A loss of resilience - the rate of recovery from a disturbance - in forests can be detected from satellite vegetation data and may act as an early warning signal to predict the death of the forest, suggests a paper published in Nature Climate Change.

Forest mortality has been observed with increasing frequency worldwide, and the resulting changes in land-cover impact ecosystems and local species. Changing temperature and precipitation patterns, due to climate change, have stressed forests in recent decades. Predicting forest death is important for prevention, but is also challenging, as the underlying mechanisms are not well understood.

Yanlan Liu, Mukesh Kumar and colleagues developed a new, observational approach for predicting forest mortality using remote sensing data of vegetation dynamics. The approach is based on observations of changes in resilience, which happens as a forest approaches a tipping point at which it may change to a different kind of ecosystem, such as shrub land. These changes can be detected from the analysis of satellite vegetation data over time, and may indicate the slow recovery of foliage after a disturbance, which occurs as resilience declines. The authors tested the approach in Californian forests and found that the early warning signal could be detected 6 - 19 months before forest death. This signal was detected prior to other signs of forest decline, such as reduced greenness.

The authors suggest this system may enable more accurate and earlier prediction of forest mortality, giving forest managers more time to prevent it by mitigating threats and restoring forest health.