The Quaternary Tectonic Framework of New Zealand and Relationship to Earthquake Hazard. Kelvin R. BERRYMAN *

Transcription

1 Journal of Geography 103(7) The Quaternary Tectonic Framework of New Zealand and Relationship to Earthquake Hazard Kelvin R. BERRYMAN * Abstract New Zealand straddles the boundary between the Australian and Pacific crustal plates. There are three main elements to the Quaternay tectonic setting of New Zealand. They are ; (i) The Kermadec-Hikurangi, west-dipping, subduction zone that extends along the east coast of New Zealand as far south as the Chatham Rise. (ii) In the southwest of the South Island, oceanic crust of the Australian plate is being subducted, at a steep angle, beneath continental crust of the Pacific plate at the Puysegur Trench. (iii) Between the two opposite-dipping subduction zones there is a transform fault system that accommodates relative plate motion between the plates. As a result of the complex tectonic setting there is a wide variety teceonic features, and variety in styles of deformation within New Zealand. The country can be divided into regions of similar tectonic style and rate, and this approach proves useful in identifying the principal components of the tectonic framework and in broadly assessing earthquake hazard. I. Introduction Earthquakes are an integral part of the natural environment in New Zealand, resulting from New Zealand's location at the collision zone between the Australian and Pacific plates- two of the great crustal plates of the earth. With an understanding of the rate of occurrence and size of past earthquakes, the hazard from future earthquakes can be assessed (e. g. Smith and Berryman, 1986), and effective measures to mitigate the effects on the community of earthquake-related hazards can be introduced, Many earthquake mitigation measures are already in place in New Zealand, most notabl y the comprehenisive building code that requires a high level of earthquake resistant design. Earthquakes must have been felt in New Zealand since its earliest colonisation by Maori, but only since about 1840 have records been adequate to make reliable estimates of the location and magnitude of these earthquakes (Fig. 1). This record is incomplete because only the larger earthquakes are likely to have been reported before ca (Eiby, 1968, 1970) due to the sparse population density in the late nineteenth and early twentieth centuries. Known pre-1930 earthquakes are, therefore, located near to centres of relatively high population density. Other smaller earthquakes that occurred in more remote areas will not have been recorded. Magnitude 6 events have been reported reliably since the mid-1940s, and since the establishment of the national seismograph network in 1964, New Zealand earth- * Insii:ute of Geological and Nuclear Sciences P. O. Box 30368, Lower Hutt, New Zealand 799

2 Fig. 1 Major elements of the Australian-Pacific plate boundary and main topographic features in the New Zealand region Stipping represents continental crust. Arrows show motion of the Pacific plate relative to the Australian plate, with lines representing the direction of underthrusting. quakes have been recorded and analysed in a consistent manner to locate all events of magnitude 4.0. The incompleteness and limited time range of the written and instrumental records of large earthquakes reduces their usefulness for long term seismic hazard assessment, particularly as the recurrence of large earthquakes in different regions of New Zealand may range from several hundred years to tens of thousands of years. Therefore, the geologic record of past major earthquakes, preserved as active fault traces and uplifted coastlines, are a valuable additional source of information. In New Zealand only moderate to large earthquakes (M `6.5) are associated with ground surface rupture, because of the relatively thick (Ca km) seismogenic crust. Fault traces preserve a record of the location and magnitude of the largest shallow earthquakes that have occurred in the recent geological past. If the history and amount of ground displacement of 800

