On Prediction of Non-linear Phenomena Related to Natural Disaster

On prediction of natural disaster, it is very important to predict rapid and discontinuous changes of non-linear phenomena (multiple regimes). As examples of multiple regimes related to natural disaster, there are various phenomena; appearance and disappearance of atmospheric blocking, transition among patterns of deep water circulation, transition between large meander and non-meander of Kuroshio path, and quasi-periodic collapse of snow avalanche. Recently, some new trial approaches to predict their behavior are conducted. In this paper, among their approaches, we introduce mainly a thermodynamic approach and its application to deep water circulation. It is a trial to understand the essence of phenomena by detecting macroscopic characteristics not depending on the details of system and is valid to predict various phenomena related to natural disaster involved in complex factors.