Using Hyperspectral Remote Sensing to Scale From Genes to Ecosystems
Christopher Doughty1, Gerard J. Allan1, Paul G. Flikkema1, Catherine A. Gehring1, Temuulen T. Sankey1, Thomas G. Whitham1
1 Northern Arizona University, Flagstaff, AZ, 86001, USA; chris.doughty@nau.edu
The genes of foundation plant species have been documented to influence communities and ecosystem processes like productivity and nutrient cycling in biomes around the world. For example, experimental forest field trials containing 1000’s of replicated tree genotypes have shown how individual genes and the genetic profiles of individual trees reveal the effects of climate change on communities. The next step is to scale up these local observations to regional, continental, and global scales. A major challenge of scaling up this community genetics approach to restore ecosystems and maintain forest and cropland production is the ability to measure changes in traits (phenotypes) of critical species at large spatial scales. A revolution of hyperspectral remote sensing has enabled nearly continental scale remote measurement of phenotypic traits such as leaf chemistry and structure. Such leaf traits are highly correlated with traits throughout the plant (such as rooting depth) through the leaf economics spectrum, and non-leaf plant traits have been accurately predicted with in situ leaf spectroscopy. Here we show how leaf spectroscopy can possibly predict genetic differences between populations of cottonwood trees in field trials.