Reproducibility Package for Spatiotemporal Dependence and Finite-Threshold Tail Co-Exceedance in Compound Drought–Heat Extremes in Peru

Abstract

This reproducibility package contains the complete computational materials supporting the reported statistical analyses. It includes the final Jupyter/Google Colab notebook, source and derived analytical datasets, statistical tables, figures, spatial outputs, software-environment information, and machine-readable result files required to reproduce the analyses. The workflow includes rank-based dependence measures, finite-threshold conditional co-exceedance, moving-block bootstrap inference, climatological-reference sensitivity analysis, autocorrelation-robust trend assessment with multiple-testing correction, change-point diagnostics, regional and exploratory spatial analyses, network-level conditional modeling, and Monte Carlo validation. When the primary exchangeable GEE specification does not provide finite coefficient inference, the documented binomial GLM with site-clustered covariance is used as the predefined fallback. Spatial interpolation outputs are provided exclusively for exploratory visualization and are not used for inferential testing. The package is intended to support transparency, verification, reproducibility, and future methodological extensions.