Files
Abstract
This dissertation determines how drainage basin characteristics, including watershed morphometry, soil parent material, and land cover, influence stormflow and channel morphology in the Southern Blue Ridge Mountains by addressing three separate but interrelated research questions. The first study uses discharge records from fifteen USGS gaged watersheds to determine the drivers of winter stormflow. This analysis indicates relief, slope, and landscape connectivity explain the majority of regional variance in winter stormflow. Additionally, continental-scale atmospheric circulation patterns (i.e., the North Atlantic Oscillation) and land cover on specific portions of the landscape (e.g., development on residuum) explain a significant amount of the variance in the hydrologic variables during dry-, average-, and wet-years. The results of this study have important implications because they suggest that the hydrologic record is not a homogeneous population and the processes that create and propagate stormflow are fundamentally different in periods of wetness versus drought. The second study uses digital soils maps from a single county and the random forest classification algorithm to map the distribution of soil parent materials in the Southern Blue Ridge Mountains. This study determined that digital soil maps can be used to create predictive functions that accurately model regional distributions of soil parent material, as long as the data are representative of the region and the land surface characteristics of the different soil parent material classes can be differentiated. The third study determined the watershed- and reach-scale characteristics that explained the most variance in the channel morphology of fifty-five cross-sections in three watersheds that were analyzed as part of the first study. These results, coupled with results of the first study, indicate that in order to predict channel morphology in mountainous terrain, there is a need to understand not only watershed size, but also the watershed- and reach-scale characteristics that influence runoff. This research illustrates the fact that a model by definition is an imperfect representation of reality, but that the wealth of digital spatial data sets accumulated over the past few decades can be very effectively applied to problems in hydrology and fluvial geomorphology.