Quantitative Morphology, Recent Evolution, and Future Activity of the Kameni Islands Volcano, Santorini, Greece

Quantitative morphology, recent evolution, and future activity of the Kameni Islands volcano, Santorini, Greece

David M. Pyle*† John R. Elliott† Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UK

ABSTRACT dimension is insensitive to the composition opens up many new avenues for future work. In of the fl ow. particular, the new data reveal details of the sur- Linking quantitative measurements of Dome-growth rates during eruptions of face morphology (on 1–100 m length-scales) of lava fl ow surface morphology with histori- the Kameni Islands in 1866 and 1939 are young dacite lava fl ows, cones and domes, from cal observations of eruptions is an important, consistent with a model of slow infl ation of a which important rheological information can but underexploited, route to understanding dome with a strong crust. Lava domes on the be extracted. In combination with a reanalysis eruptions of silicic magma. We present here Kameni Islands have a crustal yield strength of historical eruption accounts, we show how a new, high-resolution digital elevation model (4 × 107 Pa) that is lower by a factor of 2–4 this information can be used to understand the (DEM) for the intracaldera Kameni Islands, than the domes at Pinatubo and Mount St. emplacement processes of viscous silicic lavas. Santorini, Greece, which reveals the potential Helens. The dome-height model combined of high-resolution imaging (at ~1 m per pixel) with the apparent time-predictable nature Santorini Volcano, Greece of lava-fl ow fi elds by airborne light detec- of volcanic eruptions of the Kameni Islands tion and ranging laser radar (LiDAR). The allow us to suggest that should an eruption Santorini is one of the active volcanoes of the new DEM has an order-of-magnitude better occur during 2006, it will last for more than South Aegean volcanic arc. It has a rich volca- resolution than earlier models, and reveals a 2.7 yr and produce a dome ~115–125 m high. nic history, with more than 12 major explosive wealth of surface morphological information eruptions recognized over the past 250,000 yr on the dacite lava fl ows of the Kameni Islands. Keywords: digital elevation model, lava dome, (Druitt et al., 1989, 1999). Typically, episodes In turn, this provides quantitative constraints lava rheology, Aegean volcanic arc, dacite lava. of caldera formation on Santorini have been on the bulk rheology of the emplaced lava followed by extended periods of lava effusion, fl ows. When combined with a reanalysis of INTRODUCTION leading to the intercalations of andesite to dacite contemporary eruption accounts, these data lava piles and andesite to rhyodacite tephra yield important insights into the behavior A long-standing goal in volcanology is to formations that are exposed within the caldera of dacite magma during slow effusive erup- develop ways of extracting quantitative infor- cliffs at the present day (Druitt et al., 1999). tions on Santorini and elsewhere, and allow mation about past eruptions that will allow us Currently, Santorini volcano is in an effu- the development of forecasts for the style and to develop forecasts of the nature and timing of sive phase, with the focus of intracaldera vol- duration of future eruptions. future activity. In the case of lava fl ows, a key canism for the past 2200 yr being the dacitic Kameni Island lava fl ows exhibit classic area of investigation is understanding the rela- Kameni Islands, which represent the emergent surface morphologies associated with vis- tionship between the surface morphology of top of a 2.5 km3 volcano that has a basal area of cous magma: levées and compression folds. lava fl ows, and the bulk dynamics of the erupt- ~3.5 km2 and that rises 500 m from the fl oor of Levée heights and fl ow widths are consis- ing material (e.g., Hulme, 1974; Fink 1980; Kil- the fl ooded caldera (Druitt et al., 1999). These tent with a Bingham rheology, and lava burn, 2004). Understanding this link is critical islands, of which there are currently two (Palea yield strengths of 3–7 × 104 Pa. Compres- in general because of the importance of lava in and Nea Kameni, Fig. 1), have been the sub- sion folds have long wavelengths (15–25 m), planetary resurfacing, and, specifi cally, because jects of extensive petrological, geochemical, and change only a little downstream; this is it allows reconstruction of the detail of past and textural investigations over the past thirty consistent with observations of other terres- eruptions from the morphology of the emplaced years because of their unusually uniform chemi- trial silicic lava fl ows. The blocky a‘a dacite lavas, and underpins forecasts of how future cal compositions and their contrastingly het- lava-fl ow margins show a scale-invariant lava fl ows will evolve. erogeneous population of exotic xenoliths and morphology with a typical fractal dimen- Here, we present a new high-resolution digi- cognate enclaves (e.g., Nicholls, 1971; Barton sion that is indistinguishable from basaltic tal elevation model (DEM) for the volcanic and Huijsmans, 1986; Higgins 1996; Zellmer Hawaiian a‘a, confi rming that the fractal Kameni Islands (Santorini, Greece) based on a et al., 2000; Holness et al., 2005; Martin et al., light detection and ranging laser radar (LiDAR) 2006). There are considerable ongoing efforts to aerial survey carried out in April 2004. This is characterize and monitor the state of the Kameni *Corresponding author e-mail: dmp11@cam. ac.uk. one of the fi rst applications of an aerial LiDAR Islands, in particular, the seismicity (Dimitria- †Now at Department of Earth Sciences, University survey to a volcano, and the unprecedented dis et al., 2005), hydrothermal and fumarolic of Oxford, Parks Road, Oxford OX1 3PR, UK. spatial resolution that the technique offers activity, and ground deformation (Stiros and

Geosphere; August 2006; v. 2; no. 5; p. 253–268; doi: 10.1130/GES00028.1; 16 ﬁ gures, 9 tables, Data Repository 2006120.

20 25 30

Greece

40 40 N

Turkey Colombos bank

Milos Nisyros Santorini

35 35 A Colombos line

20 25 30 Megalo Vouno Cape Colombos Oia Therasia

Cape Skaros Thera

Figure 1. General location (A) and Nea Kameni Fira geological map (B) of Santorini, Kameni line and the intracaldera Kameni Islands, after Druitt et al. (1999) and Martin et al. (2006). Part A shows the shallow submarine Colombos bank (NE of Santorini), Palea Kameni which was the location of an eruption in 1650 A.D., and the “Kameni line,” which marks the NE-SW trend of a region of preferred vent locations on the Kameni Islands Profitis Akrotiri peninsula (Druitt et al., 1989, 1999). Ilias Akrotiri

2km

Kameni volcano

cycle Minoan tuff First explosive cycle

sive Therasia dome complex Cinder cones of Akrotiri peninsula Skaros shield Peristeria volcano Cinder cones and tuff ring Early centers of Akrotiri peninsula B Pyroclastic deposits Basement

Second explo

254 Geosphere, August 2006 Morphology of Kameni Island Lavas, Santorini, Greece

Chasapis, 2003; Vougioukalakis and Fytikas, bathymetric map of the caldera. The 1866–1870 than 500–1000 m over durations of 30–200 d 2005). Although the islands are currently in a eruptions were closely documented by Fouqué, (e.g., Kténas, 1926; Georgalas and Papastama- state of intereruptive repose, there is no reason based in part on his own observations, as well tiou, 1951, 1953; Georgalas, 1953), with fl ow- to suppose that there will not be future eruptions as on the many books and pamphlets that were front advance rates ranging from 10−3 ms−1 in the of a similar nature, perhaps within decades, and published within a year or two of the start of early stages to <10−4 ms−1 after 3 mo, and aver- certainly within centuries. the 1866 activity (e.g., Virlet d’Aoust, 1866; aged effusion rates on the order of 0.5–2 m3s−1 Since the extensive work of Fouqué (1879), von Seebach, 1867; Reiss and Stübel, 1868). (Fig. 2; Tables 3 and 4). The erupting lavas are, Kténas (1926, 1927), Georgalas and Liatsi- The 1866 eruptions clearly drew considerable therefore, classical examples of creeping vis- kas (1925a, 1925b, 1936a, 1936b), and Reck scientifi c interest at the time; for example, the cous fl ows, with low Reynolds numbers (imply- (1936a), which documented the course of many eruption was responsible for the earliest medical ing laminar fl ow) and moderate Peclet num- of the historic eruptions in great detail (Table 1), work on the health effects of volcanic eruptions bers, similar to those of slow-growing domes little attention has been paid to the physical form, (da Cologna, 1867). Together, the many pub- (Griffi ths, 2000). or posteruptive morphology and evolution, of the lished works that describe the major eruptions lavas and ash cones that make up the Kameni of the eighteenth, nineteenth, and twentieth cen- DATA COLLECTION Island group. Indeed, many of these original turies present an excellent basis (Table 2) from works have been overlooked, and the wealth of which to develop a quantitative analysis of the Airborne data were collected over the relevant data they contain has been forgotten. evolution of an intracaldera volcano. Kameni Islands, Santorini, in April 2004 dur- The salient features of the Kameni Islands ing an overfl ight by the UK’s Natural Environ- Historical Eruptions of the Kameni Islands eruptions and their products are summarized ment Research Council and airborne remote- in Tables 1–4. The Kameni Islands make an sensing facility (ARSF) Dornier 228 aircraft. The historical activity of the Kameni Islands is excellent case study since the eruptions have This aircraft was equipped with a WILD RC-10 well known from contemporary written records exclusively involved dacite lava (Table 2), in camera, and Cambridge University’s Airborne (Table 1). Modern compilations (e.g., Georga- contrast to the many active lava-forming sys- Laser Terrain Mapper (ALTM) system (Model las, 1962; Fytikas et al., 1990) generally begin tems that have been studied in recent decades, 3033, Optech Inc., Canada). The aircraft survey with the work of Fouqué (1879), who produced which are predominantly basaltic (e.g., Etna, took place at an altitude of 650–780 m above one of the classic modern scientifi c accounts Hawaii; Walker, 1973; Pinkerton and Sparks, sea level (asl) and at a ground speed of 60–75 of an active volcano and its evolution. Fouqué, 1976). Eruptions typically include both dome- ms−1. The ALTM LiDAR was operated at 33.3 in turn, drew on material collated by Pègues forming and mildly explosive (Vulcanian) kHz and used a side-to-side scanning mecha- (1842), who published an extended account phases, and the steady effusion of lava. During nism, which combined with the forward motion of the historical activity of the islands, and by the most closely observed eruptions of the twen- of the aircraft to provide swath coverage of the Leycester (1851), who also published a detailed tieth century, fl ows typically extended to more ground surface below. Data from an onboard

