What problem does it solve? Running the GR4J hydrological model correctly requires operational expertise scattered across R documentation, Fortran source, and practitioner knowledge: exact unit conventions (mm/day fluxes, Celsius temperatures), warm-up requirements, calibration in transformed parameter space, and recovery from silent failures like Kelvin-to-Celsius errors that produce plausible-looking but wrong discharge. ## Core Features & Use Cases - End-to-end simulation pipeline: Four validated Python tools convert CMFD/ERA5/Caravan forcing to airGR format, extract catchment parameters, execute GR4J via rpy2 with warmup and calibration, and parse outputs with NSE/KGE/PBIAS metrics. - Diagnostic recovery: 18 symptom-diagnosis-remedy triplets cover known failure modes such as unit mismatches, missing warm-up, NA forcing values, and parameter clipping. - Snow coupling: Optional CemaNeige snow module with elevation-band hypsometry for cold-region catchments where plain GR4J fails. - Use Case: Calibrate GR4J against a HYDAT or GRDC gauged basin using MSWX or Caravan forcing, then validate simulated discharge against observed streamflow with cited performance thresholds. ## Quick Start Run the preflight check with python preflight_check.py, then ask the agent to calibrate GR4J on your catchment forcing CSV and report NSE and KGE against observed discharge.
