A Covariance-Aware Framework for Spatially Resolved Exoplanet Biosignature Inference with the Solar Gravitational Lens
Year: 2026
Type: preprint
Abstract: Assessing possible life on an exoplanet requires spatial, spectral, temporal, and environmental context. We develop a covariance-aware Solar Gravitational Lens (SGL) framework for regional biosignature inference from reconstructed Einstein-ring measurements. A controlled 0.45-2.40 um Stokes-$I$ experiment uses ten illuminated cells, 128 channels, and R=70 to compare resolved and exactly projected disk spectra with common parameters and priors. Across 96 synthetic scenes, the joint gas, vegetation, and mineral block gains (1.52+/-0.03) nats after marginalizing cloud and calibration/response parameters; adding independently trained surrogate-discrepancy covariance gives (1.49+/-0.03) nats. Held-out mismatch nevertheless produces linearized shifts up to 6.08 posterior standard deviations; preserved precision does not establish unbiased recovery or evidence for life. For fixed Gaussian hypotheses, the expected log-likelihood-ratio loss under disk integration equals half the squared whitened contrast discarded by the projection. A separate 0.40-20 um continuum calculation provides absolute photon rates and observing-time estimates for an Earth-radius planet at 30 pc, observed from 650 AU. It includes source-averaged SGL amplification, finite-aperture extraction, coronal and host-star photons, detector noise, and a declared calibration floor. A reconstruction-covariance bracket reduces an 8 x 8 coadd gain from 7.77 to 3.00, costing 6.7 times the exposure in the statistical-noise limit. The framework connects regional spectroscopy and co-location tests to the physical and instrumental validation needed to assess exoplanet biospheres and possible biological activity.
Source: arXiv (Cornell University)
Author: Turyshev, Slava G.
Institutions: California Institute of Technology, Jet Propulsion Laboratory
Language: English
Subfield: Astronomy and Astrophysics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7166728871