Replication data: Leakage effects of establishing protected areas in multi-use forests in NSW, Australia

Replication data for: Cho, F. H. T., Rhodes, J. R., Maron, M., Lunney, D., Iftekhar, M. S., McIlwee, A., & Williams, B. A. (forthcoming). Leakage effects of establishing protected areas in multi-use forests. This dataset contains the pre-computed data files needed to reproduce all figures and summary tables in the paper. The study estimates the causal effect of New South Wales (NSW) state forest revocations on deforestation in adjacent areas using a Before–After Control–Impact (BACI) design with propensity-score matching. Two leakage channels are estimated: (1) leakage to private land within 10 km of a revoked state forest, and (2) leakage to neighbouring state forests within 10 km. Contents: processed_data_for_plotting.RData — all model estimates, Average Marginal Effects (AME), matched pair summaries, and spatial objects required to reproduce Figures 1–6 (103 MB) state_forests_revoked_update.gpkg — NSW state forest polygons with revocation year and woody vegetation attributes focal_sf_year.gpkg — focal state forest polygons aggregated by revocation year nsw_grid_points.parquet — NSW 10 km regular grid cell centroids (712,714 cells; 1.2 GB) matching_data/parquet/ — pixel-level covariate and deforestation panel for all NSW grid cells, split across 1,000 Parquet files The R code for full replication is available at Github. To reproduce all figures, download the data files, place them in the data/ subdirectory of the code repository, and run scripts/05_replicate_figures.R with R ≥ 4.2.