Supplementary Materials for : Spatial gradients and responses of ecosystem services to landscape patterns in railway corridors: A case study of Liaoning, China

This dataset contains the supplementary materials supporting the analysis of spatiotemporal dynamics, nonlinear determinants, and adaptive zoning of ecosystem services in railway corridors. The provided tables and figures offer detailed methodological specifications, data sources, and statistical diagnostics for the associated study.
The supplementary document includes:
Table S1 & Table S2: A comprehensive list of abbreviations and detailed descriptions of the multi-source geospatial datasets utilized in the study, including land cover, meteorological, topographic, and socioeconomic variables.Table S3: A summary of the optimal hyperparameters and model performance metrics ($R^2$) for the Extreme Gradient Boosting (XGBoost) machine learning models across the years 2000, 2010, and 2020.Table S4: Variance Inflation Factor (VIF) diagnostic results for the explanatory variables, confirming the absence of severe multicollinearity prior to spatial regression modeling.Fig. S1: Co-variation networks illustrating the complex correlations between landscape pattern metrics (AI, PD, SHDI, PARA) and control variables.Fig. S2: Visualizations of spatial non-stationarity and local driving effects of control variables (e.g., DEM, SLOPE, PRE, TEMP, GDP, POP) on ecosystem services, derived from the Multiscale Geographically Weighted Regression (MGWR) model.