Impact of Interactive Chemistry of Stratospheric Ozone on Surface Climate in Paleoclimate Simulations and Future Global Warming Projections

A series of numerical experiments on the paleoclimate simulations for the Last Glacial Maximum (LGM; 21 kyr B.P.) and for the mid-Holocene (MH; 6 kyr B.P.) were performed by using an Earth System Model (ESM) of the Meteorological Research Institute (MRI) of the Japan Meteorological Agency, as well as the experiments on future climate projections, including abrupt increase experiment to quadruple CO2 (abrupt 4xCO2), in order to investigate the impact of interactive chemistry of stratospheric ozone on the surface climate. In all paleoclimate simulations in the fifth Coupled Model Intercomparison Project (CMIP5), the ozone distribution was prescribed to the preindustrial period (PI; 1850 C.E.) level, but this assumption may not be appropriate. This is because the stratospheric ozone distribution in such paleoclimates could be modulated by the change in orbital elements of the Earth (i.e., Milankovitch cycle), change in stratospheric temperature, and/or changes in trace gas compositions of the atmosphere, and such changes in the stratospheric ozone could interact with and influence on climates through dynamical, physical, and chemical processes.

Noda et al. (2017, JGR) made interactive ozone chemistry calculations for MH and PI, and compared them with the results of the corresponding experiments in CMIP5, in which the ozone distribution was prescribed to the PI level. The contribution of the interactive ozone chemistry in a quasi-equilibrium state (the last 50 years of 110-year integrations) reveals a significant anomaly of up to +1.7 K in the Antarctic region for the annual mean zonal mean surface air temperature. The influence of the orbital element variations in MH to the stratosphere and troposphere, oceans, and sea-ice in the Antarctic region is affected by the interactive ozone chemistry in the following way:

(1) In austral autumn, the solar insolation anomaly over the Antarctic region is negative due to the change of the perihelion point during MH.

(2) It leads to cooling in the upper stratosphere, as well as a positive ozone anomaly due to a change in photo chemical equilibrium.

(3) Throughout winter, the positive ozone anomaly in the upper stratosphere slowly descends to the lower stratosphere by downward flow inside the polar vortex, because the lifetime of ozone is long during the polar night without solar insolation.

(4) In spring, enhanced heating by the positive ozone anomaly, together with the positive insolation anomaly in MH, produces a positive temperature anomaly.

(5) In association with the springtime positive temperature anomaly, the lower stratosphere exhibits the weakening of the circumpolar westerly jet, and the weaker jet extends down to the troposphere in association with Southern Annular Mode.

(6) In the southern hemisphere ocean, the near-surface westerly jet maintains the equatorward sea surface Ekman current, and the weakening of the westerly jet reduces this current.

(7) The weakening of the equatorward Ekman current suppresses the sea ice transport to lower latitudes, and promotes the sea ice retreat.

(8) The sea ice retreat causes the surface albedo decrease, and accompanied with high rates of the solar insolation that reaches the surface, results in positive surface air temperature and sea surface temperature anomaly in the Antarctic region.

The stratosphere-troposphere coupling processes in spring, (4) positive temperature anomaly and (5) weaker westerly jet, represent similar variations, albeit in opposite sign, to the response of the deepening of the Antarctic ozone hole in recent years. All the MH runs by CMIP5 models with the prescribed ozone had cold bias in sea surface temperature when compared with geological proxy data, whereas the bias is reduced in our simulations by using interactive ozone chemistry.

Noda et al. (2018, JGR accepted) investigated the impact of changes in the stratospheric ozone profile in the LGM simulations under reduced atmospheric CO2 concentrations (185 ppm) and different orbital elements. Comparison of the simulation results with interactive atmospheric chemistry with those using the prescribed ozone profile for the PI condition in CMIP5. The contribution of the interactive chemistry reveals a significant warming of zonal mean surface temperature, +0.5 K (approximately 20 %) in the tropics and up to +1.6 K in high latitudes. In the tropics, this mitigation of global cooling in the LGM climate is related to longwave radiative feedbacks associated with circulation-driven increases in the lower stratospheric ozone and in the stratospheric water vapor, and related decrease in cirrus clouds. The mechanisms are of opposite sign to and consistent with those obtained by increased CO2 simulations by Nowack et al. (2015) using the UK Met Office’s Unified Model coupled to the UK Chemistry and Aerosols model. In high-latitudes, the stronger mitigation of cooling is associated with sea-ice retreat, which has the same sign to and is consistent with our previous paleoclimate simulation of MH including interactive chemistry (Noda et al., 2017). Most previous LGM simulations in CMIP5 with the prescribed ozone profile exhibited cold bias in the tropics compared with geological proxy data, whereas this bias is reduced in our simulations through the use of the interactive ozone chemistry, although a warmer bias in the mid-latitude is enhanced.

An abrupt 4xCO2 experiment with the same ESM of MRI was performed to investigate the impact of interactive chemistry of stratospheric ozone on surface climate in future global warming projections. The result should be compared with previous results by Nowack et al. (2015) and others.

The impact of stratospheric ozone on the surface climate in MH and LGM implies that the stratospheric ozone profile is an important factor for paleoclimate simulations of any periods for which the CO2 concentrations and orbital elements vary substantially. In conclusion, we recommend that paleoclimate simulations which will be conducted in the future should use an ESM that includes interactive sea ice as well as seasonally varying ozone profiles in the stratosphere that are consistent with CO2 concentration and orbital elements, either by performing computations with an ESM with interactive ozone chemistry component or by prescribing an ozone distribution which is computed offline for the given CO2 concentration and orbital elements.