TABLE 1. HISTORICAL ACTIVITY OF THE KAMENI ISLANDS TABLE 2. TYPICAL FEATURES OF THE KAMENI DACITE LAVAS Eruption date Location References and notes Parameter Value 10 January–2 February 1950 Nea Kameni Georgalas (1953).

20 August 1939–early July 1941 Nea Kameni Georgalas and Papastamatiou (1951, 1953). SiO2 content 64–68 wt% 23 January–17 March 1928 Nea Kameni Kténas and Kokkoros (1929); Georgalas and Inferred eruption temperature 900–950 oC Liatsikas (1936b); Reck (1936c). Phenocryst content* 15% 11 August 1925–January 1926 Nea Kameni Kténas (1925a, 1925b, 1925c, 1926, 1927); Inferred viscosity†, anhydrous (4–6) × 104 Pas Washington (1926a, 1926b); Georgalas and magma, η Liatsikas (1936a); Reck (1936b). Effective viscosity† of porphyritic (6–9) × 104 Pas 26 January 1866–15 October 1870 Nea Kameni Described by Fouqué (1879); accounts of 1866 magma activity in von Fritsch et al. (1867) and Reiss and Bulk lava density† (vesicle 2680–2725 kgm–3 Stübel (1868). free), ρ 23 May 1707–14 September 1711 Nea Kameni Described by Goree (1710) and Tarillon (1715a). Eruption end date is from Tarillon (1715b); Flow characteristics misquoted as 11 September in Georgalas (1962) Typical fl ow rates§, V 10–5–10–3 ms–1 and subsequent catalogues. Effective Reynolds number#, Re 10–5–10–3 1570 or 1573 Mikra Kameni The eruption was in either 1570 or 1573 (Fouqué, Peclet number**, Pe 2 × 102–2 × 104 1879). Prandtl number††, Pr 2 × 107 1457 Palea Kameni No eruption, but a collapse of part of Palea Kameni (Fouqué, 1879). *Phenocrysts (1–2 mm) of plagioclase and 726 Palea Kameni Northeast side of Thia Island clinopyroxene (Huijsmans, 1985; Higgins, 1996). † 46–47 Palea Kameni (Thia) Eruption of Thia, about 400 m from Hiera, to form an Viscosity and density calculated using Ken island of ~5600 m circumference (Fouqué, 1879; Wohletz’s ‘Magma’ (http://www.ees1.lanl.gov/Wohletz/ Stothers and Rampino, 1983). Now Palea Kameni. Magma.htm), based on the Bottinga and Weill 19 A.D. Thia Reported by Pliny, but not regarded as an eruption (1972) model. § by later authors (e.g., Pègues, 1842; Fouqué, From fi eld observations by Kténas (1926) and 1879). Georgalas and Papastamatiou (1951, 1953). # ρ η 199–197 B.C. Hiera (or Lera) Island that formed in 4 d, with a circumference of Re = Vh/ , assuming a typical fl ow depth, h, of 2200 m (Stothers and Rampino, 1983). Presumed 20 m. κ κ –6 2 –1 to have been eroded below sea level. Fouqué **Pe = Vh/ ; , thermal diffusivity = ~10 m s . (1879) suggested Hiera was the Bancos reef, ††Pr = η/(ρκ) which was northeast of Nea Kameni and was covered with lava during the 1925–1926 eruption.

Geosphere, August 2006 255 Pyle and Elliott

TABLE 3. DEVELOPMENT OF EIGHTEENTH- AND NINETEENTH-CENTURY FLOW FIELDS ~8 km2, including a ground area of 3.86 km2. Eruption Flow length Lobe area Typical fl ow Approximate Inferred Measured points have a nominal accuracy of (m) (km2) thickness fl ow duration effusion rate 1.0 cm, 1.2 cm, and 5.9 cm in their x, y, and z (m) (d) (m3s–1) directions, respectively. The data were imported 1707–1711* 1200 0.9 > 30–60 300–600 1–2 into a geographic information system (GIS) 1866†, Aphroessa fl ow (NNW lobe) 1000 0.2 20–50 90 0.5–1.3 application (ARCMAP), and anomalous data † 1866–1869 , Giorgios fl ows 1200 1.7 30–70 600–1000 1–1.4 points (due to cloud cover) were removed. A *Accounts suggest that most lava extrusion occurred in 1707–1708 (Tarillon, 1715a, 1715b; Goree, 1710; large area of absorption due to low cloud cover Fouqué, 1879). †See Fouqué (1879) and Georgalas (1962). appeared on one fl ight line (Fig. 3), so for this area there is no precise height information. Instead, this area was patched with a 15-m-resolution digital elevation model (DEM; from Inte- TABLE 4. DEVELOPMENT OF TWENTIETH-CENTURY FLOW FIELDS gralGIS). The LiDAR data were used to con- Eruption Maximum Area Typical Duration Averaged Data source struct a high-resolution DEM by interpolating length (m) (km2) thickness (d) effusion rate (m) (m3s–1) the LiDAR points to an ~1 m post spacing using a cubic spline function and fi tting a surface to August 1925, N branch 728 0.24 60 80 2.2 Kténas (1926) these data. August 1925, E branch 1200 0.61 60 150 2.8 Kténas (1926) A hill-shade rendition of the DEM is pre- September 1939, NW and 407–620 0.13 20–30 52 0.15–0.25 Georgalas and SW lobes Papastamatiou sented in Figure 4, and a contoured DEM is pre- (1951) sented in Figure 5. In addition, maps of slope November 1939–April 1940, 320–620 0.38 20–30 95–230 0.4–0.7 Georgalas and (not shown) were generated to help visualize N, E, SW lobes Papastamatiou (1953) surface morphological features. This digital elevation model may be referenced either to the UTM (Universal Transverse Mercator) grid, or to standard latitude and longitude coordinates global positioning system (GPS) and the local detection-ranging (LiDAR) measurements were (as shown in Figs. 3 and 4). It should prove a GPS base station (at Akrotiri) were used to made. In addition, Airborne Thematic Mapper useful starting point for any future analysis of derive the precise fl ight path. First pulse and (ATM) data were collected across 11 spectral either pre-eruptive deformation of the Kameni last pulse data were recorded for each airborne bands. Islands or of posteruptive morphological evolu- laser measurement and converted to location (x, tion. For the present time, and the purposes of y, z) and intensity data by Gabriel Amable of LiDAR this paper, the DEM allows us to proceed with a the Cambridge University Unit for Landscape detailed quantitative investigation of the surface Modeling. During the overfl ight, 21 aerial pho- The processed LiDAR data comprised 4.52 morphology of a number of key aspects of the tographs and 4.52 million light-imaging and million measurements over a surveyed area of Kameni Island lava fl ows and domes.

