MAIN QUEST · JUN–OCT 2026 · RELEASED
Wahaj ·
A reproducible ML Python and Google Earth Engine framework for evaluating urban greening–energy trade-offs in desalination-dependent cities, with Jeddah as the case study.
OBJECTIVES CLEARED
- XGBoost land surface temperature model on Landsat 8, spatial cross-validation R² 0.795 on 19,650 cells, benchmarked against random forest, gradient boosting and linear regression
- Matched-contrast benchmark: −1.18 °C per greened pixel outside the built-up area, stress-tested with emissivity re-retrieval, alternative NDVI thresholds and a thermal-footprint simulation
- Counterfactual gates that test whether predicted greening effects hold up, and report where they do not
- NEGI index plus a water–energy ledger: each degree of cooling carries ~2.9–5.3 MWh/yr of desalination energy
LOOT
- Python
- XGBoost
- Google Earth Engine
- Landsat 8
- Spatial statistics
