Detection of Large Holocene Earthquakes in the Sedimentary Record of Wellington, New Zealand, Using Diatom Analysis

Abstract

New Zealand is situated on the boundary between the Pacific and Australian tectonic plates. The Wellington region lies near the southern end of the Hikurangi subduction zone and within a zone of major, active strike-slip faults. Wellington's paleoseismic and historic records indicate that large surface rupture earthquakes have occurred on these faults in the past. Development of a complete record of past large earthquakes is a high priority for the region because of the risk posed by occurrence of large earthquakes in the future. The existing paleoseismic record has been derived predominantly from studies of fault trench stratigraphy, raised beach ridges and offset river terraces. The sedimentary record of lakes and coastal waterbodies is a source of information that has not been used specifically for paleoseismic purposes in the region. Therefore investigation of Wellington's sedimentary record is used in this thesis to make a contribution to the paleoseismic record. Holocene sedimentary sequences are studied from three small, low elevation, coastal waterbodies: Taupo Swamp, Okupe Lagoon and Lake Kohangapiripiri. Sequences of between 200 and 650 cm depth were collected using a hand-operated coring device. Sedimentology and diatom microfossil content were analysed and interpreted to enable reconstruction of paleoenvironment at each site. Radiocarbon dating was used to provide chronologies for the sequences that are aged between 5000 and 7500 calibrated years before present (cal. years BP). Diatom analysis is the main tool used to reconstruct paleoenvironment and detect evidence for occurrence of past large earthquakes. To aid reconstruction of sedimentary sequences used in this project, as well as coastal sequences in New Zealand in general, a coastal diatom calibration set was constructed using 50 sites around New Zealand. Modern diatom distribution and abundance, and associated environmental variables are analysed using ordination and weighted averaging techniques. Detrended correspondence analysis arranges species according to salinity preferences and divides sites clearly into waterbody types along a coastal gradient. This analysis enables reconstruction of waterbody type from fossil samples by passive placement onto ordination diagrams. Weighted averaging regression of calibration set samples results in a high correlation (r2jack=0.84) between observed and diatom inferred salinity, and enables salinity preferences and tolerances to be derived for 100 species. This confirms for the first time that species' preferences derived in the Northern Hemisphere are generally applicable to diatoms living in the coastal zone of New Zealand. Weighted averaging calibration and the modern analogue technique are used to generate quantitative estimates of paleosalinity for fossil samples. Paleoenvironmental reconstructions of Taupo Swamp, Okupe Lagoon and Lake Kohangapiripiri indicate that each waterbody has been isolated from the sea during the late Holocene. Isolation has been achieved through interplay of sediment accumulation causing growth of barrier beaches, and coseismic uplift. Ten distinct transitions between different paleoenvironments are recognised from the three sequences. These transitions involve changes in relative sea level or water table level often in association with catchment disturbance or marine influx events. All transitions occur suddenly and are laterally extensive and synchronous within each waterbody. Quantitative estimates of paleosalinity and waterbody type are used to differentiate between large and small magnitude changes in paleoenvironment. Five transitions involve large amounts of paleoenvironmental change and provide evidence for earthquakes occurring at approximately 5200, approximately 3200, and approximately 2300 cal. years BP. Five other transitions are consistent with the effects of large earthquakes occurring at approximately 6800, 2200, approximately 1000, approximately 500 cal. years BP and 1855 AD but do not provide independent evidence of the events. Environmental transitions at Lake Kohangapiripiri clarify the timing of rupture of the Wairarapa Fault by bracketing incompatible age estimates derived from two different sites on the fault. The oldest environmental transitions recognised at Taupo Swamp and Okupe Lagoon both occur at approximately 3200 cal. years BP indicating that western Wellington was uplifted at this time. Environmental transitions are recorded at all three study sites at approximately 2300 cal. years BP indicating that the entire western and central Wellington region experienced coseismic uplift at this time. Because of the distance between sites this apparent synchroneity implies that several faults in the region ruptured at a similar time. Investigation of sedimentary sequences contributes to the existing paleoseismic record by providing additional estimates of timing for past large earthquakes, enabling estimation of the areal extent of the effects of past earthquakes, and by highlighting periods of fault rupture activity in the late Holocene.