Neogene Tectonics of Raukumara Peninsula, Northern Hikurangi Margin, New Zealand

Abstract

The Raukumara Peninsula lies in the forearc of the Hikurangi subduction zone at the boundary between the Australian and Pacific plates. Geodetic strain and active faulting show that eastern Raukumara Peninsula is currently extending at right-angles to the plate margin. Recent bathymetric mapping shows that the frontal wedge of Raukumara Peninsula is non-accretionary. Seamounts have collided with the frontal wedge, steepening the wedge toe and causing it to collapse into the Hikurangi Trough. I have used existing structural mapping in eastern Raukumara Peninsula, together with paleomagnetic sampling across the Waerenga-o-Kuri Fault, to constrain the kinematics of geologic structures and to construct cross-sections across the region. Northeast striking, active normal faults run parallel to north-east trending folds, and north-west to west striking dextral-normal faults cut across the north-easterly structural grain. Absence of reverse faults suggests that the north-east trending open folds probably formed by tilting and antithetic faulting between normal faults. A low-angle normal fault (the Whakoau Fault) forms western boundary to the region of both active normal faulting and rapid geodetic strain. This fault is probably a detachment beneath the Tertiary succession in eastern Raukumara Peninsula, on which the Tertiary rocks slide down-slope towards the Hikurangi Trough. The detachment fault follows a unit of smectite-rich mudstone within or at the base of the East Coast Allochthon and normal faults in the overlying rock mass sole into the decollement. Seismic reflection profiles and geologic cross-sections show progressive rotation of the dip of bedding with increasing age since the Late Miocene, implying that extension on margin-parallel normal faults has been taking place for at least 5 - 10 Ma. In many places, diapiric intrusion of smectitic clay up fault-planes has folded strata into tight anticlines, which are separated by broad synclines. Diapirism began in the mid - late Miocene and widespread mud volcanism continues today. The western part of Raukumara Peninsula is experiencing broad antiformal uplift, reaching a peak (3 mm/y over 125 ky) along the crest of the Raukumara Range, arcward of a belt of much slower uplift (0.5 - 1.0 mm/y) between the range front and the east coast. Along the east coast, terraces east of the Gable End Fault Belt have been raised at 2 - 4 mm/y. Active faults segment the east and west coasts of Raukumara into 20 - 30 km-wide domains of differing rates of uplift Rapid uplift of the Raukumara Range is believed to be a result of sediment subduction and underplating at the base of the crust South-eastward tilting of thrust faults at the base of the East Coast Allochthon by 10 – 30° gives an estimate of approximately 6 Ma for initiation of uplift, if Late Quaternary rates of tilting are extrapolated back into the past. Extensive new paleomagnetic sampling along the east coast of Raukumara Peninsula constrains the boundary between the unrotated Raukumara and clockwise-rotated Wairoa domains of rotation to lie between Gisborne and Tolaga Bay. My results confirm that northern Raukumara Peninsula has not rotated with respect to the Australian plate, while southern Raukumara Peninsula has rotated about a vertical axis at a rate of 2 - 3 °/Ma since the early Miocene, increasing to 7 °/Ma in the last 5 Ma. It is not clear how the relative rotations are accommodated structurally. By comparing horizontal-angle triangulation dating from 1875 - 1986 and a 1995 GPS survey, I have calculated geodetic strain across Raukumara Peninsula using least-squares regression of observations, stations coordinates and strain parameters. Rates of maximum shear strain vary spatially on a regional scale (rapid south-easterly extension in the Tertiary rocks of eastern Raukumara at a rate of (1.5 ± 0.6) x 10-7 /y over the last 120 years contrasts with negligible strain in western Raukumara) and on the scale of faults (around Poverty Bay, areas of higher than average strain coincide with active faults). In both the Gisborne and the Tolaga Bay areas, the direction of relative extension has changed by 90° in less than 100 years, switching between extension at right-angles to the margin and margin-parallel extension or margin-normal contraction. These changes in the direction of strain in eastern Raukumara Peninsula may result from local variations in porosity or fluid pressure in the clays along the decollement at