Evaluation of Land Subsidence and Ground Failures at Bicycle Basin, Fort Irwin National Training Center, California, 1992–2017

Groundwater has been pumped in the Bicycle Basin at Fort Irwin National Training Center since the 1960s, and the amount pumped has generally increased since the 1990s. After a large crack (approximately 0.5-kilometer long) formed at the surface of Bicycle Lake playa during 2005–06 in the area used as an aircraft runway, a monitoring study was initiated by the U.S. Geological Survey, in cooperation with the U.S. Army Fort Irwin National Training Center, to help determine the cause. The extent and effect of groundwater-level declines and land-surface deformation in Bicycle Basin were evaluated using a number of approaches, including water-level measurements from December 2007 to June 2017, land surveys across the playa area, interferometric synthetic aperture radar (InSAR) analyses, geophysical surveys of the playa area, and numerical experiments to test hypotheses about soil mechanical processes. A specific objective of this study was to evaluate the recent development of ground failures in the form of large, surface cracks that pose a hazard to aircraft operations on the Bicycle Lake playa airstrip. 

Another large crack and set of interconnected cracks formed at the surface of Bicycle Lake playa in 2013 following a period of inundation similar to that of 2005–06; however, the 2013 cracks formed a network rather than a single feature. Groundwater pumping resulted in more than 27 meters (90 feet) of water-level declines in wells north of the Bicycle Lake playa from 1990 to 2017, and InSAR interferograms indicated more than 400 millimeters (16 inches) of subsidence during 1993–2015. Subsidence rates calculated from InSAR interferograms were variable, temporally and spatially. Results of leveling surveys indicated differential subsidence between 2009 and 2016 along a transect across the playa to the area of maximum subsidence; there was less subsidence south of the 2005–06 crack than north of it. The steepest subsidence gradient for this time was in the area of maximum subsidence. Repeat tape-extensometer measurements from April 2009 to November 2017 across the main-runway crack indicated slight opening along part of the crack. A baseline lidar (light detection and ranging) survey of the main-runway crack was done in January 2009 to track the development of this feature. During the 95-months from January 2009 to December 2016, 0.5 meters (19.7 inches) of subsidence was observed next to the subsidence pit in the crack along the western end of the scanned area, presumably due to erosion and backfilling when the lake was inundated. Standing water in the crack along the eastern end of the scanned area prevented determining change for that part of the crack. Time-series, shaded-relief images show the progression of change in the crack from January 2009 through December 2016, with the crack closing and opening, depending on the time of most recent inundation relative to when the scans were taken. Volumetric changes in the size of the crack, calculated between the sequential lidar surveys, show that from 2009 to 2016, the cumulative volumetric change was 1.5 cubic meters (54 cubic feet). This volume is a minimum estimate because the crack volume below the water surface in the pits was not included in the calculations. Subsurface imaging of the 2005–06 main-runway crack by a series of electrical resistivity tomography surveys in 2008 and 2017 indicated that the crack could extend 5 meters (16 feet) in the subsurface. 

Electromagnetic induction surveys in 2008, 2014, and 2015 evaluated the technique for ground-failure monitoring and showed that the technique was effective at revealing anomalies correlated with the features of concern in the playa area, such as the 2005–06 crack, numerous ”healed” macropolygon features, and scattered sink-like depressions. 

Results from numerical experiments simulating water-table decline at depth indicated that the material deposits in the desaturating capillary fringe zone might not be able to transmit large enough stresses up through the overlying soil to cause cracks at the land surface. Results from simulations of desiccation in the presence of a regional tectonically induced stress field, however, tended to support the hypothesis that the combined processes could control the formation of giant desiccation macropolygons in certain areas, such as Bicycle Lake playa.