An analytical solver is developed for characterizing the coherent scattering responses of tree scenes. Realistic 3-D tree structures are first constructed using an open-source random tree generation engine. The trees are then parsed into discrete, canonical scatterers, such as cylinders and disks, and a multiray approach is applied to calculate the aggregate response of the scene, with the transmissivity of each ray determined from a cell-based representation of the computational domain. As each scatterer in the outlined framework is assigned a deterministic position, the spatial distribution of the trees and their canopy structures is fully preserved. A cell-by-cell strategy is also proposed for speeding up the calculations of the responses from small components such as secondary stems and leaves, which are expected to far outnumber those scatterers composing the trunks and primary branches. The accuracy of the analytical solver is assessed by comparing simulation results for a forest stand with solutions from a large-scale, full-wave solver. In addition, as an application of interest, the detection and imaging of a tree-obscured walking human target is demonstrated.