Advertisement

Natural Hazards

, Volume 70, Issue 1, pp 415–445 | Cite as

Tsunami inundation in Napier, New Zealand, due to local earthquake sources

  • Stuart A. Fraser
  • William L. Power
  • Xiaoming Wang
  • Laura M. Wallace
  • Christof Mueller
  • David M. Johnston
Original Paper

Abstract

Deterministic analysis of local tsunami generated by subduction zone earthquakes demonstrates the potential for extensive inundation and building damage in Napier, New Zealand. We present the first high-resolution assessments of tsunami inundation in Napier based on full simulation from tsunami generation to inundation and demonstrate the potential variability of onshore impacts due to local earthquakes. In the most extreme scenario, rupture of the whole Hikurangi subduction margin, maximum onshore flow depth exceeds 8.0 m within 200 m of the shore and exceeds 5.0 m in the city centre, with high potential for major damage to buildings. Inundation due to single-segment or splay fault rupture is relatively limited despite the magnitudes of MW 7.8 and greater. There is approximately 30 min available for evacuation of the inundation zone following a local rupture, and inundation could reach a maximum extent of 4 km. The central city is inundated by up to three waves, and Napier Port could be inundated repeatedly for 12 h. These new data on potential flow depth, arrival time and flow kinematics provide valuable information for tsunami education, exposure analysis and evacuation planning.

Keywords

Local tsunami Inundation modelling COMCOT Hikurangi subduction margin Structural damage Arrival time Evacuation 

1 Introduction

The Hikurangi subduction margin is a potential source of ‘large-to-great earthquakes’ (Berryman 2005), and the local nature of tsunami generated from such an earthquake is of significant concern to the New Zealand scientific and emergency management communities. New Zealand is also at risk of tsunami from regional and distant sources and is a member of the Pacific Tsunami Warning System. The Ministry of Civil Defence & Emergency Management (MCDEM) aims to assess the tsunami threat to New Zealand and if appropriate disseminate a national tsunami advisory or warning within 15–30 min following receipt of a Pacific Tsunami Warning Center information bulletin, watch or warning, or a GNS Science earthquake report (MCDEM 2010, p.29). However, for local tsunami, defined in New Zealand as having travel time of <1 h from source to the coastal area of interest (MCDEM 2010, pp. 20–21), 15–30 min represents a significant portion of the time available for the evacuation of coastal locations proximal to the source. Education and awareness of the local tsunami hazard are vital to enable the public to recognise the potential for tsunami following a local earthquake and act accordingly through immediate self-evacuation.

This paper enhances current knowledge of the local tsunami hazard in Napier, Hawke’s Bay, by demonstrating potential tsunami inundation due to earthquakes at the Hikurangi subduction margin. Although the city has not experienced a significant local tsunami in recorded history, there are sites in Hawke’s Bay that exhibit evidence of past tsunami in the form of high-energy marine deposits. Damaging local tsunami have been recorded elsewhere on the east coast of the North Island, most notably two local-source tsunami that affected Gisborne in 1947 (Downes et al. 2000). We provide the first published assessment of inundation in Napier based on the simulation of the full tsunami process for multiple subduction zone tsunami scenarios. Previous consideration of tsunami generated by earthquakes on the Hikurangi subduction margin (Power et al. 2008; Berryman 2005) is limited to the estimation of wave heights at-shore with empirically estimated maximum run-up heights. In order to advance tsunami mitigation and evacuation planning in coastal communities, accurate estimates of flow depth and inundation extent are now required.

The scenarios applied in this study include the worst-case tsunami for use in community preparedness and tsunami mitigation activities, but it is also important to investigate onshore impacts due to other plausible tsunamigenic earthquakes. The maximum credible earthquake is represented by a moment magnitude (MW) 9.0 rupture of the whole subduction margin. This is considered as an upper limit to plausible magnitudes of such a rupture at this margin. We also demonstrate the potential for inundation due to smaller local ruptures that have been discussed in previous studies, but have not been used in simulation of onshore inundation. We describe tsunami impacts in terms of flow depth and structural damage potential and present wave arrival times to constrain estimates of time available for evacuation.

2 Study area

Napier Territorial Authority (hereafter, Napier) is a coastal urban centre, identified at substantial risk of tsunami from previous hazards assessments (Berryman 2005; Power et al. 2008). There is a need for detailed tsunami hazard assessment to inform the development of evacuation plans and tsunami education to increase the resilience of individuals and the local community. Napier has an estimated resident population of 57,800 (at June 30 2012; Statistics New Zealand 2013). The same projections indicate that 10,200 (17.6 %) of the population are over the age of 65, which is the demographic group shown by the recent experience to be the most vulnerable to tsunami. In the 2011 Great East Japan tsunami, over 65 % of deaths in the three worst-affected prefectures were of people over 60 years of age (Nakasu et al. 2011). Napier covers an area of 106 km2, comprising residential suburbs, commercial and industrial areas, and agricultural land including orchards and vineyards. Hawke’s Bay Airport, which operates internal passenger and freight flights and would likely be required to provide access in response to an earthquake and tsunami, is situated on an area of land between 0.5 and 1.5 m above mean sea level between 250–1,700 m from the coast in Bay View. Napier Port is the fourth largest in New Zealand in terms of the number of containers handled and the second largest in the North Island based on export by volume, handling cargo including forestry products and container shipments (Port of Napier Limited 2012). Stored timber and containers on site are a potential source of fire and damaging debris if entrained in tsunami flow.

The local topography is predominantly low elevation, except for Bluff Hill, which provides an area of high ground immediately north of the city centre to maximum elevation over 100 m (Fig. 1). On the eastern shore of the city, there is a steep gravel beach and berm stretching along the coastline south from Bluff Hill to the confluence of the Tutaekuri, Ngaruroro and Clive Rivers. This berm ranges in elevation above mean sea level (MSL) from 4 m in the south and exceeding 7 m high at its northern end. North-west of Bluff Hill, the suburbs of Ahuriri and Westshore are separated by a tidal inlet and small marina. Westshore is situated on a peninsula elevated 4–6 m above MSL. Bay View is the most northern suburb of Napier, extending north around the bay. Much of the land around the present Ahuriri Lagoon and Airport was previously below sea level until uplift during the 1931 Hawke’s Bay earthquake and artificial drainage in the years since (Hull 1986). Some of this land remains below mean sea level. The Hawke’s Bay earthquake destroyed many buildings in Napier and resulted in major reconstruction in the 1930s Art Deco style. The 1930s building stock is an important factor in the city’s tourism activities. During peak tourist season (January to March), an average of 2,342 visitors stays in Napier accommodation every night (2006–2011 data; Statistics New Zealand 2012).
Fig. 1

A Study area shown in the national context. B Napier Territorial Authority in the Hawke’s Bay regional context with numbered locations indicating paleo-ecological analysis referred to in-text. The boundary of Napier Territorial Authority is shown and other locations referred to in-text are also indicated. C 10 m DEM showing topography of Napier Territorial Authority with labelled unit areas (suburbs), rivers, Napier Port and Hawke’s Bay Airport. Virtual tide gauges quoted in the text are identified

3 Hikurangi subduction margin

3.1 Tectonic setting and seismic potential

The Hikurangi Trough is the surface expression of westward subduction of the Pacific plate beneath the Australian Plate, situated approximately 50–100 km offshore of the east coast of New Zealand’s North Island (Fig. 2). The subduction margin continues to the north as the Kermadec Trench and the southern limit of the subduction is located offshore of the northern South Island, where the plate boundary transitions to strike-slip (Wallace et al. 2012). Moderate subduction thrust earthquakes (e.g. Downes 2006) and tsunamigenic slow-slip earthquakes (Downes et al. 2000) have occurred on the Hikurangi subduction margin in recorded history (since c. 1840 A.D.), but no subduction thrust earthquakes greater than MW 7.2 have been recorded.
Fig. 2

Map illustrating the shaded bathymetry of the model domain. The main map shows grid 1 with the plate boundary marked. The inset shows Hawke’s Bay and nested grids 2–4

Ansell and Bannister (1996) initially characterised the subducting slab using microearthquake seismicity, and more recent work has provided detailed images of the subduction interface configuration (Henrys et al. 2006; Barker et al. 2009). GPS and seismological observations have been used to define regions of distinctly different subduction interface behaviour and seismic potential (Wallace et al. 2009). Since the early seismological studies of the Hikurangi subduction margin, it has been suggested that the plate interface below the lower North Island (Cook Strait to Cape Turnagain) is more likely to produce large subduction thrust events in comparison with the interface below the upper North Island (north of Mahia Peninsula, Fig. 1B) (Reyners 1998; Wallace et al. 2009). Convergence of the subducting plates is taking place at rates of around 50–60 mm/year offshore of the upper North Island, while for offshore of the lower North Island, this rate is lower at 20–25 mm/year (Wallace et al. 2004).

Below the lower North Island, the plate interface is interseismically coupled (coupling coefficient: 0.8–1.0) to around 40 km deep and 90–180 km wide (Wallace et al. 2009). GPS data reveal that this part of the plate interface is building up significant elastic strain that will eventually be released in a large megathrust earthquake (Wallace et al. 2004). Wallace et al. (2009) estimate the lower North Island segment to be 230 km long and 150–185 km wide, and using Abe’s (1975) fault scaling relationships translates this to a potential event of MW 8.5–8.7 with 8–12 m of slip. If the current estimated slip rate deficit of 20–25 mm/year is steady throughout the interseismic period, this results in a proposed return period of 300–625 years (Wallace et al. 2009), although uncertainty remains around the amount of slip and recurrence interval. It is also possible that this segment of the interface ruptures in smaller (MW < 8.0) earthquakes more frequently.

At the central North Island segment, including offshore Hawke’s Bay, the plate interface currently exhibits low interseismic coupling and is dominated by aseismic slip and slow-slip events (Wallace et al. 2004; Wallace and Beavan 2010), suggesting lower potential for ‘stick–slip behaviour’, whereby elastic strain accumulates during an interseismic period (‘stick’ component; Scholz 1998), to be later released in a large earthquake (‘slip’ component). However, we cannot rule out the possibility of large or great earthquakes at any part of the margin (Wallace et al. 2009). Despite the shallow interseismic coupling extending to only 15 km depth, and the low coupling coefficient, frequent rupture of small patches of the subduction plate occurs and is inferred to be localised asperities exhibiting stick–slip behaviour, possibly related to subducting seamounts (Bell et al. 2010). Subducting seamounts have been related to recurrent earthquakes internationally, including earthquakes of around MW 7.0 with 30-year recurrence intervals offshore Tohoku (Yamanaka and Kikuchi 2004) and the MW 7.0 Gulf of Nicoya earthquake, Costa Rica (Husen et al. 2002), and have been related to local seismic coupling in an otherwise seismically decoupled subduction zone on the Tonga-Kermadec and Izu-Bonin Trenches (Scholz and Small 1997).

