Preliminary Report Regarding Coral Reef and GPS Measurements of Crustal Deformation Associated with the 2 April 2007 Gizo Great Earthquake

Preliminary Report Regarding Coral Reef and GPS Measurements of Crustal

Deformation Associated with the 2 April 2007 Gizo Great Earthquake

Frederick W. Taylor

Institute for Geophysics, The Jackson School of Geosciences

The University of Texas at Austin

Austin, Texas, U.S.A.

Allison K. Papabatu

Department of Mines, Energy and Water Resources

Ministry of Natural Resources, Honiara, Solomon Islands

With Participation by:

Richard Briggs

Department of Earth and Planetary Sciences

California Institute of Technology

Pasadena, California, U.S.A.

Abel Brown

Department of Geosciences

Ohio State University

Columbus, Ohio, U.S.A.

Abstract

The 2 April 2007 earthquake, with an epicenter southeast of Gizo, ruptured several

segments of the Solomon Islands arc convergent plate margin along the northern side of
the San Cristobal-New Britain trench extending from near the southwest corner of
Rendova Island until approximately the eastern end of Bougainville. A zone of uplift
parallels the trench and approximately overlies the seismic rupture zone. Parallel to and
along the northern side of the uplift zone is a wider zone of subsidence. Our observations
of vertical movements from coral reefs and shorelines confirm a lack of uplift of southern
Rendova and Tetepare. Instead, we found slight subsidence of southern Rendova and the
eastern end of Tetepare. This indicates that the swath of uplift terminates somewhere
west of Rendova. On this basis, we infer that rupture of the interplate thrust zone did not
propagate from the epicenter eastward past Rendova and Tetepare Islands despite there
being some aftershocks reported in this area. This observation leads us to seriously
consider the possibility that the sooner or later the interplate thrust zone beneath Rendova
and Tetepare and farther eastward could rupture and produce a large thrusting earthquake
and tsunami. Furthermore it is also possible that failure of the interplate thrust zone on
April 2 has increased stress on the fault farther east as may be indicated by slight
subsidence of southern Rendova and western Tetepare. An alternative interpretation of
reduced seismic hazard for Rendova and Tetepare Islands could emerge if our GPS
measurements on Rendova Island indicate that significant aseismic slip has occurred
beneath Rendova Island over recent years or following the April 2 earthquake. The
possibility of a future Rendova-Tetepare earthquake is supported by the existence of
fossil coral reefs on Rendova Island that document uplifts of 1-3 m during the past few

thousand years. The last such event is dated by

C at roughly 300 years ago. This

provides clear evidence that episodic uplifts have occurred in the past at Rendova and
Tetepare Islands and that it is likely that these uplifts were seismic and could have
generated large tsunamis. Some large coral blocks on the shores of Tetepare could have
been emplaced by such large tsunamis. However, we do not predict that a large
Rendova-Tetepare earthquake is imminent. We can only say that such events have
occurred and are certain to occur again sooner or later and that the 2 April event has the
potential to accelerate the timetable for new events. Ranongga also had large uplifts in
the past similar to that of 2 April 2007. However, because the interpolate thrust zone
beneath Ranongga did rupture we believe that the potential for large thrusting
earthquakes at Ranongga probably will be relatively low for the next few decades until
more elastic energy accumulates in that area.

Introduction
On 2 April 2007 a “Great” earthquake occurred along the convergent plate boundary at
the southern margin of the Solomon Islands arc with the epicenter located off the coasts
of Gizo and Parara. Great earthquakes are designated based on a Richter scale magnitude
of Ms > 8.0 and occur on average only once every few years. The earthquake rupture
propagated westward past Ranongga and the Shortlands and terminated near the eastern
end of Bougainville. This rupture crossed Simbo Ridge, a transform fault system,
separating the Australian and Solomon Sea Plates. Thus, the earthquake involved
coseismic slip of both the Australian and Solomon Sea Plates, which have different
convergence rates and directions relative to the Solomon islands. Convergence rates of
the Australian plate at the New Georgia Group are ~ 100 mm/yr toward ~N75º W while
the Solomon Sea Plate converges with the area farther west at a rate of ~130 mm/yr in a
direction nearly perpendicular to the arc trend. These very rapid tectonic plate
convergence rates have the potential to generate numerous shallow subduction zone
earthquakes if interaction between the Solomon arc and the subducting plate involves a
component of unstable slip that leads to cycles of elastic energy accumulation and release
as earthquakes.