1200 10-3 ms-1 10-4 ms-1 1925 N branch 1925 E branch 1000 August 1939 SW August 1939 NW November 1939 E branch 800 November 1939 SW branch

600

Figure 2. Flow lengths as a function of Flow length (m) 400 time for the eruptions in 1925–1926 and 1939–1941, based on ﬁ eld maps and observations in Kténas (1926) and Georgalas and 200 Papastamatiou (1951, 1953). Time-averaged ﬂ ow rates typically range between 10−5 and 10−3 ms−1 (contoured). 0 1 10 100 1000 Time elapsed (d)

256 Geosphere, August 2006 Morphology of Kameni Island Lavas, Santorini, Greece

Aerial Photographs Geological Map of the Kameni Islands ure 8C is the same view from the 2004 overfl ight. The new lava fl ows that erupted during During the overfl ight, a set of 21 color ana- There have been several modern interpre- 1939–1941 activity can be clearly seen in the logue aerial photographs were captured using tations of the surface geology and eruption two photographs, as can the minor changes in the onboard Wild RC-10 camera. Together, sequence of the Kameni Islands, based primar- crater morphology associated with the 1950 these images give a mosaic that covers the ily on the observations and interpretations of eruption, and the emplacement of the very entirety of the Kameni Islands. Images were Georgalas (1962). More recently, maps were small Liatskias dome and lavas, as described scanned digitally at high resolution, internally published by Pichler and Kussmaul (1980), by Georgalas (1953). referenced using fi ducial points, and linked to Huijsmans (1985), and Druitt et al. (1999). We A second feature of the map is that, in contrast the x, y, z coordinates of identifi able ground present an interpreted map in Figure 7, which to the earlier mapping of Pichler and Kussmaul control points on the digital elevation model. was prepared using a combination of feature (1980), we see no evidence for exposed surfi cial In turn, this allowed the photographs to be mapping (e.g., using the slopes visualized in faults on Nea Kameni. In our map, we do not corrected for distortion and terrain effects. the DEM) and visual imagery. The new map explicitly identify the areas covered or partially An orthorectifi ed DEM-referenced aerial pho- is properly georeferenced to the UTM and lati- covered by younger ash deposits (e.g., the por- tograph mosaic is shown in Figure 6. This tude-longitude grids, and is offset by 280 m (S) tions of the 1866–1870 lava fl ows covered with image is complete, save for the easternmost and 225 m (W) from the UTM coordinates of ash from the 1925–1928, 1939–1941, and 1950 part of one lobe of the 1925 lava fl ow, which the Kameni Island map of Druitt et al. (1999). activity), but these areas can be identifi ed from was obscured by a cloud. This mosaic image The main difference in the new map, com- the surface texture and coloration on the aerial demonstrates clearly the ease with which the pared to that of Druitt et al. (1999), is the inter- photograph mosaic. surface textures of the lava fl ows and ash cones pretation of the lava-fl ow sequence from the can be identifi ed, and with which fl ows of dif- 1939–1941 eruptions in the western portion Lava-Flow Surface Textures: Folding ferent age can be distinguished. We used the of Nea Kameni. This is shown in the detail in aerial photographs, combined with the DEM, Figure 8A, along with extracts from two aerial The high-resolution DEM (Figs. 4 and 6) in the following sections to develop an updated photographs of the same region: Figure 8B is reveals a wealth of lava-fl ow surface textural interpretation of the geological map of the a monochrome photograph taken in May 1944 information, most strikingly, the forms of lava- Kameni Islands. by a British reconnaissance aircraft, and Fig- fl ow levées and fold patterns. Both of these

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E Relative N laser return

5"N intensity

24'4 '45"N High : 1276

36°

36°24

Low : -4

24'30"N

'30"N

36°

36°24

Figure 3. Light detection and

'15"N 24'15"N ranging laser radar (LiDAR)

36°

36°24 return intensity map. Return intensity is qualitatively related to surface smoothness; there

4'0"N are strong returns from the

36°24'0"N old surfaces of Palea Kameni,

36°2 for example. The central strip with no return intensity is the ﬂ ight line that was affected 23'45"N by low cloud cover. The raw

36°

36°23'45"N LiDAR data (x, y, z, intensity) are available in the GSA Data Repository item accompanying 25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E this paper (see text footnote 1).

Meters 0250 500 1,000 1,500 2,000

Geosphere, August 2006 257 Pyle and Elliott

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E N

36°24'45"N

36°24'45"N Figure 4. Hill-shaded digital map of the Kameni Islands,

24'30"N based on light detection and

36° °24'30"N ranging laser radar (LiDAR) 36 data. A coarse-resolution patch (data provided by P. Moore, IntegralGIS) was used to aug- ment the missing strip. This

°24'15"N

36°24'15"N

36 rendition clearly picks out surface morphological features of the lava ﬂ ows and domes,

24'0"N including prominent levées and

36°

6°24'0"N lava channels. Note the contrast

3 between the rough surfaces of the blocky a‘a dacite lava ﬂ ows and the smoother ash cones of north-central Nea Kameni.

36°23'45"N

°23'45"N

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E

m 0250 500 1000 1500 2000 Coordinates are latitude longitude

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E Elevation (m) N above sea level Sea level 0-10

4'45"N 36°24'45"N 10 - 20

36°2 20 - 30 30 - 40

4'30"N

36°24'30"N 40 - 50

36°2 50 - 60 60 - 70

4'15"N 70 - 80

36°24'15"N

36°2 80 - 90 90 - 100 100 - 110

24'0"N

36°24'0"N

36° 110 - 120

3'45"N

36°23'45"N 36°2 Figure 5. Elevation map of the Kameni Islands, shown as elevation above mean sea level. 25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E

m 0250 500 1000 1500 2000

258 Geosphere, August 2006 Morphology of Kameni Island Lavas, Santorini, Greece

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E N

5"N Figure 6. Orthorectiﬁ ed aer-

24'4 '45"N ial photograph mosaic of the

36° Kameni Islands, April 2004. 36°24 Particularly prominent features that can be seen include the twentieth-century lava

24'30"N '30"N ﬂ ows (darker, rough surfaces

36°

36°24 in the east and west of Nea Kameni) and the discoloration, due to persistent low-grade

'15"N 24'15"N fumarolic activity of the ash-

36° covered regions of central Nea

36°24 Kameni. The modern tourist trail that crosses the northern part of Nea Kameni can also be

4'0"N

36°24'0"N very clearly seen. The two small 36°2 islands that lie between Nea Kameni and Palea Kameni are the May islands, which erupted

23'45"N in May 1866, and are now just 36° submerged. A high-resolution

36°23'45"N version of this ﬁ gure is available from the GSA Data Repos-

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E itory (see text footnote 1).

m 0250 500 1000 1500 2000

sets of features, in principle, have scales that show clearly that the ridges formed very close are thought to refl ect the rheological properties to the vent, essentially as the lava fl ow was Yhg=ρsin( α ) . (1) of the underlying fl ow, and there are widely extruded. In the case of the 1925–1926 erup- used models for both approaches (e.g., levées: tions, this much is also clear from the contem- For a levée of width w, this is equivalent to Hulme, 1974; folds: Fink and Fletcher, 1978). porary eruption reports (Kténas, 1926; Reck, The resolution of the DEM (with a point spac- 1936b). We quantify and interpret each of these Ygw= 2ραsin2 ( ) , (2) ing on the order of 1 m) means that the meter- morphological features in subsequent sections. scale blocks that are often the most obvious while for a whole fl ow width of W, surface feature of these and other silicic lavas Shapes of Lava Lobes: Levées and Folds ρgh2 in the fi eld cannot be resolved. Y = . (3) One of the prominent features of the lava- The shapes of lava fl ows, and the forms W fl ow morphology that is visible both from aerial of lava-fl ow fi elds, are complex functions photographs and the DEM is the abundance of the rheological properties of the fl uid and Dimensions of fl ow levées and surface folds of gently arcuate ridges on the fl ow surfaces, the emplacement conditions. It remains a were determined by taking transverse and lon- which are convex downstream (away from the continuing goal of physical volcanologists gitudinal (fl ow-parallel) sections across fl ows vent) and lie approximately perpendicular to to understand and quantify these functions (Table DR11). A typical cross-sectional profi le the fl ow direction. These ridges are best devel- (e.g., Griffi ths, 2000; Blake and Bruno, 2000). of a fl ow is shown in Figure 9. This exhibits oped on fl ows (such as the 1707–1711 fl ow) One widely used approximation of the behav- the classical form associated with a fl ow levée that also have well-defi ned levées. Ridges such ior of viscous lavas is that they behave as (e.g., Sparks et al., 1976). Levées are typically as these have been described from a number Bingham fl uids, with a yield strength that must of terrestrial and planetary volcanic settings be exceeded before fl ow can occur and a plastic 1GSA Data Repository item 2006120, text fi le with (e.g., Fink 1980; Gregg et al., 1998). Analogue viscosity. Assuming a Bingham rheology, one the raw LiDAR data (UTM coordinates, zone 35N), experiments, for example, using cooling wax may use the widths of fl ows and the dimen- a high-resolution aerial photomosaic of the Kameni fl ows, strongly suggest that ridge development sions of their levées to gauge the apparent yield Islands, a table summarizing eruption volumes, and happens early in the fl ow history as the fl ow strength of the fl uid (Hulme, 1974; Hulme and fi gures showing the locations of sections used for fl ow ρ shape analysis, is available online at www.geosociety. surface develops a crust (whether by cooling or Fielder, 1977). For a fl ow of bulk density and org/pubs/ft2006.htm, or on request from editing@ by crystallization; Griffi ths et al., 2003). In the depth h fl owing down a slope of angle α, the geosociety.org or Documents Secretary, GSA, P.O. case of the Kameni fl ows, fi eld observations yield strength Y is Box 9140, Boulder, CO 80301-9140, USA.

