Work

Clouds, coronal noise and the definition of resolution in solar-gravitational-lens imaging: a linear-Gaussian information map and clouds measured in DSCOVR/EPIC images of Earth

Year: 2026
Type: preprint
Abstract: In a linear-Gaussian Fisher model of solar-gravitational-lens imaging of an Earth-like surface (baseline, fixed cloud parameters), clouds dominate information loss at the largest scales and coronal photon noise at the smallest, while the barrier binding the recoverable resolution (clouds, coronal noise, the solar quadrupole, or none decisively) depends on how resolution is defined (for Gaussian-correlated clouds, because these lower the signal-to-noise ratio below unity only at intermediate scales) and on the threshold. With clouds measured in 2024 DSCOVR/EPIC images of Earth, it is clouds, coronal noise or undetermined depending also on the brightness divisor. These clouds are skewed, heavy-tailed and more correlated at 48 and 72 h than modelled. In 2024, main-configuration reconstructions of EPIC images passed through the imaging model classified below the bound did not remain below it when the bound's cloud covariance was built per daily slot, a variant identified after seeing these results. With this construction fixed in advance, none was classified below the bound in four 2023 periods (96 judgements per period, four divisor pairs); the 00:00 UTC slot had an empty limit band (not compared), and 0, 5, 4 and 7 of the 48 12:00 UTC judgements per period were undetermined. Other 2024 findings held in 2023 under pre-written criteria (the empty 00:00 UTC band only for the per-slot bound), except one holding only when pooled across periods, one applicable in only two periods, one breaking in one period, and one untestable finding. The bound belongs to the declared model, not all methods. We make no statement about whether or when a solar-gravitational-lens mission could image an exoplanet, and we do not extend the cloud properties of the observed periods to other planets or other epochs of Earth.
Author: Maeda, Yukie
Language: English
Open Access status: green
OpenAlex ID: W7225395349
Cites:
Cited by:
Related to: 10