Mediterranean Geoscience Reviews

, Volume 1, Issue 2, pp 223–242 | Cite as

Contrasting fragmentation and transportation dynamics during the emplacement of Dikkartın rhyodacitic dome; Erciyes stratovolcano, central Turkey

  • Orkun ErsoyEmail author
  • Erkan Aydar
  • Erdal Şen
  • Alain Gourgaud
Original Paper


Dikkartın is a monogenetic lava dome emplaced on the southern flank of Erciyes stratovolcano (3917 m) in central Anatolia, Turkey. Stratigraphic, granulometric, morphometric and textural variations are presented to obtain quantitative physical volcanological insight and to interpret eruption and emplacement mechanisms in Dikkartın eruptive sequence. Contrasting fragmentation and transportation dynamics in the stratigraphic sequence of the Dikkartın eruption is related to the ratio of magma and external water during each phases of the activity. We suggest four eruptive phases. The eruption begins with a dry, purely magmatic Plinian eruption with a column height of 20 km. According to the isopleth and isopach maps, the plume stretched in a NNE-SSW direction. The amount of water interacting with magma modified the course of the Dikkartın eruption. During a phase dominated by phreatomagmatism, an explosion crater and a tuff ring were formed. Exhaustion of the water in the environment and depletion of initial magmatic volatiles in the course of time resulted in extrusion of a lava dome. Sequential Fragmentation/Transport (SFT) theory has been applied on tephra samples to comment on fragmentation and transportation dynamics using grain size data. The magmatic and phreatomagmatic characters contained in each explosion sequence were numerically demonstrated. The relationship between the porosity time and the magma-water interaction time was investigated by detailed SEM analysis and roughness values calculated on volcanic ash. Surface analysis on volcanic ash suggested that most of the phreatomagmatic eruptions producing the tuff ring were dry with superheated steam but little vapour condensation occurred allowing adhesion of fine material on surfaces. Considering the c. 20 km plume height calculated for the climactic eruption column (Phase 1), the prevailing winds directed the cloud to the south, but due to the decrease in the plume height (Phase 2), the cloud was later directed to the northeast. According to the direction of seasonal winds, it is suggested that Dikkartın erupted during winter.


Erciyes Dikkartın Cappadocia Anatolia Rhyodacitic dome Fragmentation Phreatoplinian 


The Mount Erciyes is a voluminous stratovolcano in the Cappadocia Volcanic Province, extended over 3300 km2, with at least sixty-four monogenetic vents on its flanks (Şen et al. 2003). Its summit is 3917 m above sea level (relative height c. 3000 m from the Sultansazlığı basin) and majestically dominates Kayseri city, with > 1 Million inhabitants. The geological and volcanological evolution of Mount Erciyes was studied by Şen et al. (2003), who subdivided into two main evolutionary stages namely: Koç Dağ and Erciyes stages, from Pliocene–Quaternary to historical times. A caldera collapse marks the end of the Koç Dağ stage with the eruption of Valibabatepe Ignimbrites (2.52 ± 0.49 Ma, Aydar et al. 2012). The Erciyes stage started with dacitic and andesitic lavas extruded following caldera boundary and/or within collapsed caldera floor (Şen et al. 2003; Aydar et al. 2019). During the Erciyes stage, several generations of various magma types such as basaltic andesite, andesite, dacite, and rhyodacite were repeatedly erupted (Şen et al. 2003). The Mount Erciyes witnessed a very localized, but important rhyodacitic eruptions leading to emplacement of Perikartın, Karagüllü and Dikkartın domes on its northern and southern flanks. Zircon double dating results related to those lava domes (U/Th–He) yield the eruption ages as 11.4 ± 0.9 ka, 9.4 ± 1.4 ka, 7.2 ± 0.7 ka, respectively (Yurteri 2018). Besides, Sarıkaya et al. (2019) used cosmogenic surface exposure method to date those domes, and proposed average exposure ages for Karagüllü, Perikartın and Dikkartın lavas ranging between 6.4 ± 0.9 and 10.5 ± 0.9 ka, 6.3 ± 0.6 and 8.9 ± 0.5 ka, 7.6 ± 0.6 and 10.3 ± 0.5 ka, respectively. A pyroclastic flow originated from the Perikartın lava dome buried trees that were later converted to charcoal yielded an average radiocarbon (14C) age of 9728 ± 110 cal. yr B.P. (calibrated using Calib 7.1) (Sarıkaya et al. 2019).

Hamann et al. (2010) found 1 mm thick tephra, at offshore of Israel, within S1 Sapropel, that was deposited between 8970 and 8690 cal yr B.P. and attributed that marine tephra to Dikkartın eruptive sequence. Moreover, Cullen et al. (2014) found a cryptotephra layer in a Black Sea core and they chemically correlated those tephras with Perikartın or Karagüllü sequences. Each one of the Dikkartın, Karagüllü and Perikartın domes has been emplaced after important pyroclastic sequences characterized by an alternation of layers related to contrasting fragmentation and transportation dynamics (Şen et al. 2002).

In this paper, stratigraphic, granulometric, morphometric and textural variations are presented in detail, to obtain quantitative physical volcanological insight and to interpret eruption and emplacement mechanisms in Dikkartın dome. In order to better characterize the samples with respect to their size distribution and highlight possible changes in the fragmentation processes occurred during eruptions, Sequential Fragmentation/Transport (SFT, Wohletz et al. 1989) theory has been applied to tephra. Owing to the fact that every single ash particle might convey information about its own formation environments and conditions, quantitative surface descriptors on volcanic ash surfaces were calculated. We finally proposed a volcanological model concerning Dikkartın sequence and a tephra dispersal model for different plume heights and seasons.

Presentation of Dikkartin eruptive sequence

Dikkartın sequence exhibits non-welded fallout, massive pumiceous flow deposit, planar, massive to wavy pyroclastic density current deposits (surges) interlayered with fallouts, lithic rich fallout and a later stage dome extrusion (Fig. 1). Here, the term pyroclastic surge deposit has been used to designate thin (maximum 30 cm) planar-massive-wavy ash-rich bed produced from pyroclastic density currents. According to the distribution, bed forms and eruptive style, we suggest four eruptive phases for Dikkartın. Isopleth and isopach sketchings for Plinian fall deposits of the main phase (Phase 1) are illustrated in Fig. 2. According to pattern obtained we can suggest that the plume of main phase was stretched along NNE-SSW.
Fig. 1

a Geodynamical sketch map of Anatolian block, locations of major faults and study area (Mount Erciyes) placed on Digital Elevation Model (DEM) of Anatolia. (AEZ aegean extension zone; NAF north anatolian fault; EAF east anatolian fault; EF ecemiş fault; TGF tuz gölü fault. White and black arrows indicate the sense of plate motion; white half arrow show the relative motion direction on the faults), b simplified geological sketching map of Mount Erciyes after Şen et al. (2002), c generalized stratigraphic section of Dikkartın products. Grain-size distributions were given as histograms (between − 4 and 5 φ)

