What we did
RefCal's first research area asks whether effects that shape a measurement show up in real evidence, and whether the conditions under which they were measured are recorded. An earlier entry answered this partly from comparisons that matched surface readings against top-of-atmosphere reference values, a mismatch found and fixed in September. This entry rechecks it with every comparison made on a matching basis.
Three checks bound what this entry can claim:
- Viewing geometry: real sun and view angles for each acquisition, taken from Sentinel-2 product metadata at the site itself, were compared with the offset from reference values at six reference sites, on matching top-of-atmosphere values on both sides.
- Sensor spectral response: a real, independently measured reference reflectance spectrum was passed through RefCal's Sentinel-2A and Landsat 9 sensor-response models, both built from the sensors' own official calibration data.
- Atmospheric aerosol: a comparison against 87 real AERONET measurements was examined, then excluded as RefCal retrieval evidence when follow-up showed the SIAC output largely reproduced its CAMS prior.
What we saw
Viewing geometry is a real effect at more than one site. Where it appears, its sign is the same, and one shared strength describes it across sites. In the near-infrared, the step between the two viewing directions is about seven percent at RVUS and about ten percent at LCFR; BSCN shows the same sign, too weakly to be conclusive. At LCFR, geometry accounts for the site's offset; at RVUS, it accounts for the step but only part of the overall offset. The LCFR step is still present when surface readings are compared with measured surface truth. The part of the atmospheric path that depends on viewing angle is too small to explain the offsets at RVUS and GSCN.
The sensor spectral-response check passed: the same real reference spectrum produced consistent, physically sensible outputs from both sensor models.
The atmospheric comparison did not provide independent evidence that RefCal retrieved aerosol correctly, so this entry makes no atmospheric-retrieval claim from those values.
What it means
Viewing geometry is now seen at more than one site on a like-for-like basis, though the result is still preliminary, and the sensor-response models handle a real measurement consistently. Which cause explains each site's remaining offset is the subject of the P02 research area. The atmospheric comparison still marks a boundary: a passed-through prior cannot be treated as evidence that RefCal retrieved the atmosphere independently.