Global Lightning Parameterization from CMIP5 Climate Model Output

Abstract

ightning is best known as a component of severe weather, but diverse communities of researchers, such as those in atmospheric chemistry and global fire modeling, have long recognized that lightning plays a significant role in many components of the Earth system. Global climate models are unlikely to be able to simulate lightning from first principles, but the same models do simulate many parameters related to convection, including total precipitation rate, convective precipitation rate, and convective mass flux. This study combines satellite observations of lightning and CMIP5 climate model simulations to derive an empirical parameterization of monthly lightning in terms of monthly simulated convective parameters. Convective mass flux best captures the spatiotemporal distribution of observed lightning. Derived lightning seasonality is captured with 95% confidence over 69% of land but only 30% of ocean. Spatially, the correlation of derived lightning and observed lightning is 0.74. Overall, global observations suggest lightning occurs at an annual rate of 47 flashes per second, while lightning from the parameterization occurs at 44 flashes per second. A robust feature of the relationship between lightning and climate model convective parameters is that lightning flash rate increases linearly with increases in convective precipitation rate and in convective mass flux for a significant subset of the total range of those convective parameters. Namely, this linear proportionality is evident when the convective precipitation rate is less than 4?5 mm day?1 and the convective mass flux is less than 15?16 kg m?2 h?1, which account for about 90% of the values simulated by the climate models.