A Lagrangian perspective of the tropopause and the ventilation of the lowermost stratosphere

Back trajectories driven by large-scale analyzed wind fields are used to investigate troposphere to stratosphere transport (TST) in the Northern Hemisphere tropopause region, as well as the surface sources for such transport, defined in terms of the locations where each trajectory last left the atmospheric boundary layer ( log pressure height z* < 1 km). The proportion chi(BL) of those trajectories arriving in the tropopause region that have visited the boundary layer within the previous fixed time period ( typically 30 days) is determined as a function of each trajectory's final equivalent latitude and potential temperature. For a range of potential temperature surfaces ( similar to 300-380 K), chi(BL) is shown to have a sharp gradient in the extratropics indicative of a partial permeable barrier to transport that can be identified as a "Lagrangian tropopause.'' Variations in chi(BL) with equivalent latitude and potential temperature and on seasonal timescales are shown to provide a novel measure for the relative location and permeability of the tropopause barrier. Details such as the presence of a "ventilated layer'' in the northern summer extratropical lower stratosphere ( similar to 370-410 K) are clearly apparent. Directly below this layer ( 340-370 K) the tropopause barrier to transport is shown to be relatively strong, whereas below 340 K it is again more permeable. A distinction can therefore be made between "extratropical'' TST that primarily ventilates the lowermost stratosphere below 340 K and "tropical'' TST that occurs into the 370-410 K ventilated layer. The boundary layer source regions for extratropical TST are shown to correspond to those regions previously identified as sources for deep frontal uplift in the warm conveyor belt circulations of extratropical cyclones, although elevated regions such as the Himalayan plateau are also seen to be important. Tropical TST has different source regions associated with regions of active deep convection such as the western tropical Pacific and, in the northern summer, the Indian subcontinent. The source regions are, in general, found to be geographically localized, leading to the conclusion that subject to the limitations of the methodology, trace gas emissions in specific regions are substantially more likely to be transported to the lowermost stratosphere than elsewhere. The implications for the assessment of ozone depletion by very short lived halogenated species are mentioned.