Feasibility of High-Spatial-Resolution Nighttime Near-IR Imaging of Venus' Surface from a Just-Below-Cloud Platform: A Radiative Transfer Study Accounting for the Potential of Haze

In contrast with recent claims of $\sim$10~km ``scattering footprints'' from atmospheric blurring, our numerical radiative transfer modeling shows that sharp ($\sim$10~m) images of the Venus surface can be achieved at night in spectral windows free of CO$_2$ absorption found between 1.0 and 1.2~$\mu$m using a camera at 47~km altitude, just below the planet's optically thick clouds. This is in spite of the Rayleigh scattering by the dense but still semi-transparent lower atmosphere, and the potential for significant underlying hazes beneath the clouds. The thermal radiation transmitted directly to the camera forms images of spatially varying surface emissivity and/or temperature at the native sensor resolution, platform stability permitting and under reasonable seeing conditions. Quasi-isotropic Rayleigh scattering dominates in the 1.0~$\mu$m window. Combined with near-Lambertian reflections off the base of the cloud layer, the diffuse light field builds up a background radiance from surface emission averaged spatially out to several 10s of km, i.e., beyond the camera's field-of-view. At the longer wavelengths (1.1 and 1.18~$\mu$m windows), the sub-cloud atmosphere itself partially absorbs (hence less direct light), and therefore weakly emits (hence more background light), but the rapidly decreasing Rayleigh scattering compensates and contrast is maintained. In all cases, we demonstrate that the directly-transmitted surface emission component encapsulated within the native sensor resolution element ($\sim$10~m) is a significant fraction of the total radiance, and thus can be detected above the background light. Extending down to the 0.85 and 0.90~$\mu$m spectral windows, there is less direct and more background due to the enhanced Rayleigh scattering, but the resulting reduction in contrast can be mitigated by co-adding the $\sim$10~m pixels.