The plates beneath the upper North Island are less strongly coupled (coupling coefficient: 0.1–0.2) in a region extending to 15 km depth beneath Hawke’s Bay and the Raukumara Peninsula. This segment north of Mahia Peninsula is characterised by tectonic erosion and an indented toe, indicating the impact of seamounts on the subducting plate (Collot et al. 2001; Lewis et al. 1998; Pedley et al. 2010). It is believed that this segment is particularly suited to producing tsunami earthquakes that involve a large amount of rupture close to the trench characterised by slow rupture velocities, long rupture durations, low local magnitude (ML) compared with moment magnitude, and larger than expected tsunami compared with the earthquake magnitude (Kanamori 1972; Pelayo and Wiens 1992; Tanioka and Satake 1996). Two tsunami earthquakes occurred at the northern Hikurangi subduction margin in March and May 1947 (Downes et al. 2000) and prior to that in 1880 (Downes, unpublished data 2007), tentatively suggesting a return period of ~70 years for tsunami earthquakes in the Gisborne region (Power et al. 2008).

In addition to single-segment ruptures, the potential propagation of rupture across multiple segments must be considered (Power et al. 2008; Wallace et al. 2009). The occurrence of multiple-segment rupture at other major subduction zones, for example during the 2011 Great East Japan (Ishii 2011), 2010 Maule, Chile (Kiser and Ishii 2011), 2007 Solomon Islands (Taylor et al. 2008) and 2004 Sumatra (Lay et al. 2005) earthquakes, supports this possibility. Simultaneous rupture of the northern and central segment of the margin could result in an earthquake greater than MW 8.6, assuming 8 m or more of slip, while the expectation of a full margin rupture is that it would be greater than 650 km long by 100 km wide and MW 8.8 or greater (Wallace et al. 2009). There also exists the remote possibility of a rupture involving segments on both the Hikurangi and Kermadec subduction margins (Power et al. 2012a), but this extreme scenario is not investigated in this study.

3.2 Evidence of past earthquakes and tsunami

In recorded history, Napier has experienced distant tsunami in 1868 (due to an earthquake in Peru), flooding of a wharf in 1877 (earthquake, Chile) and 3 m flow depth onshore with damage to infrastructure, buildings and boats in 1960 (earthquake, Chile) (De Lange and Healy 1986). The tsunami due to the 2010 Chilean earthquake was recorded at Napier with measured peak amplitude exceeding 1.4 m in Ahuriri Harbour and economic loss of $80,000–100,000 due to closure of the Port (Hawke’s Bay Regional Council 2010). Records of tsunami generated within or proximal to Hawke’s Bay are limited. Most recently, two slow-slip earthquakes at Gisborne in 1947 (26th March: MW 7.0–7.1, and 17th May: MW 6.9–7.1) caused 10 and 6 m run-up, respectively, at the coast around Gisborne (Downes et al. 2000). There is a possible tsunami in 1937 or 1938 that may have affected Wairoa, recorded by Goff (2008). In 1931, a subaerial landslide was triggered by the Hawke’s Bay earthquake and in turn caused a localised tsunami with 15 m run-up in the Waikari River Estuary and 3 m at Wairoa (Fraser 1998). See Fig. 1B for locations referred to in this section.

In addition to the limited historic record, paleo-seismic and paleo-tsunami data are relied upon to enhance the record of pre-historic subduction earthquakes and tsunami. Goff (2008) lists around ten possible events with deposits found in Hawke’s Bay dated between the fifteenth century and 8,000 calendar years before present (yr BP), and several more that possibly occurred prior to Maori settlement. The certainty with which we can infer that deposits are due to tsunami varies significantly, but where a tsunami deposit is synchronous with an episode of sudden subsidence, we can infer that a local earthquake was a possible trigger of the tsunami. Evidence of earthquake events on the Hikurangi subduction margin exists in the form of repeated sudden subsidence and uplift events in Hawke’s Bay. At Ahuriri Lagoon, Hull (1986) identified one or more rapid subsidence events totalling 8 m of subsidence between 1,750 and 3,500 radiocarbon yr BP and one event of 1 m of subsidence at 500 radiocarbon yr BP. This was supported by further micropaleontological analysis at Ahuriri Lagoon by Hayward et al. (2006), who identified six separate subsidence events within the last 7,200 years with a recurrence interval of 1,000–1,400 years and a subsidence range of 0.5–1.8 m in each event. The approximate age of fifteenth-century tsunami deposits in Hawke’s Bay (Goff 2008) coincides with the most recent episode of (1 m) subsidence in Ahuriri Lagoon (Hayward et al. 2006) suggesting that this represents the most recent local earthquake and tsunami event for which there is geological evidence.

Cochran et al. (2006) recorded net subsidence of 4 m over the last 7,200 years at Te Paeroa Lagoon in northern Hawke’s Bay and of 6 m at Opoho, 10 km to the east. At each site, two episodes of subsidence have been dated to c. 5,550 and c. 7,100 year BP suggesting that in each case, either a single large earthquake caused synchronous vertical deformation at both sites, or that each site underwent separate incidents of localised subsidence very close together in time. At both sites, subsidence is synchronous with a high-energy marine deposit of coarse sand and gravel inferred to be due to tsunami inundation (Chagué-Goff et al. 2002; Cochran et al. 2005). This deposit extends to 2 km inland at Te Paeroa, representing a significant inundation extent. Three lakes, 60 and 30 km away from the core sites, formed at c. 7,100 yr BP (Page and Trustrum 1997) provide further evidence for an earthquake occurring with sufficient intensity to cause co-seismic landslides (Cochran et al. 2006). There are three other marine-source high-energy deposition units identified by Cochran et al. (2005) at Opoho and three at nearby Opoutama that may represent tsunami deposits but have not yet been determined to be synchronous with subsidence events.

In addition to subsidence episodes, investigations of Holocene marine uplifted terraces along the east coast of the North Island indicate episodes of rapid co-seismic uplift (Berryman et al. 1989; Clark et al. 2010), inferred as being due to rupture of local upper plate faults occurring either in conjunction with or independently of plate interface rupture (Berryman et al. 1989, 2011). Berryman et al. (2011) postulate that several anomalously young dates (with respect to the height of the terrace or age of nearby terraces) obtained for terraces in the lower North Island are due to tsunami depositing sediments on terraces which had been uplifted during previous deformation events. In particular, four marine uplift events are believed to have resulted in localised tsunami deposition along as much as 100 km of the North Island coast, at 1,463–1,670 AD, 1,282–1,408 AD, 267–449 AD and 270–405 AD. No uplifted terraces have been recognised as synchronous with the subsidence episodes identified at Opoho and Te Paeroa, as the respective records overlap by only 500 years, although there is a possibility of synchronous uplift-subsidence event at c. 5,500 year BP (Cochran et al. 2006). Dating of the known uplift episodes has not been carried out to sufficient resolution to identify whether any of the episodes occurred synchronously between sites long distances apart, which would suggest a significant plate interface rupture (Wallace et al. 2009).

Cochran et al. (2006) used forward elastic-dislocation modelling to demonstrate that the magnitude and distribution of subsidence recorded in northern Hawke’s Bay and uplift of Mahia Peninsula could be approximately replicated by several vertical deformation scenarios. They tested scenarios of 8 m slip on the plate interface, 8 m slip on the Lachlan Fault and a combination of both. Replication of the correct uplift-subsidence distribution and amount of subsidence due to the rupture of the Lachlan Fault alone are at the lower limit of that recorded in cores. Although permanent coastal deformation was replicated for the plate interface rupture with no Lachlan Fault component, uplift at the Mahia Peninsula is most likely due to a component of upper plate thrusting, rather than isolated rupture of the plate interface; therefore, a combined rupture of plate interface and upper plate faults is favoured as the cause of recorded subsidence-uplift distributions (Cochran et al. 2006). The scenarios identified by Cochran et al. (2006) inform the choice of scenarios tested in this study.

3.3 Previous tsunami hazard assessment

There have been several investigations into likely severity of tsunami generated by an earthquake on the Hikurangi subduction margin, which have provided estimates of at-shore wave heights. The first New Zealand National Tsunami Hazard Review (Berryman 2005) estimated probabilistically a median wave height of 4.5 m at Napier/Hastings from all sources to be a 1 in 500 year event. The review estimated that such an event would cause 320 deaths and 2,100 injuries in Napier and Hastings. Power et al. (2008) simulated tsunami generated by 17 rupture scenarios on the margin, producing modelled tsunami wave heights at-shore for each scenario. The largest at-shore wave heights at Napier are in excess of 5 m and are produced by an MW 9.0 rupture of the whole margin and an MW 8.2 simultaneous rupture of the plate interface offshore Hawke’s Bay and the Lachlan (splay) Fault in Hawke’s Bay. Rupture of the lower North Island segment produced wave heights of up to 3 m at Napier. Maximum wave height at Napier due to the rupture of the segment offshore of the Raukumara Peninsula was 1 m, and simulations of the 1947 slow-slip events resulted in negligible waves at Napier. It has not yet been possible to directly validate these tsunami inundation scenarios against observed or measured tsunami inundation due to the absence of well-recorded historic tsunami events at the study area (with the exception of the 1947 Gisborne earthquakes). Therefore, despite the evidence supporting the potential for such events, the results of these scenarios retain a high level of uncertainty.

Utilising detailed topographic data, Hawke’s Bay Regional Council (HBRC) produced tsunami hazard maps (Hawke’s Bay CDEM Group 2011) using an incident wave of 10 m amplitude relative to high tide, initiated approximately 20 km offshore of Napier (Craig Goodier, personal communication, April 15th 2013). Results presented here support the inundation extent of the HBRC hazard maps, but, due to the use of an incident wave rather than an earthquake source mechanism, the HBRC study is unsuitable for the estimation of wave arrival time and does not demonstrate the variability of inundation resulting from different local-source ruptures.

4 Methodology

Tsunami generation, propagation and onshore inundation are modelled using the Cornell Multi-grid Coupled Tsunami (COMCOT) model. COMCOT solves the conservative form of shallow water equations (SWE) in terms of flow velocity and volume flux within an explicit staggered leap-frog finite difference scheme (Cho 1995; Liu et al. 1998; Wang and Power 2011). A nested grid configuration is used to maximise both computational efficiency and accuracy by applying linear SWE in grids 1–3 and non-linear SWE in the onshore grid 4. In the near-shore, numerical dissipation replicates the energy dissipation of wave-breaking, although wave-breaking cannot be explicitly modelled. A moving boundary scheme tracks the moving shoreline during non-linear simulation of onshore inundation. The model has been validated in numerous analytical and physical modelling benchmark tests (Liu et al. 1995a, b; Wang and Liu 2008) and has been used in previous studies of tsunami affecting New Zealand (e.g. Power et al. 2012a, b; Prasetya et al. 2011) and several globally significant tsunami (e.g. Baptista et al. 2011; Wang and Liu 2006, 2007).

