Researchers have found a way to predict death of forests by detecting “hidden” signals of resilience using satellite vegetation data.
The Early Warning Signal (EWS) was tested in Californian forests and could detect mortality between six and 19 months before it occurs, as described in the journal Nature Climate Change.
Increasing numbers of forests are dying as a result of climate change and heat stress and associated events such as wildfires and insect outbreaks. This is having a serious impact on their biodiversity, rich ecosystems and critical role in balancing the planet’s carbon levels.
Understanding and forecasting the environmental impacts of climatic changes is a major challenge for global change scientists, according to Craig Allen, from the US Geological Survey in Los Alamos, US, and co-authors of a separate review on the topic, who add that “forecasting the impacts on forests is of particular importance”.
But until now, it’s been hard to predict forest mortality through computer models because of the complexity of the process at individual tree and ecosystem scales, particularly in a “real” setting, says corresponding author Mukesh Kumar from Duke University, Durham, in the US.
To overcome these challenges, the researchers accessed publicly available remote sensing data and applied a Bayesian Dynamic Linear model to analyse it and detect periods of “slow recovery” in the timeline of vegetation dynamics.
“Every year, the vegetation biomass generally increases during the growing season and reduces in the dormant season,” explains Kumar.
When the trees are stressed, their normal rate of vegetation recovery during the regrowth period will be slower. Such loss of resilience has been found to occur before a system reaches a tipping point, so detecting it can sound an early warning bell before obvious signs like reduced greenness appear.
The method could monitor the health of woods live to inform early protective measures. The team hopes forest managers will be able to use the predictions to assess resource risks and advocate for risk mitigation through controlled fires, removal of infested trees, selective density thinning and biological controls.
Next, they plan to test the EWS in other forests around the world where climate-induced forest mortality has occurred and refine the method for different vegetation species. They also hope to integrate the information on vegetation dynamic models to better predict water and carbon fluxes.
