Akajiaku11/Predictive-Mapping-of-Oil-Spill-Induced-Mangrove-Degradation-in-Nigeria

# Predictive Mapping of Oil Spill‑Induced Mangrove Degradation in Nigeria **Remote Sensing + Machine Learning (Python + GEE)** This repository contains a reproducible workflow for mapping and **predicting oil spill‑induced mangrove degradation** in the Niger Delta (Rivers State, Nigeria) using **Sentinel‑2 / Landsat 8**, **SRTM**, **NOSDRA oil‑spill records**, and **Gradient Boosted Decision Trees (XGBoost)** with **SHAP** explainability. > Paper context and methods adapted from the project draft provided by the authors (uploaded by the user). ## Highlights - Compute **NDVI / Red‑Edge NDVI (RENDVI)** time series (2020 → 2024) - Supervised **LULC classification** (Random Forest in **Google Earth Engine**) for 2020 & 2024 - **Oil spill hotspot severity** using Kernel Density Estimation (KDE) + **K‑Means** - Feature stack: ΔNDVI, ΔRENDVI, spill density, **ESI** rank, elevation class, LULC transition - Train **XGBoost** classifier; evaluate **Accuracy, Precision, Recall, F1, ROC‑AUC** - **Explain predictions** with **SHAP**; export risk probability map GeoTIFF + figures ## Repository structure ``` predictive-mangrove-degradation/ ├─ README.md ├─ LICENSE ├─ CITATION.cff ├─ pyproject.toml ├─ requirements.txt ├─ Makefile ├─ .gitignore ├─ .github/workflows/ci.yml ├─ configs/ │ └─ rivers_state_example.yaml ├─ data/ │ ├─ raw/ # put input rasters/vectors here │ ├─ interim/ # intermediate outputs │ └─ processed/ # final maps & model artifacts ├─ notebooks/ │ ├─ 00_quickstart.ipynb │ └─ 10_model_diagnostics.ipynb ├─ scripts/ │ ├─ gee_lulc_classifier.js # RF in GEE for 2020/2024 │ └─ prepare_shapefile_grid.py ├─ src/pmd/ │ ├─ __init__.py │ ├─ cli.py # command line interface │ ├─ io.py # loading/saving │ ├─ indices.py # NDVI/RENDVI + delta │ ├─ geoutils.py # raster/vector helpers │ ├─ spills.py # KDE + clustering │ ├─ features.p …