Finding fault

Predicting earthquakes is a highly inexact science, but a project to drill into the very heart of the San Andreas Fault hopes to discover how fault zones actually work. Andrew Lee reports.

Engineers and scientists are poised to start work on the most ambitious project ever to attempt to give people a few day’s notice that armageddon could be coming to their town.

Using the most advanced drilling and sensor technology available, it will attempt to answer fundamental questions about the behaviour of the world’s best-known earthquake site – California’s San Andreas Fault.

Fear of ‘the big one’ – a quake that would kill thousands and cause massive damage – is buried as deep in the collective psyche of California as one of the fault zones running below the state.

The San Andreas Fault Observatory at Depth (SAFOD) project, which has just received federal funding, will attempt to drill straight into the heart of the fault and place sensors there that could reveal whether earthquakes can ever beaccurately predicted.

One of the project’s leaders described SAFOD as the geo-scientific equivalent of the moon landing. If successful, it would also dispel growing scepticism over whether the huge investment in earthquake prediction technology is anything more than an expensive wild goose chase.

Dr. Steve Hickman, a senior research scientist at the US Geological Survey and one of the three project leaders of SAFOD, said the fact that it was possible to even contemplate what he admitted was a ‘highly ambitious’ project was purely down to recent advances in technology in several crucial areas.

‘This project could not have been done 20 years ago,’ said Hickman. ‘But we will be using the very latest drilling technology, and we now have the instruments to do the job. Plus we have the best people at their jobs in the world on our engineering team and if anyone can do it, they can.’

SAFOD is a major component of a much wider $219m (£130m) project called Earthscope, which will attempt to use the full range of technological tools available, including satellites, GPS, radar and high-precision seismographs, to improve scientific understanding of earthquakes, volcanoes and other natural phenomena (see sidebar below).

The scientific imperative behind SAFOD is an attempt to monitor and analyse the conditions deep inside the San Andreas fault. In its submission for funding for the project, SAFOD offered a sobering assessment of the state of scientific knowledge of deep earthquake faults. ‘We know virtually nothing about the composition of the fault at depth our current knowledge of fault zone processes is so poor that not only are we unable to make reliable short-term earthquake predictions, but we are also unable to scientifically assess whether or not such predictions are even possible.’

Hickman put it this way: ‘We don’t properly understand how earthquakes work, which is a serious drawback to understanding how they behave. A doctor can’t fix your heart unless he knows how it works, and that’s the position we are in with earthquakes.’

SAFOD, backed by its massive engineering and technological resources, is the geo-scientific community’s effort to plug that gap. ‘It is our equivalent of a moon landing,’ said Hickman.

The centre aims to gather data on the composition, strength and frictional properties of rocks in the fault, the pressures acting on them the composition of fluid and gases deep below the earth.

The site for the project is Parkfield, the centre of a magnitude 6 quake in the 1960s.Stage by stage over the next five years, the SAFOD team will drill down to the 4km depth, adding arrays of increasingly advanced instruments as they go.

The temperature and pressure at this depth will entail the instruments working in conditions of unprecedented hostility, pushing some to their limits, and in several cases requiring the development of customised systems.

For example, experimental hydraulically-clamped strain and tilt meters will be used at 2km depths, where the temperature reaches 93 degrees C. This will represent the deepest ever deployment of instrumentation of this type and a ‘formidable technical challenge,’ said Hickman.

However, the SAFOD team believes that the possible rewards – even at just half the eventual target depth – could be significant.

While the bottom of the 2km hole will display extremes of temperature and pressure, it will also be a low-noise environment. Located so close to the fault, SAFOD hopes this will enable it to detect seismic events that have never before been recordable.

In May 2005 SAFOD plans to begin drilling the 8.5in hole that will eventually reach a depth of 4km. While the first 2km will be vertical, the remainder will be drilled at an angle of 50 degrees, steering it towards the heart of the fault zone in areas where repeated micro-earthquakes are known to occur, and out the other side.

At these depths the geological features are expected to consist of highly fractured and crushed rock.

The fluid pressures on the instruments are unknown, and SAFOD will use a 15,000psi stack of blow-out preventers (BOPs) – valves designed to allow crews to regain control in the event of an underground blow-out, when liquids or gases flow chaotically inside the hole under pressure.

