Tropical middle atmosphere ozone response to the solar rotational cycle in observations and chemistry-climate simulations
Résumé
Solar spectral irradiance fluctuations associated with the Sun's 27-day rotational cycle can significantly influence ozone variability in the tropical middle atmosphere. In the stratosphere, previous observational studies based on different instruments and time periods revealed however large disparities in the magnitude of the ozone response to the 27-day solar signal. For the mesosphere, only few observational estimates of the ozone response to the 27-day solar cycle are available. This is partly due to the relative paucity of ozone measurements and the considerable amplitude of the ozone diurnal variations at these altitudes which both make the detection of solar signals in ozone difficult.The results presented here summarize the main findings of two studies that used satellite ozone data (UARS-MLS, Aura-MLS and ENVISAT-GOMOS) and chemistry-climate simulations of LMDz-REPROBUS and HAMMONIA models in order (i) to understand the discrepancies in previous estimates of the stratospheric ozone solar rotational signal and (ii) to quantify the mesosphere/lower thermosphere ozone response to the solar rotational cycle. Observational ozone response to solar irradiance variations are identified and quantified from linear correlation and regression methods combined with non-parametric statistical tests. Observational results are then compared with the results from chemistry-climate model simulations which are driven with realistic and/or idealized solar irradiance forcings.The analysis of stratosphere MLS ozone data and their comparison with LMDz-Reprobus results show that the dynamical variability modulates the ozone solar rotational signal, which partly explain the discrepancies found in previous observational studies. This dynamical variability masking effect can be overcome by considering long timeseries; we found that a minimum of a 3 year time window is needed for the 1_ uncertainty estimate of the ozone solar rotational signal to drop below 50%. The analysis of mesosphere/lower thermosphere GOMOS data allowed us to derive the first observation-based night-time ozone response to the solar rotational cycle in the range [50,110] km. The vertical distribution of this ozone response exhibits a maximum at 80 km and an abrupt increase above 100 km. Although qualitatively consistent with GOMOS observations, the results from HAMMONIA simulations significantly underestimates the magnitude of the ozone response. Possible causes for these discrepancies will be discussed.