Canyonlands InSAR

    Current models of fault growth can be grouped into two broad categories, fault growth by radial propagation and fault growth by segment linkage [Biggar and Adams, 1987; Ely, 1987; Trudgill and Cartwright, 1994; Adhikary et al., 1997; Cartwright and Mansfield, 1998]. These varied approaches to fault growth present ambiguous criteria for measuring fault length. The length of a fault can be considered to be the trace length of an individual segment, or the total length of a series of segments that interact mechanically, but do not share a common trace. A meaningful calculation of fault length is critical in determining regional strain history.

Preliminary 75 month interferogram overlaid on shaded relief illuminated from the upper right. Positive changes in line of sight distance (yellows & oranges) are interpreted as ground subsidence, whereas negative changes (blues & purples) are interpreted as uplift. Gray patches represent phase decorrelation between scenes, possibly due to vegetation cover.

    The purpose of this research is to test models of fault growth by radial propagation and segment linkage against InSAR-derived strain measurements from a series of active normal faults within Canyonlands National Park, Utah. The work will provide an independent test of these models and compliment on-going UNR-Geomechanics research into fault propagation mechanics at the Canyonlands.

FAULT PROPAGATION MECHANICS

    A fault growing by radial propagation maintains a constant ratio between maximum slip and maximum length regardless of its sense of slip, either dip- or strike-slip. Through the displacement to length (D/L) ratio, a known magnitude of fault slip can be systematically related to a predicted increase in fault length.

    The segment linkage model incorporates radial propagation with interaction and linkage of adjacent fault segments. Initially, isolated fault segments grow by radial propagation through the D/L relationship. With sufficient displacement, the isolated faults lengthen and impinge on the stress fields of adjacent segments. These stress field perturbations lead to mechanical interactions between overlapping segment tips. Consequently, further in-plane propagation of the individual segments is hindered and orthogonal propagation toward adjacent segments is promoted. Ultimately, the fault tips breach the step-over barrier between the two segments, forming a through-going fault. Thus, faults growing by segment linkage lengthen through a sequence of radial propagation and subsequent mechanical interaction and linkage with adjacent segments.

Looking down the axis of a graben. Note the downward displacement of the red sandstone unit in the center of the photo.

    Mechanical interaction between overlapping segments is the vital step in fault growth by segment linkage. Overlapping segments must interact mechanically to transition from radial propagation to through going fault linkage. Mechanical interactions between overlapping segments is manifest in asymmetric along-strike profiles of fault displacement. In the absence of mechanical interactions (i.e., during radial propagation), maximum displacement ideally occurs at the midpoint of an isolated fault segment and decreases to zero at the end points. Along interacting segments, the point of maximum displacement is skewed toward the zone of overlap between adjacent faults. This strain transfer to the overlap regions drives faulting of the step-over and through going segment linkage.

    Slip along a fault produces systematic elastic surface displacements [Pollard and Segall, 1987]. The amount strain at the surface is a function of the magnitude of slip along the fault. Therefore, the slip distribution along a fault can be inferred from post-seismic strain measurements. Accordingly the radial propagation and segment linkage models can be tested by measuring surface strain above active faults, then inferring the slip distribution along the fault plane. Faults growing by radial propagation should show surface displacements consistent with an elliptical slip distribution, while faults growing by segment linkage should show surface displacements consistent with a non-elliptical slip distribution.

Fresh ground cracking in alluvium along the floors of some graben suggest active surface deformation, potentially detectable by InSAR.

CANYONLANDS

    Present day cumulative offsets of active normal faults within the Needles District of Canyonlands National Park, Utah, are key lines of evidence supporting the radial propagation and segment linkage models. The Canyonlands faults are expressed as master and antithetic graben-bounding faults that define a 10 km by 18 km fault zone along the Colorado River. The grabens are typically 25 m - 100 m deep and 100 m - 400 m wide [McGill and Stromquist, 1975]. The floors of the grabens are sub-horizontal and are covered by loose Quaternary sediment [Huntoon et al., 1982; Biggar, 1986] and low-lying vegetation. Graben walls are typically unvegetated near-vertical exposures of Permian sandstone. The grabens range from 100 m to 6 km in length and are generally spaced 700 m to 1 km apart [Moore and Schultz, 1999]. The grabens are slightly arcuate and trend generally at 10 - 50 degrees east of north, with the more northerly-trending grabens located within the northern half of the fault zone. Ages of Quaternary sediment obtained by Biggar and Adams [1987] suggest that faulting at the Canyonlands began 60 - 65 k.y.a. Based on an estimated 25% strain across the 10 km wide fault zone, Moore and Schultz [1999] proposed an average strain rate for the Canyonlands graben of 1.5 - 2 cm/yr.

    The Canyonlands present favorable observing conditions for InSAR detection of surface deformation. Normal fault displacements are directed perpindicular to the trend of the fault. The north-northwest trend of the graben (bounding faults) will lead to west-northwest-directed displacements complimentary to the ERS 1 & 2 SAR look direction. These displacements are in a favorable orientation for detection. The estimated strain rates of 1.5 - 2 mm/yr [Moore and Schultz, 1999] are well within the detectable threshold of InSAR, given that ERS observations of the Canyonlands have exceeded a period of eight years. The graben lie within fairly flat terrain. The regional slope of the Needles District is about 4 degrees to the west or less [Moore and Schultz, 1999]. Furthermore, the Canyonlands are located within an extremely arid region (southern Utah). Consequently vegetation cover and atmospheric water vapor content are minimal.

graben are bound by high-ange normal faults. Graben floors are commonly filled with loose sand hosting low-lying vegetation.

    Leveling surveys within the Canyonlands are severely limited in aerial extent and repeat interval due to the difficult terrain (e.g., near-vertical graben walls) and arid climate. A long period of observation, a high spatial resolution, and a high sensitivity to surface change required in detecting deformation at the low predicted strain rates of 1.5 - 2 cm/yr. The continuous spatial and temporal ERS SAR coverage of the Canyonlands is well suited for this task. The ERS 1 & 2 SAR radiometers have provided continuous observations of the Canyonlands from 1993 to the present, an observation period of almost eight years. The typical 20m/pixel spatial resolution of the ERS SAR imagery, as well as the cm-scale change detection threshold of InSAR, will adequately characterize the distribution of strain along the 100 m to 6 km-long faults.

A brief stop to chat with our friends from Statoil.

    The project will provide a significant contribution to current hypotheses of fault growth by testing the segment linkage model against strain distribution for individual fault slip events. Previous workers have shown overlapping interactions amongst the Canyonlands faults using measurements of total scarp height [Schultz and Moore, 1996; Moore and Schultz, 1999]. These measurements indicate overlapping interactions for the cumulative geologic displacement across the faults, but not for individual slip events. The project will use InSAR to measure overlapping interactions on a per-event basis. Observations of overlapping interactions for individual slip events will provide a test of the segment linkage model on a per-event timescale and supplement the cumulative geologic observations.

    Many thanks to Dr. James Taranik and Don Sawatzky for making the computing resources of the Arthur Brant Laboratory for Exploration Geophysics available to this project. Additional computing resources were graciously provided by Gregg Stefanelli and the W. M. Keck Earth Sciences and Mining Research Information Center. ROI_PAC InSAR processing software provided by JPL. ERS ephemeris provided by the Delft Institute for Earth-oriented Space Research. ERS data provided by the Western North America InSAR Consortium.

Investigators:
Chris Okubo, Richard Schultz; University of Nevada, Reno
Falk Amelung; University of Miami