ABSTRACT The geology of Great Smoky Mountains National Park (GRSM) in Tennessee and North Carolina is dominated by siliciclastics and metamorphic strata. However, in the western portion of GRSM, a series of carbonate fensters (windows) expose the Lower Ordovician–age section of the Knox Group, a series of dolomite and limestone units that are partially marbleized as a result of contact metamorphism from the Great Smoky fault. The fensters create opportunities for allogenic recharge to occur at points along the contact of the surrounding insoluble strata with the underlying soluble carbonates. The combination of chemically aggressive surface recharge and vertical relief has resulted in the formation of deep caves, many of which have active streams and water resources. Though the karst is limited in extent and the number of caves is fairly small, the significance of the resources is substantial, with several of the caves in the area over 150 m in depth and at least two being major bat hibernacula. In 2017, the U.S. Geological Survey (USGS) began a study to better understand the hydrologic behavior of these karst systems through hydrologic and geochemical monitoring, groundwater tracing using fluorescent dyes, and seepage runs. Stage and water-quality instrumentation was installed in two caves in GRSM, the main stream of Bull Cave, and in a sump pool in Whiteoak Blowhole, at 173 m and 70 m below land surface, respectively. Following setup of the cave sites, dye injections were conducted to determine discharge points for four of the deep cave systems on Rich Mountain and Turkeypen ridge. Results show water in these systems has an extremely rapid travel time, with tracers detected from caves to springs in less than 24 h for each of the systems. This field guide describes the complex geology, regional hydrogeology, and unique landscape characterized by high-gradient subterranean streams, carbonate fensters, and deep caves of the GRSM karst.

The southern Appalachian crystalline core is composed of lithotectonic assemblages that are largely sedimentary in origin. Sixteen paragneiss samples from the Blue Ridge and Inner Piedmont of North Carolina and Georgia, and one sample of Middle Ordovician rocks from the Sevier-Blountian clastic wedge in the Tennessee Valley and Ridge were sampled for sensitive high-resolution ion microprobe (SHRIMP) U-Pb detrital zircon geochronology, whole-rock geochemistry, and zircon trace-element analyses. Detrital zircon ages range from Archean (~2.7 Ga) to Middle Paleozoic (~430 Ma), with a notable abundance of Mesoproterozoic zircons (1.3–0.9 Ga). Many samples also contain moderate populations of slightly older Mesoproterozoic zircons (1.5–1.3 Ga). Minor populations of Paleoproterozoic (2.3–1.5 Ga) and Neoproterozoic (754–717 and 629–614 Ma) ages occur in several samples; however, Paleozoic detrital zircons (478–435 Ma) are restricted to samples from the Cat Square terrane. Depositional periods of the metasedimentary terranes are bracketed by detrital zircon, metamorphic, and magmatic ages, and include: (1) Mesoproterozoic, (2) Neoproterozoic to early Paleozoic, and (3) middle Paleozoic. A xenolith from the ~1.15 Ga Wiley Gneiss suggests a post–~1.2 Ga period of sedimentation prior to the ~1.15 Ga Grenvillian magmatic event. Detrital zircon populations of Neoproterozoic to Middle Ordovician suggest a mixed Laurentian provenance with Amazonian and peri-Gondwanan sources deposited in divergent and convergent plate settings. Blue Ridge and Inner Piedmont detrital zircon ages, whole-rock geochemistry, lithologic assemblages, and field relationships are compatible with deposition of immature clastic material in a rift and passive-margin setting from the Neoproterozoic to early Paleozoic. Occurrence of 1.3–0.9 Ga, 1.5–1.3 Ga, and 754–717 Ma detrital zircon ages indicate a dominantly Laurentian provenance for the Cartoogechaye, Cowrock, Dahlonega gold belt, Smith River allochthon, and Tugaloo terranes. Minor Paleoproterozoic populations in these terranes suggest input from distal terranes of the Laurentian midcontinent or the Amazonian craton. Transition to a convergent plate margin in the Middle Ordovician resulted in collision of central Blue Ridge and Tugaloo terranes and recycling of material from these terranes into the Mineral Bluff Formation and Sevier Shale. Ordovician and 629–614 Ma detrital zircons from the Cat Square terrane document the first occurrence of peri-Gondwanan material, which was deposited in a convergent setting between the Laurentian margin and the accreting Carolina superterrane during the Late Silurian to Devonian.

