Search This Blog


Friday, 24 October 2014

Phase 2 Alpine Fault Drilling

Rupert Sutherland with DFDP-2 flags
Whilst researchers continue to pull together the history of past Alpine Fault earthquakes, the Deep Fault Drilling Programme is well underway in Whataroa on the West Coast of the South Island. For an introduction to this project have a look at my blog and video here, or check out the DFDP-2 Facebook page or project leader Rupert Sutherland's blog for updates over the next few weeks.

The first phase of the drilling process was to penetrate down through a thick sequence of gravel and mud left behind in the Whataroa Valley after the retreat of ice at the end of the last ice age. This was surprisingly challenging because of a thick sequence of very sticky mud that was deposited in the valley at a time when it was a deep fiord or lake.
DFDP-2 drill site   J.Thomson@GNS Science
Eventually the team struck bedrock 240 metres below the surface, and the second phase could commence. This involves drilling down towards the fault plane, thought to be about a kilometre below the rig, without trying to retrieve any large intact pieces of the rock at this stage. (That process is the goal of phase three, which will start when the geologists see from the minerals in the rock fragments that the drill is closing in on the Alpine Fault.)

DFDP-2 drill site   J.Thomson@GNS Science

This is a view of the drill site on a nice morning with Phase 2 well established and the drill at a depth of 340 metres. Behind the rig you can see the drilling mud ponds. The science labs are on the right and spare drilling rods that are added as the drill gets deeper are in the foreground.

This is a view of the site, science labs and working area. The  study the rocks being brought up by the drill, and make geophysical measurements taken by equipment that is lowered down the borehole.

Close up to the rig you can see the vertical drill rod (or pipe) that is rotating and gradually descending down the drill hole. The next rod is lined up ready for connecting when the drill is a few metres deeper. The speed of drilling is roughly 1 to 4 metres an hour at this stage, and a new drill rod is added about every 6 hours.

Next to the drill is this pond of muddy water, which is a vital part of the system used for cutting down into the rock. The mud is specially formulated to have the right viscosity and density and is sucked up by a very powerful pump. After having large particles sieved out of it, it is sent down the centre of the drilling pipe right down to the cutting face of the drill bit.

The drill bit on the right has cut through about a hundred and twenty metres of bedrock, and is about to be replaced by the nice shiny one on the left. The drilling mud is forced out of the holes that you can see, and then flows up the outside of the drill pipe back to the surface, bringing with it the rock chips and also carrying heat away from the cutting face at the same time.

This is the base of the drill rig, with a section of the rotating drill pipe visible. Drilling mud is flowing down the centre of it on its way down to the drill bit. After its return journey on the outside of the drill pipe, loaded with rock fragments, it emerges at ground level and is carried away in the pipe that extends to the right.

The drilling mud flows into a collection pond. The sieve that you see is for collecting samples of the rock fragments for analysis.
 The samples are first carefully washed of fine mud or clay.
They are then sorted by hand.
After being glued to a microscope slide, the rock samples are ground down to a thickness of 30 microns. They are then transparent and can be analysed using an optical microscope. The mineral content can then be studied in detail. As the drill gets closer to the fault, the scientists expect to be able to see changes in the types of minerals present. In this way they will be able to judge the right time to change the drilling system to phase 3 and start retrieving intact rock cores.

DFDP-2 drill site   J.Thomson@GNS Science
Finally here are a couple more views of the DFDP-2 drill site looking up the Whataroa Valley.
DFDP-2 drill site   J.Thomson@GNS Science

Thursday, 23 October 2014

Nature's Seismometers

In order to make sense of the sediment cores that can be retrieved from lakes near to the Alpine Fault such as Lake Christabel, it is worth having a think about what happens to the environment when the fault ruptures in a large earthquake.
Under normal conditions, alpine lakes fill up very slowly with sediment that is fed into them by rivers. The particles settle onto the lake bed gradually, to create a sequence of finely layered mud.
When an earthquake occurs, a number of consequences affect the landscape. The soft surface sediment on the bed of the lake gets deformed and folded, and the shallower slopes at the side of the lake collapse to create flowing avalanches (turbidites) that sweep down and across the lake floor. In the nearby mountains, large landslides occur that choke the river valleys with a chaotic mix of large and small rock fragments.
In the months and years following the earthquake, the landslide debris is gradually washed into the lake, to form a recognisable layers on top of the turbidite deposit.
Eventually, conditions return to normal, with the finely layered sediments gradually covering over all of the evidence of the earthquake and its aftermath. It may be hundreds of years before another earthquake sttikes that is near enough and strong enough to leave its mark in new layers of the lake sediment.

Now lets have a look at the real thing - an example of a sediment core that has been retrieved from a New Zealand's alpine lake.

Back in the lab at the University of Otago in Dunedin, Jamie Howarth opens a core tube to reveal the layers of sand and mud from Lake Christabel.