3 past surface fault ruptures can be determined from the study of these faults, then the magnitudes and frequency of past large earthquakes can be estimated. These data are suitable for incorporation into an assessment of the longer term earthquake hazard in New Zealand. This paper presents a subdivision of. New Zealand into tectonic provinces where the style and rates of deformation are comparable, and makes some additional remarks as to how earthquake hazard is related inces. II. Tectonic Setting to the tectonic prov- The:re are three main elements to the Quaternary tectonic setting of New Zealand (Fig. 1). They are ; (i) The Kermadec-Hikurangi, west-dipping, subduction zone that extends along the east coast of New Zealand as far south as the Chatham Rise. Subduction is broadly defined by a chipping band of seismicity (Reyners, 1989) most often located within the subducted Pacific pl.ate. Subduction related crustal thinning and ex tension is observed in the Havre Trough to the north of New Zealand, and persists into the central North Island as a continental the Taupo Volcanic 1987 ; Wright, 1992). rift in zone (Cole, 1990 ; Stern, (ii) In the southwest of the South Island, oceanic crust of the Australian plate is being subducted, at a steep angle, beneath continental crust of the Pacific plate at the Puysegur Trench. (iii) Between the two opposite-dipping subduction zones there is a transform fault system that accommodates relative plate motion between the plates. In the southwest of the South Island the structural pattern is relatively simple with a large proportion of the motion being accommodated on a single structure \he Alpine fault (Berryman et al., 1993), while to the northeast, in the Marlborough area and in eastern North Island, there are many branching faults. In addition to the complexity imposed by opposite dipping subduction zones, and contrasts in crustal thickness across the plate boundary, the proximity of the pole of rotation between the two plates means that relative plate motion has a large, but variable component of oblique motion which varies by almost 20 in orientation between East Cape and Fiordland (Fig. 1). III. Tectonic Framework As a result of the complex tectonic setting there is a wide variety tectonic features, and variety in styles of deformation within New Zealand. The country can be divided into regions of similar tectonic style and rate (Fig. 2), and this approach proves useful in identifying the principal components of the tectonic framework of the country (Berryman and Beanland, 1991), although the boundaries between these regions or "tectonic provinces" are rarely well-defined or sharp. The provinces identified are ; Western North Island Blocks and Basins In contrast to the central and eastern parts of New Zealand, this province is tectonically relatively stable as judged by the small number of known faults and their average vertical slip rates that rarely exceed 1 m/ka (Pillans, 1990). Although several moderate to large earthquakes have occurred within the Western North Island Blocks and Basins province since 1840 (Eiby, 1968 ; Smith and Berryman, 1986), none are known to have been associated with surface fault displacement. Taupo Volcanic Zone Deformation within the Taupo Volcanic Zone (TVZ) is distributed 801

4 Fig. 2 Tectonic provinces of New Zealand, as determined by active deformation rate and style across many short faults, often arranged in zones of multiple, parallel fault traces. The faults strike northeast and accommodate largely vertical movements associated with WNW- ESE-directed stretching and extension of the crust in this part of New Zealand. Three historical earthquakes associated with surface faulting have occurred in this zone in 1922 (Grange, 1932) ; 1983 (Grindley and Hull, 1986) ; and 1987 (Beanland et al., 1990). The thin and weak crust and shallow seismogenic depth of earthquakes in this province indicates that MR may be a likely maximum magnitude for surface fault related earthquakes in the TVZ. Canterbury-Otago-Southland Ranges and Basins Tectonic studies in this province indicate that the characteristic range and basin topography of Otago-Southland has developed by the growth of steeply-dipping reverse faults during the last 2 million vears (Beanland et 802

5 al., 1986). Faults strike both northeast and northwest and are commonly about 50 km in length, but it is the northeast-striking set that appears to have been most active during the last one million years. No significant historical earthquakes have occurred within the province ; indeed the area is distinctive in its lack of recorded seismicity. The faults, where studied, have average recurrence intervals ranging from several to tens of thousands of years. When earthquakes associated with surface rupture occur again in the Canterbury-Otago-Southland Ranges and Basins province, they are likely to be M 7.5 and involve surface rupture along several closely-spaced faults near the base of one of the major northeast-trending mountain ranges. Cant erbury-chathams Platform The Canterbury -Chathams Platform is a province of Bay earthquake in 1931 (Ms 7.8) resulted in sudden and permanent uplift of the coast by about 3 m (Henderson, 1933 ; Hull, 1990), these marine terraces have been interpreted to have been uplifted in similar large earthquakes during the last 7,000 years. Dated sequences of uplifted marine terraces therefore provide a record of the magnitude and timing of past earthquakes. The uplifted terraces preserved along the east coast appear to record significant vertical movement during successive earthquakes on km long buried or offshore reverse-oblique-slip faults within the upper, Australian, plate of the subduction zone. Axial Tectonic Belt Strike-slip faults of central New Zealand predominate in the Axial Tectonic Belt. In the southern South Island, the Alpine fault appears to accommodate much of the relative motion between the tectonic general tectonic stability. Minor long term subsidence is indicated from the distribution and elevation of ƒ20 ka sediments within the Christchurch metropolitan area (Brown and Weeber, 1992). Recent interpretation of offshore high-resolution seismic data along the western Chatham Rise (Barnes et al., 1993) has shown the presence of geologically young normal faults that accommodate minor extension( ƒ 2 %) within the western Chatham Rise. Hikurangi Subduction Margin Along much of the East Coast of the North Island and northeastern South Island the Pacific plate is being subducted beneath the Australian plate, and contraction is occurring in the overlying Australian plate. At many localities along the eastern coastline of the North Island there are sequences of emergent marine terraces, up to 27 in above sea level and aged 7,000 years or less. Based on geomorphic, coastal process and stratigraphic grounds (see Ota et al., 1991), plates, while in the northern South Island several faults branch from the Alpine fault to form the strike-slip fault province extending northeast through Marlborough and along the central part of the North Island, northward to the Bay of Plenty. Four of the eleven earthquakes that are known to have been associated with surface faulting since 1840 have occurred within the Axial Tectonic Province. Nelson-Westland Ranges and Basins The Nelson-Westland Ranges and Basins province lies northwest of the Alpine fault and encompasses several north-striking, high-angle ( 60 ) reverse faults and folds accommodating the dominantly contractional strain within the province. Two large earthquakes have occurred within the province in 1929 (Ms 7.8, Murchison) and 1968 (M3 7.4, Inangahua), and a high rate of seismic activity continues to the present day. A maximum of 4.5 m vertical movement and 2.1 m left-lateral slip was recorded coupled with the observation that the Hawke's along an 8 km long section of the White 803

