Causal inference reveals the dominant role of stronger climate-adaptation coupling in future potato production

README: Causal inference reveals the dominant role of stronger climate-adaptation coupling in future potato production Dataset Description This database contains the R scripts and data used to generate the figures for this study. The study systematically evaluates the potential of different adaptation measures for potato production in China under future climate change by integrating multi-crop model simulation results, climate projection data, and causal inference methods. It reveals the complex coupling effects between climate factors and adaptation measures. Code Structure Main Script generate_figures.R: Contains the complete code for generating all 6 figures in the paper. The code is organized by figure, including the following 6 main sections: Figure 1: Potato yield change trends and regional differences in historical and future periods. Figure 2: Yield change rates under different adaptation measures and the spatial distribution of optimal adaptation measures. Figure 3: The impact of climate factors on yield revealed by random forest and SHAP analysis. Figure 4: Causal effect (ATE) distribution of climate-adaptation coupling on potato yield. Figure 5: Sensitivity estimation of the impacts of climate factors, adaptation measures, and their coupling effects on yield. Figure 6: Spatial distribution and contribution decomposition of climate-adaptation coupling. Data File Structure Data is stored in the directory E:/Data and code/Data/, categorized by figure: P1/: Data related to Figure 1 P2/: Data related to Figure 2 P3/: Data related to Figure 3 P4/: Data related to Figure 4 P5/: Data related to Figure 5 P6/: Data related to Figure 6 CNLP/: Spatial boundary data for the potato cultivation region on the Chinese Loess Plateau. Output Files Figures are saved in the E:/Data and code/Figure/ directory in TIFF format. System Requirements R version: ≥ 4.0.0 System memory: Recommended ≥ 16GB Graphics device: Support for high-resolution graphics output. Dependent Packages # Data processing library(data.table) # Efficient data reading library(tidyverse) # Data wrangling and transformation # Spatial data processing library(sf) # Spatial vector data processing # Data visualization library(ggplot2) # Basic graphics system library(scales) # Graphics scale adjustment library(ggridges) # Ridge plots library(ggbeeswarm) # Beeswarm plots library(ggnewscale) # Multiple scale color system # Figure assembly library(cowplot) # Figure composition library(aplot) # Figure insertion # Others library(viridis) # Color scales library(mgcv) # GAM models Usage Instructions Basic Usage Ensure all data files are in the correct directory structure. Install all necessary R packages. Running the script will automatically generate all figures and save them to the specified directory. Custom Modifications Data paths: Modify the file paths in the fread() function to match your local data storage location. Output settings: Adjust the dimensions, resolution, and output format in the ggsave() function. Figure style: Modify the theme() parameters to adjust the figure appearance. Color scheme: Modify the scale_fill_*() and scale_color_*() functions to adjust color settings. Notes The code uses hard-coded paths; adjust them according to your local directory structure before first run. Some figures require more memory; it is recommended to run on a computer with sufficient memory. Generating all figures may take a long time (approximately 5-10 minutes).