Modelling NOx from lightning and its impact on global chemical fields

We have used a three-dimensional off-line chemical transport model (CTM) to assess the impact of lightning emissions in the free troposphere both on NOx itself and on other chemical species such as O-3 and OH. We have investigated these effects using two lightning emission scenarios. In the first, lightning emissions are coupled in space and time to the convective cloud top height calculated every 6 h by the CTM's moist convection scheme. In the second, lightning emissions are calculated as a constant, monthly mean field. The model's performance against observed profiles of NOx and O-3 in the Atlantic and Pacific ocean improves significantly when lightning emissions are included. With the inclusion of these emissions, the CTM produces a significant increase in the NOx concentrations in the upper troposphere, where the NOx lifetime is long, and a smaller increase in the lower free troposphere, where the surface NOx sources dominate. These changes cause a significant increase in the O-3 production in the upper troposphere and hence higher calculated O-3 there. The model indicates that lightning emissions cause local increases of over 50 parts per 10(12) by volume (pptv) in NOx, 200 pptv in HNO3, and 20 parts per 10(9) by volume (ppbv) ( > 40%) in O-3. In addition, a smaller increase of O-3 in the lower troposphere occurs due to an increase in the downward transport of O-3. The O-3 change is accompanied by an increase in OH which is more pronounced in the upper troposphere with a corresponding reduction in CO. The method of emission employed in the model does not appear to have a significant effect globally. In the upper troposphere (above about 300 hPa) NOx concentrations are generally lower with monthly mean emissions, because of the de-coupling of emissions from the model's convection scheme, which vents NOx aloft more efficiently in the coupled scheme. Below the local convective outflow altitude, NOx concentrations are larger when using the monthly mean emissions than when coupled to the convection scheme, because the more dilute emissions, and nighttime emissions, lead to a slower NOx destruction rate. Only minor changes are predicted in the monthly average fields of O-3 if we emit lightning as a monthly constant field. However, the method of emission becomes important when we make a direct comparison of model results with time varying data. These differences should be taken into account when a direct comparison of O-3 with measurements collected at particular times and locations is attempted. (C) 1999 Elsevier Science Ltd. All rights reserved.