Examining relationships between cloud-resolving model parameters and total flash rates to generate lightning density maps

Forecasting the electrical activity of a storm is a difficult task because of the complexity of cloud electrification processes and the unforeseeable characteristics of lightning flashes. Nowadays, very few models are able to simulate explicitly the entire life cycle of the electric charges in clouds, which is a result of the balance between the charge separation rates (microphysics) and the charge neutralization rates by lightning flashes (physics of streamer discharges). The main objective of this study is to investigate which particular dynamical and/or microphysical parameter can serve as the best proxy for the total lightning activity. With this aim, eight storms were simulated with Meso-NH and its cloud electrical scheme. This model-to-model approach guarantees a full consistency between the dynamics, the microphysics and the lightning activity. Two approaches are followed to analyse the results. The first one considers the entire domain of simulation, while the second one concentrates on individual convective cells. Linear regressions between predicted flash rates and recorded parameters or proxies (total graupel mass, updraught volume with vertical velocity higher than 10m/s, product of precipitating and non-precipitating ice mass flux and maximum updraught speed) are analysed to show the benefit of the cell-scale approach. Then, in order to evaluate the relationships, the HyMeX storm case of 24 September 2012 was simulated. The mass of graupel best represents the overall location of the lightning activity (RMSE of the flash rate density better than 10(-2)flmin(-1)km(-2)) and locates the peak lightning flash rate well. The three other proxies tend to cause false alarms due to the y-intercept value of the regression equation.