The Development of a New Lightning-Frequency Parameterization and its Implementation in a Weather Prediction Model

摘要

Based on a straightforward physical model, a new lightning parameterization has been developed:A two-plate capacitor represents the basic dipole charge structure of a thunderstorm, whichis charged by the generator current and discharged by lightning. In this approach, thegenerator current as well as the discharge strength are parameterized using the graupel-mass field.If these two quantitiesare known, and if the charging and discharging are in equilibrium, then the flash rate is uniquelydetermined. This approach remedies shortcomings of earlier theoretical approaches that relate the flashrate e.g., to generator power. No distinction is made between intracloud andcloud-to-ground discharges.In order to test this approach, polarimetric radardata were used, from which the graupel distribution in observed thunderstorms could be inferred. The lightningactivity was detected using the LINET network. The comparison betweentheoretically-predicted and measured flash rates is encouraging: Over a wide range of flash rates,the theoretical approach yields accurate results for isolated thunderstorms. Two existingparameterizations, which only use the depth of the clouds as predictor, producesubstantially less accurate forecasts.These two existing approaches, the one developed in this study, as well as a fourth one based on updraft velocity,were implemented in the convection-resolving COSMO-DE numerical weather prediction model. With this model,real-world convective scenarios were simulated. The output of the lightning scheme includes the location and timeof every simulated discharge.Testing the performance of theparameterizations with modeled convection is difficult as there is no one-to-one correspondence betweenobserved and modeled convective clouds. Where a comparison between modeled and observed flash rates of individual cloudswas possible, the results for individual cells were promising.The comparison of the bulk lightning activity over an area comprising southern Germany and adjacent countriessuggests that none of the four parameterizations captures the overall lightning activity well. This is mainlybecause COSMO-DE does not simulate the observed number of cells at the correct times.