A trait-based approach to plant ecology focuses on the morphological and physiological properties of plants that control their process and function in the ecosystem. Therefore, trait variation and dynamics help us understand functional trade-offs in the plant community and reveal patterns about functional diversity without relying on taxonomic information. 

Because many of these traits (such as carbon and nitrogen content, phenolics, and chlorophyll) have absorption features that can be detected with imaging spectroscopy, we can train trait models using field measured traits and their corresponding image spectra (Fig.1). Applying the trained trait models to hyperspectral imagery produces trait maps across regions of interest at large scales (Fig.2). These trait maps provide major potential for further investigating the linkages between traits and ecosystem response to environmental change. As such, we aim to test the predictive power of modeling trait distributions.