Featured Expert: Tong Qiu - Mapping Forest Fecundity

Mapping Forest Fecundity: Correlating Nutrient Availability and Seed Production

Understanding the factors that influence seed production in trees is one of the keys to building better models of forest growth and regeneration. Dr. Tong Qiu and his co-investigators used remote sensing data from the NEON Airborne Observation Platform (AOP) along with seed estimates from the Masting Inference and Forecasting (MASTIF) Network to correlate crown nutrient concentrations with tree fecundity. The results were published in the journal Ecology: “Remotely sensed crown nutrient concentrations modulate forest reproduction across the contiguous United States.” 

Nutrients, Seed Production and Forest Regeneration 

Tree fecundity—the ability of the tree to produce seeds or fruit—is an important element of forest dynamics and regeneration. Qiu explains, “The question behind this paper is: How does a forest regenerate after a disturbance or die back? The first process in regeneration is seed production. We know that the number of seeds a tree produces is tied to the nutrients available to the tree. So that’s where our focus is here: How does nutrient availability influence seed production and ultimately forest regeneration?” Previous studies have investigated the correlation between foliar (leaf) nutrient concentrations and seed production at small scales, often for individual species in a small geographic area or for individual trees over time. Qiu and his co-investigators wanted to explore these relationships across more species and at a much larger scale. They leveraged NEON remote sensing data in combination with the MASTIF data to explore relationships between nutrient availability and seed production at 13 NEON sites across the country. The MASTIF Network is organized by the Clark Lab at Duke University and funded in part through a National Science Foundation (NSF) grant. It is a long-term monitoring program with more than 500 plots and crop count locations. Thirteen of these are co-located with NEON field sites (see figure below). This enabled researchers to correlate NEON remote sensing data with seed count data from the same locations. 

Seed counts are conducted through a combination of seed traps and crop counting. Seed traps collect seeds that fall from trees, which are then sorted and identified by species. Crop counts are conducted by field researchers or citizen scientists using binoculars. The collected data from these methods are combined in a modeling framework to estimate seed production for individual trees and across different species. 

Remote sensing data from the NEON AOP was used to determine crown nutrient concentrations for individual trees at each site. Lidar data, which measures vegetation height and structure, was used to delineate individual trees. The Townsend lab at University of Wisconsin Madison utilized Hyperspectral data to quantify nutrient levels in leaves at the top of the canopy, including nitrogen (N), phosphorus (P), potassium (K), calcium (Ca) and magnesium (Mg). Hyperspectral data captures the reflectance of sunlight across hundreds of wavelengths from the tree canopy. Different nutrients cause specific patterns of light absorption and reflection. By analyzing these spectral signatures at various wavelengths, researchers can correlate them with nutrient concentrations in leaves. This method allows for mapping of nutrient distribution across large areas, providing insights into tree health and nutrient availability without the need for labor-intensive field sampling. 

Correlating Nutrient Availability with Seed Production 

In the study, the researchers measured five key nutrients—nitrogen (N), phosphorus (P), potassium (K), calcium (Ca) and magnesium (Mg)—to understand their influence on tree seed production. These nutrients were chosen because they play crucial roles in tree growth and reproduction. Nitrogen, phosphorus and potassium are essential macronutrients linked to plant growth, while calcium and magnesium support cellular function and structure. The study sought to determine how varying concentrations of these nutrients, especially phosphorus, impact tree fecundity for different species and in different parts of the country. 

Overall, the data showed that phosphorus had the biggest impact across the widest range of species. Most species had a negative correlation between phosphorus concentrations and seed production, meaning trees in phosphorus-rich environments produced fewer seeds overall. In fact, with the exception of a few conifer species, seed production decreased by four orders of magnitude between trees with the lowest concentrations of phosphorus and those with the highest. Why might that be? Qiu says, “We hypothesize that if you over-fertilize the tree, it will have more resources to allocate to growth rather than reproduction. So, they grow like crazy and produce a lot of leaves instead of producing more seeds.” 

Correlations with other nutrients proved to be more species-specific and less clear-cut. Overall, communities with the highest seed production had low levels of phosphorus and moderate levels of nitrogen; effects of nitrogen on seed production at a community level depend on phosphorus. However, reactions to K, Ca and Mg showed much more variation between different species of trees. 

Predicting—and Supporting—Forest Regeneration 

Qiu says the results could be used to build better models of forest regeneration or even inform land management and conservation decisions. “If we understand how different species are responding to nutrients, maybe you can apply certain nutrients strategically to ensure successful natural regeneration,” he says. “For example, we don’t just want to apply phosphorus at the same level to different types of forests. We have to tailor our fertilization strategy for different sites depending on the species composition found. It’s like precision forestry.” 

Next, Qiu hopes to expand this research beyond the NEON sites. Remote sensing data from NASA satellites could provide a broader continental-scale view of nutrient availability across the country. The researchers also want to examine how year-to-year variations in nutrient availability impact seed production. This could inform predictions of mast events, in which certain tree species produce an unusually large number of seeds in one year, followed by one or more years of low seed production. Mast events impact not only tree regeneration, but also the bird and small mammal communities that depend on seeds for sustenance. “Nutrient levels influence the quality of the habitat,” says Qiu. “There are a lot of broader impacts on the forest food web.” 

Understanding the factors that influence forest regeneration is important in an era where forests are increasingly threatened by wildfires, human activity and invasive species. This research is helping to illuminate the distinct relationships between nutrients and seed production at the species, community and landscape levels, which in turn could lead to better models of forest regeneration for different regions. “It’s helping us predict what regeneration might look like,” explains Qiu. “We can couple that with predictions of tree size and nutrient availability to generate a spatial map of seed production and regeneration capacity in different areas.” 

The study was enabled by NEON’s open-access ecology data, along with the MASTIF seed trap data collected at NEON sites. “This research would not be possible without the support of NEON data,” says Qiu. “NEON has given us the capacity to generate nutrient maps in totally different ecosystems, from western conifer stands to eastern hardwood forests. It’s helping us fill important knowledge gaps and understand how different species react across nutrient gradients.”