4.1 Model set-up

The model comprises four levels of nested grids (Fig. 2) with maximum grid cell resolution of 15 m in the onshore area of interest (Table 1). Although the DEM is at 10 m resolution, significant computational cost is associated with processing non-linear shallow water equations at 10 m cell size. We conducted model sensitivity tests to determine the influence of model grid resolution on computational expense and congruence of inundation extent in 13 simulations before selecting the appropriate resolution to apply in the model. Compared to the highest resolution model (7 m in grid 4, identical resolution in grids 1–3), simulation at 15 m resolution produces <1 % difference in inundation area and inland extent, but requires only 13 % of the computational time.
Table 1

COMCOT model domain information, showing spatial extent and cell resolution of model grids in arc seconds and metres, data source and data resolution

Model grid number

Long. extent (degrees)

Lat. extent (degrees)

Model grid resolution (arc sec)

Model grid resolution (m)

Bathymetry/topography source and resolution

1

166.0–186.0

−48.0 to −33.5

60

1,239–1,852

30 arc sec ETOPO1 updated with GEBCO08/LINZ charts/CMAP

2

176.5–177.7

−39.85 to −38.8

12

284–370

10 arc sec interpolated LINZ charts/CMAP

3

176.7–177.5

−39.75 to −39.1

2.4

57–74

10 arc sec LINZ charts/CMAP

4

176.81–177.18

−39.66 to −39.33

0.48

11–15

10 m DEM created for this study—LiDAR, LINZ Nautical charts

Grids 1–3 are based on worldwide ETOPO1 global relief data (Amante and Eakins 2009), General Bathymetric Chart of the Oceans data (GEBCO 2008) and LINZ (Land Information New Zealand 2006) charts at 10 arc seconds and 30 arc seconds resolution (Table 1). In order to achieve high-resolution inundation modelling in grid 4, we developed a seamless 10-m horizontal resolution digital elevation model (DEM) comprising Light Detection and Ranging (LiDAR) ground elevation data for onshore topography and interpolated LINZ digital bathymetric sounding depths. Original sounding depths are available at irregular spacing on the order of several kilometres distance, so we augmented these data by digitising LINZ nautical charts to provide irregularly spaced depth distances on the order of several hundred metres. To produce a regularly spaced grid of data for use in COMCOT, all bathymetry and topography data were interpolated to a 10 m grid using the ArcGIS Topo To Raster algorithm, based on the ANUDEM program (Hutchinson 1989).

Bottom friction, or surface roughness, is an important cause of tsunami flow resistance and energy dissipation, particularly in shallow water (0–10 m depth) of the near-shore and onshore areas (e.g. Myers and Baptista 2001). Onshore, land cover is used to define surface roughness, commonly in terms of Manning’s coefficient (Arcement and Schneider 1989). The influence of surface roughness on inundation extent, flow depth and velocity in high-resolution modelling has been demonstrated by comparisons between several approaches to implementation of roughness. Muhari et al. (2011) describe the available approaches, such as Topographic Model, whereby building geometry and elevation are incorporated into the DEM; Constant Roughness Model, whereby a uniform surface roughness coefficient is applied throughout the study area; and Equivalent Roughness Model, whereby spatially varying land cover and density of buildings are used to assign variable roughness coefficient in a model domain. The 15 m horizontal resolution of our model is coarser than many buildings and some streets, precluding accurate representation of individual buildings, either in a topographic model or as polygons with a high roughness coefficient surrounded by lower roughness coefficient representing streets (e.g. Gayer et al. 2010; Kaiser et al. 2011). We apply the Equivalent Roughness approach, using New Zealand Landcover Database version 2 (LCDB2; Ministry for the Environment 2009), to assign a roughness coefficient to each generalised land cover category in Napier, resulting in variable roughness across the study area (Fig. 3). Urban land cover is accounted for using a single roughness coefficient without the consideration of individual buildings.
Fig. 3

Surface roughness values (Manning’s coefficients) applied in Napier, based on New Zealand Landcover Database 2 data (LCDB2; Ministry for the Environment 2009)

We assign urban areas as n = 0.030, arable land as n = 0.019 and woodland as n = 0.026, after Wang et al. (2009). We consider the urban value appropriate for the low building density in Napier (10–30 % density in 92 % of city meshblocks and maximum building density of 67 %). This is a lower roughness coefficient than those used in other recent studies for medium-density urban areas (n = 0.059; Kaiser et al. 2011) and populated areas (n = 0.045; Kotani et al. (1998) in Muhari et al. (2011)). The values used for arable and woodland areas are also lower than those used elsewhere in the literature for various types of vegetated and forested land, which vary between 0.025 and 0.26 (Kaiser et al. 2011; Van der Sande et al. 2003). As the purpose of this study is to inform tsunami evacuation planning, a conservative assessment of inundation extent and flow depth is required, hence the use of low roughness coefficient values. Consistent with the previous studies of roughness influence, roughness coefficients applied in this study are static through time and do not account for the destruction of buildings in tsunami flow. The sea bed is represented by homogenous bed friction of n = 0.013 in all grids, although at most water depths in grids 1–3, the impact of friction is negligible.

Five ‘virtual’ tide gauges (Table 2) are used to record time-series water level and flow depth data for the 12 h of simulated flow at key areas of interest. Gauge 1 is located in the near-shore area east of Napier and gauges 2–5 are located onshore at the Port, in the eastern city and at Westshore (Fig. 1C). These virtual gauges exist only in the model to record simulated flow time-series.
Table 2

List of virtual tide gauges with their location and elevation

Gauge number

Location

Long.

Lat.

Elevation/depth relative to MSL (m)

1

Nearshore

176.93

−39.52

−10.4

2

Port

176.92

−39.47

1.8

3

City

176.92

−39.49

3.2

4

City

176.92

−39.53

1.2

5

Westshore

176.89

−39.48

2.63

4.2 Earthquake source mechanisms

The earthquake source mechanisms used in this study represent scenarios local to Hawke’s Bay that are discussed in previous studies as subduction zone earthquakes with the potential to cause significant at-shore wave heights in Hawke’s Bay or replicate geologically recorded subsidence-uplift patterns (e.g. Cochran et al. 2006; Wallace et al. 2009; Power et al. 2008). These scenarios are implemented in COMCOT as instantaneous ruptures using vertical deformation as an initial surface condition (Fig. 4). Vertical deformation is modelled using elastic fault dislocation theory (Okada 1985), and coseismic deformation is accounted for during inundation modelling.
Fig. 4

Vertical deformation applied as tsunami generation in COMCOT, with source patches delineated (grey dashed lines). A Lachlan Fault rupture using simple fault geometry and 9.0 m slip (MW 7.7); B Rupture of the plate interface offshore of Hawke’s Bay (MW 8.4); C Rupture of southern and central Hikurangi subduction margin (MW 8.8); D Rupture of the whole Hikurangi subduction margin (MW 9.0). Vertical scale differs on each image. The background value of approximately zero deformation is retained to indicate the extent of vertical deformation domain. Depth contours (km) of the plate interface model (Ansell and Bannister 1996) are shown in C

Although not considered in this initial investigation of inundation, implementation of dynamic rupture may affect arrival times and wave heights, particularly where rupture length is on the order of several hundred kilometres (Suppasri et al. 2010). Mean water depth of approximately 1,000 m between the Hikurangi Trench and the coast of the North Island gives wave celerity of approximately 99 m/s using c = gD 0.5 (where g is gravitational acceleration and D is water depth). Applying typical rupture velocities of 1–2.5 km/s to calculate the ratio of rupture velocity to wave celerity (Suppasri et al. 2010), the ratio for our study area is <40, indicating that dynamic rupture is likely to influence wave height and arrival time at the east coast of the North Island. Due to the influence of propagation direction and distance from the tsunami source on the magnitude of this effect, and the rapidly shallowing bathymetry between the Hikurangi Trench and the coast (water depth = 3,400 m offshore Hawke’s Bay; Fig. 2), further research expanding this study is required to constrain the potential influence of dynamic rupture in the context of our study area.

Tidal level at the time of tsunami arrival can be an important factor in determining inundation extent and flow depth. We test the impact of tidal level on inundation in scenarios A and D by simulating the events at mean high water (MHW = MSL + 0.75 m) and discuss the results with reference to the two tidal conditions. There remains significant uncertainty in the specific fault geometry, magnitude and spatial distribution of slip and temporal development of potential ruptures of the Hikurangi margin due to some limitations in current understanding of the Hikurangi subduction margin. Nevertheless, our application of this range of scenarios provides an extension of current knowledge of potential inundation due to local earthquake-generated tsunami.

4.2.1 Scenario A: Lachlan Fault rupture using simple fault geometry (MW 7.7)

The Lachlan Fault is an active upper plate splay fault with a west-dipping thrust fault mechanism and a listric geometry, steepening from a shallow dip of ~15° at 7–8 km depth to 55–70° in the upper 1 km, emerging at the sea floor 3–8 km offshore (Barnes et al. 2002). We use a simplified geometry to represent the Lachlan Fault, applying a uniform dip of 60°. Barnes et al. (2002) identified three segments on the fault with a total length of 80 km and acknowledged there may be a fourth segment present. We extend the length of the Lachlan Fault source to the south, to a conservative length of 139.79 km to account for the possibility of continuation of Lachlan Fault rupture onto other nearby splay fault structures further along strike (such as those related to the Cape Kidnappers anticline) as rupture of multiple splay faults beneath southern Hawke Bay could pose a greater danger to Napier than an isolated Lachlan Fault rupture.

Berryman (1993) identified four Holocene uplift events at the nearby Mahia Peninsula, with a maximum uplift of 4 m, and attributed this uplift to the movement of the Lachlan Fault. Barnes et al. (2002) projected this uplift across the Lachlan Fault and, based on the ratio of previous surface deformation to modelled subsurface displacement on nearby faults, estimated an average dip-slip displacement of 5.0–9.0 m. We define three scenarios of whole-fault rupture at the upper end of this range with respective uniform dip-slip values of 7.0, 8.0 and 9.0 m (Fig. 4A; Table 3); for the 9.0 m scenario, we simulate for both MSL and MHW conditions. We derive the seismic moment (M0) for each scenario using the relationship M 0  = μSD (where S = average slip (m), D = fault area (m2) and μ = rigidity, assumed in all scenarios to be 3*1010 Pa; Stirling et al. 2012). The magnitude of each slip scenario is MW 7.7, using M W  = (LOG(M 0 ) − 9.1)/1.5 (IASPEI 2013), which is consistent with estimates that a rupture of all segments of the Lachlan Fault could generate an event of MW 7.6–8.0. Barnes et al. (2002). This level of slip is believed to represent a recurrence interval of approximately 1,000 years, corresponding to the mean recurrence interval of the four Holocene uplift events (Berryman 1993). The most northerly segment of the Lachlan Fault has a recurrence interval of 615–2,333 years based on the ratio of surface displacement to surface slip rate 0.30–0.65 cm/year (Barnes et al. 2002).
Table 3

Scenario fault parameters

Scenario

Focal plane centre long/lat

Focal plane centre depth (km)

Focal plane depth range (km)

Length (km)

Width (km)

Strike (deg)

Dip (deg)

Rake (deg)

Max. dip-slip (m)

Max. strike-slip (m)

Seismic moment (Nm)

Moment mag. (MW)

