Tsunami hazard posed by earthquakes on the Hikurangi subduction zone interface

Abstract

To the east of the North Island the Pacific tectonic plate is being thrust beneath the Australian plate in a process known as subduction. Large tsunamis, such as the 2004 Indian Ocean Tsunami, are most frequently caused by earthquakes on plate boundaries where subduction takes place. In this project we attempt to determine whether the coast of New Zealand is at risk of tsunamis originating from earthquakes on the Hikurangi (east coast of North Island) plate boundary. To do this we examine the slow deformation of the North Island as measured by GPS, and use this to locate the regions of the plate boundary in which elastic energy is being stored for eventual release in earthquakes. We also examine the distribution of smaller earthquakes on the plate boundary, as these can also be used to identify areas in which energy is being stored.

We find that the Hikurangi plate interface can be best described in terms of three segments” the lower North Island, the Hawke Bay region, and the Raukumara Peninsula. Each of these segments has their own distinctive earthquakes characteristics, and we find that there is a potential tsunami hazard associated with each of them.

The lower North Island segment is found to be storing elastic energy over a wide area, which has the potential to be released in large tsunami-causing earthquakes.

In the Hawke Bay region the region in which the plates are storing energy is considerable narrower; however we find that the Lachlan Fault, which rises from the plate interface under Hawke Bay, is a likely source of hazardous tsunamis.

Seamounts (extinct volcanoes) attached to the Pacific plate are being subducted beneath the Raukumara Peninsula, and this process appears responsible for a characteristic type of earthquake known as a “tsunami earthquake” in which very little shaking is felt, yet dangerous tsunamis are often caused. Two earthquakes near Gisborne in 1947 were probably of this type. Tentative new evidence suggests that the average interval between these events might only be around 70 years. If true, this implies that the probability of such an earthquake occurring in the next couple of decades is higher than previously thought.

---

Technical Abstract

Subduction of the Pacific plate beneath the eastern North Island, New Zealand occurs at the Hikurangi trough. Seismological and geodetic data (Reyners, 1998; Wallace et al., 2004) indicate that areas of the Hikurangi subduction interface are “stuck”, leading to the accumulation of elastic energy which will eventually be released in future large subduction earthquakes. Although records of paleo-subduction earthquakes and associated tsunami are sparse, Cochran et al. (2006) and Hayward et al. (2006) have found convincing evidence for  repeated subsidence events in the Hawkes Bay region that may be related to large subduction zone earthquakes just offshore, beneath Hawke Bay. Rupture of the Hikurangi subduction interface may be one of the most important sources of tsunami that impact New Zealand. In this project we use the latest geodetic and seismological evidence for interseismic Hikurangi interface coupling to model vertical deformation associated with a variety of subduction zone rupture scenarios, and then use wavefield modelling to determine the nature of tsunami generated by these ruptures.

We find that the Hikurangi subduction interface divides naturally into three segments: the lower North Island, the Hawke Bay region, and the Raukumara Peninsula. While each of these has different seismogenic characteristics, we find that at least one plausible scenario results in a significant tsunami hazard for each segment.

The plate interface beneath the lower North Island is believed to be strongly coupled over a >100 km wide zone, and is currently accumulating strain that is likely to be released in one or more large earthquakes. The updip limit of such a rupture, and the recurrence interval is poorly constrained at present, so we present a variety of scenarios to illustrate the dependence on such factors. The tsunami risk to Wellington is also strongly dependent on the location of the southern termination of such a rupture, a variable currently unconstrained by (land-based) GPS data.

In the Hawke’s Bay region coupling of the plate interface does not currently reach to the same depths as in the lower North Island, a trend which if constant over the earthquake cycle suggests that the largest subduction zone earthquakes on this segment are likely to have lower magnitudes than the largest earthquakes on the lower North Island segment. However the Lachlan thrust splay fault is found to be a significant tsunami source in its own right.

The subduction margin adjacent to the Raukumara Peninsula is characterised by tectonic erosion and seamount subduction. GPS studies imply that the subduction interface is weakly coupled here, though this is probably a reflection of strong coupling at localised asperities (probably related to subducted seamounts) surrounded by uncoupled areas where the plate interface is lubricated by fluids. This environment has been associated with “tsunami earthquakes”, characterised by efficient tsunami generation and relatively weak earthquake shaking. Our modelling of the 1947 Gisborne earthquakes reproduces the observed tsunami distribution with sufficient accuracy to infer that these events were probably of the “tsunami earthquake” type. We find tentative evidence that the recurrence interval for such events might only be about 70 years, which if true implies that the probability of another such event occurring in the next few decades is higher than previously thought.

We find that the possibility of an outer-rise earthquake giving rise to a tsunami on the Hikurangi margin cannot be ignored, as a scenario event located offshore of the Raukumara Peninsula exhibits modelled wave-heights that represent a significant hazard at the coast.