Using Leaf-off LiDAR in Modeling Forest Canopy Structure and Assessing the Effect of Spatial Resolution in Landscape Analyses

In recent years, the increasing availability of airborne Light Detection and Ranging (LiDAR) has allowed landscapes to be studied in unprecedented detail. These data are primarily acquired during leaf-off canopy conditions because it is optimal for modeling the bare earth terrain in deciduous forests. However, there has been limited research investigating the utility of leaf-off data for modeling forest canopy structure and land cover. Furthermore, given the far greater spatial and temporal abundance of moderate- and coarse-resolution remote sensing data, it would be advantageous to determine when coarser data can serve as a useful proxy for finer data in landscape analyses. Thus, the objectives of this dissertation were to examine existing and novel methods to use leaf-off LiDAR data in modeling forest canopy height, canopy closure, and land cover. For each of these objectives, the ability of coarser resolution data to predict high-resolution data was investigated. This research contributes to the field of remote sensing and landscape ecology by: 1) demonstrating that leaf-off LiDAR is effective in modeling canopy height for the variety of tree species common to temperate deciduous forests in the northeastern United States; 2) developing a novel technique that is the first to successfully model canopy closure in a deciduous forest; 3) developing the first fully automated algorithm capable of accurately classifying spatially high resolution land cover across a large geographic extent (i.e. eastern Connecticut); and 4) demonstrating that moderate-resolution Landsat-based data can serve as a good proxy for high resolution data in predicting land cover areas given a sufficiently large analysis window