A: Lachlan Fault

177.72, −39.53

6.25

0.5–12.0

139.8

13.3

218.5

60.0

90.0

7.0

0.0

3.90E + 20

7.7

8.0

4.46E + 20

7.7

9.0

5.01E + 20

7.7

* B: Offshore Hawke’s Bay

177.77, −39.66

10.6

3.0–26.9

213.7–254.0

121.4–157.3

195.3–219.0

5.0–17.0

62.9–90.0

5.7

1.3

2.80E + 21

8.2

8.5

2.0

4.20E + 21

8.3

11.4

2.6

5.60E + 21

8.4

* C: Southern to central Hikurangi

176.38, −39.87

28.6

3.0–69.6

643.0–736.2

188.9–297.5

195.3–256.0

5.0–30.3

11.3–90.0

22.0

13.6

2.27E + 22

8.8

* D: Whole margin

176.38, −39.87

28.6

3.0–69.6

643.0–736.2

188.9–297.5

195.3–256.0

5.0–30.3

11.3–90.0

33.4

27.1

4.56E + 22

9.0

* Denotes that the scenario uses variable slip parameters on multiple source patches, for which this table presents a summary. A range of slip values, strike and dip angles are given for these events, and full source parameters are presented in supplementary material

4.2.2 Scenario B: Rupture of the plate interface offshore Hawke’s Bay (MW 8.2–8.4)

Cochran et al. (2006) showed that rupture of the plate interface offshore Hawke’s Bay is potentially responsible for past subduction earthquakes and tsunami in Hawke’s Bay, and Wallace et al. (2009) estimate that rupture of the full segment of the interface here could result in an MW 8.3 event. To test the potential tsunami inundation from such an event, we apply a non-uniform slip distribution on a model of plate interface geometry derived from seismicity and seismic reflection data (Ansell and Bannister 1996). The source model comprises 14,143 individual slip patches with 195–219° strike and dip of 5–8° at 3–15 km depth increasing to 10–17° at 15–27 km (see Table 3 for summarised parameters and supplementary material for full parameter table).

We assume slip within the source area of previous slow-slip events at the central Hikurangi subduction margin, which are located at the down-dip limit of shallow (<10–15 km) interseismic coupling (Wallace and Beavan 2010). Rupture occurs with an oblique thrust mechanism and strike-slip component of ±1 m. Maximum dip-slip (5.7 m) occurs at a depth of between 3 km and 5 km offshore of Mahia Peninsula where coupling coefficient is >0.5 (Wallace and Beavan 2010) and slip rate deficit exceeds 25 mm/year (Wallace et al. 2012). Slip diminishes with increasing depth, reducing to zero below the area of slow-slip at a depth of 20–27 km. The seismic moment of each patch sums to a total M0 = 2.80*1021 Nm, equivalent to MW 8.2. In order to explore the impact of increasing slip on inundation, we have applied a linear scaling relationship of 150 % and 200 % to produce three different magnitude events with an unchanged slip distribution: MW 8.2, MW 8.3 and MW 8.4 (Table 3). Rupture of this segment alone is unlikely to produce an earthquake of greater than MW 8.3–8.4 due to the size of the segment, so MW 8.4 represents an extreme case of rupture on this fault and results in maximum subsidence of −0.5 m in the southeast of Napier (Fig. 4B).

4.2.3 Scenario C: Rupture of southern and central Hikurangi subduction margin (MW 8.8)

This non-uniform slip scenario represents a combined rupture of the lower North Island segment from Cook Strait to Cape Turnagain and the central segment from Cape Turnagain to Mahia Peninsula. This type of scenario was proposed as a cause of subsidence events recorded in Hawke’s Bay (Cochran et al. 2006), and the resulting vertical deformation (−0.3 to −0.5 m in Napier) is shown in Fig. 4C. We apply non-uniform slip distribution on the model of plate interface geometry of Ansell and Bannister (1996), comprising 38,186 individual patches. The interface has a strike of 195–220° north of −41.0°S, increasing to 235–255° as subduction becomes more oblique in the southern North Island and beneath the Cook Strait. Dip angle is shallow (5–14°) between 3 km and 20 km depth, gradually steepening to 30° at 70 km depth (Table 3 and supplementary material).

Significant dip-slip (15–21 m) is applied in the slow-slip event source area beneath Hawke’s Bay at depths of 3–10 km where coupling coefficient is 0.5–1.0 (Wallace and Beavan 2010) and slip rate deficit is 15–30 mm/year (Wallace et al. 2012). The lesser amount of slip (9–15 m) applied in the southern part of the margin represents a recurrence interval of 500 years based on slip deficit derived from interseismic locking patterns from GPS data (Wallace et al. 2012). Slip generally occurs with an oblique thrust mechanism (rake of 70–90°), but there is a patch of greater strike-slip component (maximum of 13 m slip) at 30 km depth below the Kapiti coast. Summing all source patches gives M0 = 2.27*1022 Nm, equivalent to MW = 8.8.

4.2.4 Scenario D: Rupture of the whole Hikurangi subduction margin (MW 9.0)

This scenario represents the maximum credible subduction zone earthquake and an upper limit of magnitude due to the rupture of the whole margin. To approximate a worst-case scenario, we apply a non-uniform slip distribution on the same plate interface geometry (Ansell and Bannister 1996) detailed above in Sect. 4.2.3 (Table 3 and supplementary material). Our scenario locates an area of peak dip-slip (20–33 m) offshore the southern North Island at depths of 3–10 km where slip rate deficit is 20–30 mm/year (Wallace et al. 2012), but where the plates are strongly coupled (Wallace and Beavan 2010) and approximate a recurrence interval of 1,000 years. This recurrence interval is analogous to those of great earthquakes at the Japan Trench, first identified by Minoura et al. (2001). The amount of slip during the Great East Japan earthquake and tsunami (Simons et al. 2011), the 1,144-year interval since the last comparable event (Jogan, 869) and their similar inundation extents (Goto et al. 2012; Sugawara et al. 2012, 2013), suggests that both events were generated by a similar style and magnitude of earthquake.

A moderate amount of slip (generally 2–17 m) is transferred north through the central and upper North Island segments into the less coupled slow-slip event source areas. This mirrors slip propagation during the 2011 Great East Japan earthquake, when slip propagated southward into a slightly less coupled area at the southern end of the rupture (Maercklin et al. 2012). An additional patch of high slip (20–33 m) is incorporated at depths of 3–10 km at the northern end of the margin, to account for higher convergence rates (50–60 mm/year; Wallace et al. 2004). As with scenario C, slip generally occurs with an oblique thrust mechanism (rake of 70–90°), but there is an area with a greater strike-slip component (rake 11–30°) of up to 27 m at 30 km depth below the Kapiti coast due to increasing obliquity of motion down-dip. Summing all source patches gives a seismic moment of M0 = 4.56*1022 Nm, equivalent to MW = 9.0. Vertical deformation in Napier resulting from this slip distribution is between −0.6 m in the south-east of Napier and −0.3 m in the north-west (Fig. 4D). This scenario is tested under both MSL and MHW conditions.

5 Results

5.1 Flow depth and inundation extent

Flow depth and inundation extent define the tsunami hazard from each scenario, determining the area to be evacuated and providing an important control on the potential for structural damage. Flow depth is defined as the water surface level relative to ground level at the point of measurement. In calculating flow depth for each of the scenarios, co-seismic deformation has been included, so as to represent flow depth above simulated post-earthquake ground elevation. Flow velocity is also an important control on damage, but the absence of explicit modelling of buildings in this analysis precludes analysis of accurate flow velocity results in the onshore urban area and is not discussed here. The sections below present maximum inundation extent and peak flow depth in each scenario for the most-commonly inundated suburbs and inundation extents (Table 4). There are local variations in flow depth, which are evident in flow depth maps (Figs. 5, 6).
Table 4

Approximate peak flow depth and maximum inundation extent in selected suburbs

Scenario

Maximum inundation extent (m)

Maximum Flow depth (m)

Nelson/Mclean Park

Napier Port

Ahuriri

Westshore

Bay View

Hawke’s Bay Airport

Maraenui/Marewa

A: Lachlan Fault 7 m slip, MW 7.7

300

0.25

2.6

A: Lachlan Fault 8 m slip, MW 7.7

800

0.5

2.8

1.0

A: Lachlan Fault 9 m slip, MW 7.7*

1,000/1,300

0.8/1.0

3.8/4.0

–/0.25

0.7/1.6

2.2/3.4

–/<0.25

0.8/1.0

B: Plate interface offshore Hawke’s Bay MW 8.2

100

1.3

B: Plate interface offshore Hawke’s Bay MW 8.3

100

1.7

0.6

B: Plate interface offshore Hawke’s Bay MW 8.4

600

0.25

2.9

1.0

2.0

2.2

0.4

C: Southern and central Hikurangi margin MW 8.8

3,500

4.2

6.4

2.0

3.7

6.5

1.3

1.0

D: Whole Hikurangi margin MW 9.0*

4,000/5,000

4.5/5.5

6.9/7.5

6.5/6.9

4.1/4.5

7.5/8.1

2.1/2.2

2.6/3.2

Refer to Fig. 5 for inundation mapped for all suburbs. Values refer to inundation at MSL conditions, unless scenario is denoted with *, in which case the first value refers to inundation under MSL conditions and the second to MHW conditions

Fig. 5

Maximum flow depth and inundation extent 1 h after rupture in and around Napier Territorial Authority due to simulated scenarios. Legend and scale are identical for each map. A1, A2: Lachlan Fault rupture using simple fault geometry with 9.0 m slip (MW 7.7) at MSL and MHW, respectively; B Rupture of the plate interface offshore Hawke’s Bay (MW 8.4) at MSL; C Rupture of southern and central Hikurangi subduction margin (MW 8.8) at MSL; D1, D2 Rupture of the whole Hikurangi subduction margin (MW 9.0) at MSL and MHW, respectively

Fig. 6

Maximum flow depth and inundation extent 1 h after rupture in Napier city due to simulated scenarios. Legend and scale are identical for each map. A1, A2 Lachlan Fault rupture using simple fault geometry with 9.0 m slip (MW 7.7) at MSL and MHW, respectively; B Rupture of the plate interface offshore Hawke’s Bay (MW 8.4) at MSL; C Rupture of southern and central Hikurangi subduction margin (MW 8.8) at MSL; D1, D2 Rupture of the whole Hikurangi subduction margin (MW 9.0) at MSL and MHW, respectively

5.1.1 Scenario A: Lachlan Fault rupture using simple fault geometry (MW 7.7)

The MW 7.7 Lachlan Fault rupture with uniform slip of 9.0 m under MSL conditions causes maximum inundation of 1 km and flow depth of 1.0–3.0 m in the 100 m closest to the eastern shore and 3.8 m at the Port (See Fig. 5A1 for inundation mapped in the whole Territorial Authority and Fig. 6A1 for inundation mapped in greater detail in the city). When the amount of slip is reduced to 8.0 m, flow depth does not exceed 2.5 m and maximum inundation is 800 m inland from the eastern shoreline. For the 7.0 m slip scenario, maximum inundation is 300 m and flow depth does not exceed 1.5 m except at the Port, where it reaches a maximum of 2.6 m. Simulation of scenario A with 9.0 m slip at MHW results in more extensive inundation at Westshore and Ahuriri than at MSL (Fig. 5A2, 6A2). Westshore experiences flow depth up to 1.6 m in some residential areas, generally up to 0.25 m in Ahuriri, and in the eastern city, maximum flow depth is generally up to 3.5 m. This event generates at-shore wave heights of 4–5 m, comparable to a 1 in 500 year to 1 in 1,500 year tsunami based on the last National Tsunami Hazard Review (Berryman 2005).

