An international team that drilled almost a kilometre deep into New Zealand’s Alpine Fault, which is expected to rupture in a major earthquake in the next decades, has found extremely hot temperatures and high fluid pressures.
Our findings, published today in Nature, describe these surprising underground conditions. They have broad implications for understanding what happens in the buildup to a major earthquake, and may represent the discovery of a new type of geothermal energy resource.
The Alpine Fault is one of the world’s major plate boundaries and New Zealand’s most hazardous earthquake-generating fault. It runs for 650 kilometres along the spine of New Zealand’s South Island and we know that it ruptures on average every 300 years, producing an earthquake of about magnitude 8.
The last time the Alpine Fault did this was in 1717, when it shunted land horizontally by eight metres and uplifted the mountains a couple of metres. It is expected to rupture in a major earthquake in the next few decades and, even though this may not happen in the next 30 years or even 100 years, we know that the fault is at the end of its seismic cycle.
Other projects around the world have drilled into major faults, but usually just after a major earthquake. The Deep Fault Drilling Project, which involved more than 100 scientists from 12 countries, gave us an opportunity to take a close look at a fault as it builds up to its next rupture. It is the first time this has ever been done on a major fault that is due to fail in coming decades.
We drilled two holes and during our second attempt made it to 893 metres deep. As we drilled deeper, the temperature increased rapidly, at a rate of about 15 degrees Celsius per 100 metres in depth. This is much higher than the normal rate of about 3°C per 100m in depth. At a depth of 630 metres, the water at the bottom of the drill hole was hot enough to boil, if it had been allowed to rise to the surface. The high pressures at depth stop it from boiling.
The hottest boreholes on Earth are mostly found in volcanic regions. We discovered a geothermal gradient – a measure of how fast temperature increases with depth – that is similar to the hottest geothermal energy boreholes drilled into volcanoes of the central North Island; but there are no volcanoes near the Alpine Fault.
There are two processes we think explain the extreme underground conditions at our drill site. An earthquake on the Alpine Fault has two geological effects: mountains are pushed higher and the shaking breaks up rocks.
During an earthquake and over time, the fractured rocks come down in landslides and rivers carry them to the sea. This limits how high the mountains can get. This process has operated for millions of years, with the height of the mountains staying about the same. Eventually, hot rocks from great depth (about 30 kilometres deep, at 550°C) were transported to the surface quickly enough (on geological time scales) that they did not have time to fully cool. Heat is transported from depth by the rock movement.
The other process that helps explain our findings is the rock fracturing, which allows rain water and snow melt to percolate downwards into the mountains so fast that it can move heat towards the valley, where water wells up and discharges. The flow needs to be fast enough so that the heat is not lost along the way, just as a water pipe in your home moves heat from a hot water cylinder to your bath before having time to cool. Water flowing through the rock concentrates heat and raises fluid pressure beneath the valleys.
The hot, high-pressure water beneath the valleys is mostly invisible at the surface, because it mixes with shallow, cold groundwater that flows to a depth of about 50 metres at our drill site. However, most of the valleys in the region where we drilled have a few warm springs that hint at this deeper source of hot water.
The unexpected results of our research are important beyond New Zealand. Other faults around the world that we know are similar to the Alpine Fault may also have extreme conditions that have never been investigated.
Perhaps most significantly, we can now describe and estimate conditions on a geological fault that will rupture in an earthquake. This will help us to develop better computer models of earthquake rupture. It may also help us to explain how some types of geology (for example certain types of gold mineralisation) have formed as a result of similar conditions in ancient earthquakes.
The extreme underground conditions we discovered may result in substantial economic benefits for New Zealand by providing a sustainable and clean geothermal energy resource that could be used by industry and local communities. We expect that similar hot geothermal conditions exist in other nearby valleys, and maybe in some other places in the world that are geologically similar to western New Zealand.
More drilling and measurements are needed to establish the scale of this local resource, its possible uses, and if it is safe to develop.
Preliminary modelling suggests that the earthquake was caused by a rupture of a northeast-striking fault that projects to the surface offshore.
But this may be a complex event, involving several faults on the South Island.
The northern part of the South Island straddles the boundary between the Pacific and Australian tectonic plates.