Geosphere, August 2006 259 Pyle and Elliott

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E Lava dome N with crater Nea Kameni lavas 5"N Folds Mikra Kameni

24'4 lavas Dafni lavas '45"N 36° Fissures

36°24

Levées

24'30"N

'30"N

36°

36°24 Kténas lavas Lava ages 1950

24'15"N

36° 1940-1941 Agios Nikolaos 36°24'15"N Liatsikas Niki lavas lavas Fouqué, Reck dome 1939-1940 & Smith lavas 1939

4'0"N

36°24'0"N 1925-1928

36°2 1866-1870 Thia Lavas 1707-1711

23'45"N 1570-1573

36°

Nea Kameni 36°23'45"N 726 46-47 Palea Kameni

25°22'30"E 25°22'45"E 25°23'0"E 25°23'15"E 25°23'30"E 25°23'45"E 25°24'0"E 25°24'15"E 25°24'30"E

m 0250 500 1000 1500 2000 Figure 7. Interpreted geological map of the Kameni Islands, based on the new imagery and on previous maps compiled by Georgalas (1962), Huijsmans (1985), and Druitt et al. (1999). The names of selected lava ﬁ elds follow the conventions started by Fouqué (1879) and subsequent authors (Georgalas, 1962). The Liatsikas dome is the youngest extrusive element on the Kameni Islands.

12–30 m high, 30–60 m wide (Table 5), with levée width are substantially smaller. The poor (e.g., lava crust) during fl ow of the underly- outer slopes of 25–35°. While measurement agreement between the yield-strength estimates ing fl uid. Detailed descriptions and analyses errors associated with the DEM are small, more based on levée width and those based on gross- of examples of folding have been presented signifi cant errors result from the subjectivity fl ow morphology is consistent with observations by Fink and Fletcher (1978), Fink (1980), and of identifying levée margins, from the addi- elsewhere (e.g., Etna; Sparks et al., 1976) and Gregg et al. (1998), among others; the motiva- tional complications where levée dimensions most likely refl ects the fact that the processes tion for this work, again, was the possibility change downstream, or where fl ows bifurcate, important in levée formation and evolution of inferring fl ow rheology and compositional and from uncertainty in the slope down which (including accretion, avalanching, and cooling, parameters for extraterrestrial examples (e.g., the fl ow was emplaced. The apparent viscosity for example; Sparks et al., 1976) violate the Warner and Gregg, 2003). and yield-strength estimates based on lava-fl ow assumptions required for Hulme’s (1974) analy- Nine lava fl ows were selected for surface morphology should only be regarded as order- sis to work. folding analysis, based on the visible ridges of-magnitude estimates of fl ow parameters. In addition to the levées, another prominent seen in the DEM and aerial photographs. For Estimates of yield strengths derived using surface morphological feature of the Kameni each fl ow, fi ve parallel profi les, running per- the three complementary approaches are sum- lava fl ows that can be seen clearly in the DEM pendicular to the axis of the folds, spaced 2 m marized in Table 6. The yield-strength esti- and aerial photo images is folding. A typical apart, were taken along the middle sections of mates derived from the gross-fl ow morphology longitudinal section along a fl ow, corrected for each fl ow. Fold scales were investigated by fast appear to be the most internally consistent and the general slope of the fl ow fi eld, is shown in Fourier transform (FFT) analysis in MatLab suggest yield strengths on the order of 30 kPa Figure 10. This shows folding with wavelengths using a custom series of scripts to investigate (1925 fl ows) to ~70 kPa (1940 fl ows). These on a 10–100 m scale, defi ned by a train of arcu- fold wavelength characteristics. Data were values are consistent with yield-strength esti- ate ridges that runs perpendicular to fl ow direc- prepared for FFT analysis by interpolating the mates for dacites elsewhere (Hulme and Fielder, tion. Folds of these sorts have been recognized data to unit spacing, detrending the data stream, 1977; Wadge and Lopes, 1991). Yield-strength on a number of length scales on lava fl ows of and applying a cosine taper (Kanasevich, 1975; estimates based on levée height are similar, in all compositions, from basalt to rhyolite, and Bracewell, 2000) to remove the baseline shift terms of order of magnitude, to those based on the usual interpretation is that they form as a and sinc convolutions, which would otherwise whole-fl ow morphology, while those based on result of buckling of a more rigid surface layer have reduced the data quality. Typical examples

260 Geosphere, August 2006 Morphology of Kameni Island Lavas, Santorini, Greece

355500 356000 356500 Lava ages of a contoured periodogram are shown in Fig- N 1950 ure 11, which shows the relative power (arbi- 1940-1941 trary scale) of fold wavelengths along the length

4030500 4030500 1939-1940 1939 of a 1925 lava fl ow. This gives a rapid visual 1925-1928 assessment of the extent to which fold dimen- 1866-1870 sions change, or otherwise, along a fl ow. In this 1707-1711 case, the fold wavelength is ~20 m close to the 1570-1573 vent and ~30–40 m further downstream. 726 Most of the fl ow profi les share a number of Lava domes 46-47 & craters common features: folds form close to the origi- 0000 nal eruptive vent. Folding is most prominent

4030000 403 close to the center-line of the fl ow, and in many cases there is a consistent pattern of fl ow generations that develops downstream, i.e., several fold generations of differing wavelength can be recognized. Results are summarized in Table 7. A It is clear from Table 7 that most of the fl ows

355500 356000 356500 that show folding have one generation of folds m λ Coordinates in WGS84 with wavelengths ( 1) on the order of 15–25 m. UTM Zone 35N 0125 250 500 750 1000 When multiple generations of folds exist, sec- λ ond-generation folds ( 2) have wavelengths λ on the order of 25–65 m (or 1.5 ± 0.2 1), and third-generation folds have wavelengths of λ 30–110 m, or 1.8 ± 0.3 2. This pattern of low ratios (<2) between subsequent generations of folds is consistent with the few data that exist for other dacite (r = 2.1 ± 0.3) and rhyolite (1.8 ± 0.4) fl ows (Gregg et al., 1998), and contrasts strongly with the high ratios (5.1 ± 1.1) that characterize basaltic fl ows (Gregg et al., 1998). This confi rms the potential of fl ow fold analysis for the identifi cation of “silicic” fl ows by remote sensing. The scale of the folds qualitatively constrains the minimum apparent viscosity of the underlying fl uid (Fink, 1980; Gregg et al., 1998). Data B from the Kameni Island fl ows are consistent with the view that in slow, effusive eruptions of silicic magmas, crust formation on the fl ow by cooling is much faster than crustal thickening by strain, and consequently the folds that form are of long wavelength. The slow effusion of the fl ows is consistent with the observation that the folds develop close to the vent.

Planform of Kameni Dacite Lava Flows

One motivation for the study of the quantitative morphology of terrestrial lava fl ows of known composition is that it may reveal features that allow investigators to make inferences about the compositions of fl ows on distant plan- ets, simply from the remote observations of C morphology. One such approach that appears to have some promise was proposed by Bruno et Figure 8. (A) Expanded map section showing the revised interpretation of the 1939–1941 al. (1992, 1994), which is to determine the frac- lavas, with the inferred fl ow paths shown with arrows (+ symbols show the UTM grid inter- tal dimension, D, of the fl ow shape in planform. sections). The extracts from a 1944 Royal Air Force reconnaissance air photograph (TARA The technique relies on measuring the gross ACIU 42090 3080) (B) and a 2004 air photo (C) show the subtle changes in the summit cra- linear distance around an object as a function ters due to the minor explosive activity in 1950, and the emplacement of the Liatsikas dome. of the length scale of the measuring rod. On a

Geosphere, August 2006 261 Pyle and Elliott

X: 96.65 X: 190 70 Y: 68.23 Y: 68.67 X: 292.2 Y: 67.19 Levée width = 79 m 65

Vertical exaggeration x6 Flow height = 35 m 60 Flow height = 36 m

Flow width = 331 m 45

40 Levée angle = 25°

Height above geoid (sea level = 32 m above geoid)

35 X: 17.77 X: 348.8 Y: 31.9 Y: 32.24

30 0 50 100 150 200 250 300 350 400 Distance across levée (m) Figure 9. Cross section showing a classical levée structure across the 1866 SW lava ﬂ ow. Levées tend to be 15–30 m high, while ﬂ ows are 100–300 m wide (Table 5).

