analysis #### **Main analysis**- `R/analysis/sdmtmb_world_grid.R`- run global greening models for global grid dataset- `R/analysis/sdmtmb_pa_dataset.R`- run global greening models for protected area dataset #### **Sensitivity**- `R/analysis/sensitivitiy_zhu_n_depo.R` - test relationship between greening and N depo using an alternative product - `R/analysis/sensitivitiy_n_depo_trend.R` - test relationship between greening and N depo trend in europe and the USA ### Visualization - `R/viz/extract_model_preds_world_grid.R` - extract predictions from the global grid dataset models- `R/viz/extract_model_preds_pa_dataset.R` - extract model predictions from the protected areas dataset model- `R/viz/grid_figures.R` – Figures based on grid dataset (Figs 1&2, plus supplementary figures).- `R/viz/pa_figures.R` – Figures based on PA dataset (Fig. 3 + supplementary figures).- `R/viz/sensitivity_zhu_et_al_estimates.R` – Plot results of model with alternative N depositon product. ### Functions - `R/functions/move_polygon.R` – Displace polygons (e.g., to find control sites).- `R/functions/resolve_overlaps.R` – Resolve overlapping PA boundaries. --- ## Requirements- R ≥ 4.5.0- Google Earth Engine account and `rgee` setup.