Since most of large disturbances, such as earthquake and wildfire, rarely but repeatedly occur, it is crucial to reveal how the populations which experienced such disturbances overcome such disturbances for understanding the persistence of the populations. Understanding resilience, capacity to return to pre-disturbance reference, is effective to predict recovery process after disturbances. Although resilience has been developed and defined by various measure, it was much less quantified in natural populations comparing with conceptual or experimental works. This was because natural populations’ dynamics do not follow the assumption of the resilience, especially after large infrequent disturbances. For example, the course of recovery may not be straightforward. In addition, to observe the complete recovery requires too long time, conversely, can be impossible. Therefore, there is no consensus how the resilience should be quantified for natural large disturbances. The aim of this study is to review the concepts and measures of resilience and investigate which measures can be applied to the natural populations experienced large infrequent disturbances. We tried to quantify and compare resilience in various definition and measures by using long time series data of natural populations.