log-log plot of distance versus length scale, data TABLE 5. LEVÉE AND FLOW DIMENSIONS following a fractal distribution should defi ne a Flow* N† Levée height Levée width Flow width Surface slope§ trend with a linear slope of 1 – D. Bruno et al. (m) (m) (m) (o) (1992) showed that the parameter varies with 1707 16 23.1 ± 2.3 53.7 ± 6.6 224 ± 22 3.1 ± 0.7 lava type (a‘a versus pahoehoe) in basalts, but 1866 SSW 2 20.3 ± 2.0 32.9 ± 4.0 196 ± 17 4.3 ± 0.6 also showed (Bruno et al., 1994) that many 1866 SW 5 25.2 ± 3.4 61.6 ± 5.5 310 ± 16 4.2 ± 0.5 more chemically evolved lavas do not show a 1925 NE 4 12.2 ± 1.5 29.5 ± 3.4 133 ± 4 1.5 ± 0.4 1925 SE 3 18.8 ± 1.6 51.5 ± 1.6 201 ± 6 2.0 ± 0.2 consistent fractal behavior. 1925 W 5 21.2 ± 3.5 39.4 ± 9.3 119 ± 12 0 ± 0.7 The Kameni lavas present an excellent oppor- 1940B 5 18.0 ± 1.4 27.7 ± 1.8 121 ± 18 8.3 ± 2.0 tunity to test whether this model is applicable to 1940 SW 12 20.9 ± 2.1 44.9 ± 3.7 193 ± 13 2.6 ± 0.9 blocky dacite lavas. 1940 W 4 16.4 ± 2.1 25.4 ± 3.3 103 ± 1.2 6.7 ± 1.0 Lava fl ows with well-defi ned margins were *Sections across which measurements were taken are shown in Figure DR2 (see text footnote 1). digitized from the orthorectifi ed aerial photo- †N—number of sections measured. Quoted error is the standard deviation of the measurements. §Slope is slope at present day. This may not have been the slope down which the fl ow was emplaced. graphic images to obtain an outline with a spatial resolution of better than 1 m. Spatial data were exported into a program (Fractals; http:// www.nsac.ns.ca/envsci/staff/vnams/Fractal. TABLE 6. YIELD STRENGTHS, ASSUMING BINGHAM RHEOLOGY htm; Nams, 1996) to determine the gross length Yield strength as a function of length scale, for length scales of (×103 Pa) 1–100 m. Most of the analyzed lava fl ows had Flow Flow width method Levée width method Levée height method margins that could be traced for ~1 km, which 1707 64 ± 14 8 ± 4 64 ± 6 limited the upper range of length scale that could 1866 SSW 57 ± 12 10 ± 3 75 ± 7 be used. A typical plot of gross distance against 1866 SW 55 ± 15 18 ± 5 93 ± 13 1925 NE 30 ± 8 1.1 ± 0.6 23 ± 3 ruler length, confi rming a fractal distribution, 1925 SE 48 ± 8 3.4 ± 0.7 35 ± 3 is presented in Figure 12. All of the lava fl ows 1940B 72 ± 16 32 ± 15 134 ± 10 that were analyzed showed clear evidence for 1940 SW 61 ± 13 5 ± 3 58 ± 6 a fractal distribution, and, surprisingly, there 1940 W 71 ± 18 19 ± 6 107 ± 14 was no detectable difference between the fractal

262 Geosphere, August 2006 Morphology of Kameni Island Lavas, Santorini, Greece

5 dimensions of younger fl ows (<200 yr old) that form the coastline and those with margins that 4 are entirely exposed inland. The only example that was clearly nonfractal was the profi le for the 3 A.D. 726 fl ow on Palea Kameni, which presum- ably has been deeply eroded. The fractal dimen- 2 sions for all of the lava fl ows that were analyzed are plotted in a histogram in Figure 13. It is clear

end (m) 1 from this graph that the blocky a‘a dacite lavas of the Kameni Islands have a fractal dimension, D (1.067 ± 0.006, n = 10), that is indistinguish- 0 able from that of Hawaiian a‘a, at least on the 1–100 m scale. Thus, fractal dimension alone -1 may be a good discriminator for gross-ﬂ ow type (a‘a versus pahoehoe), but is not necessarily a -2 good discriminator for composition.

Deviation in height from linear tr -3 Dome Growth on the Kameni Islands

-4 One typical feature of all of the historical eruptions of the Kameni Islands is the growth of lava domes. Field observations of the chang- -5 100 200 300 400 500 600 700 800 900 1000 ing height of the summit of the 1866–1870 Figure 10. Surface buckling shown in a topographic section along the 1925 SE lava ﬂ ow, (Giorgios) dome and the 1939 (Fouqué) dome with the underlying slope removed. Note the prominent folding, which has 10–100-m-scale were recorded, respectively, by Fouqué (1879), wavelengths and 2–3 m amplitude. Data from ﬂ ows are summarized in Table 7. and Georgalas and Papastamatiou (1953).

Figure 11. Contour map of the spectral power (arbitrary scale) of the different fold wavelengths in a set of fi ve parallel profi les, spaced 2 m apart, along the length of the 1866 SW lava fl ow. Window 1 is at the fl ow toe, and 11 is at the fl ow origin. Buckling with an ~20 m wavelength can be observed near the vent; downstream, a 30–40 m wavelength becomes dominant.

Geosphere, August 2006 263 Pyle and Elliott

TABLE 7. WAVELENGTHS OF FOLDS Flow* N† Fold generation 1 Fold generation 2 Fold generation 3 (m) (m) (m) 1707 21 19.9 ± 0.9 24.8 ± 0.9 35.5 ± 2.4 1866 SSW 11 23.1 ± 0.4 32.7 ± 2.1 72 ± 5 1866 SW 11 23.5 ± 1.5 35.4 ± 1.0 1925 NE 21 15.4 ± 1.1 21.0 ± 1.3 28.4 ± 2.2 1925 SE 21 26.0 ± 1.5 48.9 ± 2.0 93 ± 19 1925 W 11 41.9 ± 1.7 63 ± 7 1940B 11 21.5 ± 1.7 28.5 ± 1.1 53 ± 6 1940 SW 11 35.5 ± 2.2 48.8 ± 4.1 1940 W 11 38.5 ± 3.1 64 ± 6 109 ± 12 *Sections along which measurements were made are shown in Figure DR1 (see text footnote 1). †Number of measurements

2.86

2.84

2.82 . Figure 12. Graph showing the relation- 2.80 ship between the length of the measuring rule (x axis) against the gross length of the planform for the 1925 N fl ow. The data can 2.78 be described by a power law with a slope of −0.054 (±0.001), which confi rms that (gross distance, m) 2.76 10 the shape of the margin of this lava fl ow is self-similar on a 1–100 m length scale (see Log 2.74 Mandelbrot, 1967). The data have a fractal Caculated gross distance dimension of 1.054. Least squares regression line 2.72 95% prediction interval for a data point

2.70 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.6 1.8 2.0

Log10 (length scale, m)

Kameni blocky a'a 12 Galapagos a'a Hawaii a'a 10 Hawaii

8 Figure 13. Comparison of fractal dimensions determined for Nea Kameni blocky a‘a fl ows with data from basaltic a‘a and 6 pahoehoe fl ows from Hawaii and Galapagos (Bruno et al., 1992, Number 1994). Fractal dimension appears to be an excellent discrimi- 4 nant of gross fl ow type, independent of composition.

