On the stochastic modeling of Lightning Occurrences by Non-Homogeneous Poisson Process, estimation and simulation performance evaluation and applications: Paris-Venezuela comparison

摘要

The characterization of lightning occurrence by Non-Homogeneous Poisson Process with multiple cyclic effects was first proposed on a ICLP 2012 paper. On that paper was presented an overview of the main reasons for proposing this approach, emphasizing the discrete nature of a lightning event seen as an epoch of a stochastic counting process. It was only presented a preliminary analysis of lightning data sets based on a multi-resolution analysis methodology that makes it feasible to show some potentials application on lighting protection analysis and atmospheric research. This paper is focused on the performance evaluation of a complete analysis tending to show its potential uses for descriptive and simulation purposes. The evaluation results show a good quality of representation supported on performance measures, and reduced deviations errors on repeated simulations. The fitted model is used to introduce the idea of Lightning Occurrence Cycle associated to each resolution of analysis and based on it are outlined some possible applications of this concept on risk analysis and atmospheric research.