A revolution in remote sensing, light detection and ranging (LiDAR) laser altimetry swath mapping, reveals the details of topographic features at such high resolution that they have transformed our understanding of tectonic forcing of the shape of the Earth’s surface. Meter-scale DEMs (digital elevation models) capture fault offsets, fault zone structure, off-fault deformation, and landscape properties at microgeomorphic scale, highlighting that the surface faithfully records the complexity and sensitivity of deformation in detail.
New techniques for surface imaging and analysis continue to inform and challenge our understanding of the Earth’s surface evolution. A Mw 7.3 earthquake in the eastern California shear zone (the 1992 Landers earthquake [Sieh et al., 1993]), for example, saw the debut of synthetic aperture radar (SAR) interferometry as a means for measuring the coseismic displacement field of an earthquake (Massonnet et al., 1993). Stunning SAR images revealed both the displacement along the fault, measurements typically made via labor-intensive field methods, as well as off-fault deformation in a 90 × 100 km area surrounding the 80-km-long surface rupture. Whereas the use of landforms as strain markers to characterize active faulting and folding was established decades earlier (e.g., Wallace, 1968), the Landers SAR images demonstrated that the Earth’s surface records earthquake deformation to a greater degree than previously known. Suddenly, time series analysis of surface imagery became a key tool for inversion of coseismic and postseismic data sets, for modeling static stress changes on faults and on neighboring faults, and a host of other geological and geophysical applications (e.g., Fialko et al., 2002).
Light detection and ranging (LiDAR) topographic surveying takes Earth surface imaging to a new level because of its resolution and its wide availability in the public domain. A search in Google Scholar using the key words “light detection and ranging” results in ∼780,000 hits—an astounding number given that the first papers in the Earth sciences to use these data started appearing after ∼1995 (e.g., Krabill et al., 1995). The National Center for Airborne Laser Mapping (NCALM), a National Science Foundation (NSF) data collection and processing facility, came into existence in 2003 (Carter et al., 2007). The applications of LiDAR data seem boundless. LiDAR data are used in fields that include engineering, planning, forestry and ecology, glaciology, geomorphology, and active tectonics. What are these data and why are they so alluring?
The principle underlying LiDAR surveying is relatively straight-forward; a laser rapidly emits light pulses that are reflected back from any object they strike. The travel time between the instrument and the object are used in combination with the instrument location and orientation to determine the absolute position of every object that reflects the light (Ackermann, 1999; Carter et al., 2007; Harding, 2000). LiDAR data used in the Earth sciences are most often collected via either ground-based terrestrial laser scanning surveys (TLS) or airborne laser swath mapping (ALSM) of large tracts. Airborne LiDAR data are collected by affixing a laser to an aircraft, controlling for the aircraft location using differential kinematic GPS (global positioning systems) with on-board and ground-based base stations, and flying a series of overlapping swaths over the survey area (Harding, 2000). Because current lasers send up to 150,000 pulses per second (Carter et al., 2007), ALSM bathes the survey area with laser pulses that produce returns off of any object between the aircraft and the Earth’s surface. Flying over a forest, for example, yields a 3 dimensional set of points that includes returns from the forest canopy, the ground surface, and from any object in between. A “point cloud” data set results from the swath mapping. Most Earth science users utilize the “bare earth” model, which is a very high-resolution digital elevation model (DEM) created from the last returns during a survey—the pulses that traveled the farthest from the scanner, penetrating vegetation, and presumably hitting the ground. An example of the spatial and vertical resolution of high quality data is provided by surveys funded and hosted by the Oregon LiDAR Consortium. The Consortium, formed by the Oregon Department of Geology and Mineral Industries in partnership with a host of public and private entities, requires data collected with eight point-per-square-meter density and <6-cm RMSE (root-mean-square error) vertical accuracy (Fig. 1) (Luccio, 2013; Madin et al., 2010). Data with this resolution generated through the consortium cover more than 16,000,000 acres in Oregon (http://www.oregongeology.org/dogamilidarviewer/; Luccio, 2013).
