Fig. 1. Summertime (June–August) NO_x (pptv) at tropopause levels. Top: in situ aircraft measurements from the NOXAR project for the year 1995. Data compiled by Grewe et al. (2001). Bottom: E39/C model data at 250 hPa from a 1990 time-slice simulation.

Fig. 2. Schematic diagram: zonal mean cross-section of the interaction of tropical lightning and mid-latitude ozone changes, and the method used to investigate this interaction. The transport of 10 000 air parcels, which were released in the region of tropical lightning (and thus of tropical lightning induced NO_x emissions) in the upper troposphere, has been simulated. 500 of these air parcels are transported into the LMS, where additional ozone is produced from the lightning induced NO_x. Due to the longer lifetime of O₃ in the LMS, the described transport pathway in combination with the chemistry along this pathway contributes significantly to the mid-latitude ozone concentration (for details see text).

Fig. 3. Composites of trajectories which start in the aircraft and lightning source regions AIR (top panels) and LIG (bottom panels) and end in the northern (left panels) and southern (right panels) LMS. All trajectories pass through the tropical (20°S–20°N) troposphere (air pressure more than 400 hPa). Mean trajectories are calculated for each path and afterwards combined into a single trajectory. Mean deviations of this trajectory are also shown (dashed). The trajectories are smoothed for clarity. The composites include 70%, 100%, 71%, and 67% of all trajectories from the source region AIR–NLMS, AIR–SLMS, LIG–NLMS, and LIG–SLMS, respectively (see Table 1). A, B, and C mark sections of the trajectory which are referred to in the text.

Fig. 4. Annual mean relative changes (%) of NO_x (left) and ozone (right) due to lightning NO_x emissions. Results are shown for the E39/C models with an upper boundary for NO_y at 10 hPa (top) and 100 hPa (middle) and for the GISS model (bottom) with an upper boundary at 110 hPa. Changes are calculated relative to a simulation without lightning emissions. All E39/C results are significant at a 99% level. In the GISS model, the light and dark shading indicates the 95% and 99% significant levels (t-test). The tropical changes of NO_x due to lightning in the two models with a low upper boundary appear quite different due to the greater abundance of tropical NO_x from non-lightning sources in the GISS model, though in fact they have similar absolute NO_x changes.

Fig. 5. Contribution (%) of the tropical transport pathway (stratosphere–troposphere-coupling, STC, see Fig. 2) to lightning induced ozone and NO_x changes at 150 and 200 hPa. This has been calculated from the difference between the E39/C simulations with the 100 hPa upper boundary and the E39/C standard version (10 hPa boundary). See text for further details.

Fig. 6. Temporal development of various quantities along the mean lightning trajectory ending in the northern LMS. (a) air pressure (hPa); (b) latitude (°); (c) lightning induced NO_x change (pptv); (d) lightning induced change of ozone production (solid), destruction (long-dashed), net-production (short dashed), and ozone destruction in the control simulation (dotted) (pptv/h); (e) ozone lifetime (days); (f) lightning induced ozone production, which is relevant for the lightning induced ozone change at the end of the trajectory (pptv/h); (g) accumulated lightning induced ozone change (ppbv), which is relevant for the lightning induced ozone change at the end of the trajectory. For details see also Appendix A.

Table 1. Share of particles (%), its density (air parcels per air mass), and share of mass within the target regions (%), which are transported from the source region AIR (150–300 hPa, 20°N–50°N) and LIG (150–300 hPa, 20°S–20°N) to the stratosphere ‘S' (air pressure<150 hPa), northern LMS ‘NLMS' (150–250 hPa, 20°N–90°N), southern LMS ‘SLMS' (150–250 hPa, 20°S–90°S), and the troposphere ‘T'