Here is a section of the core that shows the finely laminated lake sediments formed in normal conditions (on the right). In the centre you can see that the layers are slightly folded - this is the indication of an earthquake that has deformed these layers. They would have been at or just below the surface of the lake floor at the time.

Here Jamie is indicating the remains of a leaf next to the blade of the knife. This is not far below the earthquake layer, and can be used to get a radiocarbon age which will help to date the earthquake event.

This dark coarse layer is the next layer that was added to the sequence on top of the folded sediment. It is the base of an earthquake generated turbidite deposit. The material gets gradually finer to the left ('upwards') as the cloud of particles slowly settled onto the lake floor.
The section shown here is the landscape recovery phase. Dating of the base and top of this layer in several cores has shown that it can take 50 years for the landscape to recover from an Alpine Fault earthquake. During that time, hillsides are destabilised, debris flows cover flat areas near to the mountains, and rivers are prone to changing course due to being overloaded with sediment.

Finally we see the thinly layered sediment  indicating that normal conditions have returned to the lake environment.
This map shows what can be done when this research is carried out at a number of lakes along the Alpine Fault. The coloured lines (purple, orange, green etc) show earthquake records that have been identified so far in some of the lakes along the length of the fault. You can see that the last earthquake rupture (in 1717 AD) was over 300 km long. The one prior to that around 1600 AD ruptured the northern end of the fault. Information about previous earthquakes is still incomplete, but the picture is starting to become clearer. With more research, Jamie and his colleagues will be able to show a more detailed history of the last 10 Alpine Fault earthquakes including the dates, lengths of rupture and magnitudes of the events.

Friday, 17 October 2014

Lake Christabel

Lake Christabel   J.Thomson@GNS Science
This is Lake Christabel in New Zealand's South Island. It is one of the many beautiful alpine lakes  to be found close to the Alpine Fault.
Lake Christabel was formed when a huge landslide blocked the valley, thus damming the river that then backed up to form the lake.The present day outlet runs over the old landslide deposit of large chaotic boulders.

Hidden beneath the waters of Lake Christabel are very distinctive sediment layers that tell the story of huge earthquakes that rocked the nearby mountains during ruptures of the Alpine Fault. Jamie Howarth from GNS Science, and Sean Fitzsimons from Otago University, have been investigating several such lakes to read the earthquake histories.
I joined them on a recent expedition along with Delia Strong and Rob Langridge from GNS Science. The aim was to retrieve sediment cores from the lake to investigate the earthquake records. First of all a seismic survey was undertaken to find the best spots to sample on the lake bed. Sean is in the lead boat, towing a second dinghy that carries the equipment.

The survey uses an acoustic source that sends pulses down into the water. The boat is towed along so that noise interference produced by a nearby motor is avoided.

As the sound pulses are reflected back from the lake bed and its layers of underlying sediment, they are translated into a two dimensional vertical section image of the lake floor. A number of survey lines are made across the lake to give some idea of the 3 dimensional structure of the lake sediments.
Once the best locations for sampling have been chosen from the survey results, the corer is prepared with a fresh 6 metre pipe that will be pushed into the lake floor to retrieve a sediment core.

The corer is transported to the chosen point on the lake surface, and then dropped off the side of the boat once it is in position.
After being connected with several airlines which are required to control the pressure coring process, the corer is lowered the 90 metres to the lake floor. The large barrel sits at the bottom, and is sucked into the mud to create a stable platform for coring.

High pressure air is then applied to the piston which pushes the 6 metre coring pipe into the mud, releasing clouds of bubbles up to the surface. These bubbles allow Sean and Jamie to monitor what is going on with the corer at depth.

Lake Christabel Corer Retrieval J.Thomson@GNS Science
When the coring is complete, an airbag is attached to the line and filled up with air so that it  pulls the whole assembly out of the mud. The airbag bursts up to the surface from below in a spectacular fashion.

Lake Christabel Corer Retrieval J.Thomson@GNS Science
About a minute later, the corer assembly also emerges from the depths. It is not a good idea to be too close to this as it could easily sink a boat that was in the wrong place.
The corer is then plugged and loaded into the boat to be brought back to shore, with its precious cargo of sediment.

The PVC tube containing the core is then cut into 1.5 metre lengths for ease of transport. Each tube is carefully labelled to avoid any confusion  about where it was taken from and its relationship to the other samples.

Lake Christabel Flight  J.Thomson@GNS Science
Once all the sampling has been completed, the expedition is over. It takes several helicopter loads to transport the two boats, safety gear, corers, generators, samples and all our personal equipment back to the road end. The samples are then taken to Otago University for analysis. My next post will describe how alpine lakes like Lake Christabel have shown themselves to be very useful natural seismometers through this research approach.

Monday, 13 October 2014

Jumping Faults

The Alpine Fault is divided into several segments based on changes in its tectonic structure and earthquake history along the plate boundary.

The northern end of the Alpine Fault is much less straightforward in comparison to the southern and central sections. This is in the area where other faults of the Marlborough Fault System branch off the Alpine Fault and take up a large amount of the total slip. There is still a lot to find out in terms of their combined earthquake histories and how these faults interact in relation to each other.