Background
This particular earthquake was unexpected in the sense that the Solomons arc from
approximately Ranongga to the region of Nggatokae has no recorded history of large
shallow subduction zone earthquakes such as occurred on 2 April. It has been suggested
that the relative youth and high temperature of the Woodlark Basin crust would not allow
brittle failure and seismic slip. On the other hand, the low rates of seismicity along the
New Georgia arc sector could be taken as an ominous warning that plate convergence is
being stored as elastic energy rather than being released continuously as numerous small
earthquakes. It appears that the latter interpretation may be closer to the truth. In any
case, the New Georgia Group is tectonically very active with Holocene uplift rates on
Tetepare, Rendova, and Ranongga exceeding 6 mm/yr in some areas. The tectonic uplift
is attributed to strong interaction of the Woodlark Basin seafloor with the base of the arc
as it subducts because the Woodlark sea floor resists being pushed beneath the Solomons
arc. The Woodlark lithosphere is actively forming at seafloor spreading centers offshore
from New Georgia and is thus extremely young and buoyant so that it does not readily

sink into the mantle at the subduction zone. In addition, the Woodlark Basin sea floor is
very rugged and studded with topographically prominent structures such as the Simbo
Ridge and Coleman Seamount, a volcanic edifice rising ~3000 m above the surrounding
sea floor. The outer flanks of Coleman are presently thrusting beneath the arc just 20 km
off the south coast of Rendova and may significantly resist subduction with consequent
deformation of the outer arc revealed by extremely rapid uplift of southern Rendova and
Tetepare Islands. Subduction of the Simbo Ridge is the likely cause of the particularly
rapid uplift of Ranongga.

The extremely active tectonics of the New Georgia Group Islands led us to map Holocene
uplift rates (Mann et al., 1998; Taylor et al., 2005) and to install a sparse network of
seven GPS sites in the area in 1997 sites installed in Western Province with the assistance
of the Department of Mines, Energy and Water Resources. We re-measured that
network in 1998 and 2001 to obtain basic plate (Phillips, 2004). Thus, following the 2
April 2007 event, we possessed baseline studies that enable us to measure total
movement on all seven GPS sites since 2001 that will include the coseismic
displacements which are expected to include several meters of horizontal dislocation as
well as the several meters of obvious uplift reported on Ranongga Island.

Preliminary work on Ranongga, Rendova, and Tetepare Islands indicates that uplift is
accomplished in steps as great as several meters each. Corals that were uplifted and
killed by some of these events have been identified and dated by radiocarbon, especially
on Fearo Island and nearby on Rendova. For that area we identified at least four recent
uplifts with the most recent one about 300 years ago. The past four large uplifts appear to
have occurred within the past 2000 years or so. However, this is not a complete history.
First, the dating should be done by a more accurate method using

230

Th/

234

U because

radiocarbon errors are too large. Second, the geographic extent of these uplifts is not
known so Tetepare may have a distinct uplift history from Rendova. Third, there may be
earthquakes and vertical movements of other types that do not create the larger 1-3 m
steps that we have seen in the record of uplifted corals. However, the uplifted corals do
document the fact that Tetepare, Rendova, and Ranongga have undergone repeated large
uplifts that were most likely accompanied large earthquakes and that such events are
certain to occur again in the future sooner or later.

Measurements of Deformation Associated with the 2 April 2007 Event
Within eight days of the earthquake we had secured a small amount of funding from the
Jackson School of Geosciences of the University of Texas at Austin to conduct a quick
response study of crustal deformation associated with the Gizo earthquake. Dr. Kerry
Sieh of the California Institute of Technology sent Dr. Richard Briggs to assist with
measurements of coral reefs from which we can determine vertical displacements. Dr.
Michael Bevis of Ohio State University sent his student, Abel Brown, with seven GPS
receivers to occupy our seven GPS. This includes one GPS receiver that will remain
indefinitely at the World Fish Center at Nusatupe Island near Gizo to serve as a
continuous GPS station that will, among other benefits, continue to measure ongoing
crustal strain as the Earth adjusts to this large earthquake.

We began deploying the GPS receivers on 18 April and all were functioning by 23 April.
The GPS sites are located on Simbo, the southern tip of Ranongga, at Saboro on
southeastern Vella Lavella, Nusatupe Island next to Gizo, at Munda airport, at Husuzu on
mid-eastern Rendova, and at Rava Point across the Balfour Channel from Tetepare. The
GPS receivers (except Nusatupe) were removed on 8 and 10 May and will be returned to
the United States.

In addition, we also made precise measurements of emerged coral colonies and other
features that will provide a quantitative assessment of the vertical deformation, both
uplift and subsidence. We made coral measurements GPS sites and circumnavigated
Tetepare, Rendova, and Parara and examined much of the coasts of other islands
including New Georgia, Kohinggo, Kolambangora, and Vella Lavella as well as many
smaller islands. We were unable to get out to the Shortlands to evaluate vertical motions
in that region.

