Decision-grade risk and cost mapping for illegal gold mining at Crucitas, Costa Rica: prioritisation, phased remediation portfolios, and uncertainty-aware policy ranking

Description

Supporting code, raster inputs, vector data, and tabular files for the Crucitas illegal gold-mining risk and cost-mapping study, including spatial prioritisation, remediation-cost modelling, Monte Carlo uncertainty analysis, and legislative policy-pathway comparison.

Steps to reproduce

1. Download all dataset files and preserve the uploaded folder/file structure where possible, especially the “TIF FILES” folder. 2. Keep the Python scripts, CSV files, GeoPackage, field-point CSV, and TIF folder accessible from the same working directory. Unzip “CSV FILES.zip” if the CSV files are provided as a compressed archive. 3. Install a Python 3 environment with the required libraries: pandas, numpy, geopandas, rasterio, shapely, scikit-learn, scipy, matplotlib, and joblib. A conda/mamba geospatial environment is recommended because rasterio/geopandas require GDAL-related dependencies. 4. Open the Python scripts and update the local file paths. The original scripts contain Windows/OneDrive paths from the development environment; users must replace these with their own local directories. 5. Run “MODEL 1 ML.py” to prepare spatial and temporal predictor features from the raster and tabular inputs. 6. Run “MODEL 2 IMPACT SCORING.py” to generate the 0–100 spatial impact-prioritisation indices for land, water, hydrology, and combined prioritisation. 7. Run “SCIENTIFIC VALIDATION.py” and/or “MODEL 3 HG CALIBRATION.py” to evaluate the field-based triangulation using SCIENTIFIC.gpkg or article_points_with_distance_and_ring.csv. 8. Run “MODEL 4 FINANCIAL.py” to convert the prioritisation outputs into phase-timed remediation costs, present-value liabilities, credits, and policy-pathway overlays. 9. Run “MODEL 5 SENSIVILITY.py” to perform sensitivity analysis, Monte Carlo uncertainty propagation, exceedance-risk estimation, and policy-pathway uncertainty comparison. 10. Review the generated CSV, raster, and figure outputs in the corresponding output folders. Results should be interpreted as conditional decision-support outputs, not as definitive contaminant maps, legal conclusions, mineral reserve valuations, or final engineering designs.

Institutions

• Chongqing University

  Chongqing, Chongqing

Categories

Licence