Sedimentation of long-lasting wind-affected volcanic plumes: the example of the 2011 rhyolitic Cordón Caulle eruption, Chile

Abstract

Sedimentation processes and fragmentation mechanisms during explosive volcanic eruptions can be constrained based on detailed analysis of grain-size variations of tephra deposits with distance from vent and total grain-size distribution (TGSD). Grain-size studies strongly rely on deposit exposure and, in case of long-lasting eruptions, can be complicated by the intricate interplay between eruptive style, atmospheric conditions, particle accumulation, and deposit erosion. The 2011 Cordón Caulle eruption, Chile, represents an ideals laboratory for the study of long-lasting eruptions thanks to the good deposit accessibility in medial to distal area. All layers analyzed are mostly characterized by bimodal grain-size distributions, with both the modes and the fraction of the coarse subpopulation decreasing rapidly with distance from vent and those of the fine subpopulation being mostly stable. Due to gradually changing wind direction, the two subpopulations characterizing the deposit of the first 2 days of the eruption are asymmetrically distributed with respect to the dispersal axis. The TGSD of the climactic phase is also bimodal, with the coarse subpopulation representing 90 wt% of the whole distribution. Polymodality of individual samples is related to size-selective sedimentation processes, while polymodality of the TGSD is mostly related to the complex internal texture (e.g., size and shape of vesicles) of the most abundant juvenile clasts. The most representative TGSD could be derived based on a combination of the Voronoi tessellation with a detailed analysis of the thinning trend of individual size categories. Finally, preferential breakage of coarse pumices on ground impact was inferred from the study of particle terminal velocity.

Appendices

Appendix 1

Isomass maps of individual ϕ categories for unit I (cumulative A to F layers) in kg m^-2

Appendix 2

Calculation of particle terminal velocity

Particle terminal velocity was calculated with the equation of Ganser (1993) for the mean diameter within each half-ϕ size bin (assuming a Gaussian size distribution within each size bin) and averaged over sedimentation height (1 to 14 km a.s.l. for A–F, 1 to 12 km a.s.l. for H, and 1 to 10 km a.s.l. for K2; maximum plume heights were considered for all phases). Sphericity was kept at 0.9 for both clast categories (e.g., Folch 2012). Particle density was measured for individual clast sizes by Pistolesi et al. (2015) (see Fig. 9b, e, h). In Fig. 11, we show results for individual layers A to F.

Fig. 11

Componentry (left column) and terminal velocity (right column) of tephra clasts ranging from 0 to −4.5ϕ (i.e., 1–22.6 mm) for layers A to F at a locality about 15 km from vent on the dispersal axis for unit I. See also Fig. 9 for results of the other layers

Appendix 3

Calculation of the total grain-size distribution of unit I

In order to account for the missing distal data in the derivation of the total grain size distribution of the A–F cumulative layer (i.e., unit I), we have derived the mass/area of six points about 600 km from vent from the isopach maps compiled soon after the eruption by Gáitan et al. (2011) and the . Based on these two maps that also describe the distal deposit, we assigned to the six distal data a thickness of 0.1 cm (i.e., 0.8 kg m⁻² based on our most distal value of deposit density, i.e., 834 kg m⁻³; Fig. 5) and the grain size shown in Fig. 12. The two maps (based on dataset 1 and 2) are compared in Fig. 13.

Fig. 12

Grain size distribution derived for the distal points (610 km from vent; blue histograms) based on the distribution of our most distal sample at 240 km from vent (red histograms) and on the proportion of lapilli, coarse ash, and fine ash as derived from the best-fit equations in Fig. 3a, main text (i.e., Gaussian distribution with mode of 7ϕ and composed of 99 % fine ash)

Fig. 13

Maps showing the Voronoi triangulation and the zero line used for the determination of the total grain size distribution based on the Voronoi tessellation method of Bonadonna and Houghton (2005) and using the application of Biass and Bonadonna (2014) for: a dataset 1 (i.e., samples collected in our field surveys) and b dataset 2 (dataset 1 combined with six distal points derived from the isopach maps compiled soon after the eruption by Gaitán et al. (2011) and the INTA (2011))

Fig. 14

Plot of Log(erupted mass) vs. plume height (km) (above sea level) showing the minimum values (dark blue) of the goodness-of-fit measure (root mean square error, RMSE) associated with individual ϕ size categories (indicated in each plot) of the climactic phase of Córdon Caulle eruption (i.e., unit I)

Fig. 15

Best-fit plume height above sea level as derived from inversion on individual grain size categories (average value = 13.9 km). The mass along the eruptive plume in the model TEPHRA2 is described by a beta function characterized by two parameters, α and β (Connor et al. 2015, in press). For simplicity, in these inversion calculations, α is kept constant (i.e., 3) and β is left vary between 0.5 (i.e., particles are mostly released at the top of the cloud) and 2 (i.e., particles are released at a height which is 60–70 % the total height). Average best-fit value for the climactic phase of Córdon Caulle eruption is 1.8 (i.e., particles are mostly released at a height which is 70 % the total height)

Appendix 4

Inversion on grain size using the model TEPHRA2