Paleoseismicity, rates of active deformation, and structure of the Lake Jasper pull-apart basin, Awatere Fault, New Zealand

Abstract

The Awatere Fault, located in northeast South Island last ruptured during the –Mw7.5 Marlborough Earthquake of 1848. Although this remains the second most severe event to have struck Wellington in historical times, the Awatere fault has been little studied relative to other major faults in the region. This study addresses shortfalls in our knowledge of the active deformation, and the magnitude and frequency of earthquakes on the Awatere Fault. A detailed study of landforms offset by the fault indicates the rate of deformation occurring on the fault has been constant for at last the past 65,000 years. Two trenches excavated across the fault revealed evidence for at least 7 large earthquakes during the past 8500 years. The spacing of events is non-uniform, ranging between 600 and 2500 years. Careful analysis of sediments deposited in Lake Jasper, a small depression produced by a bend in the Awatere Fault trace, the fault has ruptured 14 times in the past 15,000 years. Abrupt changes in the abundance and of Diatoms, an environmentally sensitive algae, contained in the lake sediments reflect large disruptions to the local environment following a large earthquake. The size of displacements that have accompanied the last two earthquakes indicates each was of a similar magnitude.

---

Technical Abstract

The results of a detailed investigation of Lake Quaternary paleoseismicity and deformation along the easternmost part of the Awatere Fault are presented. Field mapping of the fault trace, trenching of the fault scarp, and meticulous analysis of lake sediments provide a dataset that greatly enhances our understanding of the neotectonics of this major strike-slip fault.

Detailed mapping of the active fault trace and measurement of the offset of fault displaced geomorphic features provides evidence for significant spatial and temporal patterns in the development of the Awatere Fault. Important new slip-rate data show that: (1) southwest of Dumgree Station, -12km inland from the coast, the rate of strike-slip on the Awatrere Fault has been 5-6mm.yr-1 for at least the past 65ka; (2) there has been no significant change in either the sense or rate of the Late Quaternary dip-slip on the fault; and (3) northeast of Dumgree Station, the rate of strike-slip on the Awatere Fault decelerates rapidly to <1.5mm.yr-1 near the coast. There is good evidence indicating that on some parts of the Awatere Fault, the vertical component of oblique-slip is accommodated off-fault on adjacent structures. A model is offered that invokes topographic perturbation of near-surface stresses as the mechanism controlling the development of a distinctive small-scale segmentation along portions of the Awatere Fault trace.

Two fault trenches were excavated across sag ponds that have formed along the fault scarp immediately northeast of the Lake Jasper pull-apart graben. The first trench yielded stratigraphic evidence for at least six surface rupturing events since 8330-8610 cal yr B.P., whilst the second trench preserved a record of four events since 4550-5300 cal yr B.P. The youngest recognised event is correlated with the 1848 Marlborough Earthquake (~Mw7.5), while the penultimate event is thought to have occurred at 910-1180 cal yr B.P. The maximum coseismic slip during these two events was apparently 6-7m, with both probably involving slip of the entire eastern section of the Awatere fault (>100 km). Future ruptures of the eastern section might be expected to be of an equivalent extent and magnitude.

A set of criteria for the use of diatom biostratigraphy in paleoseismic analysis of lacustrine sediments has been developed, and applied to sediment cores recovered from Lake Jasper. Distinctive changes in sedimentation and diatom assemblages are interpreted as reflecting abrupt co-seismic changes to the depth of and chemical environment within the fault bounded basin. When calibrated in time using 11 14C dates, the lake sediments preserve a potentially continuous paleoseismic record spanning the last 15,000 years. Fourteen coseismic fault ruptures are inferred, of which 7 are correlated with events recognised in the two trenches. It is possible that one or more of the sequences attributed to rupture of the basin bounding (Awatere) fault could have been storm induced or a response shaking during earthquakes on other large faults. Combined with results from the trenching studies, the resulting paleoseismic record implies a mean earthquake return time of ~1000 years, considerably shorter than might be inferred from the trench data alone (~1200-1400 years).

The geodetic strain-rate measured over a small (<10 km) survey network straddling the Awatere Fault is shown to be indistinguishable from strain rates determined from repeated GPS surveys of regional scale networks (>50 km).