Importance of fitting more than a single global atmospheric metallicity in chemical equilibrium retrievals when probing spectral features affected by disequilibrium processes such as cold-trapping. Top: brightness temperature NIRISS spectrum (black points) overplot-ted with the best fit model from chemical equilibrium retrieval analyses only fitting for a single overall metallicity (pink), or fitting for volatiles, refractories, and titanium separately (green). The inset panel shows a zoom in on the 0.6–1.1 µm region, with the data now binned to R = 100 for order 1 and R = 50 for order 2 (grey points) for better visualisation. Bottom left: retrieved TP profile in both cases (with A_g fitted) showing that the single-metallicity case prefers an inversion much deeper in the atmosphere. Bottom right: recovered abundances highlighting the importance of treating titanium as a separate parameter relative to other species in chemical equilibrium retrievals. Even when allowed to vary independently within a retrieval (‘Chem. multi’, green), the volatile ([𝒱/H], solid) and refractory ([ℛ/H], dashed) metallicities are recovered at similar enriched values while [Ti/H] (dotted) is significantly depleted in comparison. Enforcing all species to vary according to a single global metallicity (‘Chem. single’, [M/H], pink) and only allowing the C/O ratio to vary can result in some biased average of the depleted [Ti/H] and super-solar other metals to be inferred.