Two years after completion of this phase, SAFOD will attempt to carry out coring operations from the 4km hole. Using a twin-stage drilling technology called wireline coring, borrowed from the oil industry, up to four 250m offshoots will be drilled from the main hole. wireline coring allows an inner core of samples to be isolated and pulled back to the surface by cables for analysis.

The core holes could yield invaluable data on sliding sections of the San Andreas fault, including sub-faults that have ceased to be active. After initial testing, three of the core holes will be sealed off while the fourth will be used for long-term monitoring. After being fitted with a slotted steel liner, a state-of-the-art geophysical instrument package will be lowered into it via coil tubing.

Together with the instruments in the main hole, the SAFOD team hopes this will provide it with 20 years of data on conditions inside the fault. However, because of the ambitious nature of the project – and their lack of knowledge of the conditions they will find deep inside the earth – the researchers are not banking on the outcome being entirely successful.

Sophisticated ‘logging while drilling’ (LWD) instruments will sit just behind the head of the drill, providing valuable data even while the boring process is taking place.But Hickman is confident SAFOD has assembled the technology to go all the way.

‘We’re using extremely sophisticated technology developed by the oil and gas industry. Those people know how to do it,’ he said.

‘It will be difficult, and with any drilling project you can never quite work out how it is going to go. The whole point of SAFOD is that we do not know exactly what we are going to find down there, or what the conditions will be.

‘But all the way along we have assumed the worst when developing the drilling plan and built in a number of contingencies. Even if we hit pressures of 20,000psi we can deal with that. I’m confident we have thought of everyproblem that could occur.’

Hickman pointed out that SAFOD has already sunk a pilot 2.2km deep holeadjacent to where the main hole will be drilled. ‘We know all about the top 2km because we have already done it.’

By the end of the decade SAFOD and the wider Earthscope initiative hope to be providing the type of data that could firm up the notoriously shaky business of earthquake prediction.

Parkfield has been the site of major research activity for decades, but the value of what has been produced is the subject of increasing debate in the scientific community and beyond. One quake predicted to hit the region in the late 1980s is already 15 years late.

Indeed, some sceptics believe investment in earthquake prediction amounts to not much more than tipping money down a very deep drain, so little has all the effort put into it yielded to date.

‘Prediction is a contentious area,’ confirmed Roger Musson, head of seismic hazards at the British Geological Survey. ‘There are a number of leading scientists who believe prediction will never be possible – in the sense of giving the place, day, and times of forthcoming earthquakes – simply because the nature of earthquake occurrence is inherently chaotic.’

It is also difficult to know exactly how to respond to a prediction, said Musson, given the extreme social consequences.

Predicting an earthquake publicly would cause widespread panic, so researchers would be vilified for getting it wrong. Even if a quake is predicted correctly and in time to evacuate the local population, homes would still be damaged or destroyed.

‘If somebody says there is going to be a big earthquake in Los Angeles, do you evacuate the whole of the city? That’s not very easy,’ said Musson.

‘What people are more concerned with in defending communities are the processes that generate earthquakes, to estimate what sort are to be expected over a period of time. Engineers can use this to design buildings to withstand the sort of earthquakes they are going to be exposed to.’ (see sidebar below).

This, said Musson, is where projects such as Earthscope may be useful – helping researchers to understand the crustal processes involved in plates scraping by one another and where the strain is building up. This would allow them to better estimate the strength of earthquakes that areas such as California are likely to face over the next 30 years.

‘But you will certainly not be able to say there will be an earthquake next Tuesday,’ said Musson.

Hickman and the SAFOD team is making no such claims. ‘The whole earthquake prediction issue is difficult, and we are not talking about hours or days,’ said Hickman. ‘Some scientists think it is fundamentally impossible. But the only way to know for sure is to find out whether earthquakes are preceded by big signals, small signals or no signals at all.

And the only way to find that out is to be not just close to the fault, but inside it. That is where we are going with SAFOD.’

Sidebar: Earthscope: the three components

After more than a decade in the planning stage, Earthscope finally received the green light in May when the US National Science Board approved funding of up to $219m (£130m) for the project over the next five years.

