Settings
Prototype configuration. These controls change live application state.
Data
Where forecasts and observations come from.
Dataset
Deterministic seeded prototype dataset
Data refresh interval
Observation source
Forecast sources
BharatFS · NCUM · GFS · GEFS · AI Forecast Proxy
Blending
How adaptive weights are produced.
Weighting mode
Rolling skill window
90 days
Minimum model weight
5% floor per source
Uncertainty mode
Display
Defaults applied across the application.
Map layer
Default location
Default forecast horizon
System
Prototype identity and mode.
Prototype version
Problem statement
Demo mode
Keeps the deterministic dataset and the demo indicator visible
Adaptive (contextual skill)
90 days
Settings adjust live prototype state. In deployment these map to ingestion schedules, verification windows and blend configuration held server-side.