A Mixed Poisson Approach to Account for Ground Flash Density Variability in Transmission Line Lightning Performance Analysis

Lightning performance of transmission lines is commonly assessed using long-term average values of ground flash density, implicitly assuming stationary lightning conditions. However, observations from lightning detection networks have consistently shown that ground flash density exhibits non-negligible interannual variability. This paper investigates how uncertainty in ground flash density propagates into the annual number of lightning flashes to transmission lines. A mixed Poisson formulation is adopted, in which the intrinsic randomness of lightning strikes is combined with yearto-year variability of lightning activity. Sensitivity analyses are performed to quantify the impact of dispersion and median ground flash density on the predictive distributions and intervals of annual lightning exposure. Results show that even moderate interannual variability leads to a substantial widening of predictive intervals, particularly in the upper tail, indicating an increased likelihood of extreme years. The findings suggest that classical Poisson-based approaches may underestimate lightning exposure variability and motivate the use of mixed Poisson models in transmission line performance assessments.