A voyage of discovery: Exploring the atmospheres of solar system planets and exoplanets with NEMESIS

To extract, or 'retrieve' atmospheric properties from the observed radiance spectra from a planetary atmosphere requires software that can generate the expected radiances from a guessed atmospheric model, compare the radiances with those measured, determine how the model should be updated to reduce any discrepancy between the modelled and observed radiances, and then iterate these steps until these differences are minimised. One such retrieval model is NEMESIS (Nonlinear optimal Estimator for MultivariatE Spectral analySIS), which was initially developed by myself and my colleagues in the 1990s, and which has since been continually updated and enhanced. NEMESIS has now been used in more than 300 papers retrieving atmospheric properties from observed thermal and solar-reflected radiance spectra from all the planetary atmospheres in our solar system and also some beyond. NEMESIS uses the Optimal Estimation framework for atmospheric retrievals and is written in FORTRAN. Recently, more Bayesian frameworks have become computationally possible and favoured, especially for exoplanetary retrievals where prior constraints are almost entirely absent. Hence, NEMESIS has recently been updated to Python (ArchNEMESIS), and combined with PyMultiNest to allow nested sampling retrievals that can better explore the degeneracy between different atmospheric properties. I will review how NEMESIS retrievals have improved our understanding of planetary atmospheres over the last 30 years and how the development of ArchNEMESIS has breathed new life into the NEMESIS/ArchNEMESIS project.