Telling effects apart
Objective
Combine independent kinds of evidence to separate effects that any single source leaves entangled, such as sensor degradation and a change in the atmosphere.
Observed in real data Current focus See the results How stages are defined
Separating sensor and surface explanations for the one remaining site where the current test could not tell them apart.
Slow radiometric degradation of a sensor and a change in the atmosphere can produce the same shift in a single image. From one scene the two cannot be distinguished, and a model that fits the data well can still attribute the shift to the wrong cause.
Unless the two are told apart, factoring out one effect can quietly introduce another. Long time series, cross-sensor comparison and change detection all depend on getting that attribution right.
Each kind of evidence constrains a different part of the problem. Lunar observations isolate the instrument, because the Moon is a stable reference seen with no atmosphere in between. Independent atmospheric measurements constrain the atmospheric state. Changes in sun and view angle separate geometric effects, and other sensors at the same wavelength expose differences between instruments. We establish what each source reveals on its own, then what becomes separable when they are combined, and we report what remains correlated.
Each effect RefCal attributes is supported by evidence that no competing explanation fits equally well, with every assumption stated.
Each result is dated and published on its own page with its data, conditions and limits. Results that went against the objective are listed here with the rest.
-
Telling apart what caused a measurement offset, rechecked on a like-for-like basis
Correcting the reference basis did not overturn the earlier site-by-site findings, and the new evidence narrows them further. Viewing geometry is now seen at more than one site, though the result is still preliminary. The sensor and the atmosphere are ruled out as the cause of the largest remaining offset at RVUS. What is left is a smaller, sharper question: at RVUS, is the remaining difference in the reference measurement itself or in the area being compared, and at GSCN, could the atmospheric state also play a part? That is the next step in this research area.
Full result, data and credits -> -
Telling apart what caused a measurement offset, site by site
RefCal can now rule specific causes in or out at specific sites using real data, rather than treating every disagreement as one undifferentiated "the numbers are off" problem. That is real progress: it is the difference between not knowing why a reading is off and knowing, site by site, which explanations are ruled out. It also means no single treatment that applies everywhere is justified yet. Some sites still have open questions, and settling them is the next step in this research area.
Full result, data and credits ->
Established work this area builds on. Results from the programme are published separately, with their data and limits.
- Lunar calibration U.S. Geological Survey Why the Moon is a stable reference that a sensor sees without an atmosphere in between.
- Learning about physical parameters: the importance of model discrepancy (2014) J. Brynjarsdottir and A. O'Hagan Why fitting data well is not the same as identifying the physical causes behind it.
Questions about this work?
If you work on this problem, or hold reference data that bears on it, contact us.