Municipality-Scale Flood Risk Mapping in Narino, Colombia

Using Sentinel-1 SAR and Ensemble Machine Learning (2015--2025)

Cristian Espinal Maya · Santiago Jimenez Londono

School of Applied Sciences and Engineering, Universidad EAFIT, Medellin, Colombia

·

---

About

This project replicates the Antioquia flood risk mapping framework for the Department of Narino, Colombia. Narino (33,268 km^2; 64 municipalities; ~1.9 million inhabitants) presents unique challenges for flood risk assessment due to its extreme topographic variability (0--4,764 m elevation), dual precipitation regimes (bimodal Andean vs. unimodal Pacific), and high vulnerability in Pacific lowland communities.

Study Area Characteristics

Feature	Description
Area	33,268 km^2
Municipalities	64, organized in 13 subregions
Capital	San Juan de Pasto (2,527 m asl)
Elevation range	0 m (Pacific coast) to 4,764 m (Volcan Cumbal)
Major rivers	Patia, Guaitara, Telembi, Mira, Juanambu, Mayo, Sanquianga
Climate zones	Pacific lowlands (2,600--5,000+ mm/yr), Andean highlands (800--1,600 mm/yr)
Most flood-prone areas	Tumaco, Barbacoas, Olaya Herrera, Roberto Payan, Magui, El Charco

Repository Structure

.
├── gee_config.py              # Central configuration (Narino-specific)
├── scripts/                   # Processing and analysis pipeline
│   ├── 01_sar_water_detection.py
│   ├── 02_jrc_water_analysis.py
│   ├── 03_flood_susceptibility_features.py
│   ├── 04_ml_flood_susceptibility.py
│   ├── 05_population_exposure.py
│   ├── 06_climate_analysis.py
│   ├── 07_visualization.py
│   ├── 08_generate_tables.py
│   └── 09_quality_control.py
├── overleaf/                  # Manuscript (preprint format)
├── data/                      # Downloaded data (auto-created)
├── outputs/                   # Results (auto-created)
├── BIBLIOGRAPHY.md            # Research context and references
└── README.md

Data Sources

All data are open-access and processed via Google Earth Engine:

• Sentinel-1 GRD (ESA/Copernicus) -- 10 m SAR flood detection
• JRC Global Surface Water -- 38-year water dynamics
• SRTM DEM v3 -- Topographic features (30 m)
• MERIT Hydro -- HAND computation (90 m)
• CHIRPS / ERA5-Land -- Precipitation and soil moisture
• ESA WorldCover / Sentinel-2 -- Land cover and NDVI
• WorldPop -- Population density (100 m)
• FAO GAUL 2015 -- Administrative boundaries

Reproducing the Analysis

Requirements

• Python 3.10+
• Google Earth Engine account (sign up)
• Libraries: see requirements.txt

Setup

# 1. Clone and setup
cd narino_flood_risk_research
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# 2. Configure GEE credentials
earthengine authenticate
echo "GEE_PROJECT_ID=ee-flood-risk-narino" > .env

# 3. Run pipeline
python scripts/01_sar_water_detection.py   # ~3-4 hours (GEE)
python scripts/02_jrc_water_analysis.py    # ~30 min (GEE)
python scripts/03_flood_susceptibility_features.py  # ~2 hours (GEE)
python scripts/04_ml_flood_susceptibility.py        # <30 min (local)
python scripts/05_population_exposure.py   # ~1 hour
python scripts/06_climate_analysis.py      # ~1 hour
python scripts/07_visualization.py         # ~20 min
python scripts/08_generate_tables.py       # ~5 min
python scripts/09_quality_control.py       # ~10 min

Key Differences from Antioquia Framework

1. Terrain: Extreme elevation range (0--4,764 m) causes SAR shadow/layover in Andean zones. Detection accuracy is highest in Pacific lowlands.
2. Precipitation: Dual regime -- bimodal Andean vs. unimodal Pacific (one of the wettest regions globally).
3. ENSO Response: Narino receives above-normal rainfall during both El Nino and La Nina, unlike most Colombian departments.
4. Subregions: 13 subregions (vs. 9 in Antioquia) used for spatial cross-validation.
5. SAR Orbit: Consider using both ASCENDING and DESCENDING passes to reduce terrain distortion effects.

Citation

@article{EspinalMaya2026Narino,
  author  = {Espinal Maya, Cristian and Jim\'enez Londo\~no, Santiago},
  title   = {Municipality-Scale Flood Risk Mapping in {Nari\~no}, {Colombia},
             Using {Sentinel-1} {SAR} and Ensemble Machine Learning (2015--2025)},
  year    = {2026},
  note    = {Preprint}
}

License

Source code: MIT License. Manuscript and figures: CC BY 4.0.

Acknowledgments

This research builds on the open-access framework developed for the Department of Antioquia. All data and computational resources are open-access (Google Earth Engine, Copernicus/ESA Sentinel-1).