Predicting catastrophic shifts.

摘要: Catastrophic shifts are known to pose a serious threat ecology, and reliable set of early warning indicators is desperately needed. However, the tools suggested so far have two problems. First, they cannot discriminate between smooth transition an imminent irreversible shift. Second, aimed at predicting tipping point where state loses its stability, but in noisy spatial system actual occurs when alternative invades. Here we suggest cluster tracking technique that solves both problems, distinguishing catastrophic transitions identify shift cases. Our method may allow for prediction, thus hopefully prevention such transitions, avoiding their destructive outcomes.