Supplementary Information for: "Simulation of convective moistening of extratropical lower stratosphere using a numerical weather prediction model"
Video SI.1: the animation of 1 km simulation on a chosen cross-section during the convection event from 18:40 to 20:40 UTC, 25 Aug 2013. 1st panel: temperature (colour) and isentropic lines (thin white lines); 2nd panel: ice water content in logarithm scale; 3rd panel: water vapor content in ppmv (colour saturated at 20 ppmv); 4th panel: vertical wind speed (m s-1). Video SI.2: water vapor content in ppmv at the altitude of 16.5 km from 18:00 to 23:00 UTC (5 hours) within the evaluation Domain A. Figure SI.1: the transported water vapor in the simulation domain of 0.25 km (green box in Fig. 1) from 1800 to 19:00 UTC 25 Aug. The vertical advection of 10 km simulation is decomposed into two parts: grid scale advection (light blue bar) and sub-grid scale advection from KFC parameterization (darker blue bar). For the results presented in Fig. SI.1, the evaluation is performed for the simulated domain of 0.25 km grid-spacing (green box in Fig. 1) for the first hour of the evaluation time window (from 18:00 to 19:00 UTC 25 Aug for the three high resolution models, from 21:30 to 22:30 UTC 25 Aug for 10 km simulation). The tropopause levels are defined using the mean laps rate <2˚C km-1. The use of limited time window is due to the fact that the storm moves quickly to the outside of the simulation domain. Figure SI.1 shows the mass transport budget for the 4 simulations with different horizontal resolutions, we found that the direct transport of water vapor of 0.25 km simulation is higher than all the other simulations. Although the contributions from ice sublimation and turbulent mixing of 0.25 km simulation are the lowest. The contribution from ice sublimation is negative which suggests that vapor deposition rate is very high in the convection updrafts and exceeds that of ice sublimation. Although the results presented here should be viewed with caution due to the short evaluation period (1 hour), it corroborates the finding that higher resolution model tends to have a stronger direct vertical transport of water vapor and a weaker contribution from the ice sublimation. Figure SI.2: background: ice water content at the altitude of ~14 km, 1940 UTC 27 Aug; gray line: ER-2 aircraft path in 27 Aug highlighted by red dots for the areas with ice; circles: starting points of back tracking at 19:40 UTC 27 Aug; triangles: ending point of back tracking at 04:00 UTC 27 Aug. Figure SI.3: the changes of the properties of the highlighted air parcel in Fig. R2 along its back trajectory from 19:40 UTC to 03:00 UTC 27 Aug.