Age constraints on unstable landforms at a near-fault site in Central Otago, New Zealand: Groundwork for validation of seismic hazard models

Abstract

Earthquake hazard analysis uses past earthquake activity to forecast future activity. The written and instrumental record of past earthquakes and the prehistorical record of (paleo) earthquakes are used as a basis for estimating the future likelihood of earthquake occurrence and associated strong ground motions. This is the fundamental basis of earthquake hazard analysis. However, the main issue concerning earthquake hazard methodologies is the lack of ability to validate the resulting hazard estimates. Validation involves the use of some independent dataset or observations to gain confidence in the hazard estimates.

Our EQC-funded research project has been focussed on developing methods to validate the very strong earthquake motions predicted at sites near active faults by seismic hazard models in New Zealand. Earthquake motions are expected to be very strong near a causative fault (e.g. ground accelerations of greater than lg, g being the unit of gravitational acceleration 10m/sect but a number of recent large earthquakes (e.g. Chi-Chi, Taiwan) have exhibited motions considerably less than expected. Furthermore, studies in the arid western USA have identified fragile geomorphic features (precariously-balanced rocks, or PBRs) near the San Andreas Fault that have remained unchanged for 10,000 to 100,000 years. The ground motions have not been strong enough to destroy the fragile features, and this provides a valuable constraint on past ground motions. The objective of our EQC-funded research has been to apply these arid western USA studies to a near-fault site in the more humid temperate New Zealand environment. Numerous bedrock (schist) PBRs are present within 5km of the Dunstan Fault in central Otago, a fault capable of generating large earthquakes every 8,000 years on average. Our earlier EQC-funded work showed the PBRs to be too fragile to have survived the ground motions predicted for Dunstan Fault earthquakes from the national seismic hazard model. Either the ground motions have not been as strong as predicted, or the PBRs are too young to have experienced any such earthquakes. Since our earlier studies did not provide adequate age control for the PBRs the present study has largely focussed on this aspect.

We obtained exposure age data for 11 rock samples from one of the PBRs near the Dunstan Fault and from neighbouring features, and have extensively analysed the data. The data provide bedrock erosion rate estimates of about 0.5cm per 1,000 years for the PBRs. This means that they would have been considerably larger and more stable at the start of the Holocene (Holocene is the epoch that spans 10,000 years ago to the present), and therefore not experienced any Dunstan Fault earthquakes in the present fragile state, given the long time period between earthquakes on the fault. The main conclusion of our study is that PBRs in the humid-temperate environment of New Zealand are considerably younger than their arid counterparts, and thus of limited use for constraining near-fault ground motions in areas where large earthquakes do not occur for many thousands of years. Future New Zealand-based efforts should therefore move to studying fragile geomorphic features close to faults that produce large earthquakes more often (i.e. every 100 to 1,000 years). Lastly, the main international contribution arising from our results is that we have provided the first-ever constraints on the age of PBRs in a humid-temperate environment, and that these constraints may be more applicable to similar environments in the USA (e.g. Pacific Northwest) than results obtained from nearby arid environments.

---

Technical Abstract

We constrain the age of precariously-balanced rocks (PBRs) within a metamorphic terrane and temperate environment using in-situ cosmogenic isotopes and geomorphic interpretation. Previous PBR age-related studies have been almost exclusively restricted to arid environments in the western USA. Cosmogenic in-situ 10Be concentrations from 11 samples of schist and silcrete from one PBR site in the central Otago region of New Zealand imply rock surface erosion rates of the order 0.5cm/kyr, consistent with rates of landform evolution and catchment denundation determined elsewhere in the region. The PBRs may only have reached their precarious state less than 10kyr ago, and may not have experienced any earthquakes on the nearby Dunstan Fault, given the long recurrence interval of the fault (8kyr). Future New Zealand-based efforts should therefore move to studying fragile geomorphic features close to faults with much shorter recurrence intervals (100-1,000 years).

The main international contribution arising from our results is that PBRs from metamorphic terranes in humid-temperate environments similar to that of central Otago (e.g. parts of central California and Pacific Northwest) may be considerably younger than counterparts in the arid western USA.