Christelle Barthe - Cloud electrification and production of nitrogen oxides by lightning flashes
Along the lightning channel, hot temperatures are generated (~ 30,000 K) leading to the dissociation of the major components of the atmosphere: N2 and O2. When the channel cools the N and O atoms rearrange themselves to form NO molecules. In deep convection events, production of nitrogen oxides (NOx) by lightning flashes is a major way to distribute NOx directly in the mid- and upper troposphere. Nitrogen oxides are important trace gases in the atmosphere since they are a precursor of the tropospheric ozone which is an important greenhouse gas. They also control the photochemical regimes of the troposphere and the hydroxyl radical concentration which is the main oxidant of numerous chemical species. The climatic impact of lightning-produced NOx (LNOx) is all the more important as they are produced in the upper troposphere where their lifetime is increased compared to the lower troposphere. However, a large uncertainty exists in LNOx production at the global scale with estimates ranging from 2 to 20 Tg(N) yr-1.
My research is focused on the processes involved in NO production by lightning flashes at the cloud scale.
I have been developing a new lightning-produced NOx parameterization for cloud-resolving models. This parameterization consists of three parts: flash rate, spatial distribution of the lightning channel and NO production rate. First, the flash rate can be given by observations or deduced from model parameters. The flux hypothesis has been evaluated with the Weather Research and Forecasting (WRF) model on different convective cases (the 10 and 12 July 1996 STERAO storms, and the 13 July 2005 Huntsville, AL, storm). This hypothesis correlates the precipitation and non-precipitation ice mass flux product with the total lightning flash rate in a thunderstorm, suggesting that some microphysics parameters can be used as a proxy of lightning in modeling studies (Figure 1). Several sensitivity tests have also been performed to investigate the sensitivity of this parameterization to some microphysics parameters and to the horizontal resolution of the model domain.
Secondly, the spatial distribution of the lightning channel is no more volumetric as in previous parameterizations but attempts to reproduce the global morphology of a lightning flash. The NO molecules are then produced along a virtual lightning flash path in function of the pressure. This parameterization is tested on the 10 July 1996 STERAO storm. Some sensitivity tests are performed to investigate the relative impact of short duration flashes, cloud-to-ground flashes and flash length distribution on the NO production rate and spatial distribution. First results shows that using this parameterization with an average value of 120 moles of NO per flash allows to retrieve the observed NO concentrations both in the convective core and in the anvil (Figure 2). These results also suggest that previous parameterizations tended to overestimate the NO production rate per flash due to the assumption that NO molecules are instantly diluted over a large region of the cloud.
Figure 1. Observed and simulated total lightning flash rate (min-1) for the 10 July 1996 STERAO storm (top), and the 12 July 1996 STERAO storm (bottom). The right figures correspond to the mean total flash rate per 10 min interval whereas the left figures represent the total flash rate per 1 min interval. The simulated total flash rate is calculated from the non-precipitation and the precipitation ice mass fluxes product.
Figure 2. Observed (left) and simulated (right) vertical cross sections of the NO mixing ratio (pmol mol-1) 60 km downwind of the convective core at 6000 s.
ASP Spotlight October 2007
For more ASP spotlights click here http://www.asp.ucar.edu/spotlight/archive.jsp