What problem does it solve? Running the CWatM global hydrological model requires deep operational knowledge: preparing NetCDF forcing and static inputs with exact units, configuring a 20-section settings file, and recovering from silent failures like wrong precipitation conversion factors or Kelvin/Celsius mismatches that produce plausible-looking but wrong discharge. This Skill packages that expertise so an agent can execute real CWatM runs correctly. ## Core Features & Use Cases - End-to-end pipeline tooling: Seven tools cover basin static-stack construction from MERIT-Hydro/ESA-CCI-LC, HWSD/SoilGrids soil conversion, CMFD/ERA5/MSWX forcing conversion, model execution, and output parsing. - Diagnostic recovery: 28 symptom-to-diagnosis-to-remedy triplets in diagnostics/triplets.yaml map known failure modes (unit traps, routing library load failures, spin-up errors) to validated fixes. - Validation against observations: A DAG defines ranked outputs (discharge, TWS, ET, snow) with observability metadata, and a verifier script scores simulated discharge against gauge records using NSE, KGE, and PBIAS with crop_correct calibration. - Use Case: Simulate daily discharge for the Huai River at Bengbu by building the static grid, converting CMFD forcing, running CWatM with a 3-year spin-up, and scoring against observed gauge data. ## Quick Start Run python preflight_check.py in this directory to verify the CWatM binary, Python environment, and required data are available before starting any simulation.