Is there a stratospheric fountain?

Abstract. The impact of convection on the thermal structure of the Tropical Tropopause Layer (TTL) was investigated from a series of four daily radiosonde ascents and weather S-band radar observations carried out during the HIBISCUS campaign in the South Atlantic Convergence Zone in Southeast Brazil in February 2004. The temperature profiles display a large impact of convective activity on the thermal structure of the TTL. Compared to non-active periods, convection is observed to result in a cooling of 4.5°C to 7.5°C at the Lapse Rate Tropopause at 16 km, propagating up to 19 km or 440 K potential temperature levels in the stratosphere in most intense convective cases. Consistent with the diurnal cycle of echo top heights seen by a S-band weather radar, a systematic temperature diurnal cycle is observed in the layer, displaying a rapid cooling of 3.5°C on average (–9°, –2°C extremes) during the development phase of convection in the early afternoon during the most active period. Since the cooling occurs during daytime within a timescale of 6-h, its maximum amplitude is at the altitude of the Cold Point Tropopause at 390 K and temperature fluctuations associated to gravity waves do not display significant diurnal change, the afternoon cooling of the TTL cannot be attributed to radiation, advection, gravity waves or adiabatic lofting. It implies a fast insertion of adiabatically cooled air parcels by overshooting turrets followed by mixing with the warmer environment. During most intense convective days, the overshoot is shown to penetrate the stratosphere up to 450 K potential temperature level. Such fast updraft offers an explanation for the presence of ice crystals, and enhanced water vapour layers observed up to 18–19 km (410–430 K) in the same area by the HIBISCUS balloons and the TROCCINOX Geophysica aircraft, as well as high tropospheric chemical species concentrations in the TTL over land observed from space. Overall, injection of cold air by irreversible mixing of convective overshoots as proposed by Danielsen (1982) do not appear as episodic isolated features, but common and systematic events over a land convective area, that is a Stratospheric Fountain. Though the two-stages process proposed by Sherwood (2000) may also be operative, it offers a mechanism for producing the chemical, moisture and thermal properties observed in the stratosphere. The consistency between convective cooling of the TTL and weather radar echoes heights observed during the summer over South-East Brazil and the TRMM radar OPFs and LIS lightning events maximum frequencies, the latter showing also maximum events over Africa, South-East Asia, the Indonesian Islands and Northern Australia depending on the season (Liu and Zipser, 2005), suggests the existence of several "Stratospheric Fountains" over continents instead of the oceanic area of Micronesia as proposed by Newell and Gould-Stewart (1982), which appears a region of little overshoot.