Install
openclaw skills install @ruiduobao/geoskill-water-table-mappingopenclaw skills install @ruiduobao/geoskill-water-table-mappingThis skill generates regional groundwater level rasters and depth-to-water (surface-to-water-table depth) rasters from water level observations at discrete monitoring wells via spatial interpolation, applicable to groundwater contour mapping, depth zoning, and well network assessment.
Core algorithms: two interpolation methods are provided — IDW (Inverse Distance Weighting), pixel value = Σ(valueᵢ/dᵢᵖ)/Σ(1/dᵢᵖ), robust and fast; and simplified Ordinary Kriging, which uses an exponential variogram and solves the kriging system with an unbiasedness constraint to obtain the best linear unbiased estimate. Terrain constraints are supported (the water table must not exceed the surface elevation). Accuracy is assessed with leave-one-out cross-validation statistics (RMSE / MAE / R²).
pip install 'numpy' 'rasterio' 'scipy'
python geoskill-water-table-mapping.py --bbox 116.0 39.0 117.0 40.0 --method idw
python geoskill-water-table-mapping.py --bbox 116.0 39.0 117.0 40.0 --synthetic --output-dir ./out
python geoskill-water-table-mapping.py --bbox 116 39 117 40 --synthetic --method kriging --output-dir ./out
python geoskill-water-table-mapping.py --input wells.csv --bbox 116 39 117 40 --grid-size 128 --output-dir ./out
python geoskill-water-table-mapping.py --bbox 121 31 122 32 --synthetic --power 3 --n-wells 60 --quiet
python geoskill-water-table-mapping.py --bbox 113 23 114 24 --synthetic --method idw --grid-size 256 --output-dir ./out
| File | Format | Description |
|---|---|---|
water_table.tif | GeoTIFF | Groundwater level spatial raster (m, terrain-constrained) |
depth_to_water.tif | GeoTIFF | Depth-to-water raster = DEM − water level (m) |
interpolation_report.json | JSON | Cross-validation (RMSE/MAE/R²) + comparison against ground truth (synthetic) |
output-manifest.json | JSON | Run manifest |
--input: local well CSV; coordinate columns (x/lon) and water level columns (level/water_level/head) are auto-detected.--synthetic: wells with spatial gradients and observational noise + ground-truth water level field + DEM, fully offline.--synthetic mode requires no network at all.MIT
本 skill 由离散监测井点的水位观测,经空间插值生成区域地下水位栅格与埋深(水位到地表深度)栅格,适用于地下水等水位线图编制、埋深分区、井网评估等场景。
核心算法:提供两种插值方法——IDW(反距离加权),像元值 = Σ(valueᵢ/dᵢᵖ)/Σ(1/dᵢᵖ),稳健快速;简化普通克里金(Ordinary Kriging),采用指数型变异函数,求解带无偏约束的克里金方程组得最优线性无偏估计。支持地形约束(地下水位不得高于地表高程)。精度由留一法交叉验证统计 RMSE / MAE / R² 评估。
pip install 'numpy' 'rasterio' 'scipy'
python geoskill-water-table-mapping.py --bbox 116.0 39.0 117.0 40.0 --method idw
python geoskill-water-table-mapping.py --bbox 116.0 39.0 117.0 40.0 --synthetic --output-dir ./out
python geoskill-water-table-mapping.py --bbox 116 39 117 40 --synthetic --method kriging --output-dir ./out
python geoskill-water-table-mapping.py --input wells.csv --bbox 116 39 117 40 --grid-size 128 --output-dir ./out
python geoskill-water-table-mapping.py --bbox 121 31 122 32 --synthetic --power 3 --n-wells 60 --quiet
python geoskill-water-table-mapping.py --bbox 113 23 114 24 --synthetic --method idw --grid-size 256 --output-dir ./out
| 文件 | 格式 | 说明 |
|---|---|---|
water_table.tif | GeoTIFF | 地下水位空间栅格(m,受地形约束) |
depth_to_water.tif | GeoTIFF | 埋深栅格 = DEM − 水位(m) |
interpolation_report.json | JSON | 交叉验证(RMSE/MAE/R²)+ 与真值对比(合成) |
output-manifest.json | JSON | 运行清单 |
--input:本地井点 CSV,自动识别坐标列(x/lon)与水位列(level/water_level/head)。--synthetic:带空间渐变与观测噪声的井点 + 真值水位场 + DEM,完全离线。--synthetic 模式完全无网络。MIT