Comparison of sprite initiation altitudes between observations and models

Simultaneous analyses of measured sprite initiation altitudes with predicted initiation altitudes from simulations enable an examination of our understanding of the sprite initiation mechanism and the modeling techniques to simulate this mesospheric electrical phenomenon. In this work, we selected a subset of sprites optically observed from Langmuir Laboratory, NM; locations near Las Vegas, NM, in 2007 and near Portales, NM, in 2008; and a Duke University field station. The sprites were observed by high-speed imaging with time resolutions of at least 1 ms and by low light level imagers. Sprite initiation altitudes were determined by triangulation between Langmuir Laboratory and either Portales or Las Vegas, while star field analysis determined the approximate measured initiation altitudes for Duke observations. These video observations were coordinated with electromagnetic field measurements from Yucca Ridge Field Station and Duke University, respectively. With a 2-D finite difference time domain model, we simulated the lightning-driven electric fields and predict the likely altitude of sprite initiation and compare these findings with the measured initiation altitude of each sprite analyzed. Of 20 discrete sprite events analyzed, both the measured and the simulation-predicted initiation altitudes indicate that long-delayed sprites tend to initiate at lower altitude. The average discrepancy between the measurements and the simulation results is 0.35 km with a standard deviation of 3.6 km. This consistency not only confirms previous results about the relationship between sprite initiation altitude and time delay but also helps to develop confidence in the models to reveal the sprite physics.