the base of the Tertiary sequence, causing changes in the effective viscosity of the decollement layer. Such changes could lead to differential movement of the overlying rock mass, resulting in reversals in the direction of geodetic extension at the surface of the rock-slide. A microearthquake survey of part of Raukumara Peninsula was carried out during May 1993. I deployed five digital seismographs between Gisborne and Matawai, spaced about 15 km apart. Only four shallow events (≤ 20 km depth) were recorded during the 21-day survey, a very low rate of microseismicity, implying that geodetic deformation is not accumulating seismically. Summation of scalar moments of earthquakes on the subduction thrust and in the frontal wedge of Raukumara Peninsula, shows that seismicity from 1964 - 1993 has accounted for only 0.2% of the relative plate motion (if the plate convergence were dissipated entirely seismically, at least one earthquake of Mw 7 would occur every four years, or a Mw 8 quake every 130 years). Moreover, the central Hikurangi forearc of south-eastern North Island, where the plate interface is strongly coupled, has a frequency of microearthquakes around eight times greater than Raukumara. Although the 30-year sampling window may be too short to include the largest earthquakes, the seismic strain deficit strongly suggests that the relative motion of the Pacific and Australian plates is not being taken up seismically, presumably because the subduction interface is decoupled beneath Raukumara Peninsula and slip on the interface is largely accomplished by creep. I have estimated strain from geologic structures at all scales (including faults ranging in slip from 5 cm to 5 km) in eastern Raukumara Peninsula. The 637 faults measured in outcrop, on seismic lines and on geologic cross-sections appear to follow a scaling law with power-law exponent D = 1.105 for displacements between 0.01 - 100 m and D = 0.5758 for displacements between 100 - 10 000 m, which implies that both very small and very large faults contribute significantly to strain. Strain estimated from mesofaults (< 10 m offset) measured in Late Miocene rocks in cliffs round the east coast of Raukumara Peninsula from Poverty Bay to East Cape (according to the method of Peacock & Sanderson, 1993, see Chapter 6) gives a rate of maximum engineering shear strain of (3.6 ± 1.1) x 10-2 and an azimuth of relative extension of 123 ± 30°. Restoration of balanced geologic cross-sections by unfaulting and antithetic, inclined shear indicates a maximum shear strain on macrofaults (100 - 10 000 m slip) of 0.20 ± 0.14 with extension azimuth 097°. Strain for faults in the intermediate size range was derived using the power-law fault-size scaling relationship. Total strain preserved in the rocks since the Late Miocene (6 Ma, the minimum age for the deformation) is approximately (2.9 ± 2.0) x 10-1, which equates to a maximum shear strain rate of (4.8 ± 3.3) x 10-8 /y (with azimuth of relative extension = 100°), only 30% of the 120-year estimate of regional geodetic strain. This suggests that geodetic strain is currently accumulating elastically and less than 30% will be preserved as permanent, near-surface deformation of the rocks on faults and folds, the remaining 70% being released as slip on the basal decollement of the rock-slide. Critical wedge models and scarcity of seismicity in the frontal wedge suggest that the frontal wedge is not deforming at depth, whereas geodesy and geologic structures indicate that the upper few kilometres of the frontal wedge are extending rapidly towards the south-east. I propose that the surficial layers of the wedge in eastern Raukumara Peninsula are decoupled from the underlying Cretaceous rocks along a 100 - 500 m thick zone of detachment in Paleogene smectitic mudstone at 0 - 3 km depth. Surficial extension can be modelled as a rock-slide of the upper 3 km of the crust, driven by gravity in the direction of south-easterly surface gradient which is maintained by uplift of the inner forearc. Changes in thickness and hence sliding velocity give rise to internal deformation of the rock-slide. Temporal variation of geodetic strain orientation and magnitude may be a result of variations in the local porosity or fluid pressure within the decollement layer, causing changes in its effective viscosity. This could lead to temporarily changing, lateral displacement gradients and strain in the overlying rock-slide. According to the model, the effective viscosity of the basal decollement layer is estimated to be 1.8 x 10 18 Pa s.