Fig. 2

Grain size (long axis of the fragments) (isopleth-in cm) and thickness (isopach-in m) sketching of Dikkartın Tephra Fall Sequence. a Max 5 Pumice, b Max 5 Lithics, c Thickness (modified from Şen et al. 2002)

Phase 1

The Phase 1 includes the initial Plinian fall and pyroclastic-flow deposits. They are more widespread and exposed c. 15 km from the source. The wider distribution, magmatic character, and coarse particles of Plinian fall deposits indicate the higher eruption column in the beginning of the eruption. Applying the method of Carey and Sparks (1986), the height of the climactic eruption column was estimated as 20 km. The opening Phase 1 generated two stratigraphic units; D1 fallout and D2 pumiceous flow deposits (Fig. 3). The fallout deposits and flow deposits are separated with a thin ash layer (2 cm thick). The column deposited D1 fallout including grey, vesicular, moderately angular pumice clasts with maximum 58 cm diameters (Şen et al. 2002). The unit is 15 m thick at the outcrop, which is 6 km from the inferred source. This fall unit is overlain by a pyroclastic flow (D2) with ash matrix. At 3 km southeast from Dikkartın dome, it exhibits normal grading including oversized pumices (60 cm in diameter) at the bottom of the deposit (Fig. 3). In distal zones (7 and 8.5 km from the source), it shows reverse grading. Pumice clasts in the flow deposit are pink, white, grey in colour and rounded in shape. Pyroclastic flow deposit has 3 m thickness, while the upper surface is eroded in the relatively distal zone, 3 km southeast from the inferred source.
Fig. 3

Plinian fall and pumiceous flow deposits of Phase 1 at 3 km southeast from the source zone. The thin ash layer separating the deposits is more pronounced in A. Pink-coloured coarse pumice clasts exist in the lower parts of the flow deposits

In the very proximal location (northwest from the source area) a pumiceous flow deposit (D3) with pink, white and grey pumices also outcrops (Fig. 4). The pumices in this deposit are moderately rounded. The lower part of the flow is buried by debris and the observed thickness is c. 2–3 m. This unit is thought to represent the transition between Phase 1 and Phase 2.
Fig. 4

Deposits related to Dikkartın eruption 1.5 km northwest of the source area (very proximal zone)

Phase 2

The Phase 2 is related to the opening of an explosion crater and a tuff ring formation. The deposits of Phase 2 are not widely dispersed and crop out around the dome. The Phase 2 starts with a planar finely stratified layer (D4-surge) with 30 cm thickness in the proximal zone (Fig. 4). Planar stratification is defined by fine pumice lapilli. D5 is a fall deposit with pink, vesicular and moderately rounded pumices. The deposit is 20 cm thick with maximum pumice size (Mp) attaining 6 cm. Fall deposit D5 is overlain by a structureless (massive) ash layer (D6-surge) with 7 cm thickness and exhibit bomb sags (Fig. 4). D7 is a fall deposit with maximum 6 cm sized pink and grey pumices reaching 15 cm in thickness. D8 is wavy-stratified surge deposit with 15 cm thickness and exhibits antidune features and bomb sags.

Phase 3

The products of this phase outcrop in the north and northeastern part of the lava dome. D9 is a thick (15 m) sub-Plinian fall deposit overlying all units in the proximal zone (Figs. 4 and 5). The deposit is rich in lithics (62.9 wt %) and glass (19.15 wt %) and exhibits bread crust bombs with 60 cm diameters. Some bombs have radial cracks (Fig. 5). Transitions between phases indicate that the eruption was continuous and that there is no significant time interval between the phases.
Fig. 5

Lithic-rich fall deposits related to Phase 3 of Dikkartın eruption 1 km north of the source. a Coarse bombs (60 cm diameters) exist in this deposit. b Bomb with radial cracks

Phase 4

The Phase 4 is an extrusive phase where a lava dome was emplaced. Rhyodacitic dome of Dikkartın represents occasionally banded obsidian facies. It buries the previous explosion crater and reaches 2760 m in height (Fig. 6). The extruded lava flowed down c. 5 km towards south and is composed of blocky surface due to its high viscosity. This dome-flow covers an area of 11.7 km2 and corresponds to 0.82 km3 of erupted magma (Şen et al. 2002).
Fig. 6

View of the explosive deposits and lava dome from northeast through southwest


Component analysis

Component analysis included hand picking under a binocular microscope and weighting. Components have been divided into the following groups: (1) poorly vesicular pumice clasts (lithics); (2) pumice clasts with moderate vesicularity; (3) highly vesicular pumice clasts; (4) juvenile glass (obsidians); and (5) xenoliths (accessory lithics). Although free crystals were observed on SEM micrographs of particles in the size 4 φ (63 µm) and 2 φ (250 μm), they were lacking in the size of − 2φ (4 mm). Therefore, they did not participate in the component calculations.

The density of juvenile fragments was measured with a pycnometer after coating fragments with a thin impermeable film using a silicon-based spray. About ten measurements per sample were carried out in the − 2φ (4 mm) grain size. The mean values were considered. The vesicularity index (V %) was calculated using the method of Houghton and Wilson (1989):
$$V\left( \% \right)\, = \, 100 \, \left( {{\text{DRE density}} - {\text{clast density}}} \right)/{\text{DRE density}}$$
The DRE density was determined on non-vesicular, dense grains, while the average clast density is that of the most vesiculated fragments.

Grain size analysis

Grain size analyses were performed by dry sieving at full-φ intervals up to 4φ. Log-normal sorting, median and skewness parameters were calculated after Inman (1952).

Sequential fragmentation transport theory

In order to better characterize the samples with respect to their size-distribution, and highlight possible changes in the fragmentation processes occurred during eruptions, the SFT (Sequential Fragmentation/Transport) theory (Wohletz et al. 1989) has been applied to Dikkartın tephra. The grain size distributions were analysed using the Windows-based software SFT (a new version of SEQUEN from Wohletz et al. (1989)), which allows user-interactive discrimination of the subpopulations present in the distribution, and their characterization in terms of three parameters: mode, dispersion and weight fraction. Subpopulations were determined in all samples. There are several possible causes for existence of the designated subpopulations. These include (1) size populations inherited from the initial fragmentation of the magma and country rock; (2) size populations related to clast type and density (e.g., crystals, lithics, pumice and glass); and (3) size populations related to transport and deposition processes.

SEM Analysis

Samples in the 2 φ (250 μm) grain size were steeped in acetone not more than 5 min so as to remove dust adhered on particles during sieving process or transportation of samples from field to laboratory. Not to detach adherent alteration products or shards holding clues about steam abundance and energy in eruption, respectively was expected. Samples were dried in drying oven at 110 °C overnight. Volcanic ash particles were sprinkled on carbon adhesive tabs placed on an aluminium stub and coated with carbon in order to counteract grain surface charging while scanning with the electron beam. The SEM instrument (Zeiss EVO50) was used to take the whole grain and detailed surface micrographs. Ten scanning electron images were acquired from each sample.

