MOSAIC (Model Observation System for Assessment, Improvement, and Calibration)

What MOSAIC is

I build frameworks like this to turn ad hoc, one-off comparison scripts into a repeatable process. MOSAIC automates pairing ocean model output with observational data to compute standard skill metrics, flag periods and regions of poor agreement, and generate calibration diagnostics.

Compatible data sources

• Buoys
• Tide/water-level gauges
• Satellite
• ADCP

What MOSAIC can do

• Automated pairing of model output with observational data across stations and time periods
• Computes standard skill metrics
• Flags periods/regions of poor model-observation agreement
• Generates calibration diagnostics to guide parameter tuning

Screenshots

In Practice

See the Model Validation and Calibration with ML case study for how this framework was applied to cut a client’s calibration cycle from weeks to days.

Thumbnail: NOAA/National Data Buoy Center discus buoy — public domain (NOAA), via Wikimedia Commons.