6 Q b C Fig. 3 a Slip rates on individual structures within generalised regions b Epicentres of shallow earthquakes of magnitude 6.5 and greater from c Geodetic shear strain rates over the past ca. 100 years (from Hatherton, 1984) Creek fault that moved during the Murchison earthquake (Henderson, 1937 ; Berryman, 1980). Despite the present high level of seismicity in the province, paleoseismic studies suggest a very long recurrence interval for the White Creek and other similar faults in the region. IV. Rates of Movement The relative importance of individual tectonic structures within the overall framework of New Zealand can be appreciated from a com- 804

7 parison of their rates of movement (Fig. 3). Rates of movement along the Alpine fault, in the southwest of the Axial Tectonic Province, of m/ka (Berryman et al., 1993) are in close agreement with the rate predicted from plate motion models (DeMets et al., 1988). To the northeast in the Axial Tectonic Province, rates on individual faults and folds are in the range 2-8 m/ka. The sum of the displacements across this broader zone is comparable to the Alpine fault rate. Rates of movement on individual structures outside the Axial Tectonic Province reach about 3 m/ka in the Taupo Volcanic Zone, and are generally less than 1 m/ka elsewhere. In the northern part of the North Island no late Quaternary active faults are known, and here, and in parts of the Canterbury-Chathams Platform Province and the Canterbury-Otago- Southland Ranges and Basins Province rates may be less than 0.1 m/ka. The emphasis of information presented here is from Quaternary active faults and folds. It is interesting to compare how the deformation is manifest over different time periods, and in Fig. 3 comparison of slip rate on faults is made with geodetically derived strain rates for the past century, and with the distribution of major earthquakes in the past 150 years. There is, in general, good correspondence between t:he different measures of deformation, with the exception of the area over the central part of the Alpine fault. Along this section of the fault here is high geodetic strain, and rapid rates of fault movement averaged over many thousands of years. However the area is seismically very quiet, and is regarded as a region with a high probability of a future major earthque ke. V. Discussion The diverse tectonic setting of New Zealand has resulted in the development of the generalised tectonic provinces (Berryman and Beanland, 1988) and outlined above, and these provinces show a wide variation in the average slip rate from about 0.1 mm/yr on some reverse faults in Central Otago and Canterbury to perhaps 35 mm/yr on the Alpine fault. Because single event displacements on active faults are commonly in the range of 1 to 5 metres, the average recurrence time of surface faulting is proportional to average slip rates and varies from a few hundred to several thousand, and perhaps tens of thousands of years. And in a general way the future earthquake hazard is also proportional to fault slip rates and surface rupture recurrence intervals. However, there is increasing evidence that surface fault rupture on individual faults often do not have uniform recurrence intervals, and that estimation of future hazard must also account for variability the recurrence behaviour. Several sites within the Hikurangi subduction margin provide the best-dated earthquake recurrence sequences for the last 7,000 years in New Zealand, and show irregular, time-predictable behaviour (Berryman et al., 1989 ; Ota et al., 1991) in the manner described by Shimazaki and Nakata (1980). Uplift events are spaced at intervals of 300 to 1,500 years with the repeat times proportional to the amount of uplift. The paleoearthquakes identified in the Hikurangi subduction margin are interpreted to occur on reverse faults breaking through the brittle part of the upper, Australian plate. Examples of variable uplift recorded from different earthquakes at the same site in the Hikurangi subduction margin implies that coseismic rupture lengths and earthquake magni- in 805

