New Research Shows Promise for More-Accurate Lightning Prediction


This work is the latest to come from our collaboration with the Radar Meteorology Group within the Department of Atmospheric Science at Colorado State University. It is the culmination of a series of research studies on global-scale lightning variation where we looked at: (1) factors that influence variations in lightning around the world; (2) a case study of tropical lightning variability in the Indian Ocean (the origin for a persistent weather pattern that influences global circulation, the Madden-Julian Oscillation); (3) the relative importance and statistical significance of various environmental conditions that contribute to changes in lightning intensity; and now (4) methods to predict lightning more accurately within earth-system models. The World Meteorological Organization classifies lightning as an essential climate variable and it is worthy on-going study; we are pursuing funding and support to continue research on the interaction between lightning, clouds, rainfall, and air quality in collaboration with the University of Washington and the Pacific Marine Environmental Laboratory.

The abstract of our latest research is below. Thanks for reading!

-D.C. Stolz


Abstract This study investigates how lightning can be more accurately represented in global atmospheric models, using three different approaches: (1) based on environmental characteristics (potential energy, warm cloud depth, humidity, wind shear and atmospheric aerosols); (2) based on the vertical transport of ice particles; or (3) based on the peak thunderstorm height. A novel aspect of this research is that implementations of monthly lightning estimates incorporate the approximate areal extent of thunderstorms (convective area)/number of thunderstorms within model grid boxes. We document the performance of each lightning approximation method and its respective predictions for the frequency of occurrence for various lightning intensities over land and ocean using a single atmospheric model for simplicity. The results suggest that these lightning estimation approaches account for about 55%–64% of the monthly variations in a test sample of satellite lightning observations (without artificial scaling). Recent approaches to model lightning occurrence (e.g., approaches [1] and [2] above) show modest improvements in accuracy and reduced error bias compared to the conventional cloud top height approach (i.e., approach [3] above). We discuss ways to further improve monthly lightning estimates in global atmospheric models as well as the relationship between lightning and atmospheric chemistry under future climate change scenarios.


doi: https://doi. org/10.1029/2020JD033695