Roughness measurements on volcanic ash surfaces

In order to determine the surface roughness of volcanic ashes, image analyses were performed on SEM images. The image processing side is based on assuming that the grey level variation is related to the roughness variation of the surface. Namely, we use two-dimensional SEM micrographs and the shadow relief effect for estimation of real topographical representations. The advantages and disadvantages of the method here applied were discussed in Ersoy et al. (2007).

Image processing and analysis were performed using the ImageJ program (Rasband 1997). All images were converted to a sequential stack. Median filter was carried out which is the nonlinear filter more used to remove the impulsive noise from an image (Yin et al. 1996). Bandpass filter was used to filter out large structures (shading correction) and small structures (smoothing) of the specified size from the images. The suitability of bandpass filtering for assessing the surface structure size was also confirmed in previous studies (Chinga et al. 2007). SurfCharJ plugin (Chinga et al. 2003) was used for roughness analysis. About roughness, we dwelled upon two widely used statistical roughness descriptors (average roughness Ra and the skewness Rsk), which were used in previous applications on volcanic ash surfaces and provided information about vesicularity, types of pyroclasts (Wohletz et al. 1983), alteration intensity and/or fine particle abundance.



The relative abundance of components in the size − 2φ (4 mm) was determined as described in the methodology section and given in Table 1. Dikkartın samples include grey, white and pink phenocryst-rich pumices. Characteristically, the colour shifts from white to grey while the pumice clasts get denser. The obsidian clasts are vesicular or non-vesicular. Here, the term ‘xenolith’ for accessory components and ‘lithic’ for juvenile denser particles was used. The xenolith types include light–dark grey lava and reddish hydrothermally altered lava. The fact that the component types differ even in successive layers also explains that Dikkartın eruption formed an alternation of layers related to contrasting fragmentation and transportation dynamics.
Table 1

Components in samples (%)



































































MVP Moderately vesicular pumices, HVP highly vesicular pumices

Accessory lithics

All products, explosive and extrusive, exhibit similar mineralogical composition: plagioclase (An40-82 in pumices, An39-83 in lava dome), orthopyroxene (En60-66 in pumices, En59-66 in lava dome), amphiboles (tchermakitic hornblende and magnesio-hornblende in pumices and lava dome), and glass (74–75% SiO2 in pumices 76–77% SiO2 in lava dome) (Şen et al. 2002). The particles are generally vesicular regardless of the eruption types and bed-forms indicating the effect of magmatic volatiles during fragmentation or subsequent contact of water in conduit with already vesiculated and/or fragmented magma (Fig. 7). However, there is a general decreasing trend through time in the eruption sequence indicating loss of initial gas content during eruption, as also proposed by Şen et al. (2002).
Fig. 7

Evolution of vesicularity index (V %) in samples through the eruptive phases of Dikkartın

Grain size characteristics

The plots of samples on median diameter (Md φ) versus sorting (σφ) diagram were compared with fields of deposits from well-studied eruptions (Fig. 8a). Surge deposits (D4, D6 and D8) have median grain sizes (Md φ) between 1.55φ and 2.05 φ (medium-fine grained ash) and poor sorting (σφ = 2.13–2.35). Pyroclastic flow deposits (D2 and D3) have median grain sizes (Md φ) between 0 and 0.25 φ (coarse ash) and poor sorting (σφ = 2.45–3). Fall deposits (D1, D5, D7 and D9) have median grain sizes (Md φ) between -2.05 and − 3.35 φ (medium to fine lapilli). Sorting is moderate in the first Plinian fall (D1) (σφ = 1.65) but the phreatoplinian fall (D5 and D7) and the last sub-Plinian fall deposits (D9) are poorly sorted (σφ = 2.35–2.73). The deposits relating to Phase 1 have harmonic trends of Md φ and σφ (Fig. 8b) indicating better sorting with increasing median grain sizes. However, the other deposits have opposite trends for these parameters that show better sorting with decreasing median grain sizes. The skewness of the size distribution (Inman 1952) distinguishes fall samples from others (Fig. 9a). All pumice-rich fall samples are positively skewed, however, ash-rich surge samples and pumiceous flow deposits lack noticeable skewness. F1 (< 1 mm) versus F2 (< 63 µm) diagram also distinguishes fall, flow and surge deposits (Fig. 9b). The ranges for F1 values are narrow for flow and surge deposits but wider for fall deposits. In contrast, flow and surge deposits have wider ranges of F2 values than those of fall deposits. Fall deposits are uniform in F2 values. Within all surge deposits, the first surge deposit (D4) overlying the pumiceous flow deposits is closer to flow deposits, chiefly in terms of F2 values. Surge deposit D6 seems to be produced from the most intensive fragmentation mechanism according to its fine ash (< 63 µm) abundance (Zimanowski et al. 2003).
Fig. 8

Plots of sorting (σ) and median diameter (Md). a Walker diagram showing the 1% and 8% fields of pyroclastic flows (enclosed by solid lines) (Walker 1971) and falls (enclosed by dashed lines) and that of pyroclastic surge dunes (enclosed by dotted line) (Fisher and Schmincke 1984). Labelled fields cover silisic phreatomagmatic deposits: Askja layer C base surge and phreatoplinian fall deposits (Sparks et al. 1981), Wairekei (Oruanui) Formation Member 2 phreatoplinian fall deposits (Self 1983), distinct phreatomagmatic character described by Sheridan and Wohletz (1983). b Md and σ vs. stratigraphic position

Fig. 9

a Skewness versus sorting. b F1 (< 1 mm) versus F2 (< 63 µm)

Modelling of granulometric data and sequential fragmentation transport (SFT) theory

Four distinct subpopulations (A, B, C and D) were determined in samples in varying degrees of prominence after Sequential Fragmentation Transport (SFT) analysis (Fig. 10).
Fig. 10

Subpopulations determined from analysis of grain-size distributions of deposits


Best-fit polynomial curve on subpopulation fraction versus subpopulation mode (φm) define two groupings of subpopulations: coarser mode subpopulations A and B can be separated from those of fine subpopulations C and D by a line drawn at − 0.5 φ (Fig. 11a). These groupings are assigned to magmatic and phreatomagmatic origins (Wohletz et al. 1995). There is a distinct separation between two groups assigned to different fragmentation mechanisms in Dikkartın tephra. Subpopulation dispersion (or sorting, γ) is very sensitive to fragmentation mechanism. Subpopulations can be distinguished by their mode in Fig. 11b in which a fragmentation factor is plotted versus mode. The fragmentation factor chosen (expressed as φm/γ) discriminates the relative contribution of magmatic and phreatomagmatic fragmentation mechanisms in development of the size distribution of each sample (Wohletz et al. 1995).
Fig. 11

a Subpopulation fraction versus subpopulation mode (φm) indicates two main groupings of sub-populations. b Fragmentation factor (φm/γ) versus subpopulation mode

The variation in these parameters is uniform except the extreme coarse and fine tails. The anomalous dispersion values may be resulted from particle aggregation despite sample disaggregation. The coarse modes have positive values of fragmentation factor while fine modes have negative values. This discrimination separates magmatic from phreatomagmatic subpopulations and is supported by field, laboratory and theoretical observations described by Sheridan and Wohletz (1983), Wohletz et al. (1983, 1995), which show that magmatic tephra have coarser grain sizes and more negative γ values than do phreatomagmatic tephra.

