Case-based lightning flash forecast support system taking account of range of thundercloud

In order to prevent the damage caused by natural phenomena, such as rain, cloud, lightning and squall, we use meteorological radars for the observation. Demand has been increasing for more accurate information for natural phenomena, since it is becoming more important to analyze the observation data for the forecast. We have been applying a case-based retrieval algorithm to forecast lightning flashes. However, since the proposed method considers the partial local region in retrieving the similar case, a number of dissimilar cases may be retrieved. In order to solve this problem, we propose a new retrieval scheme which takes account of the range of the thundercloud. We show that this scheme is efficient by means of an example of the application.