Initial public release of the Beledweyne flood frequency mapping (FFM) pipeline accompanying the manuscript submitted to Sustainability:
"Are Flood-Risk Relocation Policies Sustainable? Decadal SAR Evidence of Re-encroachment in Beledweyne, Somalia"
What's included
- 9 Python modules under
src/ (33 functions: SAR preprocessing, flood detection, FFM, HAND, BFAST, GEE helpers, viz)
- 7 sequential Jupyter notebooks (
notebooks/01_* → 07_*) demonstrating full reproducibility
- Pre-computed FFM products (CSV summary tables) and publication figures
- Verification: NB 07 reproduces legacy FFM exactly (Max |diff| = 0.000000, Pearson r = 1.000000, 8,916,196 pixels)
Key results
- Three flood detection methods compared: Z-score, Fixed-threshold, Log-ratio
- Cross-method correlations: r(Z, Fixed) = -0.0435, r(Z, LR) = +0.6759, r(Fixed, LR) = +0.0792
Citation
See CITATION.cff in the repository root.
If you use this software or dataset, please cite both the article and this repository.