The distinct character of fragmentation factor was shown on a histogram of average values of each four subpopulations (Fig. 12a). Subpopulations A and B are dominantly formed by magmatic fragmentation while subpopulations C and D are formed by phreatomagmatic fragmentation. Magmatic and phreatomagmatic components for each stratigraphic position that can be related to progressive changes in eruptive dynamics were shown on Fig. 12b. The widely dispersed fallout layer (D1) is mostly magmatic with only magmatic subpopulations A and B. The pumiceous flow deposit (D2) overlying D1 varies between magmatic and phreatomagmatic. D3 pyroclastic flow unit, which is considered to represent the transition between Phase 1 and Phase 2 show higher phreatomagmatic character than D2 deposit. In the products of Phase 2, the surge deposits have small amounts of magmatic character. The fall deposits have important amounts of magmatic component while they include phreatomagmatic subpopulations. The reason of bimodality may be first disruption by expansion of magmatic gases to produce the coarse mode and then further fragmentation by explosive interaction with water (Self and Sparks 1978). The relative contribution of magmatic and phreatomagmatic components to Dikkartın were shown on Fig. 13a. Dikkartın has an important phreatomagmatic character (53.22%). In order to find how the water/magma mass ratio (R) varied during the course of the Dikkartın eruption, the fraction of phreatomagmatic constituents was assumed as a measure of water abundance and that for magmatic constituents as a measure of magma abundance (Wohletz et al. 1995). Their ratio was normalized to 1 where all fragmentation is phreatomagmatic. The estimated R values with stratigraphic positions in Dikkartın are given in Fig. 13b. The pyroclastic bed types were also plotted on samples. These data support field interpretations that Dikkartın varied between magmatic and phreatoplinian. The fall deposit (D1) of the main phase (Phase 1) underlying the pumiceous flow deposit (D2) is purely magmatic. The phreatomagmatic character in pumiceous flows increases from D2 to D3, and the deposit type changes to interlayered surge and fallouts. The sub-plinian fall deposit (D9) has also significant amount of magmatic components.
Fig. 12

a Histogram of average φm/γ values as a function of subpopulation. b Magmatic and phreatomagmatic components for each stratigraphic position

Fig. 13

a Pie diagram shows an abundant phreatomagmatic character (53.22%) for Dikkartın eruption. b Plot of R (water/magma mass ratios) vs. stratigraphic position


The Dikkartın shows a variety of depositional textures suggestive of multiple modes of tephra dispersal and emplacement. Fallout layers are characterized by the coarsest modes (subpopulations A and B) while flow/surge layers show a range for particle sizes including subpopulations C and D (Fig. 12b). Fine-grained surge beds D6 and D8 are dominated by subpopulation C but have an important content of subpopulation D. The massive forms and wavy stratification in these surges can be explained by modes of appropriate size for saltation transport. The first surge deposit D4 involves lesser amount of subpopulation D but higher subpopulation C and substantial amount of coarser subpopulation B. Thus, coarse-medium grained surge bed (D4) shows bedding texture suggestive of emplacement in a traction carpet (planar stratification beds).

The pyroclastic density current, which produced the pumiceous flow deposits, had features transitional between those of conventional pyroclastic surges and pyroclastic flows, chiefly for D3, which has a significant amount of phreatomagmatic components (Figs. 12b and 13b). Probably, water influx caused an unsteady pyroclastic current characterized by variable transport processes, both suspension and traction. In terms of textural features and origin, pumiceous flow deposits show similarities to the B pyroclastic density current deposit in Kos Plateau Tuff (Allen and Cas 1998). The moderate-high particle concentration (rather than dilute) of the current commonly suppressed turbulence and the development of tractional bedforms within the most proximal areas favoured rapid suspension sedimentation that formed massive beds. The fall deposits involving substantial amount of coarser subpopulations were likely transported by ballistic fallout (Wohletz et al. 1989).

Qualitative descriptions of volcanic ash surfaces

In Plinian fall deposit (D1) highly vesicular and slightly vesicular pyroclasts coexist together. Vesicles show variety of forms including elongated, tubular and ovoidal cavities (Fig. 14). The slightly vesicular pyroclasts are common and more angular with large ovoid vesicles and thick vesicle walls. Large pockets compose coalesced vesicles and small vesicles surround these pockets. Septae that remain from collapsed walls between coalesced vesicles are visible within the vesicles. Many of the vesicles are not only elongate but also flattened. The vesicle walls between elongated vesicles in vesicular pyroclasts are thin (~ 5 μm) while they are thick (~ 60 μm) between ovoid vesicles in slightly vesicular pyroclasts. Some pyroclasts exhibit crystals embedded in the glass. Fine-grained particles adhering to grain surfaces are present. The outlines of particles are angular in general.
Fig. 14

Representative SEM micrographs of ash surfaces from Dikkartın eruption

Pumiceous flow deposits (D2 and D3) has equant, subequant blocky pyroclasts with mostly ovoid vesicles (Fig. 14). Moderately thick vesicle walls (~ 20 μm) separate the vesicles. The pyroclasts also have slightly elongated vesicles and many of these vesicles are also flattened. Most of the ovoid vesicles show collapse of mutual vesicle walls and coalescence to form larger composite vesicles. Fine-particles adhering to grain surfaces are abundant indicating higher fragmentation intensities for fine material formation and steam abundance for adhesion in the eruption. Generally, the particles in pyroclastic flow deposit are rounded.

Fine adhering particles are also present on pyroclasts from D4 surge deposit (Fig. 14). Three different types of pyroclasts are present in this surge deposit; (1) vesicular pyroclasts with elongated vesicles and thin vesicle walls, (2) vesicular pyroclasts with ovoid and coalesced vesicles, and (3) slightly vesicular equant, subequant subrounded pyroclasts with few ovoid vesicles. The pyroclasts from surge deposit D4 are subrounded.

Fall deposit D5 has vesicular pyroclasts with abundant ovoid vesicles (Fig. 14). Elongated vesicles are also present but minor. Large pockets composed of coalesced vesicles and small, elongate vesicles surrounding the pockets are present. Fine particles adhere to surfaces but not abundant as in pumiceous flow and surge deposits. Fall deposit D5 has angular pyroclasts.

Surge deposit D6 involves pyroclasts with fine material on their surfaces similar in abundance to D4 (Fig. 14). Ovoid vesicles in immature form are as nicks on surfaces. They have minor depths and seem shallow. Slightly elongated vesicles unusually exist but they are also shallow with minor depths. Particles of surge deposit D6 are rounded.

In fall deposit D7, heterogeneous nature of the vesicles is obvious: there are large pockets composed of coalesced vesicles, spherical-ovoid vesicles, and slightly elongated vesicles (Fig. 14). In general, all vesicles are shallow with minor depths. Vesicular pyroclast are angular and rich in spherical-ovoid vesicles. Fine adhering material is present but minor.