5.1.2 Scenario B: Rupture of the plate interface offshore Hawke’s Bay (MW 8.2–8.4)

An MW 8.2 plate interface rupture of a single-segment offshore Hawke’s Bay is expected to cause very limited inundation to maximum flow depth of 1.3 m at the Port only. The MW 8.3 scenario results in additional inundation at Westshore, exceeding 0.6 m in residential areas. At the eastern shore, there is inundation of the first 100 m on land to a maximum depth of 0.25 m, but inundation does not occur inland of the gravel berm.

The MW 8.4 scenario (Fig. 5B, 6B) causes inundation to 600 m inland in Mclean Park with maximum flow depth of 1.8 m. Napier Port is inundated to maximum depth of 2.9 m, and the eastern part of the Airport is subject to flow depth up to 0.4 m. Ahuriri, Westshore and the northern part of Onekawa West experience flow depth up to 1.0 m. This scenario results in at-shore wave heights of approximately 4–5 m, which corresponds to a tsunami with return period between 1 in 500 years to 1 in 1,000 years (Berryman 2005). These estimated wave heights compare well with those generated by Power et al. (2008) for an MW 8.7 earthquake on the same segment. Analysis of inundation due to combined vertical deformation of scenarios A and B suggests that flow depth is similar to that of the Lachlan Fault component alone. Inundation extent is up to 600 m further inland at Bay View and up to 1 km further inland at Nelson Park with flow depth generally <0.25 m. Elsewhere, inundation extent is the same as for the Lachlan Fault component.

5.1.3 Scenario C: Rupture of southern and central Hikurangi subduction margin (MW 8.8)

An MW 8.8 rupture of the southern and central segments of the margin (Fig. 5C, 6C) results in maximum flow depth exceeding 7.5 m on the eastern side of Bluff Hill and exceeding 5.0 m in the first 100 m inland at Nelson Park and Mclean Park. All other areas of Nelson Park and Mclean Park, and large areas of Maraenui and Marewa, experience flow depth of 1.0–4.0 m. Flow depth at Ahuriri and Westshore is generally between 1.0 and 2.0 m but exceeds 3.7 m locally at Westshore. Maximum flow depth at Napier Port is approximately 6.4 m. Inundation extends over 3 km inland at Bay View with flow depth in residential areas and at the Airport generally up to 1.5 m. This event, with at-shore wave heights of 7–8 m, represents 1 in 2,500-year return period tsunami (Berryman 2005) but is greater than those generated by Power et al. (2008), largely due to the additional rupture of the central segment of the margin that is included in the present study.

5.1.4 Scenario D: Rupture of the whole Hikurangi subduction margin (MW 9.0)

Of the tested scenarios, greatest inundation and flow depth occur due to the MW 9.0 whole-margin rupture. Inundation extends 4 km inland in the city and over 3 km at Bay View (under MSL conditions). Areas in the 100 m closest to the eastern shore could expect flow depth exceeding 5.5 m in Nelson Park (Fig. 5D1, 6D1) while elsewhere on the east side of the city between flow depth is 1.5–4.5 m. Flow depth at Napier Port exceeds 6.0 m, Ahuriri is entirely inundated to 4.5–6.5 m, and Westshore flow depth is 1.0–4.1 m. In this scenario, the inland suburbs of Tamatea North and South are also inundated to a depth of <1.0 m. At Bay View, the majority of residential areas and the Airport are inundated with flow depth of 1.5–2.0 m.

Simulation of scenario D under MHW conditions increases maximum flow depth generally by up to 1.0 m. Inundation extent increases to between 4.5 and 5 km in the city (Fig. 5D2). The relative distribution of flow depth remains consistent with simulation inundation at MSL conditions. Flow depth in Nelson Park and Mclean Park is generally 2.0–4.0 m (Fig. 6D2). Marewa and Maraenui experience flow depth between 1.0 and 3.0 m, and further than 1 km inland, flow depth is consistently <1.5 m. Maximum flow depth at the Port is 7.5 m, and depth consistently exceeds 5.0 m in Ahuriri and 3.0 m in Onekawa West. Flow depth at Westshore is 1.0–4.5 m. Bay View experiences flow depth of 1.5–2.5 m in the majority of residential areas and 2.2 m at the Airport. At-shore wave heights due to this event generally exceed 8 m, which is equivalent to 1 in 2,500 year tsunami based on hazard curves generated by Berryman (2005) and of similar order to those generated by Power et al. (2008) for a MW 9.0 whole-margin rupture.

5.2 Tsunami arrival time and waveforms

Estimated arrival time of the first wave above a given threshold, relative to time of source rupture, is a key determinant in evacuation planning and public education on the need for immediate self-evacuation. Effective tsunami warning and evacuation messages should include advice to the effect ‘The first wave may arrive later and may not be the largest. Waves may continue for several hours’ (MCDEM 2010, p.43). Analysis of tsunami travel time (Fig. 7) and time-series data from five virtual tide gauges (Fig. 8) reinforces the importance of these messages in Napier and provides data with which to conduct simulation of urban tsunami evacuation. Arrival time is recorded when the waveform first exceeds 0.05 m (Table 5) and provides approximate arrival time for the area around each virtual gauge.
Fig. 7

Tsunami travel times for the North Island and northern South Island with key urban centres indicated. Arrival times are shown for waves above the 0.05 m threshold. Insets show arrival times in Hawke’s Bay. A Lachlan Fault rupture using simple fault geometry and 9.0 m slip (MW 7.7); B Rupture of the plate interface offshore of Hawke’s Bay (MW 8.4); C Rupture of southern and central Hikurangi subduction margin (MW 8.8); D Rupture of the whole Hikurangi subduction margin (MW 9.0)

Fig. 8

Comparison of simulated wave forms at key offshore and onshore gauges. Water level (m) is shown for offshore gauge 1, and flow depth (m) is shown for onshore gauges 2, 3, 4 and 5. Water level is adjusted for coseismic deformation

Table 5

Key parameters of waveforms at virtual tide gauges

Scenario

Parameter

Gauge–offshore

Gauge–onshore

1

2

3

4

5

Height (m)

t (mins)

Height (m)

t (mins)

Height (m)

t (mins)

Height (m)

t (mins)

Height (m)

t (mins)

A

Arrival time

 

33

 

38

 

39

 

37

 

n/a

First peak

3.0

35

2.0

39

0.3

40

0.6

38

n/a

n/a

12-h max

4.0

39

2.0

39

0.3

40

0.6

38

0.0

0

12-h min (offshore only)

−2.5

115

        

B

Arrival time

 

4

 

52

 

n/a

 

52

 

58

First peak

−0.4

27

1.3

53

n/a

n/a

0.4

52

0.7

59

First (+) peak

4.5

50

        

12-h max

4.5

50

1.6

59

0.0

0

0.4

52

0.7

59

12-h min (offshore only)

−1.8

130

        

C

Arrival time

 

4

 

49

 

50

 

49

 

55

First peak

−0.5

27

3.0

50

1.6

50

1.6

49

1.9

56

First (+) peak

7.4

47

        

12-h max

7.4

47

3.4

63

1.8

52

2.2

56

1.9

56

12-h min (offshore only)

−2.0

127

        

D

Arrival time

 

4

 

47

 

49

 

47

 

54

First peak

−2.1

38

3.0

50

1.7

50

1.8

49

2.0

55

First (+) peak

8.9

47

        

12-h max

8.9

47

6.6

63

2.0

52

3.0

178

2.8

65

12-h min (offshore only)

−4.7

143

        

Arrival time is relative to a threshold of 0.05 m. Height refers to water level for offshore gauge and flow depth for onshore gauges. The first peak is recorded, and if that peak is negative, the first positive peak is shown in order to demonstrate the lag time of the drawdown and peak water level. Only data due to maximum slip are shown for scenarios A and B

The south and east coasts of the North Island and the northern South Island are at risk of very short arrival times in the simulated scenarios, and Napier is just one of the several urban centres at risk. There is potential for very short arrival times (within 5–10 min) at Wellington, Gisborne and the Marlborough coast from at least one of these scenarios (Fig. 7). Minimum travel time to Christchurch from these scenarios is 51–60 min. Within Hawke’s Bay, the northern coast around Wairoa experiences similar arrival times to Napier.

Due to the west-dipping thrust mechanism of scenario A, wave arrival at gauge 1 offshore Napier is a positive motion occurring at 33 min and water level rapidly increases to 3.0 m at 35 min (Table 5). Maximum water level is 4.0 m at this gauge, occurring only 3 min later. Wave arrival onshore at the Port (gauge 2) occurs 5 min after wave arrival at gauge 1(38 min after rupture) in scenario A. Maximum flow depth occurs on the first wave (2.0 m at 39 min). For scenarios B-D, wave arrival at the Port occurs 47–52 min after rupture with a first peak of 1.3–3.0 m within 3 min of this time. Maximum water level at the Port is 1.6, 3.4 and 6.6 m, respectively, occurring at 59–63 min in each case.

Maximum flow depth in the city centre (gauge 3) and south-eastern city (gauge 4), albeit limited to 0.3 and 0.6 m, respectively, occurs at 38–40 min after rupture in scenario A. Scenarios B-D cause negative wave arrival at gauge 1 only 4 min after rupture, with peak drawdown of approximately −0.5 m at 27 min (scenarios B and C) and −2.1 m at 38 min (scenario D). In these scenarios, the first positive peak at gauge 1 occurs at 47–50 min, measuring 4.5, 7.4 and 8.9 m, respectively. The lag time between peak drawdown and peak positive wave at gauge 1 is 23, 21 and 9 min, respectively, and there are at least 43 min between nearshore wave arrival and occurrence of onshore inundation. Therefore, wave arrival and drawdown of the sea provide additional natural warnings after ground shaking, with some subsequent time to evacuate. Wave arrival at Westshore (gauge 5) occurs 6–7 min later than at the eastern side of the city. In scenario B, gauge 5 records maximum flow depth of 0.7 m at 59 min. Scenario C results in peak water level of 1.9 m at gauge 5 on the first wave at 56 min. Scenario D shows similarly rapid rise of flow depth to 2.0 m upon wave arrival at 55 min at gauge 5, with peak flow depth of 2.8 m occurring 10 min later.

