This repository contains the R code used to analyse the statistical associations among climatic variables in coastal Ghana using Partial Least Squares Structural Equation Modeling (PLS-SEM). The workflow includes data preparation, structural model estimation, bootstrapping, mediation analysis, moderation analysis, predictive assessment using PLSpredict, temporal robustness analysis across four decades, and the generation of tables and figures reported in the accompanying manuscript. The analyses were conducted entirely in the R programming environment using the seminr package and supporting R packages. The code is provided to support transparency, reproducibility, and future methodological applications in climate data analysis.