Abundance of fine adhering particles and hollows filled by fine material are characteristic in surge deposit D8 (Fig. 14). Pockets are present with ovoid and elongated vesicles. The vesicle walls have thicknesses in a wide range (5–20 μm). Pyroclasts of D8 surge are angular.

Fall deposit D9 has moderately vesicular pyroclasts with ovoid and slightly elongated vesicles and non-vesicular pyroclasts (Fig. 14). The vesicle walls are thick up to 100 μm. Fine material on particle surfaces is limited. Pyroclasts of D9 fall deposit are highly angular.

Quantitative analysis on volcanic ash surfaces

The geometrical characterization of rough profiles or surfaces is a widespread problem in various geological examples such as erosion patterns, multiphase fluid percolation in porous rocks, fractures, or stylolites (Brouste et al. 2007 and references therein). Several texture descriptors have been used to characterize the detailed surface structure of volcanic ash surfaces (Ersoy et al. 2006, 2007). The roughness of volcanic ash surfaces is associated with key features such as the vesicularity, type of vesicles and alteration products (e.g. fine material) on surfaces that are mainly affected by water/magma interaction during eruptions. Here, we calculated some statistical roughness parameters on volcanic ash surfaces in order to assess the micro-geometrical deviation related to different fragmentation mechanisms.

Ersoy et al. (2007) demonstrated the sensitivity of Ra parameters on fine adhered particles on volcanic ash surfaces, namely on the micro-roughness of the surfaces. The adhering dust on surfaces or in vesicle hollows of volcanic ash particles is common especially in phreatomagmatic eruptions. However, these fine materials may grow out from alteration of glass or fine-sized fraction of the eruption products. Both may be the results of steam in the eruption column causing alteration and adhesion of fine material to surfaces, respectively. Fine-sized juvenile material may form besides as an effect of intensive fragmentation mechanism (Zimanowski et al. 2003). Here, the surge deposits related to Phase 2 and flow deposits from Phase 1, mainly D2, have higher Ra values indicating higher micro-roughness (Fig. 15a). The adhered fine material on surfaces forms a rough surface in micro-scale. Moreover, the quantitative micro-roughness descriptions associated with adhered fine material overlaps the qualitative descriptions of surface textures cited above. The Plinian fall deposit (D1) despite its purely magmatic character (R = 0.01) involves substantial amount of adhered fine material on pyroclast surfaces indicating a steam-rich highly energetic fragmentation and eruption mechanism (Zimanowski et al. 2003). Rsk is more descriptive on larger features such as cavities (vesicles) (Fig. 15b). It has been proposed as a parameter that describes vesicularity and distinguishes pyroclasts from different fragmentation mechanisms on the basis of different type and sized cavities on the ash surfaces (Ersoy et al. 2008). Here, it has a considerable correlation with vesicularity values of samples indicating its sensibility on cavities (Fig. 16).
Fig. 15

Evolution of a average roughness (Ra) and b skewness (Rsk) through time in Dikkartın deposits

Fig. 16

The correlation between skewness (Rsk) values of ash surfaces and vesicularity index (V %)


Volcanic facies scenarios

Stratigraphic, granulometric, morphometric and textural variations were presented to obtain quantitative physical volcanological insight and to interpret eruption and emplacement mechanisms in Dikkartın eruption. According to the distribution, bedforms, and eruptive style, four eruptive phases for Dikkartın were suggested. Dikkartın sequence exhibits non-welded fallout (D1), massive pumiceous flow deposits (D2 and D3), planar, massive to wavy pyroclastic density current deposits (surges) (D4, D6 and D8) interlayered with fallouts (D5 and D7), lithic- rich fallout (D9) and a later stage dome extrusion (Fig. 1).

Şen et al. (2002) studied one pyroclastic-flow deposit in the south and east of the dome. Although no contact between the D1 and D3 deposits has been observed in the field, both pumiceous flows (D2 and D3) are thought to be flows formed by partial collapse of the Plinian plume in the main phase (Phase 1). Although field observations and the macro and micro pumice characteristics show that these two flows are similar, they are named as D2 and D3 due to the differences in the magmatic and phreatomagmatic components revealed in this study.

The fallout deposit (D1) at the beginning of main phase (Phase 1) is purely magmatic and relatively widespread. However, despite its purely magmatic character (R = 0.01), it involves substantial amount of adhered fine material on pyroclast surfaces indicating a steam-rich highly energetic fragmentation and eruption mechanism (Zimanowski et al. 2003). Formation of a tuff ring shortly after this eruption indicates a shallow explosion, namely interaction of magma with a surface or near-surface water (Wohletz and Sheridan 1983). Limited infiltration of water into the conduit above the fragmentation level may permit the occurrence of steam in the eruption column of the main phase. The fine materials on pyroclast surfaces are juvenile chips and flakes other than alteration products indicating relatively dry emplacement of deposits (superheated steam media) beside the little vapour condensation on pyroclasts.

The pumiceous flow deposits overlying the Plinian deposit have components with magmatic and phreatomagmatic origin indicating operation of both magmatic and phreatomagmatic fragmentation mechanisms during the eruption. The phreatomagmatic character increases from D2 to D3 and transits to a surge deposit (D4) that belongs to Phase 2. For the first phase in Dikkartın, a shift from an early convective to a transitional and partly collapsing regime is suggested. Eventually, also due to intense fragmentation and/or higher discharge rates, the vent widened or the eruption column became greatly overloaded with fine particles, and the partial column collapse began (Sparks and Wilson 1976; Sparks et al. 1978; Wilson et al. 1981). The relatively higher proportion of xenolith contents in Plinian fall and pumiceous flow deposits refers to a probable vent enlargement (Table 1).

The enlargement of vent probably allowed water access to the vent. Moreover, water interacting with the magma probably acted as a heat sink, further reducing the ability of the column to convect (Wilson et al. 1978; Sheridan et al. 1981). Lower rate of heat release decreased the height of column (Woods and Wohletz 1991) and transformed it into a phreatoplinian column relating to the fallouts and surge deposits in Phase 2. The scale of phenomena was small but similar in comparison with the eruption and deposition styles in the eruption of Kos Plateau Tuff (Allen and Cas 1998). Furthermore, initiation of phreatoplinian eruption after widening of the vent allowing water access to the magma resembles the case in the eruption of Bishop Tuff, California (Heiken and Wohletz 1985). The phreatomagmatic origin of both fall and surge deposits in Phase 2 and their dispersions connotes that they were derived from the same column and/or surges originated from lateral explosions derived from the same vent.

