ABSTRACTThis field trip examines the results of integrated geologic studies of the 9 August 2020, Mw 5.1 earthquake near Sparta, North Carolina, USA. The earthquake generated ~4 km of coseismic surface rupture of the Little River fault and uplifted a surface area of ~11 km2. The Little River fault is a thrust fault oriented 110–130°/45–70°SW, and mapped fault segments are en echelon with scarp heights from <5–30 cm. The epicenter is in polydeformed rocks of the Ashe and Alligator Back Metamorphic Suites in the eastern Blue Ridge. Bedrock structure formed during multiple Paleozoic orogenies; the regional foliation strikes NE-SW and dips SE (mean orientation 063°/52°SE). Mapping identified late Paleozoic veins and shear zones, a regional joint set striking 330–340° and 250–240°, and brittle faults that cut the Paleozoic foliation. Brittle faults oriented similar to the Little River fault are mapped up to 4 km along strike from the coseismic rupture along Bledsoe Creek valley, and the combined length of the Little River fault system is ~8 km. Paleoseismic trenches across the Little River fault corroborate the reactivation of an older fault by the 2020 earthquake and reveal two events during late Pleistocene (<50 ka). Surficial mapping identified several terrace deposits, including a deposit along Bledsoe Creek that yielded a 26Al/10Be isochron burial age of 0.46 ± 0.13 Ma and overlies a brittle fault, thus constraining the timing of movement of the fault at that location. Paleoliquefaction studies document soft-sediment deformation features in alluvium that may represent paleoseismic events. Collectively, these results highlight long-lived paleoseismicity of the Blue Ridge and that the 9 August 2020 earthquake reactivated an older, suitably oriented brittle fault in the bedrock. The Little River fault is an example of a previously unknown but active fault lying outside of known seismic zones with demonstrated recurrence of paleo-ruptures, raising questions about the assumption that damaging earthquakes are limited to areas of ongoing background seismicity, which is counter to seismic hazard assessments in the eastern United States.Bedrock mapping separates eastern Blue Ridge lithostratigraphy of the Lynchburg Group and Ashe and Alligator Back Metamorphic Suites into separate fault-bound packages juxtaposed over various 1.3–1.0 Ga basement rocks of the northern French Broad massif by the Gossan Lead fault.

Many studies of Earth surface processes and landscape evolution rely on having accurate and extensive data sets of surficial geologic units and landforms. Automated extraction of geomorphic features using deep learning provides an objective way to consistently map landforms over large spatial extents. However, there is no consensus on the optimal input feature space for such analyses. We explore the impact of input feature space for extracting geomorphic features from land surface parameters (LSPs) derived from digital terrain models (DTMs) using convolutional neural network (CNN)‐based semantic segmentation deep learning. We compare four input feature space configurations: (a) a three‐layer composite consisting of a topographic position index (TPI) calculated using a 50 m radius circular window, square root of topographic slope, and TPI calculated using an annulus with a 2 m inner radius and 10 m outer radius, (b) a single illuminating position hillshade, (c) a multidirectional hillshade, and (d) a slopeshade. We test each feature space input using three deep learning algorithms and four use cases: two with natural features and two with anthropogenic features. The three‐layer composite generally provided lower overall losses for the training samples, a higher F1‐score for the withheld validation data, and better performance for generalizing to withheld testing data from a new geographic extent. Results suggest that CNN‐based deep learning for mapping geomorphic features or landforms from LSPs is sensitive to input feature space. Given the large number of LSPs that can be derived from DTM data and the variety of geomorphic mapping tasks that can be undertaken using CNN‐based methods, we argue that additional research focused on feature space considerations is needed and suggest future research directions. We also suggest that the three‐layer composite implemented here can offer better performance in comparison to using hillshades or other common terrain visualization surfaces and is, thus, worth considering for different mapping and feature extraction tasks.