Fig. 3: Interpretation of the constituents of the NIRISS WASP-39b transmission spectrum.

a,b, Panel a shows the comparison of the transmission spectrum of WASP-39b from the nirHiss reduction (grey points) with respect to the best-fit reference model (black line). This model assumes an atmospheric metallicity of M/H = 1.38 (23 times the solar value), C/O ratio of 0.2 (0.55 times the solar value²⁰), K/O ratio of 0.1 (1.26 times the solar value), full day–night heat redistribution (f = 1) and flux-balanced clouds with inhomogeneous terminator coverage. Each coloured line removes a key constituent found in our best-fit reference model to demonstrate how the spectrum would change were these features not included. The removal of clouds and H₂O absorption from the reference model result in large-scale changes to the shape and depth of the transmission spectrum. Other sources of opacity with an impact on the spectrum are K, CO and CO₂. Residuals between the data and the reference model are plotted in panel b. c,d, These two panels show the molecular absorption cross-sections for a selection of gases observable within the NIRISS bandpass. Panel c highlights gases inferred by our analysis of the spectrum of WASP-39b. Panel d highlights some gases that were not identified in these data but may be present in future observations of other exoplanets. We define our errors as the 1σ uncertainties extracted from the 16th and 84th percentiles of the transit depths fit from each pipeline. (https://github.com/afeinstein20/wasp39b_niriss_paper/blob/main/scripts/figure3.py).