Forests around the world may be in danger as climate change causes shifts in insects and diseases, increases fires and changes precipitation patterns.
One study by researchers at Duke and other universities may help scientists figure out which forests are at risk. The team studied satellite images of California forests—encompassing a 15-year period and two droughts—to figure out potential warning signs before a forest’s demise.
Yanlan Liu, one of the paper’s authors and a former doctoral candidate at the Nicholas School of the Environment, said that the goal of her research was to monitor the health of global forests at risk due to climate change. By establishing a statistical standard for these forests, the study—published Oct. 7 in “Nature Climate Change”—predicted the likelihood of mortality of these forests.
“Before this study there were two main approaches to predicting mortality. One was relationships between primary stressors and mortality rate, and the other was the use of simulations to predict mortality risk,” Liu said.
Although these methods have worked to an extent in the past, she explained, they don’t do a good job of taking variation among tree species into account. Liu added that such relationships may change across space and time depending on a particular species’ adaptations.
Using the simulation model could also be problematic because it’s difficult to determine which particular stressor a tree may have died from, she said.
“Forest mortality from droughts is being widely observed across the globe. To assess the risks and devise mitigation and restoration strategies, we need to be able to predict forest mortality ahead of time,” Liu said in a news release.
Liu’s model uses tree recovery rate during a period of drought as a standard early warning signal (EWS). Her team used Geographic Information Systems to monitor 30-by-30-meter satellite pixels of forests in California over time.
From there, they took the pixel data and monitored for EWS signs of slow recovery, taking note of the natural recovery rate or background rate of different healthy tree populations as a benchmark for mortality.
By using a Normalized Difference Vegetation Index, which takes satellite data and determines if the image has live greenery in it, Liu correlated slow recovery rate in tree patches to leaf death and tree mortality over time.
“In general, if numerous patches of forest exhibit these early warning signals, then the forest as a whole is more likely to die,” she said.
However, Liu mentioned that there is no clear standard or set number of patches that have to see slow recovery rates in order to be considered high risk. Recovery rate, tree health due to climatic shifts and mortality of trees differ across species and ecosystem type, making the exact measure of mortality due to stressors like drought complex.
Get The Chronicle straight to your inbox
Signup for our editorially curated, weekly newsletter. Cancel at any time.
Liu emphasized that the method does not come without its drawbacks. She said that to improve, the team could include insects as an indicator of climate impacts. In doing so, Liu believes that their model for mortality can be better refined for future study.
For example, she noted that mortality in tree species in California, where the study took place, is often most greatly impacted by insect distributions and climate impacts on insect movement.
“Tree mortality is a very complicated thing, and we don’t exactly understand the mechanism behind tree mortality fully,” Liu said.
Despite the difficulty, she believes that statistical models like EWS are the future of forest monitoring and management.
Liu is currently working with other forest ecology students and faculty in her current position at Stanford University to discover if useful data exists in other regions that can expand her understanding of tree recovery and forest mortality.
“We also hope to work with different agencies around the world to get more live or real time reads of forest recovery rather than relying wholly on satellite data,” she said.
Although the Oct. 7 study was conducted in primarily temperate forests in California, Liu highlighted that she would like to see how her work may translate to forest mortality in the tropics and boreal forests.
Researchers still debate the extent to which increasing temperatures will affect forest populations. Liu mentioned that there is discussion around the impact of humidity and precipitation shifts on future forest dynamics. Either way, she said that researchers agree climate change will increase tree mortality, an eventuality that Liu and her team hope to monitor.
“Forests play very important roles in terms of regulating the land-atmosphere interactions and carbon budget. Right now, we don’t have a thorough understanding of to what extent hydrologic stress impacts forests,” she said.