When it comes to carbon, energy, water, and nutrient cycling, not all trees are the same. Different species, and even the same species at different phases of life, process these resources in diverse ways.
To fully understand how the dynamics of carbon, energy, water, and nutrient cycling operate at ecosystem scales, researchers construct models of forest ecosystems. These models perform best when supplied with detailed observations of the composition and structure of the forest, including how many trees of different sizes and species there are in the canopy. Measuring individual trees in a given plot of land is the most accurate way to do this, but such on-the-ground surveys cover only a limited geographic area and are labor intensive and expensive to conduct.
Using a combination of lidar and imaging spectrometry, Antonarakis et al. describe a new technique to remotely sense a forest’s tree composition using airborne observations.
In 2003, researchers flew the Laser Vegetation Imaging Sensor (LVIS) lidar equipment and an imaging spectrometer over the much-studied Harvard Experimental Forest in Petersham, Mass. Although they could not identify down to the species level, the scientists were able to group the species into ecologically defined types: early, mid, and late successional deciduous trees and early and late successional conifer trees. The technique uses the airborne measurements to determine how many trees of different sizes there are of each of the ecological classes.
Although the airborne sensors’ observations were not as detailed as those obtained from an on-the-ground survey, the remotely sensed measurements were good enough to use as inputs to a forest ecosystem model, the researchers say: Model runs using the remotely sensed measurements produced predictions that closely matched those calculated using on-the-ground measurements as inputs. (Geophysical Research Letters, doi:10.1002/2013GL058373, 2014)
—Colin Schultz, Writer