Examining development-induced geomorphic change using multi-temporal LiDAR-derived digital elevation models.

摘要

During development, topography is cut, filled, and graded, altering natural surface drainage patterns. The spatiotemporal characteristics of drainage network and hillslope changes throughout the development process are to date unknown. For this study we used multiple temporal LiDAR-derived DEMs spanning the development of two small agricultural watersheds to examine geomorphic changes caused by urban development. A third forested watershed was also examined to track geomorphic changes not attributable to development. We first present a technique to extract drainage channel heads and drainage networks through time utilizing surface tangential curvature and compare its ability to predict field verified channel heads against two other delineation techniques. Second, subwatersheds of the delineated networks were used to summarize topographic changes in the forested and developing watersheds through time. Finally, simplified hydrologic response functions are generated for each time step utilizing width functions augmented by curvature delineated networks. Landscape dissection and hydrologic connectivity increased throughout development, indicating an increased efficiency of water removal from the landscape. Substantial topographic changes manifest as variations in curvature standard deviations through time in first order catchments, highlighting redistributions of high and low gradient regions. Idealized hydrologic responses reflect increased landscape connectivity with a distinct shift toward more rapid storm response shapes throughout the development process. Unexpectedly high temporal variance observed in the forested watershed across analyses indicates inherent differences in digital topographic datasets collected on different dates. Results of this thesis highlight fundamental geomorphic changes caused by development practices while developing methods and rationale for temporal DEM-based geomorphic studies.