Forest productivity is extremely sensitive to changes in naturally variable conditions such as water availability and temperature. Yet we still lack reliable ways to predict how tree growth will respond to varying meteorological conditions, especially on timescales less than a year. This project develops a better understanding of how variability in environmental conditions affects tree growth to improve dynamic quantification of forest productivity in the United States. These new insights and improved tools help decision makers form practical, science based decisions that keep forests healthy, support rural economies, and ensure long term resource sustainability. The study supports education and career development by giving undergraduate students, graduate students, and K–12 teachers and students hands on learning experiences that explore how tree growth at daily to seasonal scales is shaped by weather variability. The daily to seasonal (i.e., intra-annual) above and below-ground tree growth data generated by this project are also accessible to those who want a clearer understanding of how forests grow and change over time.
The project aims to identify the mechanistic predictors of tree growth on very fine temporal scales, while simultaneously scaling across forest types to the spatial scale of the continental U.S. through a robust methodology. The project builds on preliminary work showing how fine-scale (daily to monthly) growth responses of individual trees respond to temperature and moisture fluctuations. Highly-sensitive, automated point dendrometers are installed at many sites across the U.S. using the National Ecological Observatory Network (NEON). These dendrometers track both above- and below-ground tree growth dynamics at time scales of hours to days, enabling mechanistic understanding of short term growth changes. The study also uses satellite remote sensing data to further test how observations from both optical and thermal remote sensing are linked with the intra-annual dynamics of tree growth. The project builds a model framework through testing both statistical methods and an established artificial intelligence (AI) approach. This project advances NSF’s priorities in Artificial Intelligence and Biotechnology.