Although the deposits of Phase 2 involve phreatomagmatic components, vesiculation is observed in all samples. Water-magma interaction was evidently restricted to shallow levels in the conduit, and largely above the level of vesiculation and probably in patches above the level of volatile driven fragmentation, chiefly for phreatoplinian fall deposits. Magma was disrupted first by expansion of magmatic gases to produce the coarse mode and then further fragmentation occurred by explosive interaction with water (Self and Sparks 1978). This scenario was also supported by the skewness of the grain size distribution (Fig. 9a). Positive skewness of fall samples can be interpreted as mixing of two populations of grain-size with different origins. These two size populations reflect breakage of the magma by two different mechanisms, initial explosive expansion of magmatic gases and after explosive interaction with water (Self and Sparks 1978). Within the fall deposits, the main Plinian fall deposit (D1) is closer to the zero value of skewness indicating a more homogeneous fragmentation mechanism relative to others.

Even though, the generation of dilute currents that form base-surge deposits from Plinian eruption column seems problematic (Dellino et al. 2004), the lateral or partial collapse of the phreatoplinian column generates laterally moving base surges (Self 1983). Xenolith fragments within the deposits were probably derived from the shallowest levels in the stratigraphy intersected at the vent and conduit. However, there is no clear distinction between the xenolith contents of relatively dry fall deposits and wetter surge deposits (Table 1). Namely, there is no relationship between water/magma mass ratios (R) during eruptions and the xenolith abundance of deposits (R2 = 0). Probably, the conduit wall stability was also important for xenolith abundance beside the water amount accessed to the vent and magma discharge rate. Proportions of juvenile lithic content do not enable us to distinguish fallout deposits from surge deposits. There is no correlation (R2 = 0.03) between the water/magma mass ratios (R) values and lithic content. However, the inverse correlation between the lithic proportions and vesicularity values (R2 = 0.46) indicates the volatile content as the dominant factor rather than magma/water interaction on lithic occurrence. The insignificant correlation (R2 = 0.17) between the water/magma mass ratios (R) and vesicularity values already designates interaction of water after vesiculation of the rhyodacitic magma. The obsidian content has correlation with water/magma mass ratios (R2 = 0.33). Probably, the water interaction in the late stage of vesiculation had a quench effect on pyroclasts forming vesicular and unvesicular obsidian particles in deposits.

The external water in eruption sequences of Dikkartın is believed only to be responsible from the intense fragmentation of already vesiculated magma thereby in determination of grain size distribution and bed forms. Parallel cases were observed on other eruptions where the ground or surface water interacted with magma after vesiculation (e.g. St. Helens, 1980 eruption, Taupo eruption (Heiken and Wohletz 1985)). Magma-water interaction causes intense fragmentation of the magma (Peckover et al. 1973; Wohletz et al. 1983; Wohletz and McQueen 1984; Zimanowski et al. 1991), produces a fine-grained assemblage of clasts in the eruption column (Walker and Croasdale 1972; Wohletz KH et al. 1983) and thus varying ratios of magma to water change the grain-size characteristics of the resulting deposits (Self 1983). During the eruptions that deposited fallouts, mass flux rates of magma was probably high that water magma interaction was less efficient yielding a large coarse component to the grain size distribution. The increase in median grain size is interpreted as a decrease in the explosive energy of eruptions. Experimental studies demonstrate that exsolution of magmatic gas tends to inhibit an efficient magma/water interaction (Zimanowski et al. 1991). Furthermore, vesicles developed before water-magma interaction facilitated intense fragmentation producing finer materials. The presence of tubular vesicles instead of ovoidal pores in the magma generates finer fragments (Martel et al. 2000), which also explains existence of tubular vesicles in surge deposits in Dikkartın eruptive sequence.

The subsequent interaction of water after vesiculation mainly prevented the occurrence of key features on pyroclast surfaces indicating phreatomagmatism. The viscous behaviour of magma also prevented viscous deformation of the melt if any turbulent mixing of the melt and water occurred after fracture. Presumably, Rsk specified the abundance of vesicles regardless of whether or not the water interaction prevailed in determination of vesicularity. The vesiculation was heterogeneous in samples; elongated, tubular and ovoidal cavities coexist in the same pyroclast as pockets. Differences in vesicle types can be explained by heterogeneous vesicle development. It is possible for vesicle growth to begin at depths of 2 or 3 km in rhyodacitic or rhyolitic melts with water content of 3% or more (Sparks 1978; Heiken and Wohletz 1985). Probably, early vesicle growth began at phenocryst surfaces, when those surfaces acted as bubble nuclei. The larger pockets with thin internal vesicle walls had a higher bulk viscosity than the surrounding melt and resisted deformation by flow (Heiken and Wohletz 1985). However, as magma rose, vesiculation began around the pockets. Closer to the surface, later stage vesicles were sheared into highly elongate, tube-like forms. At or near the surface, this heterogeneous, vesicular mass was fragmented, perhaps by an early pressure wave passed down into the vent and subsequently by water/magma interaction.

Due to heterogeneous vesicle types, classification of pyroclast surfaces according to their vesicle shapes was difficult. However, beside the cavities on surfaces, the cavity masking side of dust covering the surfaces was also highlighted by Rsk parameter. The surfaces having positive Rsk values probably have tapered ends of vesicles and these ends were perceived as peaks on surfaces. Furthermore, surfaces with masked vesicles or few vesicle surfaces caused negative Rsk values. Ra parameter confirmed adhered fine particle (dust) abundance and ran true to form by assigning micro-roughness on surfaces.

In the light of all the above-mentioned findings, the following scenario was proposed for the eruption of Dikkartın dome: The main phase deposits designate an initial magmatic Plinian-type eruption column with c. 20 km height and with high discharge rates, which widened the vent and allowed water access to the magma (Fig. 17a). Magma-water interaction caused intense fragmentation of the magma thus, the eruption column became greatly overloaded with fine particles and/or water in the column reduced the ability of the column to convect. Partial collapses of the Plinian column occurred (D2 and D3 pumiceous flows) however, the complete column collapse did not fully occur. The height of the column was decreased and the column shifted to a phreatoplinian column (Fig. 17b). The lateral or partial collapses of the phreatoplinian column generated laterally moving base surges interstratified with fall deposits (Fig. 17c). Depletion of water in the eruption environment and loss of volatiles in the magma terminated the explosion stage with a dry explosion (D9) followed by passive lava extrusion (Fig. 17d, e).
Fig. 17

Qualitative sketch illustrating eruption sequence of Dikkartın. a Opening eruption of Dikkartın involved Plinian column and widespread deposits. b The following eruption column is shorter and involves phreatomagmatic components due to interaction of water and rhyodacitic magma. The partial collapses occurred from margins of the column generating pumiceous flows with features transitional between those of conventional pyroclastic surges and pyroclastic flows. c Phreatoplinian column generates surge and fall deposits indicating amount of water interacting with magma. Depletion of water in the eruption environment and loss of volatiles in the magma terminated the explosion stage with a dry explosion (d) followed by passive lava extrusion (e)

