Impact of deep convection on the tropical tropopause layer composition in Equatorial Brazil

Abstract. This paper documents measurements of carbon monoxide (CO), ozone (O₃) and temperature in the tropical tropopause layer over Equatorial Brazil for the first time. These measurements were sampled by the balloon-borne instrument SPIRALE (Spectroscopie Infa-Rouge par Absorption de Lasers Embarqués) in June 2005 and in June 2008, both at the transition period from wet to dry season. The height of the Tropical Tropopause Layer (TTL) top and bottom determined from the chemical species profiles are similar for the two flights. Nevertheless the measured profiles of ozone and CO are different in their volume mixing ratio and shape. The larger CO values measured in the TTL in 2005 can be linked to a more intense biomass burning activity in 2005 than in 2008. We also show that both measured profiles are influenced by convection but in different ways leading to different shapes. The CO profile in 2005 is characterised by a generally smooth decrease in the TTL from tropospheric to stratospheric conditions, except for two layers of enhanced CO around 14.2 (>100 parts per billion by volume = ppbv) and 16.3 km altitude (>85 ppbv). Backward trajectories indicate that these layers come from the vertical transport by remote deep convection occurring 2 and 3 days prior to the flight, respectively. This shows that the transition period from wet to dry season is favourable for the transport of significant amounts of CO in the TTL, sometimes above the level of zero radiative heating, because of increasing biomass burning together with decaying but still important convective activity. In 2008 we focus our analysis on a 1 km deep layer, between 17 and 18 km, where both the temperature and the ozone profiles are uniform in the vertical, corresponding to a layer of well-mixed air. We show that this unusual behaviour is indirectly related to the interaction between convection and the Quasi-Biennial Oscillation (QBO), through vertically propagating gravity waves. Quasi-stationary gravity waves are likely to be produced by convective systems and certainly break in the intense wind shear that imposes the QBO at these altitudes. This conclusion is supported by the fact that the 16–18 km layer is devoid of ice particles (hence the mixing is not convective) and from backward trajectories that point towards a convective region as the origin of the air masses in this layer.