Australasian Science: Australia's authority on science since 1938

Earthquakes with the Midas Touch

The enormous Goldstrike pit in Nevada

The enormous Goldstrike pit in Nevada, USA, was formed about 40 million years ago, possibly due to ancient earthquakes.

By Steven Micklethwaite

Earthquakes are catastrophic events, but the stress changes they generate deep in the Earth mean they have not so much a silver lining, but a golden one.

Gold can be found everywhere in very small amounts. It exists as a trace element in rocks and soils; it is dissolved in seawater; and even plants have the ability to take up gold and retain it dissolved in their tissues.

But before we are tempted to dig up our gardens it is important to understand that the amounts are too small to extract and make money. Typical concentrations are just 2–3 parts per billion: that is, 2–3 grams of gold for every 1000 tonnes of garden rock.

What unique processes concentrate gold and help form mineral deposits in the first place? One answer may be as simple as the natural plumbing generated by earthquakes deep in the crust. By forming intense networks of fractures over short periods of time, earthquakes are able to focus the migration of hot, chemically-rich fluids much like a drain pipe collecting water during a downpour. Under the right conditions, these fluids will react with the rocks they are flowing through, or with other fluids that they mix with, triggering the precipitation of minerals and gold from solution.

By predicting where earthquakes are likely to generate fracturing, we have been able to test this theory and develop a novel approach to gold exploration.

Our Dynamic Earth
Earthquakes are just one form of phenomena that develop on active faults. They are seismic events radiating large amounts of energy, often catastrophically, over just a few seconds.

In order to understand and ultimately predict earthquake hazard, geoscientists have been studying some of the most infamous active faults in the world. Highly instrumented and well understood fault systems, such as the San Andreas Fault in California, represent great natural laboratories that have enabled us to develop and test new theories of how faults work.

In the past 20 years this close scrutiny has also allowed us to discover and observe a whole range of active fault behaviour in addition to earthquakes. We now know that some faults don’t just fail catastrophically but also creep, moving silently at millimetres per year, often accompanied by thousands of barely perceptible mini-earthquakes. Faults may also undergo deep, mysterious vibrations that last for minutes to hours – a process known as “tremor”. Similarly bizarre are earthquake “swarms”, where thousands of relatively small earthquakes occur in the same place in a sequence sometimes lasting weeks to months.

For our purposes the most interesting behaviour is perhaps the earthquake sequence itself. Only rarely does an earthquake occur on its own. Much more commonly earthquake events consist of a mainshock, many thousands of aftershocks and sometimes multiple foreshocks.

A little recognised fact is that a large percentage of aftershocks are predicted to a certain extent. For a start they obey the famous Gutenberg–Richter law, which expresses the relationship between the number of earthquakes in a sequence and earthquake magnitude. As this is a power–law relationship, in any earthquake sequence the number of aftershocks increases enormously for smaller and smaller aftershock magnitudes.

An example is the 2004 Sumatra–Andaman earthquake, where one very large mainshock triggered tens of thousands (if not hundreds of thousands) of aftershocks. Fortunately most of them were too small to be felt at the surface.

Another interesting aspect of aftershocks is that they display a surprising consistency in space. Many aftershocks are social things, tending to hang out together with other aftershocks in the wall rock around a fault. Their clustering is a key aspect of earthquake sequences, and the aftershocks are thought to be forming on networks of small faults around the big fault where the mainshock occurred. Geologists refer to this type of subsidiary faulting and fracturing as “damage”, and it can be readily observed around “fossil” fault systems that are now eroded and exposed at the Earth’s surface.

It is not hard to imagine that if this fault damage is being activated tens of thousands of times deep in the crust then it represents great plumbing for fluids to access. Unfortunately, knowing such things does not help predict which faults will be triggered or where fluids will flow, because fault damage is ubiquitous, at least in the top 10 km of the Earth’s crust. For this we required some additional insight to our observations.