The deformed wedge of Paleozoic sedimentary rocks in the southern Appalachian foreland fold-thrust belt is defined by the configurations of the undeformed basement surface below and the base of the Blue Ridge–Piedmont megathrust sheet above, together with the topographic free surface above the thrust belt. The base of the Blue Ridge–Piedmont sheet and undeformed basement surface have been contoured using industry, academic, and U.S. and state geological survey seismic-reflection and surface geologic data. These data reveal that the basement surface dips gently SE in the Tennessee embayment from Virginia to Georgia, and it contains several previously unrecognized normal faults and an increase in dip on the basement surface, which produces a topographic gradient. The basement surface is broken by many normal faults beneath the exposed southern Appalachian foreland fold-thrust belt in western Georgia and Alabama closer to the margin and beneath the Blue Ridge–Piedmont sheet in Georgia and the Carolinas. Our reconstructions indicate that small-displacement normal faults form beheaded basins over which thrust sheets were not deflected, whereas large-displacement normal faults (e.g., Tusquittee fault) localized regional facies changes in the early Paleozoic section and major Alleghanian (Permian) structures. These basement structures correlate with major changes in southern Appalachian foreland fold-thrust belt structural style from Virginia to Alabama. Several previously unrecognized structures along the base of the Blue Ridge–Piedmont sheet have been interpreted from our reconstructions. Large frontal duplexes composed of rifted-margin clastic and platform rocks obliquely overridden along the leading edge of the Blue Ridge–Piedmont sheet are traceable for many kilometers beneath the sheet. Several domes within the Blue Ridge–Piedmont sheet also likely formed by footwall duplexing of platform sedimentary rocks beneath, which then arched the overlying thrust sheet. The thickness and westward limit of the Blue Ridge–Piedmont sheet were estimated from the distribution of low-grade foot-wall metamorphic rocks, which were observed in reentrants in Georgia and southwestern Virginia, but are not present in simple windows in Tennessee. These indicate that the original extent of the sheet is near its present-day trace, whereas in Georgia, it may have extended some 30 km farther west. The southern Appalachian foreland fold-thrust belt consists mostly of a stack of westward-vergent, mostly thin-skinned thrusts that propagated westward into progressively younger units as the Blue Ridge–Piedmont sheet advanced westward as a rigid indenter, while a few in northeastern Tennessee and southwestern Virginia involved basement. Additional boundary conditions include temperatures <300 °C and pressures <300 MPa over most of the belt. The southern Appalachian foreland fold-thrust belt thrusts, including the Blue Ridge–Piedmont megathrust sheet, reach >350 km displacement in Tennessee and decrease both in displacement and numbers to the SW and NE. Much of the Neoproterozoic to Early Cambrian rifted-margin succession was deformed and metamorphosed during the Taconic orogeny, and it is considered part of the rigid indenter. Only the westernmost rocks of the rifted-margin succession exhibit ideal thin-skinned behavior and thus are part of the southern Appa-lachian foreland fold-thrust belt. Palinspastic reconstructions, unequal thrust displacements, and curved particle trajectories suggest that deformation of the belt did not occur by plane strain in an orogen that curves through a 30° arc from northern Georgia to SW Virginia. Despite the balance of many two-dimensional cross sections, the absence of plane strain diminishes their usefulness in quantifying particle trajectories. Coulomb behavior characterizes most individual faults, but Chapple's perfectly plastic rheology for the entire thrust belt better addresses the particle trajectory problem. Neither, however, addresses problems such as the mechanics of fault localization, out-of-sequence thrusts, duplex formation, three-dimensional transport, and other southern Appalachian foreland fold-thrust belt attributes.