The GPS data require processing before they can be interpreted and some adjustments
based on tide models will be applied to the coral measurements. However, the coral data
are sufficient to reveal some important aspects of the earthquake. The uplifted region is
the most important for later use in modeling the earthquake parameters.

Results
In general, the area on the northern side of the convergent plate boundary, or trench, is
expected to be uplifted by a thrust-type subduction zone earthquake. Indeed, we found
that uplift occurred on much of Parara, the reefs extending from the end of Parara toward
Gizo, the southern half of Gizo, all of Ranongga and areas extending into the southern
part of Vella Lavella until approximately Liapari on the east are uplifted. The most
spectacular uplift is on Ranongga, probably because it is the so close to the convergent
plate boundary and because it is being underthrust by the prominent Simbo Ridge. Thus,
the boundary between uplift and subsidence extends through the northern Vona Vona
Lagoon, through the reefs between the end of Parara and Gizo, through Gizo and over
through southern Vella Lavella and onward to the west parallel to the arc trend where it
lies beneath sea level and can not be mapped.

We also detected what we expect to confirm as slight uplift of the southwestern corner of
Rendova including the village of Hopongo. However, uplift clearly does not significantly
affect either Rendova or Tetepare Islands. Instead, much of the south coast of Rendova
appears to have subsided slightly as did the westernmost few km of Tetepare. The rest of
Tetepare appears to be unaffected by vertical deformation. However, most of the west
coast of Rendova has subsided even where it is on the same trend as the uplifted parts of
Parara. We interpret this to indicate that the seismic rupture did not propagate eastward
beneath Rendova and Tetepara. Subsidence of the northern and western parts of Rendova
are expected because of their proximity to the uplifted sea floor offshore and onshore
southern Parara even though most of southern Rendova did not uplift.

On the northern side of the uplifted zone is a broad area of subsidence that we have not
mapped in great detail due to time constraints. Moreover, the geographic pattern of

subsidence is less critical to understanding the event than is the uplift. However, it does
appear that subsidence of more than 0.5 m has affected large areas. We will have better
estimates of subsidence once our colleagues provide tidal adjustments for our
measurements of depths of living corals below water level.

Summary
Critical early measurements of the crustal deformation in the New Georgia Group
associated with the 2 April 2007 earthquake show that the rupture zone extends from near
the southwest corner of Rendova through Gizo and Ranongga toward Bougainville. The
area overlying the rupture zone underwent coseismic uplift on April 2. However, the
rupture zone does not extend eastward beneath Rendova and Tetepare Islands. The
southern coast of Rendova and the western end of Tetepare appear to have subsided.
This could denote increased stress on this part of the convergent plate boundary which
potentially increases the likelihood of a future earthquake on the interpolate thrust fault
underlying Rendova and Tetepare. It also is possible that aseismic slip occurred beneath
Rendova and Tetepare Islands and may have relieved stress and delayed the time of a
future earthquake. Our GPS measurements on Rendova may reveal whether stress was
relieved or if it increased following the 2 April 2007 event. We will report on this as
soon as possible.

We know from previous work that Rendova and Tetepare have undergone episodic
uplifts of 1 to 3 m each that may have been accompanied by tsunamis. We see this
evidence in the form of emerged coral reefs and blocks of coral that appear to have been
thrown on shore Tetepare by large waves. Such events are certain to happen in the
future. However, the question is whether such future events are in the distant or the near
future. At this point we simply do not know.

We also know from previous work that Ranongga had undergone large uplift events in
the past and that its uplift rate is very rapid. The large amount of uplift of Ranongga on 2
April 2007 probably has consumed the stored elastic energy beneath the Gizo-Ranongga
area and thus has reduced the potential for another large earthquake over the next few
decades or longer.

The most important thing that we can now do to advance our understanding of future
seismic risk in the Western Solomon Islands is to document the paleo-earthquake history
of the New Georgia Group, and the Shortlands as well. This will allow us to understand
the relationship between the present event and past events on Rendova, Tetepare, and
Ranongga. For example, did the 2 April event cause uplift similar to that of previous
uplifts of Ranongga? When in the past did Ranongga undergo uplift? Exactly when in
the past have Tetepare and Rendova undergone? This would give us some idea of when a
future earthquake and uplift might occur. We also need to determine the characteristic
tectonic blocks that undergo rupture as units or groups of units in Western Province. For
example, does Rendova Island have an uplift history distinct from Tetepare or do they
uplift at the same time? Knowing this would help us to estimate the size of earthquakes
and their potential hazard in the case of future events. Our goal is to organize a research
project that would answer these questions while also contributing to our understanding of