A Covariance-Aware Framework for Spatially Resolved Exoplanet Biosignature Inference with the Solar Gravitational Lens

Abstract

Assessing possible life on an exoplanet requires spatial, spectral, temporal, and environmental context rather than a threshold detection of one molecule or surface feature. We develop a covariance-aware Solar Gravitational Lens (SGL) framework in which the data product is a time-tagged Stokes spectral cube reconstructed from wavelength-dependent Einstein-ring measurements. The demonstrated calculation is a 0.45-2.40 um Stokes-I reflected-light simulation of an Earth-radius planet at 30 pc, observed from 650 AU with a (128 x 128) raster, 128 simultaneous spectral channels, and $R\simeq70$. A separate 0.40-20 um architecture-level calculation tracks reflected and thermal planet photons, SGL gain, solar-corona noise, instrumental backgrounds, throughput, dwell time, and reconstruction covariance. In the controlled population audit, structural forward-model mismatch preserves the block ordering gas > surface > cloud/path > mineral > calibration/SGL while reducing the combined conditional information gain to 0.83 of the matched-model value. A reconstruction-covariance bracket reduces an (8 x 8) regional coadd gain from 7.77 to 3.00, implying a 6.7-fold dwell penalty. The feasibility results are design scalings, not a mission verdict: imaging and low-resolution mapping are earlier objectives, whereas full regional spectroscopy requires simultaneous acquisition, sub-ppm effective coronal calibration, measured reconstruction covariance, and branch-specific radiometric validation. We show that the SGL offers a uniquely powerful path to surface-resolved mapping, regional spectroscopy, thermal-climate diagnostics, and co-location tests, providing spatial, spectral, temporal, and environmental context that could strengthen assessments of habitability and possible biological activity beyond disk-integrated precursor observations.