LHASA — Landslide Hazard Assessment System

Real-time landslide hazard assessment combining MSWEP near-real-time precipitation, ECMWF HRES forecasts, and a static susceptibility map to produce daily hazard grids at 0.1° resolution over South/Southeast Asia.

---

Folder Structure

Model_LHASA/
│
├── lhasa.py               # Main entry point — runs full pipeline
├── ls_precip.sh           # Merges MSWEP + HRES into one NetCDF
├── get_MSWEP.py           # Downloads MSWEP NRT daily from Google Drive via rclone
├── get_HRES.py            # Downloads ECMWF HRES 0.25° from ECMWF Open Data
│
├── static_data/           # !! YOU MUST PROVIDE THESE — not in repo !!
│   ├── hill.json          # Hillslope mask (GeoJSON, EPSG:4326)
│   ├── hist_ari.nc        # Historical ARI climatology (variable: p95)
│   └── sus_map.nc         # Landslide susceptibility map (variable: sus_map)
│
├── MSWEP_NRT/             # Auto-created — MSWEP downloaded files land here
├── ECMWF_HRES/            # Auto-created — HRES downloaded files land here
├── input_data/            # Auto-created — merged precip files land here
├── output_data/           # Auto-created — final hazard NetCDFs land here
│   └── YYYYMMDD/
│       ├── lhasa_YYYYMMDD.nc        # Day 1
│       ├── lhasa_YYYYMMDD_D2.nc     # Day 2
│       ├── lhasa_YYYYMMDD_D3.nc     # Day 3
│       └── lhasa_YYYYMMDD_D5.nc     # Day 5
│
├── docs/
│   └── SETUP_AND_RUN.md   # Full step-by-step guide (start here)
│
├── environment.yml        # Conda environment spec
├── requirements.txt       # pip requirements
├── .gitignore
├── LICENSE
└── README.md

---

Output Variables (per NetCDF)

Variable	Description	Values
ARI	7-day weighted Antecedent Rainfall Index (mm)	float32
HeavyRain	ARI exceeds historical 95th percentile	1 or NaN
ModerateHazard	Heavy rain AND susceptibility > 2	1 or NaN
HighHazard	Heavy rain AND susceptibility ≥ 4	1 or NaN

---

Quick Start

# Full pipeline — all 4 forecast days for a date
python lhasa.py 20250622 all

# Single day
python lhasa.py 20250622 1

See docs/SETUP_AND_RUN.md for complete setup instructions.

---

Dependencies

• Python ≥ 3.10
• CDO ≥ 2.0, NCO (ncatted, ncap2)
• rclone (for MSWEP download)
• See environment.yml for full Python package list

---

License

MIT License — see LICENSE.