0 1 1.04 1.08 1.12 1.16 1.2 1.24 D

264 Geosphere, August 2006 Morphology of Kameni Island Lavas, Santorini, Greece

Surprisingly, the importance (indeed, the TABLE 8. GROWTH OF DOMES DURING THE 1866 AND 1939 ERUPTIONS existence) of these data seems to have been 1866 Giorgos Dome* 1939 Fouqué Dome† overlooked, and they are summarized for ref- Growth started 2 February 1866 Growth started 13 November 1939 erence in Table 8. The dome-growth data are plotted in Figure 14. Both dome-forming erup- Elapsed time Dome height Elapsed time Dome height (d) (m) (d) (m) tions show closely similar growth rates in terms of dome height (H) with time (t), which paral- 3 20 3 23.5 lel those of the well-known lava domes of the 5 30 7 30.5 35 50 9 40.5 Soufrière volcano of St. Vincent (1979 eruption; 47 50 28 45.5 Huppert et al., 1982), and Mount St. Helens 108 70 65 56.5 (1980–1986; Swanson and Holcomb, 1990). 396 108 150 78.5 The power-law dependence of dome height with 1519 118 time is close to the t1/4 dependence predicted by *1866 dome-growth parameters from Fouqué (1879). models of domes that have growth controlled by †1939 dome-growth parameters from Georgalas and Papastamatiou (1953). a crustal yield strength (e.g., Fink and Griffi ths, 1998; Griffi ths 2000), and the best-fi t curve with a t1/4 dependence is H = 1.2t1/4. Follow- ing Griffi ths (2000), this suggests that the yield 1000 strength of the crust of the Kameni dacite domes Kameni 1866 dome is ~4 × 107 Pa, which is somewhat lower than for the domes of St. Vincent (1.5 × 108 Pa), Mount Kameni 1939 dome St. Vincent 1979 St. Helens (1.3 × 108 Pa), or Pinatubo (9 × 107 Mount St. Helens 1981 - 1986 Pa; Griffi ths, 2000). It is worth noting that while the dome-height data are consistent with a t1/4 time dependence, alternative relationships cannot yet be ruled out. For example, measurement errors are realisti- cally likely to have been of the order of 1 m, and 100 the start time of the eruption (as opposed to the time when magma emerged above sea level) is not necessarily well known. In addition, Kameni Dome height (m) domes often show a transition from endogenous growth early in the eruption, to modifi cation by Vulcanian explosions at later stages (e.g., Kténas, 1926). For these reasons, the possibility of a different time dependence (e.g., t1/3 for a fi xed dome shape and a uniform eruption rate) cannot be ruled out; and in the event of a future 10 eruption, careful quantitative observations of 1.0E+05 1.0E+06 1.0E+07 1.0E+08 1.0E+09 dome growth will be of great value. Elapsed time (s) The close similarity of the behavior of the 1866–1870 and 1939 domes allied with the uni- Figure 14. Lava dome height (m) as a function of growth time, showing data from the 1866 formity of the composition of the Kameni Islands and 1939 eruptions of the Kameni Islands (1866 Giorgios dome; 1939 Fouqué dome; Table 8), over the past 2200 yr suggests that we may use the the Soufrière volcano of St. Vincent eruption (1979; Huppert et al., 1982), and Mount St. heights of domes from earlier historic eruptions Helens (Swanson and Holcomb, 1990). The dependence of dome height on elapsed time for to infer the eruption durations. We summarize the the Kameni and other domes is close to t1/4, which is expected from analysis of domes with a results of this analysis in Table 9. The predictions crustal yield strength (Fink and Griffi ths, 1998; Griffi ths, 2000). The solid line is the best-fi t of both the t1/4 model and a time model based curve through the Kameni data with t1/4 dependence: H = 1.217t1/4. simply on the empirical fi t to the height data are consistent with reports of these early eruptions. Both models predict that the dome will grow to a height of ~20 m within 1 d, and ~30 m within 1875 of ~70 m (the 1573 dome), or 101 m (the TABLE 9. INFERRED ERUPTION LENGTHS 4–5 d. Thus, for a submarine eruption originating 1707 dome), expected eruption durations are on FROM DOME HEIGHT on the Kameni-Banco plateau (at ~20 m below the order of 0.3–1 yr and 1–3 yr, respectively. The Eruption Dome height above Inferred eruption sea level), the dome should emerge above the sur- latter is, again, consistent with the contemporary base level duration face within a day or two of the start of activity. records of the 1707–1711 eruption. (m) (d) This is consistent with descriptions of the start of Simply on the basis of observed height of the 1573 70–90 134–368 (Griffi th) the 1707 eruption (Goree, 1710; Tarillon, 1715a; Mikra Kameni (1573) dome, it is plausible that 119–296 (Empirical) Fouqué, 1879). For domes emerging from this the eruption did indeed last for only 1 yr or less. 1707–1711 100–120 560–1165 (Griffi th) 433–835 (Empirical) submarine base level to the heights recorded in This supports our suggestion that the quoted

Geosphere, August 2006 265 Pyle and Elliott

1570–1573 age range for this eruption derives Eruption vent location. Bars indicate approximate size. simply from an uncertainty in the calendrical date of the eruption, and does not reﬂ ect the true dura- Least squares regression line 4030600 tion of the event. 95% confidence interval

Future Activity of the Kameni Islands

Apart from the anomalous event of 1950, all of 4030100 the historic intracaldera eruptions for which there are at least adequate records (those since the

eighteenth century) have shared a number of Northing (m) important common features: pre-eruptive uplift 4029600 of parts of the submarine Kameni edifi ce, and the eruption of magma from at least two eruptive vents. All have also involved the early formation of lava domes, which later act as a focus for vig- orous, intermittent explosive activity, as well as 4029100 acting as the vent from which lavas emerge. There is no reason to suspect that a future eruption would not be preceded by the same general phenomena—including general uplift of the edi- 4028600 fi ce, and discoloration of the sea—and so, given 354200 354700 355200 355700 356200 356700 the current monitoring of the caldera, it is highly Easting (m) likely that the next eruption will be anticipated some days to weeks in advance. Some further Figure 15. Graph showing the areal distribution of known vents from the Kameni Islands. inferences about a future eruption on the Kameni Error bars indicate the scale of the vent, crater, or dome. There is a very tight clustering of Islands can also be gained from a consideration vents about an approximately NE-SW trend, which agrees with the Kameni line identifi ed of the vent distribution and intereruptive periods. previously by Druitt et al. (1989). This trend is thought to mark the locus of a tectonically (fault?) controlled magmatic plumbing system, and future activity is therefore likely to be Vent Distribution initiated at a point, or points, along the same trend.

The distribution of volcanic vents across the Kameni Islands strongly suggests that there is an underlying tectonic control on the supply of magma toward the surface. As previous authors intereruptive shield-building phase, which is The relationship between eruption length and (e.g., Fytikas et al., 1990; Druitt et al., 1989, a typical feature of the evoluti on of Santorini intereruptive interval for the last 5 eruptions of 1999) noted, the vent pattern defi nes a narrow over the past 300,000 yr (e.g., Druitt et al., the Kameni Islands is shown in Figure 16. If we NE-SW trend (Fig. 15) that extends northeast 1999). Despite the considerable work that has regard the 1950 eruption as anomalously short, of Nea Kameni to include the submarine high been completed on the dating of past erup- and interpret this as a minor extrusion follow- (including the former Bancos bank), which tions of Santorini (Druitt et al., 1999), several ing the 1939–1942 activity, the notable feature divides the fl ooded caldera basin into two parts. of the major explosive eruptions still lack a is that the remaining four events lie along a line Many of the prehistoric explosive eruptions also good age control. Consequently, there are insuf- of best fi t described by: eruption length (days) have tephra dispersal patterns that are consistent fi cient data from which to quantify the pattern = 7.4 × (intereruptive interval, yr) + 578. This with vent locations along the same trend (Druitt of intereruptive shield-building periods, for striking relationship is consistent with a model et al., 1989, 1999), and the locations of previ- example, using a rank-order approach (Pyle, for the Kameni Islands of a constant time-aver- ous (pre-1866) and current hydrothermal vents 1998). Instead, we can use the known ages of aged deep supply of magma, and with erup- around the Kameni Islands also lie along this a few key eruptions to infer the mean interval tion lengths that are determined by available same trend. Purely on the basis of the activity between all events. The last phases of seven magma volume. Estimates of erupted volumes of the past 1000 yr, one would anticipate that explosive eruptions were separated, on aver- (Table DR1, see footnote 1) are consistent the next intracaldera eruption would start in age, by 28 k.y. of repose. Within this sequence, with this model; the largest recent eruption the region between Palea Kameni (UTM zone major lava shields were constructed between (1866–1870) followed the longest pre-eruptive 35N coordinates 40291, 3542) and northern Nea ca. 67 and 55 ka (the Skaros shield) and ca. 50 repose period. If this empirical time-dependent Kameni (40302, 3568), and, as long as the erup- and 21 ka (the Therasia dome complex; Druitt relationship holds in the future, then the next tion lasted more than 2 mo, it would involve two et al., 1999). These crude constraints imply that Kameni Island eruption will last for more than or more eruptive vents. the shield-building phase might last for tens of 2.7 yr (in 2006) to more than 4 yr (in 2070). thousands of years. Assuming a mean interval Forecasting Future Events between explosive eruptions of 20–30 k.y., then CONCLUSIONS from Poisson statistics, there is a 50% probabil- The present state of the Kameni Islands can ity that the post-Minoan shield-building phase A combination of high-resolution digital most simply be interpreted as a part of the will last between 14 and 23 k.y. mapping and aerial photography with archived

266 Geosphere, August 2006 Morphology of Kameni Island Lavas, Santorini, Greece

REFERENCES CITED

2000 Barton, M., and Huijsmans, J.P.P., 1986, Post-caldera dacites from the Santorini volcanic complex, Aegean Sea, Greece: An example of the eruption of lavas of near-constant composition over a 2200 year period: 1600 1866 Contributions to Mineralogy and Petrology, v. 94, 1707 p. 472–495, doi: 10.1007/BF00376340. Blake, S., and Bruno, B.C., 2000, Modelling the emplace- 1200 ment of compound lava ﬂ ows: Earth and Planetary Science Letters, v. 184, p. 181–197, doi: 10.1016/ S0012-821X(00)00278-8. 800 1925 Bottinga, Y., and Weill, D.F., 1972, The viscosity of magmatic silicate liquids: A model for calculation: 1939