Tephra dispersal scenarios

Some marine tephra layers found in the Mediterranean (Hamann et al. 2010) and in Black Sea cores (Cullen et al. 2014) have been attributed to Holocene aged Dikkartın and Perikartın or Karagüllü eruptive sequences of Mt. Erciyes. Cullen et al. (2014) compared the chemical analysis of their tephra with tephra of those three eruptive sequences, and confirm the southward dispersal of Dikkartın tephra toward the Mediterranean basin. Besides, they claim that Perikartın or Karagüllü tephra are present as cryptotephra in their Black Sea marine core proving northward dispersal of related eruption cloud. Since the wind direction has a major effect on the tephra distribution, the TephraProb model of Biass et al. (2016) was applied under present day wind conditions (source NOAA), assuming that the wind regime was quite similar to that of present day. We modelled tephra dispersal for different plume height and used the winds of whole 2017 year (Fig. 18) for a possible scenario of Mt. Erciyes related eruption. We can conclude that should the plume was high, c. 20 km, the prevailing winds might have directed the cloud towards south. In that case, low altitude plumes would have been directed towards northeast. The tephra accumulation and probability patterns are coherent with our Dikkartın isopleths. Besides, we demonstrate here that tephra dispersal is under the control of seasonal winds (Fig. 19), and that Dikkartın eruption must have been occurred during winter. On the other hand, Perikartın/Karagüllü tephra were dispersed under dry season winds, basing on the wind conditions of years of 2017.
Fig. 18

Probability of Tephra Accumulation for different plume heights on Google Earth Image. NOAA wind data for the year of 2017 have been used

Fig. 19

a Wind velocity and wind directions versus atmospheric height on Mt. Erciyes, b rose diagram of the wind directions, c dry season plume dispersal possibilities for 20–30 km plume height and the place of Black Sea marine core, d wet season plume dispersal possibilities for 20–30 km plume height and the place of the Mediterranean marine core


Dikkartın is a monogenetic lava dome emplaced on the southern flank of Erciyes stratovolcano (3917 m) in Cappadocia, central Turkey. After stratigraphic, granulometric, morphometric and textural analysis on the eruption products of Dikkartın lava dome, we suggest four eruptive phases. The eruption begins with a dry, purely magmatic Plinian eruption with a climactic column height of c. 20 km (Phase 1). During Phase 2, which is dominated by phreatomagmatism, an explosion crater, a phreatoplinian eruption column and a tuff ring were formed. Depletion of water in the eruption environment and loss of volatiles in the magma terminated the explosion stage with a dry explosion (Phase 3) followed by passive lava extrusion (Phase 4).

Contrasting fragmentation and transportation dynamics in the stratigraphic sequence of the Dikkartın eruption is related to the ratio of magma and external water during each phase. The vesiculation of the magma probably occurred earlier than the water interaction and thereby facilitated efficient fragmentation of the melt. The heterogeneous vesicle shapes and ineffectiveness of water interaction on vesicle development make classification of pyroclasts difficult on the surface quantification basis. However, pyroclast surfaces were quantified according to their micro- and macro-roughness. The unaltered surfaces of juvenile particles suggest that most of the phreatomagmatic eruptions producing the tuff ring were dry with superheated steam but little vapour condensation occurred allowing adhesion of fine material on surfaces.

Considering the c. 20 km plume height calculated for the climactic eruption column (Phase 1), the prevailing winds directed the cloud to the south, but due to the decrease in the plume height (Phase 2), the cloud was later directed to the northeast. This scenario was also supported by isopleths generated for the Plinian deposits of Dikkartın eruption. According to the direction of seasonal winds, it is suggested that Dikkartın erupted during winter.



We thank to the anonymous reviewers and editor Attila Çiner for their help that substantially improved this paper.

Compliance with ethical standards

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.