In 1964, a concrete wall was built across part of a paddock next to the Maruia River, near Springs Junction (see yellow dot on the map above). The wall is 24 metres long, about 1.5 metres high, and at first sight seems pointless, standing alone and unconnected with any other structure. I visited this location recently with Rob Langridge, earthquake scientist at GNS Science, 50 years after the wall was built.

The wall was built directly across the Alpine Fault by scientists who wanted to test whether it would be gradually pulled apart by slow sideways creep along the fault. As you can see - it has suffered no damage due to any gradual movement since it was built.This very clear finding is in accordance with our present understanding that most New Zealand active faults are locked. They do not gradually creep between rupture events, but do all their moving in sudden jumps - during earthquakes.

Right next to the experimental wall, there is an overgrown stream channel that has been offset sideways by about 10 metres along the line of the fault.

Some years ago, a series of pits were excavated to assess the age of the offset river features. In one pit a piece of buried wood was found and then radiocarbon dated, showing that the surface is about 1200 years old. This means that the 10 metre offset has occurred since this time, giving an annual slip rate (rate of movement) of the alpine fault about 8 mm at this location.  This compares with about 27mm per year for the central and southern sections of the Alpine Fault, further south.

The last rupture here at Springs Junction in about AD 1600 offset a nearby river terrace by about 1.5 metres. This suggests that at least two earthquakes will have accumulated the 10 metres of offset of the stream channel.

Tuesday, 7 October 2014

A Ticker Tape Record of Alpine Fault Earthquakes

The famous NASA image of New Zealand's South Island clearly shows the trace of the Alpine Fault along the straight western edge of the Southern Alps.

This oblique Google Earth view of the West Coast shows the relative uplift on the eastern side of the fault that has created the Southern Alps. The Hokuri Creek location indicated on the image is the site of a very important record of Alpine Fault earthquakes going back for the last 8000 years. A significant feature on this part of the fault is that it has actually been uplifting on its western side, instead of to the east as it does along most of its length.

These diagrams show how the amazing record of earthquakes was created at Hokuri Creek.

The first picture shows how the low fault scarp blocks the stream, ponding the water and creating a swamp. This gradually fills up with carbon rich plant material (peat) so that eventually the surface of the swamp becomes level with the top of the scarp and the river flows straight across it.
When an Alpine Fault earthquake occurs, the fault uplifts by about a metre, creating a space into which the creek brings loose rock debris (gravel, sand and silt) washed down the river from landslides caused by the groundshaking.
A new swamp develops once the land stabilises until the peat layer again reaches the level of the top of the scarp.

Another Alpine Fault earthquake uplifts the scarp, and a new earthquake debris layer is deposited, adding another record to the fault rupture history. The great thing about peat layers is that they provide plenty of carbon material for radiocarbon dating.The time of past earthquakes is at the horizon where the peat changes upward into landslide derived sand and silt, so by taking samples at this layer, the earthquakes can be dated.

At Hokuri Creek, by a quirk of fate, the river found a new outlet several hundred years ago, eroding downwards and exposing the sequence of peats and earthquake debris layers like pages in a book.This does mean however that the record of the two or three most recent earthquakes is not available here. You can see how  scientists used a ladder to access good sampling points.

Amazingly, they were able to trace 24 earthquakes going back over the last 8000 years. Records of the two or three most recent 'quakes, missing from this sequence, have been found in a more recent study at the nearby John O'Groats swamp. Scientists were able to recover several sediment cores there that complete the sequence.
Carbon dating gives a date range within which the most likely date is at the peak of the probability curve. (To understand radiocarbon dating have a look at this GNS Science video. ) The dating results that you can see on this graph show how the Alpine Fault at Hokuri Creek has been rupturing in a very regular cycle over the last  8000 years. The intervals do vary - from 140 to 510 years, but the average is 330 years. In fact the most common actual interval between Alpine Fault earthquakes is 300 years, which is sobering when you realise that the last Alpine Fault rupture was in 1717, just under 300 years ago.

EQ histories compiled by U.Cochran@GNS Science
To give you some idea of the significance of this Alpine Fault history, here is a comparison with three other major transform faults. This shows clearly how the Alpine Fault exhibits by far the most regular earthquake behaviour of a big fault anywhere in the world.

This map shows the location of Hokuri Creek in the southern section of the Alpine Fault. The earthquake record it provides tells us a history of large  events in the southern and central section of the fault.

In northern South Island, the Alpine Fault divides into a number of separate faults know as the Marlborough Fault System. The slip rate (average movement) on the Alpine Fault drops down from about 27mm per year to less than 10, with the remainder being taken up by the other faults nearby. This means that the Hokuri Creek history cannot be applied to the northern (Marlborough) end of the Alpine Fault.

This project has been led by GNS Scientists Kelvin Berryman, Ursula Cochran and Kate Clark. The media release about it can be found here, or go to the GNS Science website learning pages here.