Eruption length (d) American Journal of Science, v. 272, p. 438–475. 400 Bracewell, R., 2000, The Fourier transform and its applications: New York, McGraw-Hill International, 618 p. 1950 Bruno, B.C., Taylor, G.J., Rowland, S.K., Lucey, P.G., and 0 Self, S., 1992, Lava fl ows are fractals: Geophysical 015020050 100 Research Letters, v. 19, p. 305–308. Bruno, B.C., Taylor, G.J., Rowland, S.K., and Baloga, S.M., Intereruption interval (yr) 1994, Quantifying the effect of rheology on lava-fl ow margins using fractal geometry: Bulletin of Volcanol- Figure 16. Relationship between eruption length and intereruptive interval for the last ogy, v. 56, p. 193–206, doi: 10.1007/s004450050028. fi ve eruptions of the Kameni Islands (1707–1711, 1866–1870, 1925–1928, 1939–1941, da Cologna, L., 1867, De l’infl uence des emanations volcaniques sur les etres organises, particulierement etudiée 1950). The 1950 eruption appears to be anomalously short, and may be best interpreted a Santorin pendant l’eruption de 1866: Paris, Adrien as a minor extrusion following the 1939–1942 activity. The remaining four events lie along Delahaye, 161 p. (in French). Dimitriadis, I.N., Panagiotopoulos, D.G., Papazachos, C.B., a line of best fi t described by: eruption length (days) = 7.4 × (intereruptive interval, yr) + Hatzidimitriou, P.M., Karagianni, E.E., and Kane, I., 578. Based on this relation, the next Kameni Island eruption is expected to last for at least 2005, Recent seismic activity (1994–2002) of the 2.7 yr (in 2006, interval = 56 yr, eruption length = 991 d). Santorini volcano using data from local seismological network, in Fytikas, M., and Vougioukalakis, G.E., eds., South Aegean volcanic arc: Developments in Volcanology, v. 7, p. 185–203. Druitt, T.H., Mellors, R.A., Pyle, D.M., and Sparks, R.S.J., 1989, Explosive volcanism on Santorini, Greece: Geological Magazine, v. 126, p. 95–126. contemporary eruption reports reveals a wealth domes on the Kameni Islands have a crustal yield Druitt, T.H., Edwards, L., Mellors, R.M., Pyle, D.M., of detail relating to the emplacement of viscous, strength (4 × 107 Pa) that is lower by a factor of Sparks, R.S.J., Lanphere, M., Davies, M., and Bar- reiro, B., 1999, Santorini volcano: Geological Society blocky dacite a‘a lava fl ows on the Kameni 2–4 than the domes at Pinatubo (1991), Mount of London Memoir 19, 178 p. Islands, Santorini, Greece. St. Helens (1981–1986), or St. Vincent (1979). Fink, J., 1980, Surface folding and viscosity of rhyolite Lava fl ows from recent eruptions of the The dome-height model enables us to infer fl ows: Geology, v. 8, p. 250–254, doi: 10.1130/0091- 7613(1980)8<250:SFAVOR>2.0.CO;2. Kameni Islands exhibit the classic surface the durations of early historic eruptions on the Fink, J.H., and Fletcher, R.C., 1978, Ropy pahoehoe: morphologies associated with viscous a‘a lava- Kameni Islands. This, combined with the appar- Surface folding of a viscous fl uid: Journal of Volca- nology and Geothermal Research, v. 4, p. 151–170, fl ow emplacement: levées and compression ent time-predictable nature of volcanic eruptions doi: 10.1016/0377-0273(78)90034-3. folds. Levée structures, tens of meters wide of the Kameni Islands since 1573 A.D., allows us Fink, J., and Griffi ths, R.W., 1998, Morphology, eruption and tens of meters high, develop close to the to suggest that the next eruption of the Kameni rates and rheology of lava domes: Insight from labora- tory models: Journal of Geophysical Research, v. 103, vent. Channelized fl ows within the levées show Islands will last for more than 2.7 yr (in 2006) or p. 527–545, doi: 10.1029/97JB02838. prominent ridges with ~20–40 m wavelength more than 4 yr (by 2070 A.D.), and may involve Fouqué, F., 1879, Santorin et ses éruptions: Paris, Masson et and 1–4 m amplitude. Ridges show variable, the formation of a dome of ~115–125 m and Compagnie, 440 p. (in French). Fytikas, M., Kolios, N., and Vougioukalakis, G., 1990, Post- but limited, evolution in terms of scale down- 130–145 m height, respectively. Minoan activity of the Santorini volcano: Volcanic haz- stream, with secondary and tertiary fold wave- ard and risk, forecasting possibilities, in Hardy, D.A., ACKNOWLEDGMENTS Keller, J., Galanopoulos, V.P., Flemming, N.C., and lengths only 50–100% longer than the fi rst- Druitt, T.H., eds., Thera and the Aegean world III, v. 2: formed folds. Levée heights and fl ow widths are London, The Thera Foundation, p. 183–198. Data were collected during the Natural Environ- Georgalas, G.C., 1953, L’éruption du volcan de Santorin en consistent with a Bingham rheology, and lava ment Research Council (NERC) Airborne Remote 4 1950: Bulletin Volcanologique, ser. II, v. 13, p. 39–55. yield strengths are on the order of 3–7 × 10 Pa. Sensing Facility (ARSF) campaign to the eastern Georgalas, G.C., 1962, Catalogue of the active volcanoes Analysis of the shapes of fl ow edges con- Mediterranean in April 2004. Light detection and and solfatara fi elds of Greece: Rome, International fi rms that the blocky a‘a dacite lava fl ows show ranging laser radar (LiDAR) data were expertly pro- Association of Volcanology, v. 12, p. 14–29. cessed in Cambridge, Department of Geography Unit Georgalas, G., and Liatsikas, N., 1925a, Sur la nouvelle a scale-invariant morphology, on length scales for Landscape Modelling, by Gabriel Amable. We éruption du volcan de Santorin (août 1925): Comptes of 1–100 m, with a typical fractal D value of thank David Cobby, Carl Joseph, Ivana Barisin, and Rendus Hebdomadaires des Seances de l’Academie des Sciences, v. 181, p. 425–427. 1.067 ± 0.006, which is indistinguishable from the ARSF team, and George Vougioukalakis (IGME), Georgalas, G., and Liatsikas, N., 1925b, Sur la nouvelle Hawaiian a‘a. On these short length scales, at Stathis Stiros, and Aris Chasapis for assistance with éruption du volcan de Santorin (1925): Comptes Ren- least, there may be limitations to the use of frac- data collection; and Vikki Martin for discussion. Pyle dus Hebdomadaires des Seances de l’Academie des thanks staff of the Earth Sciences and University Sciences, v. 181, p. 1147–1149. tal dimension for inferring lava composition. Libraries, Cambridge; the British Library; and Keele Georgalas, G., and Liatsikas, N., 1936a, Die historische Ent- Dome-forming eruptions of the Kameni University’s ‘Evidence in Camera’ for their able assis- wickelung des Dafni-Ausbruches 1925–1926, in Reck, Islands in 1866–1870 and 1939–1940 showed tance with archive material. Elliott thanks Shell (UK) H., ed., Santorin: v. 2: Berlin, Dietrich Reimer, p. 1–96 for support toward fi eld costs, and Zoë Rice for fi eld (in German). similar patterns of behavior, with progressive Georgalas, G., and Liatsikas, N., 1936b, Der Nautilus Aus- assistance. Part of this work formed John Elliott’s bruch, in Reck, H., ed., Santorin: v. 2: Berlin, Dietrich increases in dome height with time (height = fi nal-year undergraduate dissertation. We thank Ricky 1/4 Reimer, p. 328–340 (in German). 1.2t ) that are consistent with a model of slow Herd and Steven Anderson for their detailed and help- Georgalas, G.C., and Papastamatiou, J., 1951, Uber den infl ation of a dome with a strong crust. Lava ful reviews. Ausbruch des Santorinvulkanes von 1939–1941. Der

