- For the first time ever, researchers have mapped the underground network of microbes connecting forest trees around the world using an enormous data set of more than 1.1 million forest plots.
- Mapping the forest microbe network required global collaboration and high computing capabilities.
- The new maps confirm patterns that have been long suspected. For example, arbuscular mycorrhizal fungi dominate forests in the warmer tropics while ectomycorrhizal fungi are more widespread in colder boreal and temperate forests.
- The predicted maps are, however, limited by the geographic coverage and sampling density of trees across the world. While the coverage is good in developed countries, it is relatively poor in developing countries like India, China and countries in the tropical region, the researchers say.
Underneath the ground, hidden from view, there’s a massive network of fungi and bacteria hard at work in close partnership with tree roots. In exchange for food from the trees, the microbes help transfer nutrients from the atmosphere or soil to the trees and back. They also influence what kinds of trees grow where, affect how resilient trees are to diseases, and respond to disturbances.
For the first time ever, researchers have mapped this underground network of microbes connecting forest trees around the world using an enormous data set of more than 1.1 million forest plots. These plots collectively contain information on more than 30 million individual trees and 28,000 tree species from more than 70 countries.
Compiling this data set was no easy feat. First, researchers generated a global map of forest trees, by reaching out to forest scientists and foresters all around the world, nearly 200 of whom then contributed ground-sourced forest measurements from their study areas. The 2015 study estimated there were about 3 trillion trees on Earth.
With this humongous amount of tree-related information collated in the Global Forest Biodiversity Initiative (GFBI) database, the researchers then started exploring various questions related to the world’s forests. Understanding how the underground network of symbiotic microbes that connect forest trees was distributed around the world was one of them.
There is a wide diversity of intimate associations between trees and microbes. The researchers of the latest study in Nature focused on three kinds that are most widespread: ectomycorrhizal fungi, arbuscular mycorrhizal fungi, and nitrogen-fixing bacteria. Ectomycorrhizal fungi (EM), which tend to associate with trees like pines, eucalyptus, beeches and firs, stick to the surface of tree roots, and create vast branching networks that forage for nutrients from decomposing leaf litter and soil organic matter. Arbuscular mycorrhizal (AM) fungi also transfer nutrients, but penetrate deep inside root cells and form smaller networks. These fungi tend to associate with trees like redwoods, maples and cedars. Nitrogen-fixing bacteria associate with plants like legumes and convert atmospheric nitrogen to forms such as ammonia, which plants can use more readily.
“We know that the remarkable diversity of trees globally is not well-captured by grouping all species into one of three tree ‘types,’” Richard Phillips, an associate professor of biology at the University of Indiana, who was not involved in the study, told Mongabay. “But this type of assumption is appropriate when modeling forests at the global scale because we need some way to simplify all of the global variation in tree species form and function.”
To map the forest microbe network, the researchers used machine learning to try and figure out how the three different tree-microbe associations correlated with various environmental factors, such as temperature, rainfall, soil chemistry and topography. Their training data set of trees came from the on-ground measurements within the GFBI database, and they used published literature to assign each tree in the training data set to a type of symbiotic microbe.
“For each single tree, we know its location, we know what symbiotic state it is related to, and we know all the related climate conditions, and soil types,” Jingjing Liang, co-lead author of the study, co-founder of the GFBI, and an assistant professor at Purdue University, told Mongabay.
Based on the relationships uncovered by the machine-learning algorithm, the researchers then ranked the importance of all the variables from the highest to the lowest. They also used the relationships to extrapolate and predict the distribution of tree-microbe associations, filling in spatial gaps where data on trees and microbes are poor or unavailable. “That’s how we found that for EM and AM types, it is decomposition and climate variables that matter the most. However for nitrogen fixers, soil chemical and physical properties also have a major role,” Liang said.
Their maps confirm patterns that have been long suspected. For example, arbuscular mycorrhizal fungi dominate forests in the tropics, where warmer temperatures decay organic matter quickly, allowing the fungi to form smaller networks. Ectomycorrhizal fungi, by contrast, are more widespread in colder boreal and temperate forests, where organic matter takes longer to decompose. The researchers also estimated that some 60 percent of all tree stems on Earth are ectomycorrhizal, despite associating with only 2 percent of overall plant species.
“Before this hard data, knowledge of these patterns was limited to experts in mycorrhizal or nitrogen-fixer ecology, even though it is important to a wide range of ecologists, evolutionary biologists and earth scientists,” Kabir Peay, assistant professor of biology, and co-author of the study, said in a statement.
The predicted maps are, however, only as good as the quality and quantity of data that is out there, Liang said. “There is limitation in terms of geographic coverage and sampling density across the world,” he added. “In developed countries, like U.S., Canada and western Europe we have very good coverage, but in developing countries, especially those with high biodiversity like India, China and countries in the tropical region, the GFBI data coverage or sampling density is relatively poor.”
This disparity, Liang said, partly came down to the costs of taking on-ground measurements of forest trees. “The per tree cost of measuring a tree in developing countries is much higher than that in a temperate region like in the U.S., simply because of the infrastructure out there,” he said. “In the U.S., for example, people can travel by road with comfort and get access to forests in relatively short amount of time, but in remote areas in the global south it can take days to hike to a spot even before they are able to measure a tree.”
Despite these limitations, the global maps are important contributions, Phillips said.
“The assumption here is that the ‘training data’ used to develop the statistical model can be extrapolated to predict vegetation patterns in areas outside of the training data,” he said in an email. “However, it’s important to note that the accuracy of the map is not what matters most — at least in my opinion. Rather, the important contribution is that we now have a few different global maps of forest tree-symbioses (a few other research groups have also produced maps) that can be used in global modeling studies of our planet.”
Liang and colleagues have already taken a stab at making a prediction using their global map. They modeled how the symbiotic associations might change by 2070 if carbon dioxide emissions continue to rise, and found that there would be a reduction in EM-associated trees. And since 60 percent of the trees across the world are EM-associated, their loss could mean a large amount of CO2 being released into the atmosphere.
Global maps like these, based on ground measurements, are hard to come by and require global collaborations, Liang said. “I see this project as a success of not just the first few authors but of the global community that contributed to the GFBI.”
Such studies also rely on high computational capabilities. “All the big data married with these high computing capabilities made this study possible. Otherwise we would not be able to analyze 30 million trees in such a short amount of time,” Liang said.
Banner image of a rainforest in Costa Rica by Rhett A. Butler/Mongabay.
Steidinger, B. S., Crowther, T. W., Liang, J., Van Nuland, M. E., Werner, G. D. A., Reich, P. B., … & Herault, B. (2019). Climatic controls of decomposition drive the global biogeography of forest-tree symbioses. Nature, 569(7756), 404.
Correction: An earlier version of the article quoted Dr. Liang saying that “precipitation and temperature matter most” for EM and AM type associations. It has been changed to “decomposition and climate variables”.