A recent paper published in Lithosphere by Barth et al. (2012, v. 4, p. 435–448) provides wonderful illustration of the power of high-resolution LiDAR topographic data. Profound lineaments mark the topographic expression of strike slip faults such as the San Andreas fault in California (Wallace, 1990) and the Alpine fault in New Zealand, the focus of the Barth et al. study. Shuttle Radar Topography Mission DEM (30-m resolution) data reveal the remarkable linearity of the Alpine fault (http://photojournal.jpl.nasa.gov/catalog/PIA06661). In contrast, DEMs generated from LiDAR along a 20 km-long reach of the Alpine fault between the Waiho River at Franz Joseph and the Whataroa River to the northeast illustrate the nonlinear, step-like surface expression of the fault at scales of 1–10’s of meters. Whereas the scale-dependent nature of the topology fault is well known (Norris and Cooper, 1995), the unique contribution of the Barth et al. (2012) study is that the new high-resolution data allow for the kinematic relationship between strike- and dip-slip faults to be characterized at the scale of 10’s of meters. Not only did these new data reveal the detailed structural relations along the fault segments, dozens of new fault segments and folds were discovered. The data and analysis suggest that through-going surface rupture of the Alpine fault involves transfer of slip between neighboring strike-slip sections of the fault via dip-slip faulting and associated hanging-wall folding in newly discovered structural wedges.
A few early papers recognized the significance of LiDAR technology in understanding patterns of active faulting, which the Barth et al. (2012) paper adds to the large and rapidly growing body of literature based on LiDAR data. The first group to use ALSM data conducted a LiDAR survey as part of the post-seismic response to the Mw 7.1 Hector Mine earthquake in southern California in 1999 (Hudnut et al., 2002). Hudnut et al.’s (2002) paper represented a breakthrough because they demonstrated that high-resolution topographic surveying using ALSM yielded topographic data with an accuracy and resolution that exceeded what could be generated using traditional methods. Moreover, the comprehensive mapping along the full length of the rupture showed that ALSM-based measurement of offset agreed well with field measurements and allowed for visualization in three dimensions. A second key paper appeared the following year. Haugerud et al. (2003) showed that LiDAR sees through trees and finds previously unknown active fault traces. “High-Resolution Lidar Topography of the Puget Lowland, Washington” found, for the first time, splays of the Seattle fault system, which last had an earthquake ∼1100 years ago (Bucknam et al., 1992), provided topographic control for location of trench sites, and spawned a flurry of papers on the earthquake hazards, structure, and paleoseismic history of numerous faults throughout the Puget Lowland (e.g., Johnson et al., 2004). Finally, the “B4” LiDAR surveyed the southern San Andreas and San Jacinto fault systems in order to provide a high-resolution pre-earthquake topographic data set (Bevis et al., 2005). The southern segment of the San Andreas fault was chosen because it is the only reach of the fault that has not ruptured historically and because more than 400 years have apparently elapsed since the last earthquake (Weldon et al., 2005). Capturing the details of the surface rupture and the topographic change following the next southern San Andreas earthquake was the central objective of the B4 project, an experimental design that foreshadowed emerging cutting-edge research that exploits pre- and post-seismic surveys along faults (Nissen et al., 2012; Oskin et al., 2012).
LiDAR data sources are proliferating and much of the data are in the public domain. The Puget Sound LiDAR Consortium pioneered open access for these data (http://pugetsoundlidar.org/). The Oregon LiDAR Consortium followed the open access model (http://www.oregongeology.org/sub/projects/olc/). The National Center for Laser Altimetry (NCALM) maintains instruments for use by research investigators, has a seed program for graduate student research projects, and a litany of other products and services for the research community (http://www.ncalm.cive.uh.edu/). NCALM, regional consortia, and other LiDAR databases now collaborate with OpenTopography, which is an NSF-supported data warehouse that distributes analysis software, houses tutorials and short courses, and other products in addition to distributing data (http://www.opentopography.org/). The data sets are large, but the analysis programs are becoming increasingly optimized to handle these data. Tools available through OpenTopography allow users to visualize LiDAR point clouds and process bare earth DEMs from the original data. OpenTopography facilitates processing large data sets through the San Diego Super Computer Facility.
