<p>Photolysis of molecular oxygen (O<sub>2</sub>) maintains the stratospheric ozone layer, protecting living organisms on Earth by absorbing harmful ultraviolet radiation. The atmospheric oxygen level has not always been constant, and has been held responsible for species extinctions via a thinning of the ozone layer in the past. On paleo-climate timescales, it ranged between 10 and 35% depending on the level of photosynthetic activity of plants and oceans. Previous estimates, however, showed highly uncertain ozone (O<sub>3</sub>) <sub></sub>responses to atmospheric O<sub>2</sub> changes, including monotonic positive or negative correlations, or displaying a maximum in O<sub>3 </sub>column around a certain oxygen level. Motivated by these discrepancies we reviewed how the ozone layer responds to atmospheric oxygen changes by means of a state-of-the-art chemistry-climate model (CCM). We used the CCM SOCOL-AERv2 to assess the ozone layer sensitivity to past and potential future concentrations of atmospheric oxygen varying from 5 to 40 %. Our findings are at odds with previous studies: we find that the current mixing ratio of O<sub>2</sub>, 21 %, indeed maximizes the O<sub>3</sub> layer thickness and, thus, represents an optimal state for life on Earth. In the model, any alteration in atmospheric oxygen would result globally in less total column ozone and, therefore, more UV reaching the troposphere. Total ozone column in low-latitude regions is less sensitive to the changes, because of the “self-healing” effect, i.e. more UV entering lower levels, where O<sub>2</sub> photolyzes, can partly compensate the O<sub>3</sub> lack higher up. Mid- and high-latitudes, however, are characterized by ±20 DU ozone hemispheric redistributions even for small (±5 %) variations in O<sub>2</sub> content. Additional regional patterns result from the hemispheric asymmetry of meridional transport pathways via the Brewer-Dobson circulation (BDC). We will discuss the different ozone responses resulting from changes in the BDC. These effects are further modulated by the influence of ozone on stratospheric temperatures and thus on the BDC. Lower O<sub>2 </sub>cases result in a deceleration of the BDC. This renders the relation between ozone and molecular oxygen changes non-linear on both global and regional scales.</p>

Photolysis of molecular oxygen (O2) sustains the stratospheric ozone layer and is thereby protecting living organisms on Earth by absorbing harmful ultraviolet radiation. In the past, atmospheric O2 levels have not been constant, and their variations are thought to be responsible for the extinction of species due to the thinning of the ozone layer. Over the Phanerozoic Eon (last ~500 Mio years), the O2 volume mixing ratio ranged between 10% and 35% depending on the level of photosynthetic activity of plants and oceans. Previous estimates, however, showed ambiguous ozone (O3) responses to atmospheric O2 changes, such as monotonically positive or negative correlations, or displaying a maximum in the O3 column around a certain O2 level. Here, we assess the ozone layer sensitivity to atmospheric O2 varying between 5% and 40% with a state-of-the-art chemistry-climate model (CCM). Our findings show that the O3 layer thickness maximizes around the current mixing ratio of O2, 21%, under the contemporary boundary conditions. Lower or higher levels of O2 result globally in a reduction of total column O3 and, therefore, allow more harmful ultraviolet radiation (UV) to reach the surface. At low latitudes, the total column O3 is less sensitive to O2 changes, because of the known “self-healing” effect. In mid and high latitudes, however, there is an equator-to-pole variation between high and low O2 cases. Particularly, polar cap O3 is more sensitive to O2 with changes up to 20 DU even for small O2 perturbations. These variations are modulated by the radiative impact of O3 on stratospheric temperatures and on the strength of the Brewer-Dobson circulation (BDC), indicating chemistry-radiation-transport feedback. High O2 cases result in an acceleration of the BDC, and vice versa. This renders the relationship between O3 and molecular O2 to be non-linear on both global and regional scales, however still maximizing global mean total O3 column at contemporary O2 levels.