8 tudes along the causative faults may be variable also. It seems that coseismic rupture boundaries are neither spatially nor temporally persistent. Slip along one segment of the fault zone may trigger slip at adjoining segments or on faults in adjoining tectonic provinces, and these can be quite closely spaced in time. For example, the 1931 Hawkes Bay earthquake was followed by the M, 6.9 Wairoa earthquake in 1932, by the Ms 7.6 Pahiatua earthquake in 1934 and by M, 7.0 and 7.2 Wairarapa earthquakes in 1942 (Dowrick and Smith 1990). This sequence of earthquakes may represent coseismic strain release on several structures along the plate boundary that was initiated by the 1931 earthquake. Future sequences may begin at a different locality and involve different structures with different segment lengths, and perhaps different provinces. The clustering of earthquakes is increasingly being recognised along other faults in the world, in addition to that seen in the W estland- Nelson and the Hikurangi subduction margin provinces in New Zealand in the historic period, and in the Hikurangi subduction margin and Central Otago provinces in the paleoseismic record. Precise dating at Pallet Creek along the San Andreas fault by Sieh et al., (1989) indicates that temporal clustering of earth- quakes occurs along this part of the fault. The El Asnam fault that generated a destructive M, 7.3 earthquake in 1980 in northern Algeria has a history of 9 previous earthquakes during the last 6,000 years (Meghraoui et al., 1988). Two groups of earthquakes appear to have occurred at intervals of over, 1,000 years, but individual earthquakes within the clusters of earthquakes occurred at year intervals. Temporal clustering of earthquakes indicates that seismic hazard from a fault may be time dependent, but the probability of occurrence is not simply a function of the elapsed time since the last earthquake. The potential seismic hazard along the active fault segment and its adjoining segments would be much greater after the fault zone has entered a period of activity, than when it was quiescent. VI. Conclusions In New Zealand geological data recording the timing of past major earthquakes, as recorded by surface fault traces, adds significantly to an understanding of tectonic provinces in the country, and to estimates of future earthquake hazard. Except for the absence of major earthquake activity on the central section of the Alpine fault, and higher than expected activity in Nelson-Westland (compared with geological data), there is, in general, broad agreement between geological, seismological, and geodetic data as to where the most active deformoation zones of the country are. With the exception of the Taupo Volcanic Zone, the maximum earthquake magnitudes likely in various parts of the country are similar, so future earthquake hazard may be approximately to either fault slip rates or recurrence of surface faulting in a region. scaled intervals However there is increasing data that suggest there may be substantial variability in the recurrence of large earthquakes on individual faults in several of the tectonic provinces of New Zealand, and that clustering of activity may be common. This phenomena places severe constraints on the utility of average recurrence intervals to estimate future hazard. Thus fault behaviour models have important implications for the assessment of seismic hazard based on geological studies. Attempts at modelling of fault behaviour is an advance on earlier assumptions that large earthquake temporal recurrence on a given fault is a random or poisson 806