To get started learning how to visualize and analyze of LiDAR data, OpenTopography has materials from a number of excellent short courses and tutorials archived on their education and training page (http://www.opentopography.org/index.php/resources/). A particularly good place to start is the short course taught by Ian Madin of the Oregon Department of Geological and Mineral Industries; Ralph Haugerud, U.S. Geological Survey; Chris Crosby, Open Topography, San Diego Supercomputer Center; and Mike Oskin of the University of California, Davis (UC Davis) at the 2009 Annual Meeting of the Geological Society of America. Topics covered in this tutorial include characteristics of a raw data set, DEM creation, visualization strategies, and data acquisition and quality. A second valuable short course geared toward using LiDAR data in active fault studies is the “Imaging and Analyzing Southern California’s Active Faults with High-Resolution Lidar Topography” tutorial from the course held at UC Davis in 2011. In addition to covering DEM creation and visualization, this course introduces the user to LaDiCaoz, which is a software package developed by Ramon Arrowsmith and his students at Arizona State University to quantify lateral displacements across faults from LiDAR DEMs (Zielke and Arrowsmith, 2012). Tools for back slipping and offset reconstruction are one of the notable capabilities unique to LaDiCaoz and of use for making high accuracy, high precision measurements of offset. A third useful course focuses on applications of LiDAR analysis in geomorphology. The “New Tools in Process-Based Analysis of Lidar Topographic Data” short course held in 2010 at the University Corporation for Atmospheric Research (UCAR) by Dorothy Merritts (Franklin and Marshall College) and Noah Snyder (Boston College) included modules for geomorphic analysis, in addition to covering active fault characterization. Topics in geomorphic analysis include extraction of river channel properties and habitat information, techniques for data filtering and smoothing, and change detection strategies.
What are the frontiers of LiDAR analysis in tectonically active landscapes? Clearly, studies of active fault traces such as that in the Barth et al. (2012) study will continue to provide new insight in fault kinematics, fault zone structure and evolution, and fault slip accrued on earthquake and longer timescales. Quantitative analysis of fault offsets at sub-meter scale are now possible along the length of any fault system surveyed with LiDAR, which allows for measurement of slip in the last earthquake event along faults where those data do not exist (Haddad et al., 2011). LiDAR data analyzed with tools such as LaDiCaoz enable new measurement of offset in key historical earthquakes. A study that revisited the 1857 Fort Tejon earthquake on the south-central San Andreas fault, for example, found additional offsets to those from conventional field study and indicated that the coseismic offsets from the last earthquake along the Carrizo Plain were about half the widely accepted values (Fig. 2A–C) (Sieh, 1978; Zielke et al., 2010, 2012). LiDAR data also permit direct measurement of landscape evolution and response times to tectonic perturbations (Hilley and Arrowsmith, 2008), tectonic geomorphologic phenomenon that are only inferred from numerical models at present (Whipple and Meade, 2006). Given that topographic scale is one variable that influences landscape response time (Whipple, 2001), structural-topographic features such as pressure ridges along strike slip faults act as “mini” mountain ranges to explore hillslope, channel, and drainage basin response to tectonic forcing. The great advantage is that drainage basins evolve on length scales of meters to tens of meters and timescales of hundreds to thousands of years. Finally, before and after earthquake differential LiDAR surveys, as anticipated by the B4 experiment along the San Andreas fault system, represents a very exciting frontier. A “before” LiDAR survey along the El Mayor–Cucapah fault fortuitously allowed for a differential analysis of the surface affects associated with a Mw 7.2 earthquake in Baja California in 2010 (Fig. 2D–E) (Oskin et al., 2012). Oskin et al. (2012) used differencing of the before and after point cloud data to characterize both the on- and off-fault deformation resulting from the earthquake. The analysis revealed the subtleties and complexities of the surface rupture as well as blind faulting and distributed deformation that conventional post-earthquake studies were unlikely to resolve. As explained in the Oskin et al. (2012) paper, the pre-earthquake data are of low resolution, data with very accurate elevations despite having low point density. One can only imagine what will be learned when an earthquake happens along a fault with high-resolution before and after LiDAR data.
Several people contributed to this review and are thanked. Trevor Waldien (OSU) created the images in Figure 1. Data used in Figure 1B are from the DOGAMI Oregon LiDAR Consortium. Olaf Zielke provided Figure 2A–C and Mike Oskin provided Figure 2D–E. Ramon Arrowsmith, Eduardo Guererro, Nick Legg, and Chris Madden read and commented on an earlier draft of the manuscript.