The jostling between these plates pushes up rocks that create mountains including the Southern Alps and the beautiful Seaward Kaikoura Range, one of New Zealand’s most rapidly uplifting mountain ranges.
The plate motion forces the oceanic crust of the Pacific plate beneath the Australian plate on thrust faults, and also causes the plates to slide laterally with respect to one another on strike-slip faults.
The region affected by the recent earthquake has been one of the most seismically active in New Zealand over the past few years, including earthquakes that occurred as part of the Cook Strait earthquake sequence in 2013. It is likely that these sequences are related given their close spatial and temporal association.
The preliminary analysis strongly suggests that most of the energy release during this earthquake was sourced from the rupture of a roughly 200km-long fault system. This fault system is aligned northeast and dips to the northwest, beneath the northern part of the South Island. It coincides roughly with the subduction thrust in this area.
The potential for large earthquakes on the subduction fault in the lower North Island and upper South Island of New Zealand was recently highlighted by GNS Science, New Zealand’s geological survey. It published evidence for two similar events in the Blenheim area roughly 520-470 years ago, and 880-800 years ago.
Given its setting, this latest earthquake may be structurally complex, involving a mixture of plate boundary thrusting, lateral slip on strike-slip faults, and thrusting within the Pacific plate close to the epicentre, some 15km northeast of Culverden.
The largest aftershocks suggest a mixture of thrusting and strike-slip movements.
Because the fault system was large, and the earthquake apparently started at the southwest end of the fault and propagated to the northeast, the seismic energy was released over a period of up to two minutes.
Large earthquakes produce more long period wave energy than smaller events. The 2011 Christchurch earthquake contained a lot of high-frequency energy and very strong ground accelerations, exposing more than 300,000 people to very strong to intense ground shaking.
In contrast, this recent earthquake was manifested in Christchurch as lower-frequency rolling, and due to the sparse population density in the earthquake region, roughly 3,000 people in the upper South Island experienced strong ground shaking equivalent to the Christchurch earthquake.
Reports are emerging of at least one major fracture in the ground surface that could be related to strike-slip faulting in the Clarence region.
More traces may yet be found given the complexity of the earthquake. Tide gauge analysis will help to understand if a similar trace offshore caused the tsunami.
The earthquake has also triggered liquefaction in coastal areas and in susceptible sediments, and landsliding of up to a million cubic metres along steep susceptible cliffs in the northern South Island.
Most of this damage is probably caused by strong ground shaking, which causes weak ground to move en masse and has resulted in numerous slips and road closures in the central and northern South Island.
But the tidal triggering of earthquakes has been investigated since the 19th century and remains a challenging and controversial field.
Small amplitude and large wavelength tidal deformations of the Earth due to motions of the sun and moon influence stresses in Earth’s lithosphere.
It is possible that, for active faults that are imminently close to brittle failure, small tidal force perturbations could be enough to advance rupture relative to the earthquake cycle, or to allow a propagating rupture to travel further than it might otherwise have done.
But the specific time, magnitude and location of this or any other large earthquake has not been successfully predicted in the short-term using tidal stresses or any other possible precursory phenomenon.
Deliberately vague predictions that provide no specific information about the precise location and magnitude of a future earthquake are not predictions at all. Rather, these are hedged bets that get media air time due to the romantic misinterpretation that they were valid predictions.
Most earthquake scientists, including those that research tidal triggering of earthquakes, highlight the importance of preparedness over attempts at prediction when it comes to public safety.
To this end, GNS Science uses a system of operational earthquake forecasts to communicate earthquake risk to concerned New Zealand residents during an aftershock sequence such as we are now entering.
These forecasts are based on earthquake physics and statistical seismology. The current operational forecast indicates an 80% probability of:
A normal aftershock sequence that is spread over the next few months. Felt aftershocks (e.g. M>5) would occur from the M7.5 epicentre near Culverden, right up along the Kaikoura coastline to Cape Campbell over the next few weeks and months.
This aftershock sequence will probably (98%) include several large aftershocks (some greater than magnitude 6 have already occurred), and for each magnitude 6 aftershock we expect 10 more magnitude 5 aftershocks over the coming days and weeks.