Archives / 2012 / November

Polygons, columns and joints

Over on her Georney‘s blog, Evelyn Mervine has recently posted a nice piece with some spectacular images of columnar jointing. This seemed like a good opportunity to dust off some field photos, with some more examples of polygonal joint sets in lavas from a variety of settings, to illustrate the diversity of forms that cooling-contraction joints may take in volcanic rocks.

The first example is a late Pleistocene lava flow from the Afar, Ethiopia.


Berihu Abadi, with an example of columnar jointing in a late Pleistocene basalt, Afar, Ethiopia, exposed in the face of recently-opened fracture. Fieldwork carried out as part of the NERC-funded Afar consortium.

Young lavas in the Afar are predominantly fissure-fed basalts, erupted across the topography and air-cooled. Columnar jointing is pervasive in the upper surfaces of these young lava flows, and seems to develop immediately beneath the surficial glassy and vesicular crust that forms on the top surfaces of the modern lava flows.


Uppermost surface of a fresh basaltic lava flow, Afar, Ethiopia. Field of view is about 30 cm. Beneath the surface layer formed by the scales of glassy, vesicular lava, the lava is weakly polygonally-jointed (not visible in this picture). Fieldwork carried out as part of the NERC-funded Afar consortium.

Moving across to Europe, and southern Spain. Here, in the Cabo de Gata, there are some fabulous examples of polygonally-jointed andesitic lava domes, emplaced in a shallow submarine environment. These are Miocene in age, and formed in a transient volcanic arc, which has now been faulted against the Spanish mainland.

At Playa Monsul, polygonally-jointed dykes can be found criss-crossing a fabulous series of hyaloclastite bodies. Monsul is also well known as the location in Indiana Jones and the Last Crusade, where Sean Connery fends off an aerial attack with the help of an umbrella and a flock of seagulls.


Polygonal jointing in the vertical face of a dyke, intruding a submarine sequence of wet sediment and hyaloclastite lavas. Polygons are typically 10-20 cm across. Photograph taken on a University of Oxford undergraduate field trip.

The southern-most cape of the Cabo de Gata has a wealth of collonaded andesites; presumably submarine domes.


Columnar joints in an andesite ‘dome’ of Miocene age, southernmost Cabo de Gata, Spain. This locality offers some spectacular views both of polygonal joints in planform, and also large-scale structures showing the fanning of columns that reveal the orientations of the original cooling surfaces. Location visited on an undergraduate field trip.

The final examples of polygonal jointing come from higher (southerly) latitudes, and higher elevations, where jointing has developed as a result of ice- or snow-contact volcanism. The first example is from the flanks of volcano Osorno, Chile, where characteristic hackly jointing has developed on the outer margins of an andesitic lava flow. This differs from the regular pattern typical of columnar jointing, and is thought to be one characteristic of rapid quenching of lava.


Jose Naranjo, SERNAGEOMIN, and the hackly-jointed lavas from volcan Osorno, Chile. Fieldwork carried out as part of a NERC-funded project by Sebastian Watt.

A little further South, again in Chile, and here is another classic example of a small sub-glacial ridge of andesite, this time near the summit of volcan Apagado, on the Hualaihue peninsula.


Complex polygonal jointing in a ridge of andesite, from the margins of volcan Apagado, Chile. Fieldwork carried out as a part of a NERC-funded PhD project by Sebastian Watt.

A final example is from volcan Sollipulli, a little further North in the Chilean Lake District. This example is of a dyke, thought to have intruded along the contact between a glacier and the volcanic edifice, near the present-day crater edge. Here, the portion of the dyke that was originally in contact with the ice has developed a very strong platy fabric, and is falling apart to form a scree of what looks like slate.


Platy jointing developed in the margins of an originally ice-bounded dyke, volcan Sollipulli, Chile. View from above, looking down the dyke margin. Platy scree now partially fills the channel which was once filled with ice. View – about 1 m across. Fieldwork was carried out as a part of a IAVCEI-funded project on the hazards of snow-capped volcanoes.