Further waves and reflections occur around the coastline for several hours causing significant fluctuations in water level. Scenario D shows five periods of drawdown greater than −2.0 m over 7 h at gauge 1, and wave heights of 2 m are recorded as much as 12 h after rupture. Minimum water level (−5.3 m) occurs at 143 min, followed by a further peak water level exceeding 4 m at 175 min. Despite several later wave arrivals, the high beach and gravel berm at the eastern shore (Fig. 1C) provide protection to the city centre. The berm prevents onshore inundation due to later wave arrivals, but also mitigates the impact of initial waves of significant height. For example, following rupture of the whole margin, gauge 1 records a peak water level of 8.3 m on the first wave, which is mitigated to flow depth of 1–3 m in Nelson Park and Mclean Park, immediately onshore. Despite the presence of the berm, the largest wave at gauge 4 occurs at 178 min (the third wave arrival), which highlights the importance of not returning to the inundation zone after the first wave. A greater number of waves inundate the more-exposed Westshore Peninsula and the Port, with three waves greater than 1.0 m affecting the Port in 9 h due to scenario C and at least nine waves in 9 h due to scenario D with a period of approximate 60 min. Westshore is inundated by five waves during the 6.5 h following a whole-margin rupture. Inundation reaches its maximum extent in much of the city within 60 min of rupture in scenarios A and B and reaches close to its full extent after approximately 70 min in scenarios C and D. The simulations also suggest that standing water remains for up to 12 h, in some cases, exceeding a metre for much of that time, which may have implications for emergency first responders.

5.3 Structural damage potential

An estimated 92 % of structures in the study area are of 1–2 storey light timber construction, 3 % are 1–2 storey reinforced concrete shear wall, and 3 % are 1–2 storey concrete masonry (Cousins 2009; King et al. 2008; King and Bell 2009). The high proportion of light timber structures is typical of the national building stock (Shelton and Beattie 2011). Onekawa West and Nelson Park have lower proportion of light timber construction as these suburbs contain a greater proportion of industrial buildings and commercial buildings, respectively. Nelson Park, the main commercial and civic services area, has a higher proportion of concrete masonry and RC shear wall construction.

Comparison of simulated flow depth in Napier with a large tsunami fragility data set from the Great East Japan tsunami (Ministry of Land Infrastructure Transport and Tourism 2012; Suppasri et al. 2013) and fragility curves derived from the American Samoa 2009 and Chile 2010 tsunami (Mas et al. 2012) shows that there is significant potential for structural damage from the simulated scenarios. In the absence of detailed data on tsunami resistance of New Zealand structures, these fragility data provide an approximation of likely damage. We do not consider here the likely damage to structures in Napier due to the initial earthquake. There is likely to be increased vulnerability of some buildings to tsunami loading due to ground-shaking induced weakening of structures and the presence of loose debris that may be entrained in the tsunami flow.

Maximum flow depth in our simulations exceeds the depth threshold of 2 m, at which damage to timber frame building stock becomes significant enough to cause collapse, based on the 2011 data and earlier international data sets reviewed by Suppasri et al. (2013). Figure 9 shows the probability of moderate damage, major damage and collapse of one-storey timber frame buildings and two-storey reinforced concrete building for scenario D using fragility curves from Suppasri et al. (2013) and Mas et al. (2012). The figure also shows the probability of moderate damage to timber frame structures in each of the other three simulated scenarios. For both RC and timber buildings, there is greater than 90 % probability of moderate damage (defined as ‘Slight damages to non-structural components’, ‘Possible to be use after moderate reparation’; Suppasri et al. 2013) in much of the inundated area. Damage potential remains similar for buildings of the same construction with one or two storeys, but the probability of collapse is significantly reduced when a building of a particular construction type is three storeys in height.
Fig. 9

Damage potential maps providing comparison between: i. timber frame building damage levels due to a whole-margin rupture using fragility curves from Japan 2011 (Suppasri et al. 2013) (AC) and Chile 2010 (Mas et al. 2012) (D); ii. RC building damage levels due to a whole-margin rupture using fragility curves from Japan 2011 (Suppasri et al. 2013) (EG) and American Samoa 2009 (Mas et al. 2012) (H); iii. probability of moderate damage to timber frame building due to each scenario using fragility curves from Japan 2011 (Suppasri et al. 2013) (A, IK)

Probability of major damage (‘Heavy damages to some walls but no damages in columns’, ‘Possible to be use after major reparations’; Suppasri et al. 2013) in scenario D exceeds 50 % for both construction types within 1,000 m of the coastline (which includes many of the commercial, industrial and civic structures, the Port and Airport) and is generally 10–50 %, further inland. Probability of collapse (‘Destructive damage to walls (more than half of wall density) and several columns (bend or destroyed)’, ‘Loss of functionality (system collapse), Non-repairable or great cost for retrofitting’; Suppasri et al. 2013) of timber buildings exceeds 40 % within 1,000 m of the shore and is generally <30 %, further inland. In contrast, probability of collapse of RC buildings is <30 % for most of the inundated area. Damage potential based on Mas et al. (2012) is comparable to those of Suppasri et al. (2013), showing slightly higher probability of collapse for both construction types. Figures 9I–K show the damage potential for one-storey timber frame building in scenarios A-C, for comparison against scenario D. The limited inundation extent of scenarios A and B precludes the potential for damage at large distances inland. In scenario A, probability of moderate damage exceeds 60 % within 150 m of the shore but rapidly diminishes to <10 % further inland. In scenario B, probability of moderate damage is >80 % within 150 m of the shore and exceeds 60 % locally between 150 and 600 m inland. Scenario C shows greater inland attenuation of damage potential than scenario D, but there remains a probability of >80 % of moderate damage for areas closest to the coast.

6 Conclusions

Deterministic analysis of local tsunami generated by subduction zone earthquake sources proximal to Napier, Hawke’s Bay, New Zealand, demonstrates the potential for extensive inundation with flow depth sufficient to cause major damage to structures. We demonstrate the variability of inundation in Napier due to different local earthquake scenarios in addition to the maximum credible earthquake. The scenarios analysed are based on geodetic and geophysical characteristics of the Hikurangi subduction margin, and geological evidence for significant pre-historic subduction earthquake and tsunami.

Rupture of multiple segments of the plate interface causes the most extensive inundation, to 4 km inland, and greatest flow depth generally up to 3 m in the city centre, but between 4.5 and >8.0 m in Ahuriri, close to Bluff Hill and at the Port. At the upper limits of their potential magnitude range, ruptures of the Lachlan splay fault and the plate interface offshore Hawke’s Bay also have the potential to cause onshore inundation with sufficient flow depth to cause moderate damage to timber frame and reinforced concrete structures. In the lower magnitude ranges of these events, Napier Port is at risk, but inundation is relatively limited in the rest of the city and extensive damage unlikely. At mean high water, flow depth can increase by the order of 0.5 m and inundation by several hundreds of metres compared to tsunami occurring when the tide is at mean sea level. The maximum simulated flow depths result in high probability of moderate to major damage of timber frame and RC structures in large areas of the city, particularly within 1 km of the coast. There is low probability that RC buildings would suffer collapse due to the maximum simulated flow depths, but there remains a moderate probability of the tsunami causing collapse of timber frame buildings. The results of this analysis are encouraging for the consideration of a vertical evacuation strategy in Napier, which would likely use RC buildings as refuges, but suggest high economic losses among residential building stock, high entrainment of debris in the tsunami flow and a high probability of casualties.

Onshore inundation commences only 37 min after rupture of the Lachlan Fault and at close to 50 min in the other scenarios, suggesting that there is time for evacuation to high ground (Bluff Hill) if evacuation is started immediately. Distances of up to 4 km to the inland extent of inundation in scenarios C and D are likely to present significant challenges for large numbers of people to travel in the time available, and Bluff Hill remains the closest area of land to the city centre that does not become inundated. Evacuation simulations will provide greater detail on this issue. The occurrence of multiple waves onshore reinforces the need for education on staying out of the inundation area until receiving an ‘all clear’ message. Although it would be impossible for those experiencing ground shaking in Napier and any other coastal areas to determine tsunami potential in real time, the short arrival times promote the need to consider ground shaking a natural warning of imminent tsunami. Tsunami earthquakes, during which ground shaking may not be felt onshore, are not considered amongst these scenarios and future research should investigate the potential impact of such events.

The data generated in this study provide dynamic flow input for agent-based evacuation simulations and provide detailed inundation scenarios from which to analyse tsunami exposure and casualty potential. Although we provide an enhanced view of the variability of potential inundation in Napier, several aspects of this work require further research to continue improving tsunami hazard assessment in Napier. This study does not address the frequency of local tsunami hazard due to the current absence of data with which to confidently constrain dates of past ruptures and recurrence intervals of local earthquake sources. We incorporate co-seismic vertical deformation in this study, but we do not consider the potential for the gravel berm to be destabilised during ground shaking or tsunami inundation, nor do we consider earthquake-induced structural damage and liquefaction prior to inundation. These should be included in future hazard assessments. Probabilistic tsunami hazard analysis incorporating spatially and temporally variable slip distribution should be carried out to investigate the impact of source parameters on inundation. Finally, onshore inundation should be simulated with buildings modelled explicitly to analyse flow velocities and accurately assess the structural impact of tsunami loads.

Notes

Acknowledgments

We thank Craig Goodier and Hawke’s Bay Regional Council for the provision of LiDAR data, and Ursula Cochran for early review of the paper. We sincerely thank the three reviewers for providing detailed comments which helped to improve this article. This research was supported by public research funding from the Government of New Zealand. Credits for figures using an Esri ArcGIS basemap layer: GEBCO, NOAA, National Geographic, DeLorme, and Esri (Fig. 1); Esri, i-cubed, USDA, USGS, AEX, GeoEye, Getmapping, Aerogrid, IGN, IGP, and the GIS User Community (Figs. 5, 6, 9); World Shaded Relief, copyright ESRI 2009 (Figs. 4, 7).

Supplementary material

11069_2013_820_MOESM1_ESM.txt (1 mb)
Supplementary material 1 (TXT 1,024 kb)
11069_2013_820_MOESM2_ESM.txt (2.6 mb)
Supplementary material 2 (TXT 2,624 kb)
11069_2013_820_MOESM3_ESM.txt (2.6 mb)
Supplementary material 3 (TXT 2,657 kb)

