Wetland Change Monitor
Monitors wetland extent, inundation frequency, and land cover transitions from multi-year water occurrence rasters. Identifies degradation, recovery, and human encroachment patterns.
CLI Usage
python scripts/wetland_change_monitor.py --water-frequency water_freq.tif --years 2015 2016 2017 2018 2019
python scripts/wetland_change_monitor.py --water-frequency wf.tif --years 2018 2019 2020 --frequency-thresholds 0.25 0.75
python scripts/wetland_change_monitor.py --water-masks masks.tif --years 2015 2020 --change-schema degradation recovery stable
数据下载
本 skill 可自动从 Microsoft Planetary Computer 下载水体数据 (无需 API key):
# 自动下载 JRC GSW (water frequency) + Sentinel-1 GRD (water mask)
python scripts/wetland_change_monitor.py \
--bbox 116.30,39.85,116.45,39.95 \
--date-range 2020-06-01,2020-09-30 \
--years 2015 2020 \
--output-dir ./wcm-auto
下载的内容:
- JRC Global Surface Water (
jrc-gsw) — 历史水体频率，替换--water-frequency - Sentinel-1 GRD (
sentinel-1-grd) — SAR 水体掩膜，替换--water-masks
下载元数据（data_source, fetched_at, collection, jrc_downloaded, sentinel1_downloaded）会写入 output-manifest.json。
Parameters
| Parameter | Description |
|---|---|
--water-frequency |
Multi-band water frequency raster (JRC-style, bands = years) |
--water-masks |
Alternative: multi-band binary water masks (1=water, 0=dry) |
--years |
List of years corresponding to bands |
--frequency-thresholds |
Two thresholds for classification (default: 0.25 0.75) |
--change-schema |
Change types to detect (default: degradation recovery stable) |
--include-sar |
Include SAR-based water detection (optional) |
--min-area |
Minimum patch area in hectares (default: 0.5) |
--output-dir |
Output directory (default: wetland-output) |
Output
| File | Description |
|---|---|
wetland_type.tif |
Wetland type classification raster |
inundation_frequency.tif |
Inundation frequency raster (0-1) |
wetland_change.geojson |
Vectorized change patches with type and confidence |
transition_matrix.csv |
Year-to-year transition matrix |
wetland-report.json |
Detailed statistics and summary |
output-manifest.json |
Machine-readable manifest |
Exit Codes
| Code | Meaning |
|---|---|
| 0 | Success |
| 2 | Argument error |
| 3 | Dependency missing |
| 6 | Validation error |
| 7 | Processing failure |
数据下载
本 skill 可自动从 Microsoft Planetary Computer 下载数据 (无需 API key):
python wetland_change_monitor.py --bbox 116,39,117,40 --date-range 2024-06-01,2024-06-30 --output-dir <tmp>
--bbox W,S,E,N: WGS-84 边界框 (西, 南, 东, 北)--date-range START,END: 日期范围 (YYYY-MM-DD,YYYY-MM-DD)--aoi-file <path.geojson>: 替代 --bbox 的 GeoJSON 多边形--cache-dir <path>: 缓存目录 (默认 ~/.geoskill_cache)
当用户只给 --bbox + --date-range (没有 --image) 时，skill 自动下载数据。
当用户给 --image 时，走原文件路径 (向后兼容)。