Predicting the Unpredictable
Such an insight has emerged, and from a slightly surprising direction. Our planet is constantly under stress as its great tectonic plates grind past one another. It is the build up of these stresses that ultimately lead to an earthquake on a fault. Faults are subject to shear stress parallel to their plane, and they are also squeezed by forces pressing across them. When the shear stresses become greater than the strength of the fault rock, or when the pressing stresses relax a little, an earthquake is suddenly initiated.

Both the pressing stress and the shear stress, when summed together, are known as Coulomb stress, and this is redistributed by the earthquake. Indeed, any earthquake is likely to cause some rock volumes to increase in Coulomb stress while other nearby rock volumes will develop stress shadows where Coulomb stress has decreased. Could areas of increased Coulomb stress be where aftershocks and future earthquakes are triggered?

This idea was spectacularly verified in 1992, when a series of earthquakes hit southern California in the USA. The magnitude 6.1 Joshua Tree earthquake was followed 3 months later by the magnitude 7.4 Landers earthquake to the north. Just 3 hours 26 minutes after that the magnitude 6.4 Big Bear earthquake struck 20 km west of the Landers event. Geophysicists Geoffrey King, Ross Stein and Jian Lin calculated the areas where Coulomb stress had increased with each subsequent rupture, and were surprised to see that the aftershocks occurring around these events were triggered in zones of stress increase. Furthermore, each earthquake mainshock seemed to have influenced the location of the next earthquake mainshock.

The reason why this discovery was so surprising is that the stress changes in question are remarkably small. The typical shear strength of a rock under compression at 1 km depth is equivalent to around 300 atmospheres pressing down on you. But the stress changes transferred by earthquakes that trigger aftershocks and generate damage are in the order of between one and five atmospheres.

One possible explanation is that the crust in areas of active faulting is close to failure at all times. In these circumstances even small stress changes imposed on the rock around a fault can lead to failure.

In effect the crust is “stressed out” in the same way that a picture can be hung badly on a wall – the picture may have been up for days or weeks, stressed under its own weight and close to falling, so it only takes a small vibration or a slight change in temperature for the picture to fall.

From Earthquakes to Exploration
So how does all this relate to gold deposits? Myself and Stephen Cox at the Australian National University decided to study the indirect influence that stress changes may have had on gold mineralisation, using fossil fault systems that are no longer active but were in the ancient past. The idea was a simple one. If we could calculate stress changes from earthquake events in the past, we could identify which portions of their fault damage were repeatedly activated, allowing the migration of crustal fluids.

To test the idea we worked closely with a number of gold mining companies. This allowed us to calculate stress changes around fault systems where gold deposits were already known to exist.

As we suspected, we found that regions of positive Coulomb stress change matched quite closely with existing gold deposits. In regions like the Mount Pleasant and St Ives goldfields in Western Australia or the Carlin goldfield in Nevada, USA, we not only matched known distributions of gold deposits but also gave predictions about where new gold deposits may be found.

The results of our work seem to indicate that ancient slip events on faults triggered stress changes and damage, allowing fluids to migrate up through the damage networks and ultimately mineralising the Earth’s crust. Although this explanation seems a little convoluted, it is supported by a common feature – that many gold deposits all over the world are located on small faults next door to very large faults. This geometric relationship seems somewhat reminiscent of aftershocks and damage around a mainshock.

Coulomb stress changes are no silver bullet for exploration, although perhaps they have golden highlights. Nevertheless, it is fun to speculate, if a little controversially, where such ideas may take us in the future.

Apart from the obvious use of Coulomb stress calculations as an exploration tool, perhaps we could also create our own gold deposits if ethical and environmental concerns could be addressed. There are some unique places, like the geothermal springs in the Taupo Volcanic Zone in New Zealand, where we know that hot waters are precipitating fine particles of gold at the surface. Sophisticated chemical probes, testing those same waters at depths of more than 1 km, have found significant levels of gold in solution.

Will we ever be able to tap such gold? Perhaps we will be able to form our own gold deposits by repeatedly triggering earthquakes in a controlled manner, and then letting nature run her course.

Steven Micklethwaite is a Senior Research Fellow with the ARC Centre of Excellence in Ore Deposits at the University of Tasmania.