Leos19100/Sentinel-DisasterAssessment

Research project for MSc Transport at Imperial College: Temporal analysis of infrastructure and roads in sub-Saharan Africa in post-disaster situations # Flood Impact Assessment — Niger–Benue Floods, Nigeria (October 2022) Google Earth Engine–based remote-sensing analysis of the October 2022 Niger–Benue floods in Nigeria, built as the code appendix to an Imperial College London MSc Transport Research Project. The analysis starts as a tightly-scoped, statistically validated case study around Lokoja (Kogi State) and is progressively extended — to road-network and population impact, to a multi-year temporal-anomaly analysis, and finally to all 37 states and territories of Nigeria. ## Contents - Overview - Repository layout - Study area and event window - The notebooks - Data sources - Key results - Getting started - Limitations and validation - Outputs reference - Methodology references ## Overview The project uses Sentinel-1 (SAR), Sentinel-2 (optical), CHIRPS (precipitation), WorldPop (population) and OpenStreetMap (road network) data on Google Earth Engine to answer four progressively broader questions about the same flood event: 1. **Where did it flood, and how confident can we be about it?** (`Index_Analysis.ipynb`) 2. **What did that flooding do to roads and people?** (`Impact_Assessment.ipynb`) 3. **Was this actually unusual, and did the landscape recover?** (`Temporal_Evolution.ipynb`) 4. **Does any of this hold up outside Lokoja, across an entire country?** (`National_Flood_Assessment.ipynb`) Each notebook is self-contained — it re-derives the flood masks it needs directly from Sentinel-1/Sentinel-2 rather than depending on another notebook's saved output — while staying numerically consistent with the others, since they are meant to be read together as one appendix. ## Repository layout ``` . ├── README.md ├── requirements.txt ├── notebooks/ │ ├── Index_Analysis.ipynb # 1. flood detection + threshold validation │ ├── Impact_Assessment.ipynb # 2. road network + population exposure │ ├── Temporal_Evolution.ipynb # 3. multi-year anomalies + recovery │ └── National_F …