We performed simulation experiments on cloud-to-ground lightning by using the Lightning Simulator which was incorporated in the CReSS (Cloud Resolving Storm Simulator) in respect to the cases in which LLS (Lightning Location System) observed cloud-to-ground lightning. Using the results of the reproductive experiments, we examined the parameter sensitivity of charge neutralization process to cloud-to-ground lightning.

The Lightning Simulator has “first stage” and “second stage” of the lightning parameterization. The first stage is mainly modeling flash development of the lightning channel, and the second stage is mainly modeling the charge neutralization process in thunderstorm cloud. In the charge neutralization process, it is necessary to set values of the three-parameters; chan r , the minimum value of charge density in regions where we assume grid points of charge neutralization, neut r , the magnitude of charge density that is not available as a lightning flash (and that is, too small to neutralize charge), and fr , the rate of the available charge density neutralized at a grid point by a lightning flash.

From the results of the experiments, we find that in respect to cloud-to-ground lightning frequency obtained from the lightning simulation, chan r and neut r had low sensitivity and fr had high sensitivity in cases in which the results reproduced sufficiently developed clouds. That is to say, while changing the value of chan r or neut r did not vary the cloud-to-ground lightning frequency, it decreased with increasing the value of fr . The percentage of cloud-to-ground lightning frequency in all lightning frequency was little affected by changing of the parameter value of charge neutralization process. This suggests that the parameter value of charge neutralization process hardly acts on the pattern of charge distribution in thunderstorm cloud.

Although the previous study using this Lightning Simulator had been hardly able to simulate the cloud-to-ground lightning polarity which LLS observed, the results of our experiments properly simulated the cloud-to-ground lightning polarity which LLS observed. This suggests that if we perform experiments in suitable vertical resolution and can simulate sufficiently developed thunderstorm, the Lightning Simulator can properly simulates realistic cloud-to-ground lightning polarity.