UC 7: Observations reported and exchanged involve multiple sensors

[joergklausen]

multi-sensor configuration. Some parameters like snow on ground (SOG), temperature, wind (on marine platforms), pressure (on marine platforms) have multi-sensor setup (multiple instruments measuring the same parameter) for spatial representation and/or redundancy. In this case, you have 2 or more unique instruments (serial number) all measuring the same parameter from which an official value is selected for products.

hydrological stations with 2 instruments (different method) measuring water level. Official data series are combined using measurements from both instruments.

[gaochen-larc]

AGAGE presents a similar situation. The project offers a "recommended" file type containing data combined from multiple instruments (which may be of different types). According to the NASA AGAGE documentation, "Where a compound has been measured on multiple instruments at a site, the records have been combined into a single timeseries, with the timings of transitions between instruments decided by the AGAGE team." Data provenance is tracked via an instrument_type variable, which links each data point to its corresponding instrument within the time series.

[gaochen-larc]

Gemini assessment based on the case that the same parameter measured by multiple instruments/sensors, which is a common practice for critical or developing measurements.

Gemini comparison summary: This use case demonstrates how WMDR 2 bridges the gap between physical reality (what hardware is bolted to the platform) and logical data products (the official number published to users). While wmdr2-bundled.json only gives you a flat equipment inventory, WMDR 2 provides the structural logic to explain how distinct sensors, serial numbers, and physical methods come together to create a single official dataset.

Multi-Sensor Challenge	Bundled GeoJSON Schema (wmdr2-bundled.json)	WMDR 2 Domain Model (WMDR2.pdf)
Physical Serial Number Tracking	Supported (flat list in instruments array).	Supported (linked via Instrument to specific ObservingConfigurations).
Multi-Method Provenance (e.g., Radar + Pressure)	Conflates methods; cannot explain how two different techniques feed one series.	Maps 1 ObservationSeries to multiple ObservingConfigurations with distinct methods.
Designation of "Official" Product	Cannot distinguish raw backup streams from the selected official product.	Explicitly assigns officialStatus (e.g., primary, backup, official) per stream.
Spatial Array / Redundancy Grouping	No native grouping; array sensors sit in a flat list with all other station gear.	Groups physical sensors natively using instrumentSet.
Blending Algorithm Context	Forces selection rules and processing history out-of-band or into data file headers.	Captured in metadata via ReportingProcedure.levelOfData and ObservingProcedure.

[gaochen-larc]

Gemini assessment based on AGAGE case that the parameter data selected from multiple instruments/sensors for each data point. This is a common practice to ensure highest data quality.

Gemini assessment summary:
The Operational Requirement: International networks like AGAGE stitch decades of observations across multiple instrument generations (e.g., ALE -> GAGE -> Medusa) into a single "Recommended" time series, using an internal variable (like instrument_type) to track row-by-row provenance.

wmdr2-bundled.json (Passive Inventory): Lists the physical instruments statically, but cannot represent the stitched "Recommended" stream in the discovery metadata layer. It forces 100% of the composite provenance onto the data file itself, meaning users cannot discover how the series was constructed without downloading and opening the file first (failing FAIR Principle A2).

WMDR 2 UML Model (Relational Pipeline): Decouples the composite product (ObservationSeries) from individual hardware setups (ObservingConfiguration). The instrument_type variable in the NetCDF file simply acts as an operational pointer to pre-declared metadata objects, keeping the full multi-instrument history and processing tier fully discoverable prior to data download.

AGAGE "Recommended" File Capability	Bundled GeoJSON (wmdr2-bundled.json)	WMDR 2 Domain Model (WMDR2.pdf)
Distinguishing "Recommended" vs. "Individual" Files	Cannot distinguish continuous combined products from raw individual streams.	Explicitly separates raw streams from composite streams via distinct ObservationSeries objects.
Point-by-Point instrument_type Linkage	NetCDF instrument_type points to a flat equipment list without stream context.	NetCDF instrument_type maps directly to explicit ObservingConfiguration entities in metadata.
Transition Strategy Provenance	Unrecorded in metadata; transition dates exist only in text release notes or file headers.	Captured structurally via ObservingProcedure.strategy and time-bounded ObservingConfigurations.
FAIR Principle A2 (Independent Access)	Fails: Requires opening the NetCDF file to know which instruments built the recommended series.	Passes: Metadata independently exposes the full multi-instrument timeline before data download.