Jennifer Eccles

Drilling reveals fluid control on architecture and rupture of the Alpine fault, New Zealand

Rock damage during earthquake slip affects fluid migration within the fault core and the surrounding damage zone, and consequently coseismic and postseismic strength evolution. Results from the first two boreholes (Deep Fault Drilling Project DFDP-1) drilled through the Alpine fault, New Zealand, which is late in its 200–400 yr earthquake cycle, reveal a >50-m-thick “alteration zone” formed by fluid-rock interaction and mineralization above background regional levels. The alteration zone comprises cemented low-permeability cataclasite and ultramylonite dissected by clay-filled fractures, and obscures the boundary between the damage zone and fault core. The fault core contains a <0.5-m-thick principal slip zone (PSZ) of low electrical resistivity and high spontaneous potential within a 2-m-thick layer of gouge and ultracataclasite. A 0.53 MPa step in fluid pressure measured across this zone confirms a hydraulic seal, and is consistent with laboratory permeability measurements on the order of 10?20 m2. Slug tests in the upper part of the boreholes yield a permeability within the distal damage zone of ?10?14 m2, implying a six-orders-of-magnitude reduction in permeability within the alteration zone. Low permeability within 20 m of the PSZ is confirmed by a subhydrostatic pressure gradient, pressure relaxation times, and laboratory measurements. The low-permeability rocks suggest that dynamic pressurization likely promotes earthquake slip, and motivates the hypothesis that fault zones may be regional barriers to fluid flow and sites of high fluid pressure gradient. We suggest that hydrogeological processes within the alteration zone modify the permeability, strength, and seismic properties of major faults throughout their earthquake cycles.

Extreme hydrothermal conditions at an active plate-bounding fault

Temperature and fluid pressure conditions control rock deformation and mineralization on geological faults, and hence the distribution of earthquakes. Typical intraplate continental crust has hydrostatic fluid pressure and a near-surface thermal gradient of 31 ± 15 degrees Celsius per kilometre. At temperatures above 300–450 degrees Celsius, usually found at depths greater than 10–15 kilometres, the intra-crystalline plasticity of quartz and feldspar relieves stress by aseismic creep and earthquakes are infrequent. Hydrothermal conditions control the stability of mineral phases and hence frictional–mechanical processes associated with earthquake rupture cycles, but there are few temperature and fluid pressure data from active plate-bounding faults. Here we report results from a borehole drilled into the upper part of the Alpine Fault, which is late in its cycle of stress accumulation and expected to rupture in a magnitude 8 earthquake in the coming decades. The borehole (depth 893 metres) revealed a pore fluid pressure gradient exceeding 9 ± 1 per cent above hydrostatic levels and an average geothermal gradient of 125 ± 55 degrees Celsius per kilometre within the hanging wall of the fault. These extreme hydrothermal conditions result from rapid fault movement, which transports rock and heat from depth, and topographically driven fluid movement that concentrates heat into valleys. Shear heating may occur within the fault but is not required to explain our observations. Our data and models show that highly anomalous fluid pressure and temperature gradients in the upper part of the seismogenic zone can be created by positive feedbacks between processes of fault slip, rock fracturing and alteration, and landscape development at plate-bounding faults.

Fault Zone Guided Wave generation on the locked, late interseismic Alpine Fault, New Zealand

Fault Zone Guided Waves (FZGWs) have been observed for the first time within New Zealands transpressional continental plate boundary, the Alpine Fault, which is late in its typical seismic cycle. Ongoing study of these phases provides the opportunity to monitor interseismic conditions in the fault zone. Distinctive dispersive seismic codas (~7–35 Hz) have been recorded on shallow borehole seismometers installed within 20 m of the principal slip zone. Near the central Alpine Fault, known for low background seismicity, FZGW-generating microseismic events are located beyond the catchment-scale partitioning of the fault indicating lateral connectivity of the low-velocity zone immediately below the near-surface segmentation. Initial modeling of the low-velocity zone indicates a waveguide width of 60–200 m with a 10–40% reduction in S wave velocity, similar to that inferred for the fault core of other mature plate boundary faults such as the San Andreas and North Anatolian Faults.

Bedrock geology of DFDP-2B, central Alpine Fault, New Zealand

ABSTRACT During the second phase of the Alpine Fault, Deep Fault Drilling Project (DFDP) in the Whataroa River, South Westland, New Zealand, bedrock was encountered in the DFDP-2B borehole from 238.5–893.2 m Measured Depth (MD). Continuous sampling and meso- to microscale characterisation of whole rock cuttings established that, in sequence, the borehole sampled amphibolite facies, Torlesse Composite Terrane-derived schists, protomylonites and mylonites, terminating 200–400 m above an Alpine Fault Principal Slip Zone (PSZ) with a maximum dip of 62°. The most diagnostic structural features of increasing PSZ proximity were the occurrence of shear bands and reduction in mean quartz grain sizes. A change in composition to greater mica:quartz + feldspar, most markedly below c. 700 m MD, is inferred to result from either heterogeneous sampling or a change in lithology related to alteration. Major oxide variations suggest the fault-proximal Alpine Fault alteration zone, as previously defined in DFDP-1 core, was not sampled.

Petrophysical, Geochemical, and Hydrological Evidence for Extensive Fracture-Mediated Fluid and Heat Transport in the Alpine Fault's Hanging-Wall Damage Zone

Fault rock assemblages reflect interaction between deformation, stress, temperature, fluid, and chemical regimes on distinct spatial and temporal scales at various positions in the crust. Here we interpret measurements made in the hanging-wall of the Alpine Fault during the second stage of the Deep Fault Drilling Project (DFDP-2). We present observational evidence for extensive fracturing and high hanging-wall hydraulic conductivity (∼10−9 to 10−7 m/s, corresponding to permeability of ∼10−16 to 10−14 m2) extending several hundred meters from the faults principal slip zone. Mud losses, gas chemistry anomalies, and petrophysical data indicate that a subset of fractures intersected by the borehole are capable of transmitting fluid volumes of several cubic meters on time scales of hours. DFDP-2 observations and other data suggest that this hydrogeologically active portion of the fault zone in the hanging-wall is several kilometers wide in the uppermost crust. This finding is consistent with numerical models of earthquake rupture and off-fault damage. We conclude that the mechanically and hydrogeologically active part of the Alpine Fault is a more dynamic and extensive feature than commonly described in models based on exhumed faults. We propose that the hydrogeologically active damage zone of the Alpine Fault and other large active faults in areas of high topographic relief can be subdivided into an inner zone in which damage is controlled principally by earthquake rupture processes and an outer zone in which damage reflects coseismic shaking, strain accumulation and release on interseismic timescales, and inherited fracturing related to exhumation.

Accounting for Modelling Errors in Parameter Estimation Problems: The Bayesian Approximation Error Approach

Many parameter estimation problems are highly sensitive to errors. The Bayesian framework provides a methodology for incorporating these errors into our inversion. However, how to characterise the errors in a way that can be efficiently utilised remains a problem in many inversions. Recently the Bayesian approximation error method has been utilised as a systematic way of characterising errors that arise from inaccuracies in the model. We describe the Bayesian approximation error method and demonstrate its use in a homogenisation example. In this example, it is shown that the coarse scale homogenised parameter can be estimated by accounting for the significant modelling error using the Bayesian approximation error method. This modelling error arises from inverting using a model that does not account for the fine scale and has a coarse finite element discretisation.