Ensemble sufficiency for stochastic wildfire automata: convergence sweep scripts and derived tables

Code and derived data supporting the article "How Many Monte Carlo Runs Does a Stochastic Wildfire Automaton Need? Fire-Size-Dependent Ensemble Sufficiency and a Measured Calibration Saving". The record contains the convergence sweep that measures how agreement and burned area stabilize as the ensemble grows, the calibration-loop cost measurement at fixed and adaptive sizes, the jackknife and bootstrap resampling behind the reported rank correlation, and the derived tables they produce. The primary UAECOB incident registry is not redistributed here, because its terms of use rest with the Bogota fire service. The deposited pipeline regenerates the calibrated event set from it once a requester has obtained it. How to run it Unzip the archive and run any script from code/. Paths resolve against the record itself through code/rutas.py, so nothing points at the machine where the work was done. Figures are written to figures/. The record is self-contained. It carries the calibration and simulation module under Tesis/implementaciones/, Apache 2.0, so there is no external repository to fetch first. Licensing, which is split on purpose The source code in code/, together with the vendored modules, is released under the Apache License 2.0, in the file LICENSE. The derived tabular data in data/ are released under Creative Commons Attribution 4.0 International, in the file LICENSE-DATA.txt. The split is deliberate, so that the code can be reused in a downstream product without the attribution obligations of the data travelling with it. Contents code/ holds 5 scripts, 512 lines in total. run_ensemble_sufficiency_v2.py. The convergence sweep, 8 independent ensembles at each of 12 sizes for each of the 8 calibrated events, writing the per-size mean and spread of kappa, intersection over union and burned area. run_calibration_saving.py. Measures the cost of a real calibration loop at fixed and at adaptive ensemble size, on the device. run_bootstrap_ci.py. Bootstrap and jackknife intervals for the headline quantities, including the rank correlation between fire size and sufficient ensemble. make_prediction_table.py. Builds the error and coefficient-of-variation table and the predicted-versus-observed comparison of the sizing rule. make_figures_v2.py. Regenerates the convergence, degeneracy, mean-drift and saving figures. data/ holds 8 derived tables. ensemble_convergencia_v2.csv. The full sweep, mean and spread of each metric at every ensemble size for every event. mstar_v2.csv. Sufficient ensemble size per event, as a power of two and interpolated. p8_bootstrap_ci.csv. Interval and p value of each headline quantity. p8_calibration_saving.csv. Rollout and differential-evolution times at fixed and adaptive size. p8_error_cv_table.csv. Kappa error and area coefficient of variation at three ensemble sizes. p8_multimetric.csv. Range of the sufficient size under each metric considered. p8_pred_vs_obs.csv. Observed against predicted sufficient size for the 6 non-degenerate fires. p8_spearman_jackknife.csv. The rank correlation recomputed with each fire left out in turn. What is not in this record, and why Primary UAECOB incident registry. Terms of use rest with the Bogotá fire service. The deposited pipeline regenerates the calibrated event set from it once a requester has obtained it. Environment Python 3.12 with PyTorch 2.11 on CUDA 12.8, numpy, pandas and scipy. A graphics processing unit is required. Results were measured on an NVIDIA RTX 5050. Random seeds are fixed throughout and recorded in each script. A note on language The scripts carry Spanish-language comments and the tables carry Spanish column names, because Spanish is the working language of the project. The lists above give the English purpose of every file, and the README included in the archive carries a glossary of the recurring column names.