Geosphere, August 2006 267 Pyle and Elliott

Kténas Ausbruch: Bulletin Volcanologique, v. 11, Hebdomadaires des Seances de l’Academie des Sci- Mount Etna, Sicily: Geology, v. 4, p. 269–271, doi: p. 1–37. ences, v. 181, p. 563–565. 10.1130/0091-7613(1976)4<269:CAFOLL>2.0.CO;2. Georgalas, G.C., and Papastamatiou, J., 1953, L’éruption du Kténas, C.A., 1926, L’éruption du volcan des Kaménis Stiros, S.C., and Chasapis, A., 2003, Geodetic monitoring of volcan de Santorin en 1939–1941. L’éruption du dome (Santorin) en 1925. Part I: Bulletin Volcanologique, the Santorini (Thera) volcano: Survey Review, v. 37, Fouqué: Bulletin Volcanologique, v. 13, p. 1–38. v. 7–8, p. 3–64. p. 84–88. Goree, F., 1710, A relation of a new island, which was raised Kténas, C.A., 1927, L’éruption du volcan des Kaménis Stothers, R.B., and Rampino, M.R., 1983, Volcanic erup- up from the bottom of the sea on the 23d of May 1707, (Santorin) en 1925. Part II: Bulletin Volcanologique, tions in the Mediterranean before A.D. 630 from writ- in the Bay of Santorin, in the Archipelago: Philosophi- v. 11–12, p. 7–49. ten and archaeological sources: Journal of Geophysical cal Transactions of the Royal Society of London, v. 27, Kténas, C.A., and Kokkoros, P., 1929, Le dôme parasitaire de Research, v. 88, p. 6357–6371. p. 354–375. 1928 et l’évolution du volcan des Kaménis (Santorin): Swanson, D.A., and Holcomb, R.T., 1990, Regularities in Gregg, T.K.P., Fink, J.H., and Griffi ths, R.W., 1998, Bulletin Volcanologique, v. 19–22, p. 87–97. growth of the Mount St. Helens dacite dome, 1980– Formation of multiple fold generations on lava fl ow Leycester, E.M., 1851, Some account of the volcanic group 1986, in Fink, J.H., ed., Lava fl ows and domes: Inter- surfaces: Infl uence of strain rate, cooling rate and lava of Santorin or Thera, once called Callistê, or the Most national Association of Volcanology and Chemistry of composition: Journal of Volcanology and Geothermal Beautiful: Journal of the Royal Geographical Society, the Earth’s Interior, Proceedings in Volcanology, v. 2, Research, v. 80, p. 281–292, doi: 10.1016/S0377- v. 20, p. 1–38. p. 3–24. 0273(97)00048-6. Mandelbrot, B., 1967, How long is the coast of Britain? Tarillon, 1715a, Relation en forme de journal de la nouvelle Griffi ths, R.W., 2000, The dynamics of lava fl ows: Annual Statistical self-similarity and fractional dimension: isle sortie de la mer dans le Golfe du Santorin, in Fleu- Review of Fluid Mechanics, v. 32, p. 477–518, Science, v. 156, p. 636–638. rian D’Armenonville, Th.C., ed., Nouveaux memoires doi: 10.1146/annurev.fl uid.32.1.477. Martin, V.M., Holness, M.B., and Pyle, D.M., 2006, Textural des missions de la compagnie de Jesus, dans le Levant, Griffi ths, R.W., Kerr, R.C., and Cashman, K.V., 2003, Pat- analysis of magmatic enclaves from the Kameni Islands, Paris: Paris, Nicolas Le Clerc, p. 126–161 (in French). terns of solidifi cation in channel fl ows with surface Santorini, Greece: Journal of Volcanology and Geother- Tarillon, 1715b, Extrait d’une lettre écrite de Santorin, in cooling: Journal of Fluid Mechanics, v. 496, p. 33–62, mal Research v. 154, p. 89–102, doi: 10.1016/j.jvolgeo Fleurian D’Armenonville, T.C., ed., Nouveaux mem- doi: 10.1017/S0022112003006517. res.2005.09.021. oires des missions de la compagnie de Jesus, dans le Higgins, M.D., 1996, Magma dynamics beneath Kameni Nams, V.O., 1996, The V Fractal: A new estimator for fractal Levant, Paris: Paris, Nicolas Le Clerc, p. 162–173 (in volcano, Thera, Greece, as revealed by crystal dimension of animal movement paths: Landscape Ecol- French). size and shape measurement: Journal of Volcanol- ogy, v. 11, p. 289–297, doi: 10.1007/BF02059856. Virlet d’Aoust, T., 1866, Histoire des Kaïmênis ou des iles vol- ogy and Geothermal Research, v. 70, p. 37–48, Nicholls, I.A., 1971, Petrology of Santorini volcanic rocks: caniques nouvelles du golfe de Santorin, dans l’archipel doi: 10.1016/0377-0273(95)00045-3. Journal of Petrology, v. 12, p. 67–119. de la Grèce: Paris, Loignon et Cie, 21 p. (in French). Holness, M.B., Martin, V.M., and Pyle, D.M., 2005, Informa- Pègues, A., 1842, Histoire et phénomènes du volcan et des von Fritsch, K., Reiss, W., and Stübel, A., 1867, Santorin. tion about open-system magma chambers derived from îles volcaniques de Santorin: Suivis d’un coup d’oeil The Kameni islands: London, Trübner and Co., 7 p. textures in magmatic enclaves: The Kameni Islands, sur l’état moral et religieux de la Grèce moderne: Paris, von Seebach, K., 1867, Der Vulkan von Santorin nach Santorini, Greece: Geological Magazine, v. 142, Imprimerie Royale, 667 p. (in French). einem Besuche im Marz und April 1866 geschildert: p. 637–649, doi: 10.1017/S0016756805001172. Pichler, H., and Kussmaul, S., 1980, Comments on the geo- Sammlung gemeinverstandlicher Vorträge, ser. II, Huijsmans, J.P.P., 1985, Calcalkaline lavas from the vol- logical map of the Santorini Islands, in Doumas, C., no. 38, 32 p. (in German). canic complex of Santorini, Aegean Sea: Geologica ed., Thera and the Aegean world, v. 2: London, The Vougioukalakis, G.E., and Fytikas, M., 2005, Volcanic haz- Ultraiectina, v. 41, 316 p. Thera Foundation, p. 413–426. ards in the Aegean area, relative risk evaluation, moni- Hulme, G., 1974, The interpretation of lava fl ow morphol- Pinkerton, H., and Sparks, R.S.J., 1976, The 1975 sub- toring and present state of the active volcanic centers, ogy: Geophysical Journal of the Royal Astronomical terminal lavas, Mount Etna: A case history of the in Fytikas, M., and Vougioukalakis, G.E., eds., South Society, v. 39, p. 361–383. formation of a compound lava fi eld: Journal of Aegean volcanic arc: Developments in Volcanology, Hulme, G., and Fielder, G., 1977, Effusion rates and rheol- Volcanology and Geothermal Research, v. 1, p. 167– v. 7, p. 161–183. ogy of lunar lavas: Philosophical Transactions of the 182, doi: 10.1016/0377-0273(76)90005-6. Wadge, G., and Lopes, R.M.C., 1991, The lobes of lava fl ows Royal Society of London, ser. A, v. 285, p. 227–234. Pyle, D.M., 1998, Forecasting sizes and repose times on Earth and Olympus Mons, Mars: Bulletin of Volca- Huppert, H.E., Shepherd, J.B., Sigurdsson, H., and Sparks, of future extreme volcanic events: Geology, v. 26, nology, v. 54, p. 10–24, doi: 10.1007/BF00278203. R.S.J., 1982, On lava dome growth, with application to p. 367–370, doi: 10.1130/0091-7613(1998)026<0367: Walker, G.P.L., 1973, Lengths of lava fl ows: Philosophi- the 1979 lava extrusion of the Soufrière of St. Vincent: FSARTO>2.3.CO;2. cal Transactions of the Royal Society, ser. A, v. 274, Journal of Volcanology and Geothermal Research, v. 14, Reck, H., ed., 1936a, Santorin. Das Werdegang eines Insel- p. 107–118. p. 199–222, doi: 10.1016/0377-0273(82)90062-2. vulkans und sein Ausbruch, 1925–1928. Ergebnisse Warner, N., and Gregg, T., 2003, Evolved lavas on Mars? Kanasevich, E., 1975, Time sequence analysis in geophys- einer Deutsch-Grieschischen Arbeitgemeinschaft: Observations from south-west Arsia Mons and Saban- ics: Alberta, University of Alberta Press, 363 p. Berlin, Dietrich Reimer, 3 vol. (in German). caya volcano, Peru: Journal of Geophysical Research, Kilburn, C.R.J., 2004, Fracturing as a quantitative indi- Reck, H., 1936b, Dynamik und morphogenese des Dafni- v. 108, p. 5112–5127, doi: 10.1029/2002JE001969. cator of lava fl ow dynamics: Journal of Volcanol- Ausbruches als beispiel einer Staukuppenbildung mit Washington, H.S., 1926a, Santorini eruptions of 1925: Geo- ogy and Geothermal Research, v. 132, p. 209–224, Stromerguss, in Reck, H., ed., Santorin, v. II: Berlin, logical Society of America Bulletin, v. 37, p. 349–384. doi: 10.1016/S0377-0273(03)00346-9. Dietrich Reimer, p. 114–207 (in German). Washington, H.S., 1926b, L’eruzione di Santorino del 1925: Kténas, C.A., 1925a, L’éruption du volcan de Santorin: Reck, H., 1936c, Der sekundäre vulkanismus des Bulletin Volcanologique, v. 7–8, p. 65–73. Comptes Rendus Hebdomadaires des Seances de Dafni-ausbruches, in Reck, H., ed., Santorin, v. II: Zellmer, G., Turner, S., and Hawkesworth, C.J., 2000, l’Academie des Sciences, v. 181, p. 376–377. Berlin, Dietrich Reimer, p. 208–264 (in German). Timescales of destructive plate margin magmatism: Kténas, C.A., 1925b, L’éruption du volcan de Santorin. Reiss, W., and Stübel, A., 1868, Geschichte und Beschreibung New insights from Santorini, Aegean volcanic arc: Caractères morphologiques: Comptes Rendus Heb- der vulkanischen Ausbrüche bei Santorin von der Earth and Planetary Science Letters, v. 174, p. 265– domadaires des Seances de l’Academie des Sciences, ältesten Zeit bis auf die Gegenwart: Heidelberg, 281, doi: 10.1016/S0012-821X(99)00266-6. v. 181, p. 518–520. F. Bassermann, 201 p. (in German). MANUSCRIPT RECEIVED BY THE SOCIETY 19 JULY 2005 Kténas, C.A., 1925c, Les phénomènes explosifs de Sparks, R.S.J., Pinkerton, H., and Hulme, G., 1976, REVISED MANUSCRIPT RECEIVED 11 JANUARY 2006 l’éruption du volcan de Santorin: Comptes Rendus Classifi cation and formation of lava levées on MANUSCRIPT ACCEPTED 14 FEBRUARY 2006

268 Geosphere, August 2006