COS 267-3 - Remote sensing for ecological insight: linking hyperspectral data products with physiological drought stress

Abstract: Remote sensing has dazzling potential to change the way ecologists gather and interpret biological data. With NASA’s upcoming Surface Biology and Geology mission expected to launch in 2026 - providing ecologists with global high resolution hyperspectral data - the time to ensure that we have clear linkages between remote sensing data products and biological processes is now. Using high spatial and temporal resolution hyperspectral data, we tested whether hyperspectral data could be used to accurately identify individual tree drought stress (as a precursor to observable drought response) in a blue oak woodland. Specifically, we tested if in-situ and remotely sensed measurements of canopy water content could predict leaf water potentials. During fortnightly AVIRIS hyperspectral overflights, we collected data on physiological drought status (predawn and midday water potential), leaf water content, and leaf morphological traits across an environmental gradient and compared these with remotely-sensed estimates of canopy water content (CWC).

We found that there were consistent relationships between leaf water content and water potential across time of day, geographical space, and season in in-situ measurements, but that these relationships changed as seasonal drought progressed. This suggests that the mapping between remotely sensed plant water content and actual plant water stress is context dependent based on plant physiology alone. While leaf-level traits could be linked to CWC measured remotely, the underlying mechanisms behind these relationships require further study. Our results provide promising evidence that, when combined with an understanding of physiological mechanisms and an appreciation of species-specific phenology, hyperspectral data can be used to identify drought-related stress and to predict drought response.