Extended Data Fig. 8: Comparison of NIRISS SOSS to NIRSpec PRISM before and after the estimated dilution correction.

Shaded regions indicate the wavelength ranges that experience saturation after the number of groups indicated in bold. A dilution correction cannot be computed for the region that experiences saturation after one group (dotted lines); however, we show an example correction using the correction value for the regions experiencing saturation after two groups. The probability density of the residuals is displayed in the right panel, alongside a Gaussian distribution with zero mean and a standard deviation equal to the median error on the residual (gray shading). We note that the downwards peak around 1.3 micron is likely due to this spectral region having the lowest resolution across the instrument, resulting in the highest counts in pixels at these wavelengths such that these detector columns get closer to saturation within the first group. Due to insufficient non-linearity correction within the JWST pipeline and the first-group effect, lower-than-expected flux is measured in this region, resulting in deeper transit events.