References

  1. Abe K (1975) Reliable estimation of the seismic moment of large earthquakes. J Phys Earth 23:381–390CrossRefGoogle Scholar
  2. Amante C, Eakins BW (2009) ETOPO1 1 Arc-minute global relief model: procedures, data sources and analysis. NOAA Technical Memorandum NESDIS NGDC-24, 19 p, March 2009. http://www.ngdc.noaa.gov/mgg/global
  3. Ansell JH, Bannister SC (1996) Shallow morphology of the subducted Pacific plate along the Hikurangi margin, New Zealand. Phys Earth Planet In 93:3–20CrossRefGoogle Scholar
  4. Arcement GJ, Schneider VR (1989) Guide for selecting manning’s roughness coefficients for natural channels and flood plains. United States Geological Survey Water-supply Paper 2339, Denver, CO. 38p. http://www.fhwa.dot.gov/bridge/wsp2339.pdf
  5. Baptista MA, Miranda JM, Omira R, Antunes C (2011) Potential inundation of Lisbon downtown by a 1755-like tsunami. Nat Hazards Earth Syst Sci 11(12):3319–3326. doi: 10.5194/nhess-11-3319-2011 CrossRefGoogle Scholar
  6. Barker DHN, Sutherland R, Henrys S, Bannister S (2009) Geometry of the Hikurangi subduction thrust and upper plate, North Island, New Zealand. Geochem Geophys Geosyst 10(2). doi:  10.1029/2008GC002153
  7. Barnes PM, Nicol A, Harrison T (2002) Late Cenozoic evolution and earthquake potential of an active listric thrust complex above the Hikurangi subduction zone. New Zealand. Geol Soc Am Bull 114(11):1379–1405CrossRefGoogle Scholar
  8. Bell R, Sutherland R, Barker DHN, Henrys S, Bannister S, Wallace L, Beavan J (2010) Seismic reflection character of the Hikurangi subduction interface, New Zealand, in the region of repeated Gisborne slow slip events. Geophys J Int 180(1):34–48CrossRefGoogle Scholar
  9. Berryman KR (1993) Age, height, and deformation of Holocene marine terraces at Mahia Peninsula, Hikurangi Subduction Marginm, New Zealand. Tectonics 12(6):1347–1364. doi: 10.1029/93TC01542 CrossRefGoogle Scholar
  10. Berryman KR (2005) (compiler) Review of Tsunami Hazard and Risk in New Zealand. Institute of Geological & Nuclear Sciences Limited. Client report 2005/104. Lower Hutt, New Zealand. http://www.civildefence.govt.nz/memwebsite.nsf/Files/Tsunami_Hazard_report/$file/Final_Hazard_and_Risk_Report-web.pdf
  11. Berryman KR, Ota Y, Hull AG (1989) Holocene paleoseismicity in the fold and thrust belt of the Hikurangi subduction zone, eastern North Island, New Zealand. Tectonophysics 163(3–4):185–195CrossRefGoogle Scholar
  12. Berryman KR, Ota Y, Miyauchi T, Hull AG, Clark K, Ishibashi K, Iso N, Litchfield N (2011) Holocene paleoseismic history of upper-plate faults in the Southern Hikurangi Subduction Margin, New Zealand, deduced from marine terrace records. Bull Seismol Soc Am 101(5):2064–2087CrossRefGoogle Scholar
  13. Chagué-Goff C, Dawson S, Goff JR et al (2002) A tsunami (ca. 6300 years BP) and other Holocene environmental changes, northern Hawke’s Bay, New Zealand. Sediment Geol 150(1–2): 89–102Google Scholar
  14. Cho Y-S (1995) Numerical simulation of tsunami and run-up. Dissertation. Cornell UniversityGoogle Scholar
  15. Clark K, Berryman K, Litchfield N, Cochran UA, Little T (2010) Evaluating the coastal deformation mechanisms of the Raukumara Peninsula, northern Hikurangi subduction margin, New Zealand and insights into forearc uplift processes. N Z J Geol Geophys 53(4):341–358CrossRefGoogle Scholar
  16. Cochran UA, Berryman KR, Mildenhall DC, Hayward BW, Southall K (2005) Towards a record of Holocene tsunami and storms for northern Hawke’s Bay, New Zealand. N Z J Geol Geophys 48(3):507–515CrossRefGoogle Scholar
  17. Cochran UA, Berryman KR, Zachariasen J et al (2006) Paleoecological insights into subduction zone earthquake occurrence, eastern North Island, New Zealand. Geol Soc Am Bull 118(9–10):1051–1074CrossRefGoogle Scholar
  18. Collot J-Y, Lewis KB, Lamarche G, Lallemand SE (2001) The giant Ruatoria debris avalanche on the northern Hikurangi margin, New Zealand; results of oblique seamount subduction. J Geophys Res-Solid 106(9):19271–19297CrossRefGoogle Scholar
  19. Cousins WJ (2009) RiskScape–development of a default assets model for Hawke’s Bay. GNS Science Consultancy Report 2009/50. Lower Hutt, New Zealand. 29pGoogle Scholar
  20. De Lange WP, Healy TR (1986) New Zealand tsunamis 1840–1982. N Z J Geol Geophys 29(1):115–134CrossRefGoogle Scholar
  21. Downes GL (2006) The 1904 MS6.8 MW7.0–7.2 Cape Turnagain, New Zealand, earthquake. B New Zealand Soc Earthq Eng 39(4):182–207Google Scholar
  22. Downes GL, Webb T, McSaveney M et al (2000) The 26 March and 17 May 1947 Gisborne earthquakes and tsunami: implications for tsunami hazard for the east coast, North Island, New Zealand. In: Proceedings of the international tsunami workshop tsunami risk assessment beyond 2000: Theory, Practice and Plans, pp 55–67Google Scholar
  23. Fraser RJ (1998) Historical Tsunami Database for New Zealand. Dissertation. Waikato University, HamiltonGoogle Scholar
  24. Gayer G, Leschka S, Nöhren I, Larsen O, Günther H (2010) Tsunami inundation modelling based on detailed roughness maps of densely populated areas. Nat Hazards Earth Syst Sci 10(8):1679–1687. doi: 10.5194/nhess-10-1679-2010 CrossRefGoogle Scholar
  25. Goff JR (2008) Tsunami Hazard Assessment for Hawke’s Bay Region. Christchurch, New Zealand. NIWA Client Report: CHC2008-021. National Institute of Water & Atmospheric Research Ltd, Christchurch, New ZealandGoogle Scholar
  26. Goto K, Chagué-Goff C, Goff JR, Jaffe BE (2012) The future of tsunami research following the 2011 Tohoku-oki event. Sediment Geol 282:1–13. doi: 10.1016/j.sedgeo.2012.08.003 CrossRefGoogle Scholar
  27. Hawke’s Bay CDEM Group (2011) Latest hazard maps inform tsunami preparedness. http://www.hbemergency.govt.nz/PicsHotel/HawkesBayCDEM/Brochure/Hazards/TsunamiHazardMaps.pdf
  28. Hawke’s Bay Regional Council (2010) Chile tsunami event impacts in Hawke’s Bay February 2010, asset management group technical report. http://docs.niwa.co.nz/library/public/Chile-tsunami-2010.pdf
  29. Hayward BW, Grenfell HR, Sabaa AT et al (2006) Micropaleontological evidence of large earthquakes in the past 7200 years in southern Hawke’s Bay, New Zealand. Quat Sci Rev 25(11–12):1186–1207CrossRefGoogle Scholar
  30. Henrys S, Reyners M, Pecher I, Bannister S, Nishimura Y, Maslen G (2006) Kinking of the subducting slab by escalator normal faulting beneath the North Island of New Zealand. Geology 34:777–780CrossRefGoogle Scholar
  31. Hull AG (1986) Pre-AD 1931 tectonic subsidence of Ahuriri Lagoon, Napier, Hawke’s Bay, New Zealand. N Z J Geol Geophys 29(1):75–82CrossRefGoogle Scholar
  32. Husen S, Kissling E, Quintero R (2002) Tomographic evidence for a subducted seamount beneath the Gulf of Nicoya, Costa Rica: the cause of the 1990 Mw = 7.0 Gulf of Nicoya earthquake. Geophys Res Lett 29(8):74–79. doi: 10.1029/2001GL014045 CrossRefGoogle Scholar
  33. Hutchinson MF (1989) A new procedure for gridding elevation and stream line data with automatic removal of spurious pits. J Hydrol 106(3):211–232. doi: 10.1016/0022-1694(89)90073-5 CrossRefGoogle Scholar
  34. IASPEI (2013) Summary of magnitude working group recommendations on standard procedures for determining earthquake magnitudes from digital data. http://www.iaspei.org/commissions/CSOI/Summary_WG_recommendations_20130327.pdf
  35. Ishii M (2011) High-frequency rupture properties of the M w 9.0 off the Pacific coast of Tohoku Earthquake. Earth Planets Space 63(7):609–614CrossRefGoogle Scholar
  36. Kaiser G, Scheele L, Kortenhaus A, Løvholt F, Römer H, Leschka S (2011) The influence of land cover roughness on the results of high resolution tsunami inundation modeling. Nat Hazards Earth Syst Sci 11(9):2521–2540. doi: 10.5194/nhess-11-2521-2011 CrossRefGoogle Scholar
  37. Kanamori H (1972) Mechanism of tsunami earthquakes. Phys Earth Planet In 6:346–359CrossRefGoogle Scholar
  38. King AB, Bell R (2009) Riskscape project: 2004–2008. GNS science consultancy report 2008/247. Lower Hutt, New Zealand. 153pGoogle Scholar
  39. King AB, Cousins WJ, Heron DW, Matcham I, Pringle R, Bell R, Reese S, Schmidt J, Henderson R (2008) Regional RiskScape: a multi-hazard loss modelling tool. In: 4th International i-Rec Conference 2008: building resilience: achieving effective post-disaster reconstruction. 2008, Christchurch, New Zealand, p. 11Google Scholar
  40. Kiser E, Ishii M (2011) The 2010 Mw 8.8 Chile earthquake: Triggering on multiple segments and frequency-dependent rupture behavior. Geophys Res Lett 38(7). doi:  10.1029/2011GL047140
  41. Kotani M, Imamura F, Shuto N (1998) Tsunami runup simulation and damage estimation by using geographical information system. Proc Coast Eng 45: 356–360 (in Japanese). http://www.journalarchive.jst.go.jp/english/jnltop.en.php?cdjournal=proce1989
  42. Land Information New Zealand (2006). Nautical Chart NZ561–Approaches to Napier. http://www.linz.govt.nz/hydro/charts/nz202-chart-catalogue/catalogue.aspx?Chart=561&page=index.aspx
  43. Lay T, Kanamori H, Ammon CJ et al (2005) The great Sumatra-Andaman earthquake of 26 December 2004. Science 308(5725):1127–1133. doi: 10.1126/science.1112250 CrossRefGoogle Scholar
  44. Lewis KB, Collot J-Y, Lallemand SE (1998) The dammed Hikurangi Trough: a channel-fed trench blocked by subducting seamounts and their wake avalanches (New Zealand-France GeodyNZ Project). Basin Res 10(4):441–468CrossRefGoogle Scholar
  45. Liu PLF, Cho Y-S, Briggs MJ, Kanoglu U, Synolakis CE (1995a) Runup of solitary waves on a circular island. J Fluid Mech 302:259–285CrossRefGoogle Scholar
  46. Liu PLF, Cho Y-S, Yoon SB, Seo SN (1995b) Numerical simulations of the 1960 Chilean tsunami propagation and inundation at Hilo, Hawaii. In: Tsuchiya Y, Shuto N (eds) Tsunami: progress in prediction, disaster prevention, and warning. Kluwer Academic Publishers, Netherlands, pp 99–115CrossRefGoogle Scholar