9 process, but for the moment probability estimates for future hazard are still based largely on uniform fault behaviour and the time elapsed since the last event. Fault and earthquake hazard probabilities may be seriously in error in tectonic provinces where fault behaviour is non-uniform. Acknowledgements I take this opportunity to acknowledge the assistance of the Tokyo Geographical Society in presenting this paper at the International Symposium on Quaternary Environmental Changes in the Pacific Region, held in Tokyo in November, The content of this paper draws widely on work of colleagues at the New Zealand Institute of Geological and Nuclear Sciences, notably Alan Hull, Sarah Beanland, and Russ Van Dissen, and I would like to acknowledge their contribution to the ideas expressed in this paper. References Barnes, P. M., Nodder, S. D. and Wright, I. C. (1993) : Seismotectonic studies on the submerged New Zealand continental shelf : A status report. Proc. N. Z. Nat. Soc. for Earthquake Engineeri'ng Conference, Wairakei, March 1993, Beanland, S., Berryman, K. R., Hull, A. G. and Wood, P. R. (1986) : Late Quaternary deformation at the Dunstan fault, Central Otago, New Zealand. Roy. Soc. New Zealand Bull., 24, Beanland, ;3., Blick, G. H. and Darby, D. J. (1990) : Active normal faulting in a back-are basin : Geological and geodetic characteristics of the 1987 Edgecumbe Earthquake, New Zealand. J. Geophys. Res., 95, Berryman, K. R. (1980) : Late Quaternary movement on White Creek fault, South Island, New Zealand. New Zealand. J. Geology and Geophysics, 23, Berryman, K. R., Beanland, S., Cooper, A. F., Cutten, H. N. C., Norris, R. J. and Wood, P. R. (1993) : The Alpine fault, New Zealand : Variation in Quaternary structural style and geomorphic expression. Ann. Tectonicae Suppl to Vol VI, Berrymm, K. R. Ota, Y. and Hull, A. G. (1989) : Holocene paleo seismicity in the fold and thrust belt of the Hikurangi subduction zone, eastern North Island, New Zealand. Tectonophys., 163, Berryman, K. R. and Beanland, S. (1991) : Variation in fault behaviour in different tectonic provinces of New Zealand. J. Structural Geology, 13, Brown, L. J. and Weeber, J. H. (1992) : Geology of the Christchurch urban area. Scale I : 25,000. Institute of Geological and Nuclear Sciences Geological Map 1. Institute of Geological and Nuclear Sciences Limited, Lower Hutt, New Zealand, 1 sheet +104 p. Cole, J. W. (1990) : Structural control and origin of volcanism in the Taupo volcanic zone, New Zealand. Bull. Volcanol., 52, DeMets, C., Gordon, R. G. and Argus, D. F. (1988) : Intraplate deformation and closure of the Australia- Antarctica-Africa plate circuit. J. Geophys. Res., 93, Dowrick, D. J. and Smith, E. G. C. (1990) : Surface wave magnitudes of some New Zealand earthquakes Bull. the New Zealand Nat. Soc. Earthq. Eng., 23, Eiby, G. A. (1968) : An annotated list of New Zealand earthquakes, New Zealand J. Geology and Geophysics, 11, Eiby, G. A. (1970) : Seismic regions of the South Island of New Zealand. Trans. Roy. Soc. New Zealand, 8, Grange, L. I. (1932) : Taupo earthquakes, Rents and Faults formed during earthquake of 1922 in Taupo District. New Zealand J. Science and Technology, XIV-3, Grindley, G. W. and Hull, A. G. (1986) : Historical Taupo earthquakes and earth deformation. Roy. Soc. New Zealand Bull., 24, Hatherton, T. (1984) : Earthquakes. In Speden, I. G. and Crozier, M. J. (compilers) : Natural Hazards in New Zealand. New Zealand National Commission for UNESCO, Henderson, J. (1933) : Geological aspects of the Hawke's Bay earthquakes. New Zealand J. Science and Technology, 15, Henderson, J. (1937) : The west Nelson earthquakes of 1929 (with notes on the geological structure of west Nelson). New Zealand J. Science and Technology, 19-2, Hull, A. G. (1990) : Tectonics of the 1931 Hawke's Bay earthquake. New Zealand J. Geology and Geophysics, 33, Meghraoui, M., Philip, H., Albarede, F., Cisternas, A. (1988) : Trench investigations throuh the trace of the 1980 El Asnam thrust fault : Evidence for paleoseismicity. Bull. Seismological Society of America, 78, Ota, Y., Hull, A. G. and Berryman, K. R. (1991) : Coseismic uplift of Holocene marine terraces in the Pakarae River area, eastern North Island, New 807

10 Zealand. Quat. Res., 35, Pillans, B. (1990): Pleistocene marine terraces in New Zealand : A review. New Zealand J. Geology and Geophysics, 33, Reyners, M. (1989): New Zealand seismicity : An interpretation. New Zealand J. Geology and Geophysics, 32, Shimazaki, K. and Nakata, T. (1980): Time-predictable recurrence model for large earthquakes. Geophys. Res. Lett., 7, Sieh, K., Stiuver, M. and Bullinger, D. (1989): A more precise chronology of earthquakes produced by the San Andreas fault in southern California. J. Geophys. Res., 94, Smith, W. D. and Berryman, K. R. (1986): Earthquake hazard in New Zealand ; Inferences from seismology and geology. Roy. Soc. New Zealand Bull., 24, Stern, T. A. (1987) : Asymmetric back-arc spreading, heat flux and structure associated with the Central Volcanic Region of New Zealand. Earth and Planetary Sci. Lett., 85, Wellman, H. W. (1967) : Tilted marine beach ridges at Cape Turakirae, N. Z. J. Geosci., Osaka City Univ., 10, Wright, I. C. (1992) : Shallow structure and active tectonism of an offshore continental back-arc spreading system : The Taupo Volcanic Zone, New Zealand. Marine Geol., 103, (Accepted 18 November, 1994) 808