At this stage, we will provide 4 non-blind (known ground truth) forward models and 4 blind models of similar composition.
This stage is mainly for model comparison and de-bugging.
This stage will see fully blind retrievals only with forward models of varying complexities. The forward models will reflect a range of chemical and cloud parameterisations and sources from a range of atmospheric forward models. Each participant will receive a score (recorded on a leaderboard) based on the accuracies of the retrieved parameters as well as forward model fit.