Spark-Towers/multi-hazard-agricultural-climate-risk: v0.9.2 – Multi-hazard maize risk code

Summary First archived release of the code used in the article: S. Grosse, T. Berchoux, H. Belhouchette, M. Check, N. Baghdadi (2024). Assessing Multi-Hazard Risk Dynamics in South African Maize Production: A Framework Integrating Hazard Interactions, Temporal and Spatial Variability, and Irrigation Contexts. This repository provides the scripts used to analyse the spatio-temporal dynamics of climate-induced hazards for maize in South Africa, their interactions, and the role of irrigation in mitigating risk. What this release includes Python scripts for: Climatic hazard profiling (drought, heatwaves, cold spells, excessive rainfall, disease risk) City-level clustering based on irrigation coverage Construction of multi-hazard risk scores OLS models linking hazard metrics and yield variability Random Forest models predicting yield from climatic and hazard conditions Refactored code in scripts/ and original backup scripts in scripts/bak/ requirements.txt listing all Python dependencies Project structure compatible with reproducing the main analyses in the paper Note: The data/ and results/ folders are not tracked in version control. Users must provide their own input data in the expected formats. Folder structure data/ – raw and processed data (not included) scripts/ – main Python scripts for modelling and analysis scripts/bak/ – original non-refactored scripts results/ – plots, metrics, and model outputs (not included) requirements.txt – list of Python dependencies README.md – documentation and usage instructions