Abstract
The structure of tree canopies affects turbulence in the atmospheric boundary layer, and light attenuation, reflection and emission from forested areas. Through these effects, canopy structure interacts with fluxes of heat, water, CO
2
, and volatile organic compounds, and affects patterns of soil moisture and ecosystem dynamics. The effects of canopy structure on the atmosphere are hard to measure and can be studied efficiently with large-eddy simulations. Remote sensing images that can be interpreted for biophysical properties are prone to errors due to effects of canopy structure, such as shading. However, the detailed 3-D canopy structure throughout a large spatial domain (up to several km2) is rarely available.
We introduce a new method, namely the virtual canopy generator (V-CaGe), to construct finely detailed, 3-D, virtual forest canopies for use in remote sensing, and atmospheric and other environmental models. These virtual canopies are based on commonly observed mean and variance of biophysical forest properties, and a map (or a remotely-sensed image) of leaf area, or canopy heights, of a canopy subdomain. The canopies are constructed by inverse 2-D Fouriertransform of the observed spatial autocorrelation function and a random phase. The resulting field is expanded to 3-D by using empirical allometric profiles. We demonstrate that the V-CaGe can generate realistic simulation domains.