  47. Liu PLF, Woo SB, Cho Y-S (1998) Computer programs for tsunami propagation and inundation. Cornell University, Ithaca, p 60Google Scholar
  48. Maercklin N, Festa G, Colombelli S, Zollo A (2012) Twin ruptures grew to build up the giant 2011 Tohoku, Japan, earthquake. Scientific Reports 2. doi: 10.1038/srep00709
  49. Mas E, Koshimura S, Suppasri A, Matsuoka M, Matsuyama M, Yoshii T, Jimenez C, Yamazaki F, Imamura F (2012) Developing Tsunami fragility curves using remote sensing and survey data of the 2010 Chilean Tsunami in Dichato. Nat Hazards Earth Syst Sci 12(8):2689–2697. doi: 10.5194/nhess-12-2689-2012 CrossRefGoogle Scholar
  50. MCDEM (2010). Tsunami Advisory and Warning Plan. Supporting Plan [SP01/09] Revised May 2012. Ministry of Civil Defence and Emergency Management Wellington, New Zealand. pp 64Google Scholar
  51. Ministry for the Environment (2009). New Zealand Land Cover Database version 2. http://koordinates.com/layer/1072-land-cover-database-version-2-lcdb2
  52. Ministry of Land Infrastructure Transport and Tourism (2012) For summary of the tsunami-affected urban reconstruction technique research study from the Great East Japan Earthquake. http://www.mlit.go.jp/toshi/toshi-hukkou-arkaibu.html
  53. Minoura K, Imamura F, Sugawara D, Kono Y, Iwashita T (2001) The 869 Jogan tsunami deposit and recurrence interval of large-scale tsunami on the Pacific coast of northeast Japan. J Nat Disaster Sci 23(2):83–88Google Scholar
  54. Muhari A, Imamura F, Koshimura S, Post J (2011) Examination of three practical run-up models for assessing tsunami impact on highly populated areas. Nat Hazards Earth Syst Sci 11(12):3107–3123. doi: 10.5194/nhess-11-3107-2011 CrossRefGoogle Scholar
  55. Myers EP, Baptista AM (2001) Analysis of Factors Influencing Simulations of the 1993 Hokkaido Nansei-Oki and 1964 Alaska Tsunamis. Nat Hazards 23:1–28CrossRefGoogle Scholar
  56. Nakasu T, Takeuchi K, Tanaka S, Yoneyama K, Fujiwara N, Morita T, Sagara J (2011) Meta and longitudinal analyses of high death rates of some particular municipalities in GEJET. IRDR conference 2011 disaster risk: integrating science and practice, Beijing, China. http://www.irdrinternational.org/wp-content/uploads/2011/11/Conference-Booklet_final.pdf
  57. Okada Y (1985) Surface deformation due to shear and tensile faults in a half-space. Bull Seismol Soc Am 75(4):1135–1154Google Scholar
  58. Page MJ, Trustrum NA (1997) A late Holocene lake sediment record of the erosion response to landuse change in a steepland catchment, New Zealand. Z Geomorphol 41:369–392Google Scholar
  59. Pedley KL, Barnes PM, Pettinga JR, Lewis KB (2010) Seafloor structural geomorphic evolution of the accretionary frontal wedge in response to seamount subduction, Poverty Indentation, New Zealand. Mar Geol 270(1):119–138CrossRefGoogle Scholar
  60. Pelayo AM, Wiens DA (1992) Tsunami earthquakes: slow thrust-faulting events in the accretionary wedge. J Geophys Res-Solid Ea 97(11):15321–15337CrossRefGoogle Scholar
  61. Port of Napier Limited (2012) Napier port annual report 2012. Napier, New ZealandGoogle Scholar
  62. Power WL, Reyners M, Wallace LM (2008) Tsunami hazard posed by earthquakes on the Hikurangi subduction zone interface. GNS Science, Science Report 2008/40 Lower Hutt, New ZealandGoogle Scholar
  63. Power WL, Wallace LM, Wang X, Reyners M (2012a) Tsunami Hazard posed to New Zealand by the Kermadec and Southern New Hebrides subduction margins: an assessment based on plate boundary kinematics, interseismic coupling, and historical seismicity. Pure Appl Geophys 169(1–2):1–36CrossRefGoogle Scholar
  64. Power WL, Wang X, Lane E, Gillibrand P (2012b) A probabilistic tsunami hazard study of the auckland region, part I: propagation modelling and tsunami hazard assessment at the shoreline. Pure Appl Geophys. doi:  10.1007/s00024-012-0543-z
  65. Prasetya G, Beavan J, Wang X, Reyners M, Power WL, Wilson K, Lukovic B (2011) Evaluation of the 15 July 2009 Fiordland, New Zealand Tsunami in the Source Region. Pure Appl Geophys 168(11):1973–1987CrossRefGoogle Scholar
  66. Reyners M (1998) Plate coupling and the hazard of large subduction thrust earthquakes at the Hikurangi subduction zone, New Zealand. New Zeal J Geol Geop 41(4):343–354CrossRefGoogle Scholar
  67. Scholz CH (1998) Earthquakes and friction laws. Nature 391:37–42. doi: 10.1038/34097 CrossRefGoogle Scholar
  68. Scholz CH, Small C (1997) The effect of seamount subduction on seismic coupling. Geology 25(6):487–490CrossRefGoogle Scholar
  69. Shelton R, Beattie G (2011) Timber Framed Buildings and NZS 3604. BRANZ. Judgeford, New Zealand. http://www.timberdesign.org.nz/files/00556%20Roger%20Shelton.pdf
  70. Simons M, Minson SE, Sladen A et al (2011) The 2011 magnitude 9.0 Tohoku-Oki earthquake: Mosaicking the Megathrust from seconds to centuries. Science 332(6036):1421–1425. doi: 10.1126/science.1206731 CrossRefGoogle Scholar
  71. Statistics New Zealand (2012) Territorial authority (TA) by total accommodation: April 2012. http://www.stats.govt.nz/browse_for_stats/industry_sectors/accommodation/ accommodation-pivot-tables.aspx
  72. Statistics New Zealand (2013) Estimated resident population, territorial authority areas, at 30 June 2006, 2011, and 2012. http://www.stats.govt.nz/~/media/Statistics/Browseforstats/SubnationalPopulationEstimates/HOTPYe30Jun12/snpe-at30jun12-all-tables-prov.xls. Accessed 16 April 2013
  73. Stirling M, McVerry G, Gerstenberger M, Litchfield N, Van Dissen R, Berryman KR, Barnes PM et al (2012) National seismic hazard model for New Zealand: 2010 update. Bull Seismol Soc Am 102(4):1514–1542. doi: 10.1785/0120110170 CrossRefGoogle Scholar
  74. Sugawara D, Goto K, Imamura F, Matsumoto H, Minoura K (2012) Assessing the magnitude of the 869 Jogan tsunami using sedimentary deposits: prediction and consequence of the 2011 Tohoku-oki tsunami. Sediment Geol 282:14–26. doi: 10.1016/j.sedgeo.2012.08.001 CrossRefGoogle Scholar
  75. Sugawara D, Imamura F, Goto K, Matsumoto H, Minoura K (2013) The 2011 Tohoku-oki Earthquake Tsunami: similarities and Differences to the 869 Jogan Tsunami on the Sendai Plain. Pure Appl Geophys 170(5):831–843. doi: 10.1007/s00024-012-0460-1 CrossRefGoogle Scholar
  76. Suppasri A, Imamura F, Koshimura S (2010) Effects of the rupture velocity of fault motion, ocean current and initial sea level on the transoceanic propagation of tsunami. Coast Eng J 52(2):107–132. doi: 10.1142/S0578563410002142 CrossRefGoogle Scholar
  77. Suppasri A, Mas E, Charvet I, Gunasekera R, Imai K, Fukutani Y, Abe Y, Imamura F (2013) Building damage characteristics based on surveyed data and fragility curves of the 2011 Great East Japan tsunami. Nat Hazards 66(2):312–341. doi: 10.1007/s11069-012-0487-8 CrossRefGoogle Scholar
  78. Tanioka Y, Satake K (1996) Fault Parameters of the 1896 Sanriku tsunami earthquake estimated from tsunami numerical modeling. Geophys Res Lett 23(13):1549–1552CrossRefGoogle Scholar
  79. Taylor FW, Briggs RW, Frohlich C et al (2008) Rupture across arc segment and plate boundaries in the 1 April 2007 Solomons earthquake. Nat Geosci 1(4):253–257CrossRefGoogle Scholar
  80. Van der Sande CJ, de Jong SM, de Roo APJ (2003) A segmentation and classification approach of IKONOS-2 imagery for land cover mapping to assist flood risk and flood damage assessment. Int J Appl Earth Obs 4(3):217–229. doi: 10.1016/S0303-2434(03)00003-5 Google Scholar
  81. Wallace LM, Beavan J (2010) Diverse slow slip behavior at the Hikurangi subduction margin, New Zealand. J Geophys Res 115:B12402. doi: 10.1029/2010JB007717 CrossRefGoogle Scholar
  82. Wallace LM, Beavan J, McCaffrey R, Darby D (2004) Subduction zone coupling and tectonic block rotations in the North Island, New Zealand. J Geophys Res 109:B12406. doi: 10.1029/2004JB003241 CrossRefGoogle Scholar
  83. Wallace LM, Reyners M, Cochran UA et al (2009) Characterizing the seismogenic zone of a major plate boundary subduction thrust: Hikurangi Margin, New Zealand. Geochem Geophys Geosyst 10:Q10006. doi: 10.1029/2009GC002610 Google Scholar
  84. Wallace LM, Barnes P, Beavan J et al (2012) The kinematics of a transition from subduction to strike-slip: an example from the central New Zealand plate boundary. J Geophys Res 117:B02405. doi: 10.1029/2011JB008640 Google Scholar
  85. Wang X, Liu PLF (2006) An analysis of 2004 Sumatra earthquake fault plane mechanisms and Indian Ocean tsunami. J Hydraul Res 44(2):147–154CrossRefGoogle Scholar
  86. Wang X, Liu PLF (2007) Numerical simulation of the 2004 Indian Ocean tsunami - coastal effects. J Earthq Tsunami 1(3):273–297CrossRefGoogle Scholar
  87. Wang X, Liu PLF (2008) Numerical simulation of tsunami runup onto a three-dimensional beach with shallow water equations. In: Liu PLF, Yeh H, Synolakis CE (eds) Advanced numerical models for simulating tsunami waves and runup. World Scientific Publishing Co. Pte. Ltd. pp 249–253Google Scholar
  88. Wang X, Power WL (2011) COMCOT: a tsunami generation propagation and run-up model. GNS Science, Science Report 2011/43. Lower Hutt, New ZealandGoogle Scholar
  89. Wang X, Prasetya G, Power WL, Lukovic B, Brackley H, Berryman KR (2009) Gisborne District council tsunami inundation study. GNS Science Consultancy Report 2009/233, Lower Hutt, New ZealandGoogle Scholar
  90. Yamanaka Y, Kikuchi M (2004) Asperity map along the subduction zone in Northeastern Japan inferred from regional seismic data. J Geophys Res-Solid 109(B7). doi:  10.1029/2003JB002683

Copyright information

© Springer Science+Business Media Dordrecht 2013

Authors and Affiliations

  • Stuart A. Fraser
    • 1
  • William L. Power
    • 2
  • Xiaoming Wang
    • 2
  • Laura M. Wallace
    • 3
  • Christof Mueller
    • 2
  • David M. Johnston
    • 1
    • 2
  1. 1.Massey UniversityWellingtonNew Zealand
  2. 2.GNS ScienceLower HuttNew Zealand
  3. 3.University of TexasAustinUSA

Personalised recommendations