  1. Allen SR, Cas RAF (1998) Rhyolitic fallout and pyroclastic density current deposits from a phreatoplinian eruption in the eastern Aegean Sea. Greece. J Volcanol Geotherm Res 86(1–4):219–251CrossRefGoogle Scholar
  2. Aydar E, Schmitt AK, Cubukcu HE, Akın L, Ersoy O, Şen E, Duncan RA, Atıcı G (2012) Correlation of ignimbrites in the central Anatolian volcanic province using zircon and plagioclase ages and zircon compositions. J Volcanol Geotherm Res 213–214:83–97CrossRefGoogle Scholar
  3. Aydar E, Sen E, Sarikaya MA, Kuzucuoglu C (2019) In the footprint of Strabon: Mount Erciyes Volcano: The Roof of Central Anatolia and Sultansazlığı basin. In: Kuzucuoğlu et al. (eds) Landscapes and Landforms of Turkey Springer-Verlag Book Serie, 565–576Google Scholar
  4. Biass S, Bonadonna C, Connor L, Conor C (2016) TephraProb: a Matlab package for probabilistic hazard assessments of tephra fallout. J Appl Volcanol 5:10. CrossRefGoogle Scholar
  5. Brouste A, Renard F, Gratier JP, Schmittbuhl J (2007) Variety of stylolites morphologies and statistical characterization of the amount of heterogeneities in the rock. J Struct Geol 29:422–434CrossRefGoogle Scholar
  6. Carey S, Sparks RJS (1986) Quantitative models of the fallout and dispersal of tephra from volcanic eruption columns. Bull Volc 48:109–125CrossRefGoogle Scholar
  7. Chinga G, Gregersen Ø, Dougherty R (2003) Paper surface characterization by laser profilometry and image analysis. Microsc Anal 84:5–7Google Scholar
  8. Chinga G, Johnssen PO, Dougherty R, Lunden-Berli E, Walter J (2007) Quantification of the 3-D micro-structure of SC surfaces. J Microscopy 227(3):254–265CrossRefGoogle Scholar
  9. Cullen VL, Smith VC, Arz HW (2014) The detailed tephrostratigraphy of a core from the south-east Black Sea spanning the last ∼ 60 ka. J Quat Sci 29(7):675–690CrossRefGoogle Scholar
  10. Dellino P, Isaia R, La Volpe L, Orsi G (2004) Interaction between particles transported by fallout and surge in the deposits of the Agnano-Monte Spina eruption (Campi Flegrei, Southern Italy). J Volcanol Geotherm Res 133:193–210CrossRefGoogle Scholar
  11. Ersoy O, Chinga G, Aydar E, Gourgaud A, Cubukcu HE, Ulusoy I (2006) Texture discrimination of volcanic ashes from different fragmentation mechanisms: a case study, Mount Nemrut stratovolcano, eastern Turkey. Comput Geosci 32:936–946CrossRefGoogle Scholar
  12. Ersoy O, Gourgaud A, Aydar E, Chinga G, Thouret JC (2007) Quantitative SEM analysis of volcanic ash surfaces: application to the 1982-83 Galunggung eruption (Indonesia). Geol Soc Am Bull 119(5):743–752CrossRefGoogle Scholar
  13. Ersoy O, Aydar E, Gourgaud A, Bayhan H (2008) Quantitative analysis on volcanic ash surfaces: application of extended depth-of-field (focus) algorithm for light and scanning electron microscopy and 3D reconstruction. Micron 39:128–136CrossRefGoogle Scholar
  14. Fisher RV, Schmincke HU (1984) Pyroclastic Rocks. Springer-Verlag, Berlin, p 472CrossRefGoogle Scholar
  15. Hamann Y, Wulf S, Ersoy O, Ehrmann W, Aydar E, Schmiedl G (2010) First evidence of a distal early Holocene ash layer in Eastern Mediterranean deep-sea sediments derived from the Anatolian volcanic centers. Quat Res 73:497–506CrossRefGoogle Scholar
  16. Heiken G, Wohletz KH (1985) Volcanic Ash. University of California Press, Berkeley, California, p 245Google Scholar
  17. Houghton BF, Wilson CJN (1989) A vesicularity index for pyroclastic deposits. Bull Volc 51:451–462CrossRefGoogle Scholar
  18. Inman DL (1952) Measures for describing the size distribution of sediments. J Sediment Petrol 22:125–145Google Scholar
  19. Martel C, Dingwell DB, Spieler O, Pichavant M, Wilke M (2000) Fragmentation of foamed silicic melts: an experimental study. Earth Planet Sci Lett 178:47–58CrossRefGoogle Scholar
  20. Peckover RS, Buchanan DJ, Ashby DETF (1973) Fuel-coolant interactions in submarine volcanicm. Nature 245:307–308CrossRefGoogle Scholar
  21. Rasband WS (1997) ImageJ: National Institutes of Health. Bethesda, Maryland, USA.
  22. Sarıkaya MA, Çiner A, Zreda M, Şen E, Ersoy O (2019) Chlorine degassing constrained by cosmogenic 36Cl and radiocarbon dating of early Holocene rhyodacitic lava domes on Erciyes stratovolcano, central Turkey. J Volcanol Geotherm Res 369:263–275CrossRefGoogle Scholar
  23. Self S (1983) Large-scale phreatomagmatic silisic volcanicm: a case study from New Zealand. J Volcanol Geotherm Res 17:433–469CrossRefGoogle Scholar
  24. Self S, Sparks RSJ (1978) Characteristics of widespread pyroclastic deposits formed from the interaction of silisic magma and water. Bull Volcanol 41(3):196–212CrossRefGoogle Scholar
  25. Şen E, Aydar E, Gourgaud A, Kürkçüoğlu B (2002) La phase explosive précédant l’extrusion des domes volcaniques: exemple du dome rhyodacitique de Dikkartin Dag, Erciyes, Anatolie centrale, Turquie (Initial explosive phases during extrusion of volcanic lava domes: example from rhyodacitic dome of Dikkartin Dag, Erciyes stratovolcano, Central Anatolia, Turkey). C R Geosci 334(1):27–33CrossRefGoogle Scholar
  26. Şen E, Kürkçüoğlu B, Aydar E, Gourgaud A, Vincent PM (2003) Volcanological evolution of Mount Erciyes stratovolcano and origin of Valibaba Tepe ignimbrites (Central Anatolia, Turkey). J Volcanol Geotherm Res 125:225–246CrossRefGoogle Scholar
  27. Sheridan MF, Wohletz KH (1983) Hydrovolcanism: basic considerations and review. J Volcanol Geotherm Res 17:1–29CrossRefGoogle Scholar
  28. Sheridan MF, Barberi F, Rosi M, Santacroce R (1981) A model for Plinian eruptions of Vesuvius. Nature 289:282–285CrossRefGoogle Scholar
  29. Sparks RSJ (1978) The dynamics of bubble formation and growth in magmas: a review and analysis. J Volcanol Geotherm Res 3:1–38CrossRefGoogle Scholar
  30. Sparks RSJ, Wilson L (1976) A model for the formation of ignimbrite by gravitational column collapse. J Geol Soc London 132:411–452CrossRefGoogle Scholar
  31. Sparks RSJ, Wilson L, Hulme G (1978) Theoretical modeling of the generation, movement and emplacement of pyroclastic flows by column collapse. J Geophy Res 83:1727–1739CrossRefGoogle Scholar
  32. Sparks RSJ, Wilson L, Sigurdsson H (1981) The pyroclastic deposits of the 1875 eruption of Askja, Iceland. Philos Trans R Soc Lond. Series A, Mathematical and Physical Sciences 299 (1447):241-273Google Scholar
  33. Walker GPL (1971) Grain-size characteristics of pyroclastic deposits. J Geol 79:696–714CrossRefGoogle Scholar
  34. Walker GP, Croasdale R (1972) Characteristics of some basaltic pyroclastics. Bull Volc 35:303–317CrossRefGoogle Scholar
  35. Wilson L, Sparks RSJ, Huang TC, Watkins ND (1978) The control of volcanic column heights by eruption energetics and dynamics. J Geophys Res 83:1829–1836CrossRefGoogle Scholar
  36. Wilson L, Sparks RSJ, Walker GPL (1981) Explosive volcanic eruptions-IV. The control of magma properties and conduit geometry on eruption column behavior. Geophys. J. R. Astron. Soc. 63:117–148CrossRefGoogle Scholar
  37. Wohletz KH et al (1983) Mechanisms of hydrovolcanic pyroclast formation: size, scanning electron microscopy, and experimental studies, in Sheridan MF et al. eds: explosive Volcanism. J Volcanol Geotherm Res 17:31–63CrossRefGoogle Scholar
  38. Wohletz KH, McQueen RG (1984) Volcanic and stratospheric dust-like particles produced by experimental water-melt interactions. Geology 12:591–594CrossRefGoogle Scholar
  39. Wohletz KH, Sheridan MF (1983) Hydrovolcanic explosions II. Evolution of basaltic tuff rings and tuff cones. Am J Sci 283:385–413CrossRefGoogle Scholar
  40. Wohletz KH, Sheridan MF, Brown WK (1989) Particle size distributions and the sequential fragmentation/transport theory applied to volcanic ash. J Geophys Res 94:15703–15721CrossRefGoogle Scholar
  41. Wohletz KH, Orsi G, de Vita S (1995) Eruptive mechanisms of the Neapolitan Yellow Tuff interpreted from stratigraphy, chemistry, and granulometry. J Volcanol Geotherm Res 67:263–290CrossRefGoogle Scholar
  42. Woods AW, Wohletz KH (1991) Co-ignimbrite eruption columns. Nature 350:225–227CrossRefGoogle Scholar
  43. Yin L, Yang R, Gabbouj M, Neuvo Y (1996) Weighted Median Filters: a Tutorial. IEEE T CIRCUITS-I 43(3):157–192Google Scholar
  44. Yurteri E (2018) Petrological investigation of dacitic domes, dome-flows and their related pyroclastic deposits of the Erciyes Stratovolcano. Hacettepe University, Institute of Science, MSc thesis, 152pGoogle Scholar
  45. Zimanowski B, Frölich G, Volker L (1991) Quantitative experiments in phreatomagmatic explosions. J Volcanol Geotherm Res 48(3–4):341–358CrossRefGoogle Scholar
  46. Zimanowski B, Wohletz K, Dellino P, Büttner R (2003) The volcanic ash problem. J Volcanol Geotherm Res 122:1–5CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Department of Geological EngineeringHacettepe UniversityAnkaraTurkey
  2. 2.Université Clermont Auvergne, LMVAubièreFrance

Personalised recommendations