Reproducibility archive for "Fine-fuel dryness outperforms ten-hour fuel moisture for wildfire ignition: a case-crossover comparison of four indices"

Data and code to reproduce all tables and figures in the paper "Fine-fuel dryness outperforms ten-hour fuel moisture for wildfire ignition: a case-crossover comparison of four indices". Four formulations of fuel dryness are compared as predictors of wildfire ignition in the Republic of Korea using a case-crossover design: effective humidity (the operational term), an equilibrium-moisture index evaluated instantaneously, ten-hour fuel moisture integrated at its nominal time constant, and the hourly Fine Fuel Moisture Code of the Canadian Fire Weather Index. The analysis covers 5,783 ignitions and 22,936 matched control observations, 1997-2013. CONTENTS Derived data (13,106 quality-controlled ignition records; the analysis table holding all four formulations under three temporal conventions; the surface station list), the full processing pipeline, and a single verification script. REPRODUCE Run 01_reproduce.py. It recomputes all 51 reported quantities, regenerates Figures 1-4, and compares against the published values, exiting 0 only if every quantity agrees within tolerance. Requires Python >= 3.9 with numpy, pandas, statsmodels and matplotlib. SOURCE DATA Raw hourly surface observations (order 10^8 station-hours) and the source ignition shapefile are not included for reasons of size and provenance. Both are public: the ignition inventory is issued by the Korea Forest Service and the surface observations by the Korea Meteorological Administration. Retrieval and conversion scripts are included, so the derived data can be rebuilt from the originals. See 00_README.md for file descriptions, column definitions, and known characteristics of the data.