Listed as one of the highest-priority initiatives on the US scientific agenda, Earthscope will attempt to knit together a host of disparate technologies to provide a dramatic advance in understanding of the North American continent ‘by exploring its 3D structure in space and time.’

In addition to SAFOD, it will be made up of three further distinct components.USArray will cover the US with a network of 400 broadband seismometers, creating a grid to provide continuous, real-time data on local seismic activity.

The seismometers are transportable and will be moved every few years to create an observational record of about 2,000 different sites.

Alongside the transportable array of larger seismometers, the Earthscope team will use more than 2,000 portable high-frequency sensors which can be sent to areas of particular interest to carry out short-term intensive observations.

The existing National Seismic Network of permanent monitoring stations will also be beefed up by 40, bringing the total to 130.

The second major component of Earthscope, called the Plate Boundary Observatory (PBO), will be used to study the 3D ‘strain field’ created in the western North American and Pacific region by deformation of its geological plate boundaries.

PBO will build a network of GPS receivers linked to ultra-sensitive borehole strainmeters stretching from Alaska to Mexico at intervals of 200km, providing a map of deformation of the continent to an accuracy of better than 1mm.

For the final piece in its jigsaw, Earthscope is relying on NASA. The project team hopes to work with the agency to launch a satellite mission to provide Interferometric Synthetic Aperture Radar (InSar) coverage of earthquake-prone regions.

With its wide geographical sweep, an InSar satellite would be able to provide continuous strain and shift measurements over broad areas of the continent.

According to Earthscope, images generated by the satellite would ‘provide unique insights into the mechanics of fault loading and earthquake rupture.’ It would also give researchers a tool for mapping subsidence caused by oil extraction.

Sidebar: Building on anti-quake devices

Minimising human casualties by protecting high-rise towers, bridges, roads and major pipelines from damage in the event of an earthquake is arguably just as important, if not more so, as trying to use a crystal ball to predict when and where the next quake will strike.

Existing buildings without sufficient strength to withstand an earthquake can be toughened up by covering masonry walls with wire mesh jackets, or by gluing glass fibre reinforcement strips and plates to walls using epoxy, said Robin Spence, a structural engineer at Cambridge Architectural Research. ‘The idea is to put enough sheer strength into the walls to resist the earthquake forces,’ he said.

Researchers are also attempting to use auto-adaptive controls to allow buildings to protect themselves from earthquakes, according to Mete Sozen, professor of structural engineering at Purdue University in the US.

‘Ideally we can put in electro-hydraulic systems, which would react in response to the ground motion to change the ability of the building to dissipate energy.’

A widespread network of sensors would be installed within buildings, to detect the ground motion caused by an earthquake that could lead to damaging vibrations, he said. ‘Instrumentation is getting less and less expensive, to the point where it will eventually be possible to have dense instrumentation of buildings, with sensors connected using wireless networks, so we will have a much better knowledge of what buildings do during earthquakes, and be able to control them.’

The sensors would communicate the information they collect on the motion of the ground to the electro-hydraulic actuators, or pumps, fitted to the frames of the building. These actuators would use algorithms to calculate the level of pressure they must exert on the building frames to counteract the oncoming motion of the earthquake, to prevent the frames distorting and destabilising the structure.

With adequate funding for research in the area, the technology could be ready to be fitted to buildings in earthquake zones within 10 years. But not enough is being invested in the field, particularly after September 11, said Sozen. ‘9/11 has affected the flow of research support in the US.’

However, deaths in the US caused by tremors remain relatively rare, while thousands have died as a result of major earthquakes in countries such as India and Turkey. So more money also needs to be invested in creating simple, cheap and efficient methods for evaluating and protecting the huge number of buildings vulnerable to earthquakes within developing or poorer countries, said Sozen. ‘That is where the hurt is going to be.’ – Helen Knight.

For up to the minute news on SAFOD, see <a href=’ San+Andreas+Fault+Observatory+Depth&hl=en&lr=&ie=UTF-8&safe=off&sa=G&scoring=d’>Google SAFOD News</a>.

To be sent up to the minute emails on SAFOD from Google, click <a href=’ San+Andreas+Fault+Observatory+Depth’>here</a>.