Seismotectonic Study of the Central Alpine Fault Region South Island, New Zealand

Abstract

This thesis deals with the seismotectonics of the central Alpine Fault region, about equal weight being given to seismology and to tectonics. It was found that the Alpine Fault is a seismic boundary separating relatively high seismic activity on the southeast side from low activity on the northwest side, and that the number of epicentres decreases exponentially with increasing distance from the fault on both sides. On the northwest side, most of the epicentres can be accounted for by epicentral location error. The low activity is considered to be due to the relatively strong massive old rocks. On the southeast side, the relatively high activity is considered to be due to a high uplift rate and the weak rocks. The high uplift rate results in a high geothermal gradient and consequently reduces the thickness of the brittle layer of the crust and the overall strength of the crust. Thus a high-strain zone and relatively high seismic activity is expected. The exponential decrease away from the fault is unexplained. The b-value in the Central Region, possibly because of its lower stress level, is larger than those in the regions to the northeast and southwest. Composite focal mechanisms for about 30km .southeast of the Alpine Fault trace are predominantly thrust; further southeast they become predominantly strike-slip. Both are in general agreement with the known active faulting. There is a fairly consistent compression axis almost horizontal and striking southeast for both types of focal mechanisms. It is shown that the Kakapo Fault is a major active dextral fault that branches from the Hope Fault and extends WSW for about 120km to the Alpine Fault. It is younger but similar in most respects to the other Marlborough Faults. The fault is only about 250kyr old, has an average dextral rate of 6.8mm/yr for the last 20Kyr, and has a vertical faulting rate (up on the north side) of l-1.4mm/yr for the last 14Kyr as indicated by displaced lake terraces at Lake Sumner. A dextral drag rate of 34-40mm/yr within 15km southeast of the Alpine Fault is suggested by a 10km bend in the 250Kyr-old Kakapo Fault and by a similar 34km bend in the 1Myr-old Hope Fault. It is shown that provided a region is uniformly mountainous, contours of landslide density are analogous to earthquake isoseismals. By landslide contours it is shown that the 1929 Arthurs Pass earthquake was caused by movement on the Kakapo Fault. An earlier earthquake and a 500yr recurrence interval is inferred from a 500yr-old landslide at Lake Minchin. Previously for the eight major South Island earthquakes since 1826 five were associated with faulting and now the number is six. In contrast with the only published explanation, the dextral movement on the Marlborough Faults is considered to be transformed into dextral movement on the Alpine Fault and into narrowing across the Southern Alps. The rates of both progressively increase to the southwest as the Marlborough Faults successively join the Alpine Fault. Continuing narrowing is accommodated by uplift and erosion and by the growth of the Southern Alps mountain roots. Consequently the Alps are still rising and are not in "steady-state" as has been suggested. The dextral component of the relative Indian/Pacific plate movement is about 42mm/yr in the Central Region. Almost all is taken up in the about 15km width that includes the Alpine Fault and a 15km-wide belt southeast of the fault. About 34mm/yr is drag within the 15km-wide belt which appears to be mostly aseismic. About l0mm/yr is faulting on the Alpine Fault which is interpreted to be associated with major earthquakes. From the average of the most recent 8m dextral displacement, the recurrence interval for major earthquakes is about 800yr. Assuming the brittle crust is about 5km thick and the whole length of the central Alpine Fault breaks in a major earthquake the magnitude will be about Ms7.5.