Correction to: Nature Astronomy https://doi.org/10.1038/s41550-023-02016-7, published online 6 July 2023.

In the version of the article originally published, all column density values were a factor of 10 too low. Consequently, “two orders” has been changed to “one order” in the abstract and the text, and the values have been corrected on the y axes of Fig. 3 and the color scales in Fig. 4. In addition, in the section “The spatial distribution of CO and HCN”, the CO and HCN masses reported in the text were a factor of 10 too low and the loss factors for CO and HCN were a factor of 10 too high; they have been corrected accordingly. Specifically, CO is now shown to have a meridionally uniform column density of 1.86 ± 0.52 × 1016 cm–2 and the total mass of SL9-derived CO is 5.47 ± 0.26 × 1014 g. This corresponds to a loss factor of 0.9 ± 0.3 since 1995–1998. HCN has a uniform column density of 22.6 ± 5.7 × 1013 cm–2, which is one order of magnitude lower than that measured 6.5 years after the SL9 impacts. The total mass of HCN is 5.0 ± 0.1 × 1012 g. This corresponds to a loss factor of 5.0 ± 3.0 when compared with the 1995–1998 period, or 12.0 ± 3.5 when compared with the value derived in 2000 from the Cassini flyby data. The conclusions of the paper remain unchanged.