Spectral Bands & FWHM

[mattsymbios]

Spectral bands (1.10) requires the central wavelength for all PFS. For AR it additionally requires the FWHM. Can this be aligned?

[mattsymbios]

In order to promote interoperability between different datasets, we need more information than just the central wavelength of spectral bands.

Proposal to LSI-VC-15: Decide whether we adjust the SR, ST, and other PFS as needed to include at a minimum: – FWHM at Threshold – Full Spectral Response Function at Goal

Needs discussion and decision.

[strobpr]

Sampling characteristics such as spatial, spectral, or temporal response functions are fundamental to describe and analyse any (geo-)spatio-temporal observation. Therefore they must be documented within the metadata of any acquired observation (sample, pixel). However, to include all detail with every observation result is obviously not feasible, and might also not be necessary where:

• groups of observations share a characteristic (e.g. because they have been acquired by the same or very similar sensors, like spectral bands)
• the function can be reduced through a model (e.g. gaussian)
• a simplification is sufficient to deal appropriately with the observation result (e.g. same observation time for a whole scene)
• the detailed description of the function is externalised and only a label or identifier is used

Providing the necessary metadata with large quantities of observations (sets or collections) is therefore a matter of grouping them efficiently (e.g. in grids or cubes where response functions can be attached to dimensions) and by creating human and machine readable repositories where they can be retrieved (like e.g. the projection parameters which are summarised and standardised in EPSG codes). A similar authoritative library for codes representing the full parameters for the most commonly used sensors would be something CEOS could work on.