Deriving forest canopy parameters for backscatter models using the AMAP architectural plant model

Abstract : A new approach using an architectural plant model to feed various theoretical scattering models is presented as a better interpretation of future remote sensing data acquired over natural media. The method is based on the architectural plant model (AMAP), which integrates knowledge of botanical growth processes and real plant measurements. AMAP is encapsulated in a flexible interface software called AMAP2SAR that allows one to (1) simulate a three-dimensional (3-D) plant such as a tree, (2) transform the tree into a collection of cylinders, and (3) feed theoretical models such as radiative transfer (RT) models. The method is illustrated by an example of Austrian black pine plantations in southern France. Simulated characteristics of black pines are validated for stands up to 50 years old and for a given environment. The results show the ability to derive classical forest parameters as well as those needed for electromagnetic models (such as geometry) as a function of age. Vertical profiles of canopy elements are derived and point out the vertical heterogeneity of the stands after they are 20 years old for parameters having an impact on the backscatter such as diameter and number of branches. Consequently, the crown layer variability with age and canopy depth should be considered in RT models. An RT model is modified in order to take account of accurate canopy descriptions and deal with encouraging modeling results at Cand L-band over the same test site.