While there is general agreement that these patterns of jointing form because of the contraction of magma as it cools, there is not yet a concensus model for what it is that controls the size of the polygonal joints, or the typical number of sides. Theoretically, it is argued that a hexagonal jointing pattern would be the lowest-energy, and favoured, solution. But in reality, cooling rates might be too fast for full energy minimisation, and this might explain why many polygonal joints have fewer than six sides. This is the conclusion of the most complete study to date, in which Gyoergy Hetenyi and colleagues argue that field evidence points to two major controls being the size of the cooling body, which influences how fast it cools, and composition, which influences the physical properties of the cooling magma.

Chilean volcanoes: shaken, but not always stirred?

November 7th marked the 175th anniversary of one of the largest earthquakes to have struck northern Patagonia. The earthquake, which is estimated to have had a magnitude of 8, had an epicentre close to Valdivia, and was accompanied by significant ground shaking and subsidence as far south as Chiloe island, and a major tsunami that reached Hawaii.  The eyewitness reports of the time have been well documented. From a geological perspective, the key feature of the 1837 earthquake is that it occurred along a section of the plate boundary that has ruptured repeatedly, with great earthquakes in 1575, 1737 and, most significantly, in 1960  – which, with a magnitude of 9.5, is still the largest recorded earthquake globally. The 1837 earthquake struck just two years after the great Concepcion earthquake, of February 1835, which was exquisitely documented by Charles Darwin, among others.  Because of the location, adjacent to the long southern Chilean volcanic arc, and the frequency of large earthquakes in this region, both the 1835 and 1837 earthquakes have become critical pieces of evidence for the ongoing question of whether, and how, large earthquakes might lead to small triggered volcanic eruptions. Historical records from the region that include maps, expedition reports and navigational charts mean that the record of past eruptions in the southern parts of the central valley of Chile extend back into the 16th century and the earliest Spanish colonists.

Osorno map view from 1747

Map of southern Chile, and the northern part of Chiloe island extracted from ‘A new and accurate map of Chili, Terra Magellanica, Terra del Fuego etc.’ compiled by Emanuel Bowen in 1747. Active volcanoes of Osorno (vul. of Osorno) and, probably, Hornopiren (vul. of Quechucabi) are shown. From the Bodleian libraries, University of Oxford.

The 1837 Valdivia earthquake was followed by reported eruptions, on the same date, both at volcan Osorno, and Villarrica. Both volcanoes had already been in a state of activity on and off in the months or years prior to the earthquake, and both have long historical records of activity, so the observations are not necessarily surprising. One of the challenges of testing for cause and effect when it comes to possible earthquake-triggered eruptions is the likelihood of false reporting that arises from the natural tendency of people to conflate all sorts of observations and speculations in the aftermath of major events, like earthquakes. For example, both Villarrica and Osorno have also been recorded as having erupted shortly after the great earthquake of 1575, but neither observation is necessarily secure.

Volcan Osorno, overlooking Lago Llanquihue, Chile

Volcan Osorno, overlooking Lago Llanquihue, Chile

In contrast to the reports from 1837, the 1960 earthquake did not appear to have any major consequences for either system. Osorno has now been dormant since the late 19th century, while activity at Villarrica has rumbled on into the 21 st century.

Volcan Villarica, map view from 1759

Map of volcan Villarrica, from 1759. Reproduced in ‘Cartografia hispano colonial de Chile’, published in 1952 to mark the centenary of the birth of Don Jose T Medina. From the Bodleian libraries, University of Oxford.

Volcan Villarrica

Volcan Villarrica, with steam plume, viewed from Pucon.

Work is still in progress to investigate the consequences of the most recent great earthquake in the region: the Maule earthquake of February 2010. It remains possible that reported changes in activity at Villarrica in March 2010, seen in thermal infra-red satellite imagery, and subsequent eruptions of Planchon-Peteroa and Puyehue – Cordon Caulle may ultimately be linked to the rejuvenating effects of the earthquake, but this remains to be properly tested.

Of course, this is a question that is mainly of academic interest (in terms of understanding how eruptions are triggered), since most of the eruptions documented to have occurred in the immediate aftermath of great earthquakes are very small, and are most likely to occur at systems which have already been in eruption. The consequences of these eruptions are usually negligible, compared to the effects of the large earthquakes themselves. In recognition of the frequency of these potentially devastating earthquakes, the Chilean authorities (ONEMI, Oficina Nacional de Emergencia) are today holding an earthquake simulation across the schools of Santiago, as part of the programme of national preparedness for future emergencies.


Get every new post on this blog delivered to your Inbox.

Join other followers: