Cohesional behaviours in volcanic material and the implications for deposit architecture

Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 1 2 Cohesional behaviours in volcanic material and the implications for deposit architecture. 3 1 Nemi Walding, 1 Rebecca Williams, 2 Pete Rowley, 3 Natasha Dowey 4 1 School of Environmental Sciences, Energy and Environment Institute. 5 University of Hull, 6 Cottingham Road, Hull. HU6 7RX 7 2 University of Bristol 8 3 Sheffield Hallam University 9 Email: n.walding-2021@hull.ac.uk 10 ORCID: 11 N. Walding (0000-0002-4584-1319) 12 R. Williams (0000-0002-8203-9982) 13 P. Rowley – (0000-0002-8322-5808) 14 N. Dowey – (0000-0002-9231-4781) 15 16 17 Acknowledgements: N.W was supported by the EU Horizon 2020 Programme (Project GEOSTICK 712525). We thank Gilbertson et al., (2020) at the University of Bristol for the use of the fluidisation chamber. Ulrich Küppers is thanked for suppling the volcanic material used in these experiments. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 18 Abstract 19 Pyroclastic Density Currents (PDCs) are hazardous, multiphase currents of heterogeneous volcanic material 20 and gas. Their high mobility can be partially attributed to fluidisation mechanisms. Moisture (as liquid or gas) 21 can enter a PDC through external (e.g., interaction with bodies of water) or internal (e.g., initial eruptive 22 activity style) processes and presence of moisture can be recorded within distinct deposit layers. We use 23 analogue experiments to explore the behaviour of volcanic material with increasing moisture percentages 24 from 0.00 – 10.00%. Our results show that: 1) the cohesivity of ignimbrite material changes with the addition 25 of small amounts of moisture; 2) Small increases in moisture content change the flow behaviour from a free- 26 flowing material to a non-flowable material; 3) changes in moisture can affect the formation of gas escape 27 structures, and fluidisation profiles, 4) gas flow through a deposit can lead to a moisture profile and resulting 28 mechanical heterogeneity within the deposit and 5) where gas escape structure growth is hindered by 29 cohesivity driven by moisture, pressure can increase and release in an explosive fashion. This work highlights 30 how a suite of dynamic and varied gas escape morphologies can form within the deposit resulting from 31 moisture content heterogeneity, explaining variation in gas escape structures as well as providing a potential 32 mechanism for secondary eruptions. 33 Key Words: Cohesion, Gas escape, Fluidisation, Secondary eruptions, Pyroclastic Density Currents, Deposit Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 34 Introduction 35 Pyroclastic density currents (PDC) are hazardous, multiphase, rapidly moving, often high-temperature currents 36 of heterogeneous volcanic material and gas. The high mobility of PDCs has been attributed to fluidisation 37 (Sparks, 1976; 1978; Wilson, 1984; Branney and Kokelaar, 1992, 2002; Breard et al., 2023) where the upward 38 movement of gas supersedes the force of gravity and supports the flow (Sparks, 1976; Branney and Kokelaar, 39 2002; Cocco et al., 2014). The ability of a current to fluidise and flow can be described by its ‘flowability’, which 40 depends upon interparticle forces (Van der Waals, electrostatic or capillary forces). These forces can be 41 influenced by bulk composition and material physical properties such as particle size, density, shape, and 42 moisture content (Rios., 2006; Leturia et al., 2014). 43 Fluidisation in PDCs can be initiated during formation and maintained throughout the course of the flow by 44 continued fluidisation mechanisms such as substrate evaporation (i.e., steam generated from interaction with 45 surfaces with moisture content or bodies of water), bulk self-fluidisation or ambient air entrainment (Sparks, 46 1978; Branney and Kokelaar, 2002; Chedeville and Roche, 2015; Breard et al., 2023). Sedimentation fluidisation 47 (or, hindered settling) and particle-self fluidisation is where compaction from particle settling and sorting 48 causes interstitial fluid movement. On or after deposition, the material will defluidise which can lead to 49 segregation of material through gas escape structures, which form concentrated areas of particle sorting (i.e., 50 fines depleted elutriation pipes) and segregation of particles (Wilson, 1980; Cas and Wright, 1991). 51 Previous analogue investigations into fluidisation behaviours of volcanic material and segregations structures 52 have been completed on dry (0%; Wilson 1980; 1984) and saturated (80+-15%; Roche et al., 2001) natural 53 ignimbrite (herein defined as the deposits of PDCs, typically consisting of poorly-sorted ash and pumice or 54 scoria (Giordano & Cas 2021). Experiments completed by Wilson (1980; 1984) used non-cohesive, poorly- 55 sorted ignimbrite mixtures, and added an influx of gas into the deposit. This resulted in poor fluidisation 56 behaviours and the formation of gas escape structures determined by particle size and density. Roche et al 57 (2001) explored contrasting aqueous fluidisation in a water-saturated deposit of volcanic material. These 58 results determined that fluid-escape pipes form readily with low water flux and contain localised segregation 59 of particle sizes and densities. From both experiments, we can determine that natural ignimbrite material will 60 demonstrate aggregative (i.e., inhomogeneous; Branney and Kokelaar, 2002) fluidisation resulting from 61 particle size and density range, irrespective of the fluidising medium. 62 Understanding how moisture impacts powder material has important industrial applications. Experiments 63 have explored fluidisation behaviours of industrial material with the addition of small volumes of moisture 64 (i.e., through adsorption of ambient humidity). With the introduction of moisture into a material, Van der 65 Waals forces are no longer dominant, and liquid bridges connect particles through capillary cohesion; resulting 66 in poor fluidisation behaviours (Wormsbecker and Pugsley, 2008; Ludwig et al., 2020; Yehuda and Kalman, 67 2020). A study by Wormsbecker and Pugsley (2008) looked at gas fluidisation behaviours on a semi-saturated 68 (30, 20, 15 and 5 wt.% moisture) powder. Results showed a dominant change in fluidisation behaviour Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 69 associated with the addition of moisture, which were observed in conjunction with the drying states of the 70 material from 30 to 5 wt.%. 71 The experiments detailed herein experimentally assess the impact of the addition of small volumes of 72 moisture within natural volcanic material. We identify the resulting variations in terms of fluidisation and 73 particle segregation behaviours. 74 Moisture in PDCs and their resulting deposits 75 Moisture (i.e., water vapour or liquid water) can enter a PDC system during formation at source or as they 76 propagate (Error! Reference source not found.). Eruption columns can be water-rich due to 77 phreatomagmatism (Self and Sparks, 1978; Hurwitz, 2003; Houghton et al., 2015) or atmospheric conditions 78 (Vecino et al., 2022). During transport, internal clasts of juvenile magma will exsolve and release water vapour 79 and other volatiles. Experiments have highlighted how magmatic clasts may still hold residual water content of 80 0.6 – 0.8 wt. % during transport (Sparks et al., 1978). 81 82 Fig.1 a PDC interacting with sources of moisture across a landscape which have the potential to enter the PDC 83 system and resulting deposits. 84 Externally, moisture may be introduced through a dynamic mix of atmospheric (e.g., humidity; Pepin et al., 85 2017; Camuffo, 2019), topographic (e.g., height; Barclay et al., 2006; Duane et al., 2008; Hartman, 2016), 86 climatic (e.g., global location; Barclay et al., 2006) and meteorological conditions (e.g., precipitation). 87 Furthermore, periods of intense rainfall have been suspected and observed to affect the onset of volcanic 88 activity (Barclay et al., 2006; Sahoo et al., 2022 and references therein). Matthews et al., (2009) documented 89 that within 24 hours of heavy rainfall, the probability of lava dome collapse at Soufriere Hills Volcano, 90 Montserrat (during the period 1998-2003) increased, resulting in higher moisture availability to the resulting 91 PDCs. 92 Interaction with external bodies of water (i.e., streams, lakes, sea, snow; Dartevelle et al., 2002; Cole et al., 93 1998; 2002), water saturated substrate (Moyer and Swanson, 1987; Brown and Branney, 2013; Gilbertson et Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 94 al., 2020) or by the incorporation of vegetation (as observed in Mount Pelé, 1902; Mount St Helens, 1980; 95 Montserrat, 2002 and Fuego Volcano, 2018 ) can also contribute to moisture within the flow system. 96 Therefore, we expect moisture content in PDCs to be highly variable in time and space. 97 The presence of moisture within PDCs can be demonstrated through varying degrees within an ignimbrite 98 deposit. Moisture has been linked to the formation of wet ash aggregates (e.g., pellets) in ignimbrite deposits 99 (Brown et al., 2010), by the presence of elutriation pipes that are derived from areas of evaporating moisture 100 (i.e., vegetation or water-laden sediments) or by secondary hydroeruptions forming in deposits overlying 101 moisture-rich areas (e.g., Mount St. Helens; Moyer and Swanson, 1987). The influence of these relatively small 102 additions of moisture into a PDC system has been largely ignored in analogue and experimental studies, due to 103 the difficulty associated using and controlling the characteristics of cohesive material. Therefore, prior to the 104 experiments, material is often dried to remove any residual moisture (Druitt et al., 2004, 2007; Girolami et al., 105 2008; 2015). 106 Capillary Cohesion 107 The presence of moisture in a PDC, or in a subsequent deposit, will result in cohesional forces within the 108 interparticulate space. A PDC can reach temperatures > 1000°C and the resulting deposit can maintain high 109 temperatures for extended periods of time (Dufek, 2016; Riehle et al., 1995), and it has been assumed that at 110 these temperatures the dominant cohesive forces will be electrostatic and Van der Waals forces (Branney and 111 Kokelaar, 2002). However, with increasing distance and entrainment, temperatures will lower (Benage et al., 112 2016) and the introduction of moisture will likely lead to the formation of capillary bridges (‘capillary 113 condensation’; Ma et al., 2019), resulting in a change of the dominant interparticulate forces . This is observed 114 by Telling et al., (2013), where electrostatic attraction has been observed to be dominant only where humidity 115 was lower than 71% and by Chigira and Yokoyama (2005), where capillary cohesion became the dominant 116 cohesive force with the addition of moisture into a granular material. 117 Previous studies have shown that an increase in water content and moisture leads to a drastic change in the 118 physical properties of a bulk material. For example, in sands, capillary forces were seen to affect the tensile 119 strength of a material until reaching a water saturated state (Kim and Sture, 2008; Chen et al., 2021). 120 Therefore, at lower temperatures, it is highly likely that the introduction of moisture into the dynamic 121 (pyroclastic current) and static (deposited sedimentary packages from a current) regions will introduce 122 variations in material properties (e.g., as a sedimentary package). Changes in tensile strength may determine 123 how resistant a material is to shear and erode and are important in understanding the flow properties of a 124 material (Pierrat and Caram, 1997; LaMarche et al., 2016). Within a PDC deposit, such changes may also 125 influence defluidisation through gas escape. 126 Here, we investigate how capillary cohesion, through the introduction of water, may affect volcanic material in 127 a static state. We present the results of analogue experiments that test how changes in cohesion of volcanic Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 128 material affects its fluidisation. Our results provide new and novel insights into the variation of gas escape 129 behaviours in a defluidising PDC deposit. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 130 Methodology 131 Source material and Sample Preparation 132 A range of unconsolidated material collected in 2009 from the 2006 Tungurahua, Ecuador, eruptions (provided 133 by U. Küppers, LMU Munich) has been subjected to a range of characterisation tests to elucidate flowability 134 properties and variations with moisture content. 135 The ignimbrite material is dark grey/brown and andesitic in chemistry (Eychenne et al., 2012) and samples 136 were sieved into varying size fractions, with specific particle size distributions created for the suite of analysis. 137 All samples were dried in an 80°C oven for 24 hours to ensure the removal of residual and adsorbed moisture 138 and agglomerations were broken-up by sieving prior to addition of water. For the series of characterisation 139 tests, water was added to the samples based on weight percentage (0.00, 0.25, 0.50, 1.00, 2.50, 5.00, 7.50, 140 10.00 %). Samples were stirred thoroughly to ensure a homogeneous moisture distribution. 141 Material characterisation and cohesive behaviour tests 142 Particle Size Analysis 143 Particle analysis of the ignimbrite material was undertaken using a CAMSIZER X2. This uses particle imaging to 144 build particle shape and size characteristics for dry samples. Particles were sieved prior to using the CAMSIZER 145 with samples <1000 µm were used. Any results from the CAMSIZER erroneously returned as >1460 µm were 146 removed. 147 Geldart’s Classification of Powders 148 Geldart (1973) classified powders into four distinctive groups (A-D) defined by fluidisation behaviours, which 149 vary dependent on particle size and density; resultant flowability properties are described as ‘very poor’ to 150 ‘excellent’. Group C, the finest material (< 20 µm), is dominated by interparticulate forces. Group D (>1 mm) 151 requires an increased gas velocity to fluidise. Group C and D present passable to very poor fluidisation 152 behaviours and would typically express as slugging, channelling, and spouting behaviours (Leturia et al., 2014). 153 Group A (30 – 100 µm) and B (100 µm – 1mm) powders show the best fluidisation behaviours overall and are 154 most likely to have good flowability, typically expanding under fluidisation. 155 Volcanic materials used in these experiments (Fig. 2) have particle size distributions from 2.5 to 1000 µm. and 156 should readily exhibit fluidisation behaviours typically of Groups A and B in Geldart’s classification. 157 Bulk and Tapped Density 158 Bulk and tapped density measurements describe the mass and volume ratio of a material, without and with 159 packing respectively (Amidon et al., 2017). Tapped density experiments remove interparticulate voids. The 160 differences within the bulk and tapped density measurements correspond to the cohesive properties of the 161 particles (Deb et al., 2018) and can be affected by shape and size of material (Amidon et al., 2017). Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 162 Bulk and tapped density were calculated for dry sample herein to characterize cohesive behaviour prior to the 163 addition of water (method adapted from USP: Bulk and Tapped Density of Powders, 2015). 164 Bulk density was obtained by pouring 100g of the volcanic material into a 250mL cylinder and levelling where 165 needed. The unsettled volume was measured, and bulk density calculated using Equation 1. This procedure 166 was completed three times per sample. 167 168 169 𝑚 = mass (g) 𝜌𝑏= 𝑚 [1] 𝑉0 𝑉𝑜 = unsettled apparent volume (mL) 170 The cylinder was tapped at 150 taps/min, with volume measured every minute until leveled. Using the 171 unsettled apparent volume and final tapped volume, the tapped density (Eq. 2), Carr’s Index (Eq. 3) and the 172 Hausner ratio (Eq. 4) could be calculated. 𝜌𝑡= 𝑚 173 174 175 176 [2] 𝑉𝑓 𝑚 = mass (g) 𝑉𝑓 = final tapped volume (mL) (Moondra et al., 2018) The Carr’s Index and Hausner Ratio are indicative of flowability and interparticulate behaviours (Hausner, 177 1981) and are a useful tool in determining a materials ability to fluidise and flow (Table 1) . 178 The Carr’s Index measures the strength and compressibility of a material (Equation 3; Moondra et al., 2018). 𝐶𝐼 = 100 ( 179 𝜌𝑡 −𝜌𝑏 𝜌𝑡 ) [3] 180 The Hausner Ratio determines how certain powders will behave i.e., flowability and fluidisation (Equation 4, Yu 181 and Hall, 1994; Abdullah and Geldart, 1999). 182 183 𝐻𝑅 = 𝜌𝑡 𝜌𝑏 184 185 186 Table 1 Relationship between Carr’s Compressibility Index, Hausner Ratio, and flowability behaviours. From (Gorle and Chopade, 2020). [4] Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology CI HR Flowability ≤10 1.00 – 1.11 Excellent 11 – 15 1.12 – 1.18 Good 16 – 20 1.19 – 1.25 Fair 21 – 25 1.26 – 1.34 Passable 26 – 31 1.35 – 1.45 Poor 32 – 37 1.46 – 1.59 Very Poor > 38 >1.60 Very Very Poor 187 188 Angle of Repose 189 The Angle of Repose (AoR) refers to the static friction coefficient and the angle of internal friction, and can be 190 investigated through static (funnel) and dynamic (rotating cylinder drum) methods (Al-Hashemi and Al- 191 Amoudi, 2018). AoR results are attributed to understanding the flowability of a material (Table 2). 192 Table 2 Flowability based on angle of repose results (Al-Hashemi and Al-Amoudi., 2018). Flowability Angle of Repose (°) Very free flowing <30 Free flowing 30 - 38 Fair to passable flow 38 - 45 Cohesive 45 - 55 Very Cohesive (non-flowing) >55 193 194 To determine the static angle of repose (SAoR) for each experiment, 100 g of material was released from a 195 funnel held 3.5 cm over a circular platform (Av diameter = 12 cm). The height of the cone was measured, and 196 the angle of repose calculated using Equation 5 (Al-Hashemi and Al-Amoudi, 2018). Where material did not 197 release freely from the funnel, material was lightly agitated. If the height of the cone reached the base of the Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 198 funnel, then the funnel was incrementally moved vertically to accommodate the growing cone. This was 199 repeated three times for each experiment. 200 201 202 203 ℎ = height (mm) 𝑆𝐴𝑜𝑅 (°) = 𝑡𝑎𝑛−1 2ℎ 𝐷 [6] 𝐷 = base diameter (mm) 204 Dynamic Angle of Repose (DAoR) was determined by rotating 100 g of material in a clear cylindrical drum at a 205 constant rate (Smith, 2020). This was recorded on video and critical angle (the maximum angle prior to 206 collapse) measurements analyzed using ImageJ (Schneider et al., 2012). This was completed three times. 207 Fluidisation Behaviour Tests 208 Experiments to determine the fluidisation behaviours of the volcanic material with increasing moisture 209 contents were completed using a rectangular, near-2D fluidisation chamber with a porous base (following 210 Gilbertson et al., 2020). Material (200.0 g) was measured, a weight % of water was added, and the 211 homogeneous sample was then placed into the chamber and carefully levelled. A manometer probe recorded 212 basal pore pressure changes during each experiment. Gas velocity (L/min) was increased incrementally until 213 either a stable, channelised bubbling fluidisation state was achieved, or large amounts of winnowing or 214 pressure build-up occurred. To limit the effects of drying from basal air flow, experiments were carried out 215 with gradual increases in gas flow rate (0.8 – 3.00 L/min/min for dry sediments and 6.80 – 11.00 L/min/min for 216 moisture added sediments) over a period of 01:11 – 23:51 minutes. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 217 Results 218 Particle Size Analysis 219 The experiments were completed using volcanic material (V1 – V6). All samples are moderately to very well 220 sorted (Fig. 2). Samples V1, V4, V5 and V6 were sieved into desired particle size distributions, whereas samples 221 V2 and V3 were kept as natural volcanic particle size distributions (ranging from >74 - 300 µm). Particle size 222 and sorting characterisation of each sample is presented in Table 3. As highlighted above, all samples fall into 223 groups A and B of Geldart’s classification, indicating that they should display excellent to fair flowability 224 (Geldart, 1973). 225 226 227 Table 3 Table showing particle size mean (logarithmic), particle size median (log), particle range, fines content, geometric mean, logarithmic (Φ) Method of Moments used for Mean, Sorting, Skewness, and Kurtosis. Sauter mean diameter calculated from Breard et al (2019). Geldart Group (1973) based on size of particle. 228 Particle Particle Particle Size Size Size Fines Material Mean ̅) ∅ (𝒙 Median ∅ Sauter Geometric Sorting Sorting Content Index (%) (𝝈) ∅ 35.76 0.428 Well 21.41 0.710 19.57 Range (µm) ( 𝝈𝑮 ) Skewness Kurtosis (𝑺𝒌) ∅ (𝑲) ∅ Geldart Mean Mean (mm) (µm) Group 2.5 – V1 3.776 3.734 297.3 1.891 17.00 0.07 72.93 A Moderate -0.524 2.576 0.11 107.5 A, B 0.868 Moderate -0.562 4.140 0.12 115.0 A, B 0.19 0.252 Very well 8.536 0.15 153.5 A, B 0.11 0.758 Moderate - 0.103 2.480 0.42 347.3 A, B 0.05 0.445 Well 8.277 0.73 557.1 A, B 15 V2 3.215 3.320 425 5- V3 3.118 3.140 1000 20 – V4 2.703 2.710 650 0.141 10 – V5 1.508 1.568 1000 10 - V6 0.833 0.812 1000 0.589 Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 229 230 Fig. 2 Particle size analysis of volcanic material. 231 Bulk/Tapped Density 232 Table 4 Loose and tapped bulk density, the Hausner ratio, Carr Index and Flowability. Loose Bulk Material Tapped Bulk Density -3 Density (kg m ) Hausner Ratio Carr Index Flowability -3 (kg m ) V1 1310 1420 1.08 7.73 Excellent V2 1320 1550 1.18 15.13 Good V3 1380 1610 1.17 14.28 Good V4 1320 1420 1.07 6.59 Excellent V5 1280 1440 1.12 10.68 Excellent Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology V6 1180 1370 1.15 13.37 Good 233 234 The bulk and tapped densities were calculated for volcanic samples ranging in sizes from finest (V1 – 3.8 ϕ) to 235 coarsest (V6 – 0.8 ϕ). With increasing particle size, bulk and tapped densities generally decrease (Table 4). 236 Material flowability, as determined by the Hausner Ratio and Carr Index, is good (V2, V3, V6) and excellent (V1, 237 V4, V5) under the 0% moisture conditions. The change in flowability between V5 and V6 likely reflects the 238 large increase in geometric mean from 347 (V5) to 557 (V6) (Table 3). V1, V4 and V5 all show excellent 239 flowability, presumed to be related to the smaller particle range in V1 and V4 and having a low fines content in 240 V5 (0.11%) (Table 3). 241 Angle of Repose 242 Static angle of repose (SAoR) increases with increasing across all volcanic samples (V1-6; Fig. 3). For the 0% 243 moisture condition the SAoR ranges from 21° (V2, V4, V5) to 23° (V1, V3). Interestingly, these results show that 244 under 0% moisture conditions, the SAoR is broadly similar (within 2°) regardless of particle size or sorting (Fig. 245 4a). 246 247 Fig. 3 Representative static angle of repose (SAoR) cone formation of V1 – V6. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 248 When increasing moisture contents to 5% the SAoR values increase to approximately double those achieved with 249 0% moisture, approaching 42° (V5, V6) to 47° (V4). However, this relationship is not linear with increasing 250 moisture content (Fig 4a). All materials show a rapid increase in SAoR with moisture to around 25°. However, 251 beyond a moisture content of 0.5% a division is evident between the finer and course mixtures; those with Sauter 252 mean diameters below 0.3 mm (V4, V3, V2) quickly increase to SAoR values of ~45° at moisture contents of 1-2 253 %, before plateauing and becoming invariant with additional moisture content. Mixtures with higher Sauter 254 diameters (V1, V5, V6) show a more gradual increase in SAoR with moisture content. V5, with a Sauter mean 255 diameter of 0.42 mm has somewhat intermediate behaviour, while V6 with a Sauter diameter of 0.73mm shows 256 a more linear relationship for SAoR with moisture between 0.5 – 5%. This indicates that SAoR shows distinct 257 sensitivity to increase in water in the materials and that relatively small weight percentages can produce very 258 different cohesivities within the mixtures It is notable that fines-rich mixtures are particularly sensitive to 259 moisture related cohesion, notably at <2 wt. %. 260 261 262 263 Fig. 4 a) SAoR for volcanic material with varying moisture percentages (0.00, 0.25, 0.50, 1.00, 2.50 and 5.00% with standard deviation error bars. b) DAoR critical angle of volcanic material with varying moisture percentages (0.00, 0.25, 0.50, 1.00 and 2.50%). 264 Figure 4 also shows the relationship pf Dynamic Angle of Repose experiments (Fig 4b). Generally, and similar to 265 the SAoR result, there is an increase in the DAoR with increasing moisture. However, in experiments with 266 increasing moisture levels (> 2.50%) the material was observed to clump, slide, and stick to the outer walls of 267 the drum, complicating the results. Nonetheless, it is important to observe that the Sauter mean relationships 268 detected within the SAoR experiments are not replicated in the DAoR setting. 269 Fluidisation experiments 270 Fluidisation behaviours were described via sidewall video analysis of the fluidisation chamber. The visual 271 observation of gas escape structures (i.e., bubbling, channeling, pocketing, explosive channeling, cracking (Fig. 272 5 a-e) and gas velocity measurements have been recorded at varying moisture levels (Fig. 6 a - f). Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 273 274 275 276 277 278 279 280 281 282 283 284 285 286 Fig. 5 a-e shows examples of the descriptive structures from across the experiments. 287 Bubbling gas escape (Fig. 5a) is seen initially in most experiments, where gas bubbles rise from the influx of gas 288 within the deposit. With increasing gas flux, this can lead to channeling, where material is sorted through 289 vertical channel or via pipe structures forming within the deposit (Fig. 5b). Drying profiles that migrate through 290 the deposit are shown in Fig. 5c and as this migrates with non-uniformity in the vertical deposit, formation of 291 areas of wet lobes and dry pockets can be observed (Fig. 5c). Where the material is dry and bubbling, this is 292 referred to as pocketing. Explosive channeling can also be observed in some experiments (Fig. 5d), as the 293 material dries, the upper wet deposit inhibits gas escape and causes a pressure increase and the subsequent 294 release (Online Resource 1). Finally, under the highest moisture contents, material does not form any of the 295 previous gas escape structures outlined above, instead, pressure builds until the deposit fractures into cracks 296 where gas can easily permeate through (Fig. 5e; Online Resource 2). 297 298 Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 299 300 Fig. 6 A-F. Fluidisation profiles of V1 – V6 with increasing moisture (0.00 – 10.00 %). Symbols show gas escape structure formation. 301 0.00% Moisture 302 At 0.00% moisture for samples with moderate sorting (i.e., V2, V3, V5), fine material escapes through gas 303 escape channels (Fig. 5b) in the lower area of the deposit. The observation of minimum bubbling (Umb) is first 304 seen in the upper fine fraction of the deposit at Umb 0.11 (V2), 0.08 (V3), and 0.42 (V5) cm/s. There is often a 305 separation of fines bubbling in the upper layer, a mid-area of coarse channeling (Umc) at 0.13 (V2) and 0.10 (V3) 306 cm/s as fines are being elutriated, and a coarse material layer at the base of the deposit. Bubbling only affects 307 the finer material. 308 In volcanic mixtures that are well to very well sorted (i.e., V1, V4) bubbles rise uniformly throughout the whole 309 deposit with a Umb of 0.07 (V1) and 0.19 (V4) cm/s. Within the more coarse, well sorted, material (V6) bubbles Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 310 in a sluggish motion from the base of the deposit, with mostly bubbling (Umb 1.60 cm/s) occurring in the upper 311 half of the deposit and channeling in the lower. This reflects the slight particle size variation of material used, 312 and therefore the Umb of the coarser material (Fig. 5b). 313 0.25% Moisture 314 At 0.25% moisture contents, similar behaviours are observed for V3 (Umb 0.069 cm/s), V5 (Umb 0.22 cm/s) and 315 V6 (Umb 1.25 cm/s) as described for 0.00% moisture. For V2 (Umb 0.15 cm/s), V1 (Umb 0.13 cm/s) and V4 (Umb 316 0.15 cm/s), bubbling begins at the base of the deposit. However, as the surrounding wet deposit begins to dry, 317 this dry material becomes incorporated into the bubbling deposit. In V2, we again see a separation of 318 channeling and bubbling in the lower and upper deposit. 319 0.50% Moisture 320 At 0.50% moisture, a drying profile can be observed throughout most of the deposit (V4, V5, V6). In the V4 321 sample, as drying at the base moves throughout the deposit, dry material begins to bubble (Umb 0.28 cm/s), 322 and pressure slowly increases. This is released suddenly (explosive channeling) at 0.54 cm/s through a large 323 channel which cuts through the moist, upper part of the deposit. As the surrounding wet material then begins 324 to dry, it then becomes incorporated into the bubbling deposit. In the V5 sample, the drying profile forms 325 lobes of wet material and pockets of dry material. The dry pockets slowly grow until reaching the upper 326 deposit and begin to bubble (Umb 1.04 cm/s). With continued drying as the experiment progresses, similar 327 behaviours to the 0.25% and 0.00 % moisture levels experiments are observed. After the drying profile has 328 moved through the deposit of V6, similar behaviours to 0.25% and 0.00% moisture levels are observed (Umb 329 1.60 cm/s). 330 For the V3 material, channels of coarser material begin to slowly move towards the surface. Material begins to 331 dry and is then incorporated into the bubbling deposit (Umb 0.14 (V3). 332 1.00% Moisture 333 At 1.00% moisture, V1, V2, and V4 show material at the base of the deposit drying in pockets. The dry material 334 begins to bubble (Umb 0.35 (V1), 0.49 (V2), 0.42 (V4) cm/s) and as the surrounding wet material begins to dry, it 335 is incorporated into the bubbling deposit. In V5 and V6, a distinctive drying profile moves throughout the 336 deposit. Again, this creates dry pockets of bubbling material (Umb 1.32 (V5), 1.81 (V6) cm/s) and wet lobes. In 337 V3, pressure slowly builds as gas velocity is increased. Pressure is suddenly released through the formation of 338 an explosive channel (0.35 cm/s). The dry deposit then begins to bubble (Umb 0.35 cm/s) and is slowly 339 incorporated into the surrounding drying material. 340 2.50% Moisture 341 At 2.50%, behaviours of V4 show similar results to 1.00% moisture content: as the base dries, bubbling pockets 342 are formed (Umb 0.70 cm/s) in-between lobes of wet material. In V4, pressure builds until it is suddenly 343 released through an explosive channel (2.15 cm/s). Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 344 5.00% Moisture 345 At 5.00% V2 shows the deposit drying at the base which forms drying and bubbling (Umb 0.35 cm/s) in pockets, 346 and wet lobes. 347 7.50% Moisture 348 At 7.50%, a clear drying profile forms through the V4 deposit, cracks begin to form and move through the 349 deposit until reaching the top and collapsing into pieces (3.82 cm/s). As gas was moving through cracks, there 350 was no dramatic rise and release in pressure. 351 10.00% Moisture 352 Finally, at 10.00%, V3 forms a clear drying profile within the deposit. Pressure builds before being released 353 suddenly at 3.82 cm/s. This forms a large crack in-between wet material. V6 shows a clear drying profile, as 354 pressure slowly rises as small pockets eventually form and dry material begins to bubble (4.17 cm/s). 355 356 Key Observations These fluidisation experiments clearly demonstrate how small additions of water into pyroclastic material can 357 greatly impact fluidisation behaviours and resulting gas escape structures of a defluidising volcanic deposit. 358 Two key observations are apparent in the experiments: 1) the drying profile, and 2) pressure build up and 359 release. 360 The dynamics of the drying profile, as the moisture content is impacted by the fluidising gas, exert a strong 361 control on the distribution of gas escape features, with variations across the grain size of the materials. 362 As gas flux is increased, a drying profile can move from the base to the top of the deposit. The drying profile 363 forms more easily within the coarser materials (V3 – V6). The profile initially rises uniformly across the bed, 364 before becoming irregular as it reaches the top of the deposit. These profiles are noted as they highlight 365 vertical and lateral moisture heterogeneity within the deposit and their irregular structure determines the 366 formation of drying pockets and wet lobes (Fig 5c). At low moisture percentages (< 2.50%) the drying pockets 367 bubble and the wet lobes begin to dry before being incorporated into the pockets. However, at high moisture 368 contents (> 2.5%) moisture rich lobes remain throughout the experiment, even at high gas velocities. This 369 shows that within a defluidising deposit, a drying profile will lead to lateral and vertical variations in moisture. 370 In experiments with moisture contents of 0.50 – 10.00%, explosive channelling (V3, V4) and cracking (V3, V4) 371 can occur. Across the experiments with 0.50 – 5.00% moisture contents, a wet impermeable cap was observed 372 to form above the drier underlying deposits, with progressive drying of the vertical profile. Pressure builds 373 under the cohesive cap and continues to rise with increasing gas velocity. This eventually results in explosive 374 channeling and a sudden basal pressure drop as the overburden pressure is exceeded. In higher moisture 375 levels (5.00 - 10.00%) experiments, the deposit does not dry as a relatively uniform rising profile. Instead, 376 pressure builds as the gas velocity is increased until cracks form in the deposit. These cracks act as effective 377 gas escapes and release the pore pressure. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 378 Discussion 379 The impact of moisture on ignimbrite material and PDC behaviour is relatively understudied, with previous 380 detailed investigations having traditionally focused on dry (Wilson, 1980) and saturated (Roche et al., 2001) 381 extremes. However, direct observations have shown that variable amounts of moisture can enter a PDC 382 system (Cole et al., 1998, 2002; Lipman, 2019; Vecino et al., 2022) and accretionary lapilli and ash pellets are 383 believed to provide evidence for the presence of moisture within PDCs (Branney and Kokelaar, 2002; Brown 384 and Branney, 2004; Druitt, 2014). Our results show that with increasing moisture content: 1) the cohesivity of 385 ignimbrite material alters drastically, even with very small weight-percent amounts (> 0.50%); 2) an increase in 386 moisture can change flow property behaviours from a free flowing to a non-flowing material; 3) changes in 387 moisture affects fluidisation profiles and gas escape structures; 4) a defluidising deposit can lead to a drying 388 profile, and therefore lateral and vertical heterogeneity within the deposit; and 5) pressure can increase where 389 gas escape is hindered by moisture, which can cause dramatic releases of pressure in an explosive fashion. 390 Here we discuss the implications of these findings for the flow behaviour of a PDC and variations in deposit 391 sedimentology and broader architecture. 392 Material Behaviours 393 These experiments show that in a dry (0%) sample, material with a grainsize of 5 – 1000 µm has low cohesivity. 394 The cohesive nature of dry material heavily depends on the fine ash content of the material. For example, 395 angle of repose is highest in samples V1 and V3 which have some of the highest % of fines (Table 3.1). 396 However, results indicate that sorting plays a key role; V5 has one of the lowest fines contents (0.11 %) but has 397 the same SAoR value as V2 which has a high fines content of 21.41 %. V2 and V5 have a large size range, similar 398 sorting indexes (V2 – 0.710, V5 – 0.758 ϕ) and the lowest kurtosis values (V2 – 2.58, V5 – 2.48 ϕ). When dry, 399 all samples have good flowability and are said to be free flowing (i.e., when fluidised the deposit demonstrates 400 bubbling and channeling throughout the deposit). Results from the Bulk/Tapped density experiments show 401 that V1 and V4 exhibit some of the best flowing behaviors. V1 is well sorted and displays the largest volumes 402 of fines (35.76%) whereas V4 is very well sorted and has one of the lowest volumes of fine material (0.11%). 403 The excellent flowability seen in V4 may result from its sorting and resulting packing behaviour, which is 404 known to affect flow behaviour (Breard et al., 2023). The DAoR results show contrasting behaviours versus the 405 SAoR results when comparing against the Sauter mean diameter. This suggests that particle size controls are 406 likely to be more important in understanding the remobilisation of static deposits than in the flowability of 407 particles already in motion. 408 Such an observation has wide ranging implications. For example, Breard et al (2023) suggests that long run-out 409 distances in block and ash flows (BAFs) were a result of large degrees of fragmentation, with the current 410 becoming more fines rich, and subsequently the deposit displaying higher packing. These particle size changes 411 result in a dynamically evolving flow where fines formation and increasing packing behaviour reflect elevated 412 pore-pressure within the flow (Breard et al., 2023). Our experiments (V1, V4) show that both fines content and 413 packing can contribute to good flowability behaviours, with implications for run-out distances. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 414 The material with the larger volume of fines (V1, V2) is shown to exhibit more cohesive behaviours with 415 increasing moisture (i.e., higher SAoR angle). V4, which is more well sorted, has the highest increase in SAoR 416 values. Our work demonstrates how important the role of moisture can be, at small amounts, in changing 417 flowability behaviours. We can build a hypothesis that the addition of moisture into a PDC during propagation, 418 with increasing fragmentation and packing, can be a factor in controlling run-out distances – higher moisture 419 contents reduce flowability so may reduce maximum runout distances, particularly in flows with enhanced 420 fragmentation. 421 We can then assume that during deposition from a PDC the fine ash fraction of a deposit, and material with 422 increased packing, may be more influenced by water – and therefore hold a higher moisture content during 423 deposition. Within a material with >30% volume of fines, the stress forces begin to be dominated by the fine 424 fraction (Li et al., 2020; Breard et al., 2023). A large volume of fines, both with and without moisture, may 425 dramatically alter the deposition and the preservation potential of these layers. This has implications for 426 understanding deposit architecture behaviours (e.g., gas escape structures) or extents (e.g., erosion of ash-rich 427 layers). 428 Gas Escape Structures 429 A variety of gas escape structures were observed in the fluidisation experiments, with many related to 430 moisture content. Here we define three main types of behaviour (Table 5). In Type 1 (0.00 - < 0.50%) we see 431 partial fluidisation and segregation of heterogenous material through bubbling and channeling. In material 432 with a smaller size range, small vertical bubbling occurs across the entirety of the deposit. During Type 2 (0.50 433 – 5.00%), an irregular drying profile develops and moves through the deposit from the base. As the drying 434 profile grows, dry pockets of bubbling material begin to form in between irregular lobes of wet material. 435 Explosive channeling also occurs, which releases pressure and facilitates quicker drying of the whole deposit. 436 Finally, during Type 3 (7.50 – 10.00%), similar lobe and pocket structures are formed to Type 2 but are 437 accompanied by cracking processes, where fractures in the wet material form to accommodate rapid gas 438 escape. 439 Table 5 Varying types of behaviour of gas escape observed with increasing moisture in volcanic material. Type 1 Type 2 Type 3 Moisture Range 0.00 – 0.25% 0.50 – 5.00% 7.50, 10.00% Bubbling Yes Yes Yes Channeling Yes Yes No Drying Profile No Yes Yes Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology Pocketing No Yes Yes Explosive Channeling No Yes No Cracking No No Yes 440 441 Roche et al., (2001) investigated the water fluidisation behaviour of ignimbrite material where material was 442 saturated and subjected to an increase in fluid velocity. Similarly, the findings of Wilson (1980; 1984) and the 443 experiments herein demonstrate gas escape structures forming from aggregative behaviour. The gas escape 444 structures were all fines depleted and rich in dense and coarse material. However, the aqueous gas escape 445 structures (pipes) were observed to form at lower fluid velocities than the aerated structures (Fig. 7). 446 447 448 449 Fig.7 showing minimum gas velocity of gas escape structure formation for dry, moisture-influenced, and saturated deposits. Wilson (1980) values (hourglass) based on first formation of pipes (0% wt.). The values of our results (circles) are from first formation of gas escape structure seen in V1-V6 at varying moisture Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 450 451 percentages (i.e., bubbling at lower %’s, explosive cracking at the highest %’s). Roche et al., (2001) values (cross) are on initial pipe formation (80% wt.). 452 Figure 7 shows that by increasing moisture within a sample, higher gas velocities are required for aggregative 453 fluidisation. However, between 10 and 80% moisture there is a change in the dominant fluidising medium, 454 from gas to water. Instead of impeding early fluidisation structures, a large increase in moisture leads to more 455 regular structures forming. This can be explained by changing particle-water states with increasing moisture. 456 Aggregative fluidisation mechanisms will result in the segregation of particles through gas escape structures, 457 where fines are winnowed. The nature of segregation will depend on the particle concentration and the size, 458 shape, density, and relative proportions of clasts (Sparks, 1976; Wilson, 1980; 1984; Branney and Kokelaar, 459 2002). We find that the moisture content of the deposit also controls this process; segregation structures can 460 change dynamically with drying or become hindered with increasing moisture influence. This is due to our 461 material being in a predominantly capillary state (Kim and Hwang., 2003; Kim and Sture, 2008). At higher levels 462 of moisture, particles reach a more saturated state, are completely supported by capillary bonds and 463 fluidisation behaviours are no longer inhibited (as seen in Roche et al., 2001; Kim and Hwang., 2003). We 464 observe that even small influences of moisture (as low as 0.50% by weight) into volcanic material may control 465 the formation and nature of gas escape structures. 466 Application to natural deposits 467 Our results show introducing moisture into volcanic materials may cause changes in gas escape morphology. 468 Gas escape structures have been recorded and described extensively within field volcanological literature (e.g. 469 Fisher and Schmincke, 1984; Pacheco-Hoyos et al., 2020). They have been described as pods and pipes 470 displaying single or branching patterns, or as lenticular, curvilinear, and crescentic shaped (Wilson, 1980; 471 Branney and Kokelaar, 2002; Pacheco-Hoyos et al., 2020). They can be spatially arranged within individual 472 layers or can move through multiple layers and are often fines depleted. Our results demonstrate varied 473 morphologies, including vertical channels, sub-vertical cracks, and pods (created by moisture-rich lobes and 474 dry pockets). 475 Changes in gas escape structures in ignimbrites are thought to be dominated by heterogeneity within 476 ignimbrite material (e.g., size, density, shape etc.; Wilson, 1984; Pacheco-Hoyos et al., 2020). We propose that 477 varying moisture levels will also influence changes in gas escape morphology and may explain circumstances 478 where morphological changes are observed when other conditions appear unchanged. More detailed 479 documentation of morphology of field examples may allow for improved interpretations of depositional 480 environment. 481 Morphological changes in gas escape structures also have implications for drying profiles and resulting 482 heterogeneities within a deposit. In our experiments, the formation of a drying profile demonstrates both 483 vertical and lateral variation due to an undulating contact between wet and dry material resulting in vertical 484 and lateral changes in the tensile strength of a deposit. This may have implications for erodibility, 485 remobilisation (as well as preservation potential), and subsequent deposition of certain layers within a deposit. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 486 Mechanism for secondary eruptions 487 Secondary eruptions in ignimbrite deposits form due to the interaction between water and hot material (Van 488 Westen and Daag, 2005). When in contact with a hot ignimbrite, steam will expand and cause sudden 489 explosive decompression. Secondary eruptions form large craters (20 – 80m depth), can remobilize large 490 volumes of ignimbrite material and can occur for years after the initial eruption (the 1991 Mount Pinatubo 491 generated secondary eruptions for up to a year; Riehle et al., 1995; Van Westen and Daag, 2005). Riehle et al., 492 (1995) modelled cooling, degassing and compaction behaviours within ignimbrites. High temperatures were 493 most likely to remain elevated within deposits >50 m thick, with temperatures cooling mostly by groundwater 494 and some influence from rainfall. Keating et al (2005) modelled that the addition of water on a hot deposit can 495 result in exceeded pore pressure, in turn exceeding the overburden pressure. This can result in secondary 496 eruptions. 497 Moyer and Swanson (1987) described three styles of secondary eruptions - passive degassing (least explosive), 498 ash fountaining and explosive cratering (most explosive) - controlled by thermal energy and the permeability 499 of the overlying material. Analogue experiments investigating the mechanisms of secondary hydroeruptions 500 have been explored by Gilbertson et al (2020). They identified that vertical changes in size fractions, and 501 therefore a vertical profile of minimum fluidisation velocities, resulted in secondary hydroeruptions. In these 502 experiments, a deposit capped with coarser material formed an upward doming bed leading to an explosive 503 release of material. This was due to a drag-induced system. The fine-particle layer below acted as a lower 504 minimum fluidisation layer that was unable to fluidise the overlying, coarser layer, resulting in pressure 505 increase and release. 506 Secondary eruptions are associated with subaerial PDC deposits, and also with sub-aqueous PDCs (Krakatau, 507 1883; Mandeville et al., 1996) and sub-marine sediments (Hovland et al., 2002; Rogers, 2015; Cojean et al., 508 2021) and Martian impact craters (Boyce et al., 2012). Martian pits in impact craters are thought to be formed 509 by water loss from exsolving vaporizing water, creating streams of gas and the funneling out of fine ejecta 510 towards the surface (Boyce et al., 2012). Pockmarks are circular craters that form in the sub-aqueous 511 environment: deep marine, lakes, lacustrine (Cojean et al., 2021). They can form from the upward movement 512 of fluid (liquid or gas), the trapping of the fluid by an impermeable cap, or from gradual accumulation of the 513 fluid until the gas is released. The fine mud particles within these environments can be extremely cohesive, 514 and therefore display a very high tensile strength, which is what can form the impermeable overburden 515 material in these environments (Rogers, 2015). Pockmarks have also been known to form in association with 516 slides and slumps (Hovland et al., 2002). 517 Analogous to all the mechanisms for secondary hydroeruptions is an active moving pore pressure (i.e., 518 vaporization in meteorites) as well as a vertical variation in the permeability of a material (i.e., fine, or mud- 519 rich material, minimum fluidisation velocities). Results from our experiments show that increasing moisture 520 levels within the fluidised deposit can lead to impermeable layers forming through drying. By increasing 521 moisture throughout our experiments, we exhibit passive degassing (0%), ash fountaining (> 0.50% wt.) and Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 522 explosive cratering (> 0.50% wt.) behaviours as described in Moyer and Swanson (1987). After lithostatic 523 pressure of the impermeable wet cap is overcome, explosive channeling (> 0.50% wt.) and cracking (> 7.50% 524 wt.) occurs (named explosive cratering by Moyer and Swanson., 1987). Similarly, to these works, our results 525 herein demonstrate the impact of intermediate permeability on secondary eruptive styles. We argue that the 526 change from passive degassing to explosive cratering is not only a reflection of thermal energy in the system 527 (as our material cooled significantly throughout the length of experiment), but also by internal degassing of a 528 partially fluidised deposit. 529 Critically, our results suggest a potential new mechanism for secondary eruptions that form in a moisture 530 influenced material (Fig. 8a). In our experiments, the addition of water during deposition results in increased 531 cohesion and resulting tensile strength. As the deposit dries from the base, we see a shift in gas escape as the 532 material begins to dry and bubble. In our model (Fig. 8a), the upper moisture rich layer inhibits passive de- 533 gassing and leads to increased pore-pressure. With increasing pressure in the deposit, the overburden strength 534 of wet material is compromised. The result is a sudden pressure release by explosive channeling and cracking, 535 which mimics similar behaviours seen in secondary eruptions in ignimbrite deposits. Additionally, in a dry 536 deposit later moistened by water (i.e., precipitation), the upper moisture-rich layers of material will create an 537 overall denser material (Fig. 8b). 538 Application to field examples 539 Secondary eruptions were observed following the Mt St. Helens 1980 and Mt Pinatubo 1991 events (Keating, 540 2005) and were attributed to variations in the permeability of ignimbrite deposits caused by the presence of 541 water (e.g., rainfall and lacustrine environments) (Moyer and Swanson, 1987; Manville et al., 2002). It is 542 thought that high pressure towards the base of these ignimbrites, caused by vaporization of water, led to low- 543 permeable layers preventing the balancing of pore pressures throughout the deposit, which resulted in 544 explosive depressurization (Keating, 2005). Keating (2005) suggests that after emplacement, hydrological re- 545 establishment may begin to occur and interaction with hot overlying ignimbrite material may result in the 546 formation of secondary hydroeruptions. 547 Our moisture influenced model may provide an explanation for the observations of secondary eruptions in 548 deposits that have aggraded with the presence of water (e.g., secondary eruptions followed the previous 549 location of the Rogue River; Druitt and Bacon, 1986) (Fig. 8a) and that have interacted with rain (e.g., Mt 550 Pinatubo, Daag and Westen, 1996) (Fig. 8b). Rainfall may create a moisture-rich cap to the deposit that is 551 impermeable to degassing from the lower deposit. The increased moisture from the rain would result in an 552 increased cohesivity, and therefore tensile strength, of the material. With gas escape inhibited, pressure may 553 continue to build until the overburden pressure is reached, and degassing is then allowed to escape through a 554 secondary eruption in the deposit. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 555 556 Fig. 8 moisture-influenced model of secondary eruption formation by a) a defluidising wet deposit and b) a 557 defluidising dry deposit with external influences of water. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 558 Conclusion 559 This work highlights the importance of moisture introduced into PDCs both in terms of PDC flow dynamics as 560 well as the characteristics of the resultant deposits. Our results show that: 1) the cohesivity of ignimbrite 561 material changes drastically, even though relatively small additions of moisture amounts (> 0.50%); 2) an 562 increase in moisture can entirely alter flow property behaviour from a free flowing to a non-flowing material; 563 3) changes in moisture impact fluidisation profiles and gas escape structures; 4) a defluidising deposit can lead 564 to a moisture profile, and therefore lateral and vertical heterogeneity within the deposit, and 5) pressure can 565 increase where gas escape is hindered by cohesive substrates driven by moisture content, resulting in 566 secondary eruptions. This builds on previous models of secondary eruptions in deposits and supports the idea 567 that they are formed because of the development of an impermeable capping layer, here created by the 568 addition of moisture. This work further proposes that moisture within a defluidising deposit profile may hinder 569 or change the formation of gas escape structures, which can then also lead to pressure increase and release, 570 with significant implications for the interpretations of the sedimentary structures within the deposits. Overall, 571 the work has shown the critical role of moisture within PDC dynamics and the implications for the erodibility of 572 a material, preservation potential and the broader understanding of deposit architecture. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 573 Reference 574 575 Abdullah, E.C. and Geldart, D. (1999) ‘The use of bulk density measurements as flowability indicators’, Powder Technology, 102(2), pp. 151–165. Available at: https://doi.org/10.1016/S0032-5910(98)00208-3. 576 577 578 Amidon, G.E., Meyer, P.J. and Mudie, D.M. (2017) ‘Particle, Powder, and Compact Characterization’, in Developing Solid Oral Dosage Forms. Elsevier, pp. 271–293. Available at: https://doi.org/10.1016/B978-012-802447-8.00010-8. 579 580 581 Barclay, J., Johnstone, J.E. and Matthews, A.J. (2006) ‘Meteorological monitoring of an active volcano: Implications for eruption prediction’, Journal of Volcanology and Geothermal Research, 150(4), pp. 339– 358. Available at: https://doi.org/10.1016/j.jvolgeores.2005.07.020. 582 583 584 Beakawi Al-Hashemi, H.M. and Baghabra Al-Amoudi, O.S. (2018) ‘A review on the angle of repose of granular materials’, Powder Technology, 330, pp. 397–417. Available at: https://doi.org/10.1016/j.powtec.2018.02.003. 585 586 587 Benage, M.C., Dufek, J. and Mothes, P.A. (2016) ‘Quantifying entrainment in pyroclastic density currents from the Tungurahua eruption, Ecuador: Integrating field proxies with numerical simulations’, Geophysical Research Letters, 43(13), pp. 6932–6941. Available at : https://doi.org/10.1002/2016GL069527. 588 589 Boyce, J.M. et al. (2012) ‘Origin of small pits in Martian impact craters’, Icarus, 221(1), pp. 262–275. doi: 10.1016/j.icarus.2012.07.027. 590 Branney, M. and Kokelaar, P. (2002) Pyroclastic Density Currents and the Sedimentation of Ignimbrites. 591 592 593 Branney, M.J. and Kokelaar, P. (1992) A reappraisal of ignimbrite emplacement: progressive aggradation and changes from particulate to non-particulate flow during emplacement of high-grade ignimbrite. SpringerVerlag, pp. 504–520. 594 595 596 Breard, E.C.P. et al. (2019) ‘The Permeability of Volcanic Mixtures—Implications for Pyroclastic Currents’, Journal of Geophysical Research: Solid Earth, 124(2), pp. 1343–1360. Available at: https://doi.org/10.1029/2018JB016544. 597 598 Breard, E.C.P. et al. (2023) ‘The fragmentation-induced fluidisation of pyroclastic density currents’, Nature Communications, 14(1), p. 2079. Available at : https://doi.org/10.1038/s41467-023-37867-1. 599 600 601 Brown, R.J. et al. (2010) ‘Origin of accretionary lapilli within ground-hugging density currents: Evidence from pyroclastic couplets on Tenerife’, Bulletin of the Geological Society of America, 122(1–2), pp. 305–320. Available at: https://doi.org/10.1130/B26449.1. 602 603 604 605 Brown, R.J., and Branney, M.J. (2013) ‘Internal flow variations and diachronous sedimentation within extensive, sustained, density-stratified pyroclastic density currents flowing down gentle slopes, as revealed by the internal architectures of ignimbrites on Tenerife’, Bulletin of Volcanology, 75(7), pp. 1–24. Available at: https://doi.org/10.1007/s00445-013-0727-0. 606 607 608 609 2014 The United States Pharmacopeial Convention (2015) Bulk density and tapped density of powders - US pharmacopeia (USP), 〈616〉 Bulk Density and Tapped Density of Powders. Available at: https://www.usp.org/sites/default/files/usp/document/harmonization/gen-chapter/bulk_density.pdf (Accessed: 02 June 2023). 610 611 Camuffo, D. (2019) ‘Theoretical Grounds for Humidity’, in Microclimate for Cultural Heritage. Elsevier, pp. 43– 59. Available at: https://doi.org/10.1016/B978-0-444-64106-9.00003-1. 612 613 Cas, R.A., and Wright, J.V. (1991) ‘Subaqueous pyroclastic flows and ignimbrites: an assessment’, Bulletin of Volcanology, 53(5), pp. 357–380. Available at: https://doi.org/10.1007/BF00280227. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 614 615 616 Chédeville, C. and Roche, O. (2015) ‘Influence of slope angle on pore pressure generation and kinematics of pyroclastic flows: insights from laboratory experiments’, Bulletin of Volcanology, 77(11), p. 96. Available at : https://doi.org/10.1007/s00445-015-0981-4. 617 618 Chen, D. et al. (2021) ‘Critical Shear Stress for Erosion of Sand-Mud Mixtures and Pure Mud’, Frontiers in Marine Science, 8. Available at: https://doi.org/10.3389/fmars.2021.713039. 619 620 621 Chigira, M. and Yokoyama, O. (2005) ‘Weathering profile of non-welded ignimbrite and the water infiltration behaviour within it in relation to the generation of shallow landslides’, Engineering Geology, 78(3–4), pp. 187–207. Available at: https://doi.org/10.1016/j.enggeo.2004.12.008. 622 623 Cocco, R., Reddy, S.B. and Knowlton, K.T. (2014) Back to Basics Introduction to Fluidization. Available at: www.aiche.org/cep. 624 625 626 Cojean, A.N.Y. et al. (2021) ‘Morphology, Formation, and Activity of Three Different Pockmark Systems in PeriAlpine Lake Thun, Switzerland’, Frontiers in Water, 3, p. 666641. Available at: https://doi.org/10.3389/frwa.2021.666641. 627 628 Cole, P.D. et al. (1998) Pyroclastic flows generated by gravitational instability of the 1996-97 Lava Dome of Soufriere Hills Volcano, Montserrat, pp. 3425–3428. Available at: https://doi.org/10.1029/98GL01510. 629 630 Cole, P.D. et al. (2002) ‘Deposits from dome-collapse and fountain-collapse pyroclastic flows at Soufriere Hills Volcano, Montserrat’, Geological Society, 21, pp. 231–262. 631 632 Daag, A & Westen, C.J. (1996). Cartographic modelling of erosion in pyroclastic flow deposits of Mount Pinatubo, Philippines. ITC Journal. 2. 110-124. 633 634 Dartevelle, S. et al. (2002) ‘Origin of the Mount Pinatubo climactic eruption cloud: Implications for volcanic hazards and atmospheric impacts’, Geology, (7), p. 663. 635 636 Deb, P.K. et al. (2018) ‘Particulate level properties and its implications on product performance and processing’, Dosage Form Design Parameters, pp. 155–220. doi:10.1016/b978-0-12-814421-3.00005-1. 637 638 Druitt, T. H., Bruni, G., Lettieri, P., and Yates, J. G. (2004), The fluidization behaviour of ignimbrite at high temperature and with mechanical agitation, Geophys. Res. Lett., 31, L02604, doi:10.1029/2003GL018593. 639 640 641 Druitt, T. & Avard, Geoffroy & Bruni, G. & Lettieri, Paola & Maez, F.. (2007). Gas retention in fine-grained pyroclastic flow materials at high temperatures. Bulletin of Volcanology. 69. 881-901. 10.1007/s00445-0070116-7. 642 643 644 Druitt, T.H. (2014) ‘New insights into the initiation and venting of the Bronze-Age eruption of Santorini (Greece), from component analysis’, Bulletin of Volcanology, 76(2), p. 794. Available at: https://doi.org/10.1007/s00445-014-0794-x. 645 646 647 Druitt, T.H. and Bacon, C.R. (1986) ‘Lithic breccia and ignimbrite erupted during the collapse of Crater Lake Caldera, Oregon’, Journal of Volcanology and Geothermal Research, 29(1–4), pp. 1–32. Available at: https://doi.org/10.1016/0377-0273(86)90038-7. 648 649 650 Duane, W.J. et al. (2008) ‘General characteristics of temperature and humidity variability on Kilimanjaro, Tanzania’, Arctic, Antarctic, and Alpine Research, 40(2), pp. 323–334. Available at: https://doi.org/10.1657/1523-0430(06-127) [DUANE]2.0.CO;2. 651 652 Dufek, J. (2016) ‘The Fluid Mechanics of Pyroclastic Density Currents’, Annual Review of Fluid Mechanics, 48, pp. 459–485. Available at : https://doi.org/10.1146/annurev-fluid-122414-034252. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 653 654 655 Eychenne, J. et al. (2012) ‘Causes and consequences of bimodal grain-size distribution of tephra fall deposited during the August 2006 Tungurahua eruption (Ecuador)’, Bulletin of Volcanology, 74(1), pp. 187–205. Available at: https://doi.org/10.1007/s00445-011-0517-5. 656 657 Fisher, R.V. and Schmincke, H.-U. (1984) ‘Pyroclastic flow deposits’, Pyroclastic Rocks, pp. 186–230. doi:10.1007/978-3-642-74864-6_8. 658 659 Geldart, D. (1973) Elsevier Sequoia SA, Lausanne-Printed in the Netheriands Types of Gas Fluidization, Powder Technology, pp. 285–292. 660 661 Gilberston, M.A. et al. (2020) ‘A Fluidisation Mechanism for Secondary Hydroeruptions in Pyroclastic Flow Deposits’, Frontiers in Earth Science, 8. Available at: https://doi.org/10.3389/feart.2020.00324. 662 663 Girolami, Laurence & Druitt, T. & Roche, Olivier & Khrabrykh, Z.V. (2008). Propagation and hindered settling of laboratory ash flows - art. no. B02202. Journal of Geophysical Research. 113. 10.1029/2007JB005074. 664 665 666 L. Girolami, T.H. Druitt, O. Roche. (2005). Towards a quantitative understanding of pyroclastic flows: Effects of expansion on the dynamics of laboratory fluidized granular flows, Journal of Volcanology and Geothermal Research, Volume 296, Pages 31-39, ISSN 0377-0273, https://doi.org/10.1016/j.jvolgeores.2015.03.008. 667 668 G.Giordano, Ray A.F. Cas. (2021). Classification of ignimbrites and their eruptions, Earth-Science Reviews, Volume 220, https://doi.org/10.1016/j.earscirev.2021.103697 669 670 671 Gorle, A.P. and Chopade, S.S. (2020) ‘Liquisolid Technology: Preparation, Characterization and Applications’, Journal of Drug Delivery and Therapeutics, 10(3-s), pp. 295–307. Available at: https://doi.org/10.22270/jddt.v10i3-s.4067. 672 Hartmann, D.L. (2016). Global physical climatology (Vol. 103). Newnes. 673 674 Hausner, H.H. (1981) ‘Powder characteristics and their effect on powder processing’, Powder Technology, 30(1), pp. 3–8. Available at: https://doi.org/10.1016/0032-5910(81)85021-8. 675 676 677 Hovland, M., Gardner, J.V. and Judd, A.G. (2002) ‘The significance of pockmarks to understanding fluid flow processes and geohazards’, Geofluids, 2(2), pp. 127–136. Available at: https://doi.org/10.1046/j.14688123.2002.00028.x. 678 679 680 Houghton, B., White, J.D.L. and Van Eaton, A.R. (2015) ‘Phreatomagmatic and Related Eruption Styles’, in The Encyclopaedia of Volcanoes. Elsevier, pp. 537–552. Available at: https://doi.org/10.1016/B978-0-12385938-9.00030-4. 681 682 683 Hurwitz, S. (2003) ‘Groundwater flow, heat transport, and water table position within volcanic edifices: Implications for volcanic processes in the Cascade Range’, Journal of Geophysical Research, 108(B12). Available at: https://doi.org/10.1029/2003jb002565. 684 685 Keating, G.N. (2005) ‘The role of water in cooling ignimbrites’, Journal of Volcanology and Geothermal Research, 142(1-2 SPEC. ISS.), pp. 145–171. Available at: https://doi.org/10.1016/j.jvolgeores.2004.10.019. 686 687 Kim, T.-H. and Hwang, C. (2003) ‘Modelling of tensile strength on moist granular earth material at low water content’, Engineering Geology, 69(3–4), pp. 233–244. doi:10.1016/s0013-7952(02)00284-3. 688 689 Kim, T.-H. and Sture, S. (2008) ‘Capillary-induced tensile strength in unsaturated sands’, Canadian Geotechnical Journal, 45(5), pp. 726–737. Available at: https://doi.org/10.1139/T08-017. 690 691 692 LaMarche, C.Q. et al. (2016) ‘Linking micro-scale predictions of capillary forces to macro-scale fluidization experiments in humid environments’, AIChE Journal, 62(10), pp. 3585–3597. Available at: https://doi.org/10.1002/aic.15281. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 693 694 695 Leturia, M. et al. (2014) ‘Characterization of flow properties of cohesive powders: A comparative study of traditional and new testing methods’, Powder Technology, 253, pp. 406–423. Available at: https://doi.org/10.1016/j.powtec.2013.11.045. 696 697 698 Li, W.C. et al. (2020) ‘Effects of stress state and fine fraction on stress transmission in internally unstable granular mixtures investigated via discrete element method’, Powder Technology, 367, pp. 659–670. Available at: https://doi.org/10.1016/j.powtec.2020.04.024. 699 700 Lipman, P.W. (2019) ‘When ignimbrite meets water: Mega scale gas-escape structures formed during welding’, Geology, 47(1), pp. 63–66. Available at: https://doi.org/10.1130/G45772.1. 701 702 703 704 Ludwig, B., Millington-Smith, D., Dattani, R., Adair, J. H., Posatko, E. P., Mawby, L. M., Ward, S. K., & Sills, C. A. (2020). Evaluation of the hydrodynamic behaviour of powders of varying cohesivity in a fluidized bed using the FT4 Powder Rheometer®. Powder Technology, 371, 106–114. https://doi.org/10.1016/j.powtec.2020.05.042 705 706 Ma, Y. et al. (2019) ‘Modelling the effect of moisture on the flowability of a granular material’, Meccanica, 54(4–5), pp. 667–681. Available at: https://doi.org/10.1007/s11012-018-0901-8. 707 708 709 Manville, V., Segschneider, B. and White, J.D.L. (2002) ‘Hydrodynamic behaviour of Taupo 1800a pumice: implications for the sedimentology of remobilized pyroclasts’, Sedimentology, 49(5), pp. 955–976. Available at: https://doi.org/10.1046/j.1365-3091.2002.00485.x. 710 711 712 713 Matthews, A.J., Barclay, J. and Johnstone, J.E. (2009) ‘The fast response of volcano-seismic activity to intense precipitation: Triggering of primary volcanic activity by rainfall at Soufrière Hills Volcano, Montserrat’, Journal of Volcanology and Geothermal Research, 184(3–4), pp. 405–415. Available at: https://doi.org/10.1016/j.jvolgeores.2009.05.010. 714 715 716 Moondra, S. et al. (2018) ‘Bulk Level Properties and its Role in Formulation Development and Processing’, in Dosage Form Design Parameters. Elsevier, pp. 221–256. Available at: https://doi.org/10.1016/B978-0-12814421-3.00006-3. 717 718 Moyer, T.C. and Swanson, D.A. (1987) SECONDARY HYDROERUPTIONS IN PYROCLASTIC-FLOW DEPOSITS: EXAMPLES FROM MOUNT ST. HELENS, Journal of Volcanology and Geothermal Research, pp. 299–319. 719 720 721 Pacheco-Hoyos, J.G., Aguirre-Díaz, G.J. and Dávila-Harris, P. (2020) ‘Elutriation pipes in ignimbrites: An analysis of concepts based on the Huichapan Ignimbrite, Mexico’, Journal of Volcanology and Geothermal Research, 403. Available at : https://doi.org/10.1016/j.jvolgeores.2020.107026. 722 723 724 725 Pepin, N.C. et al. (2017) ‘A comparison of simultaneous temperature and humidity observations from the SW and NE slopes of Kilimanjaro: The role of slope aspect and differential land-cover in controlling mountain climate’, Global and Planetary Change, 157, pp. 244–258. Available at: https://doi.org/10.1016/j.gloplacha.2017.08.006. 726 Pierrat, P. and Caram, H.S. (1997) ‘Tensile strength of wet granular materials. 727 728 Riehle, J.R., Miller, T.F. and Bailey, R.A. (1995) ‘Cooling, degassing and compaction of rhyolitic ash flow tufts: a computational model’. 729 Rios, Maribel. (2006) 'Developments in powder flow testing', Pharmaceutical Technology, 30 (2), pp. 38-49. 730 731 Roche, O., Druitt, T.H. and Cas, R.A.F. (2001) ‘Experimental aqueous Fluidization of ignimbrite’, Journal of Volcanology and Geothermal Research, p. 14. 732 733 Rogers, R. (2015) ‘Hydrate-associated seafloor instabilities’, Offshore Gas Hydrates, pp. 189–219. doi:10.1016/b978-0-12-802319-8.00006-1. Non-peer reviewed EarthArXiv Pre-print – Submitted to Bulletin of Volcanology 734 735 Sahoo, S. et al. (2022) ‘Eruption cycles of Mount Etna triggered by seasonal climatic rainfall’, Journal of Geodynamics, 149. Available at: https://doi.org/10.1016/j.jog.2021.101896. 736 737 Schneider, C.A., Rasband, W.S. and Eliceiri, K.W. (2012) ‘NIH Image to ImageJ: 25 years of image analysis’, Nature Methods, 9, pp. 671–675. 738 739 Self, S. and Sparks, R.S.J. (1978) ‘Characteristics of Widespread Formed by the Interaction of Silicic Magma and Water Pyroclastic Deposits’, Bulletin of Volcanology, 41, pp. 196–212. 740 741 Smith, G. M. (2020). Propagation of aerated pyroclastic density current analogues: flow behaviour and the formation of bedforms and deposits. (Thesis). University of Hull. 742 743 744 SPARKS, R.S.J. (1976) ‘Grain size variations in ignimbrites and implications for the transport of pyroclastic flows’, Sedimentology, 23(2), pp. 147–188. Available at: https://doi.org/10.1111/j.13653091.1976.tb00045.x. 745 746 Sparks, R.S.J. (1978) ‘Gas release rates from pyroclastic flows: a assessment of the role of fluidisation in their emplacement’, Bulletin Volcanologique, 41(1), pp. 1–9. Available at: https://doi.org/10.1007/BF02597679. 747 748 Telling, J., Dufek, J. and Shaikh, A. (2013) ‘Ash aggregation in explosive volcanic eruptions’, Geophysical Research Letters, 40(10), pp. 2355–2360. Available at: https://doi.org/10.1002/grl.50376. 749 750 751 Van Eaton, A.R. and Wilson, C.J.N. (2013) ‘The nature, origins and distribution of ash aggregates in a large-scale wet eruption deposit: Oruanui, New Zealand’, Journal of Volcanology and Geothermal Research, 250, pp. 129–154. Available at : https://doi.org/10.1016/j.jvolgeores.2012.10.016. 752 753 Vecino, M.C.D. et al. (2022) ‘OPEN Aerodynamic characteristics and genesis of aggregates at Sakurajima Volcano, Japan’, Scientific Reports, p. 14. 754 755 756 van Westen, C.J. and Daag, A.S. (2005) ‘Analysing the relation between rainfall characteristics and lahar activity at Mount Pinatubo, Philippines’, Earth Surface Processes and Landforms, 30(13), pp. 1663–1674. Available at: https://doi.org/10.1002/esp.1225. 757 758 Wilson, C.J.N. (1980) THE ROLE OF FLUIDIZATION IN THE EMPLACEMENT OF PYRO-CLASTIC FLOWS: AN EXPERIMENTAL APPROACH, Journal of Volcanology and Geothermal Research, pp. 231–249. 759 760 761 Wilson, C.J.N. (1984) ‘The role of fluidization in the emplacement of pyroclastic flows, 2: Experimental results and their interpretation’, Journal of Volcanology and Geothermal Research, 20(1–2), pp. 55–84. Available at: https://doi.org/10.1016/0377-0273(84)90066-0. 762 763 764 Wormsbecker, M. and Pugsley, T. (2008) ‘The influence of moisture on the fluidization behaviour of porous pharmaceutical granule’, Chemical Engineering Science, 63(16), pp. 4063–4069. Available at: https://doi.org/10.1016/j.ces.2008.05.023. 765 766 Yehuda, T., & Kalman, H. (2020). Geldart classification for wet particles. Powder Technology, 362, 288–300. https://doi.org/10.1016/j.powtec.2019.11.073 767 Yu, A.B. and Hall, J.S., 1994 ‘Packing of fine powders subjected to tapping’, p. 10. 768 769 Zeng, K. et al. (2017) ‘Combined effects of initial water content and heating parameters on solar pyrolysis of beech wood’, Energy, 125, pp. 552–561. Available at: https://doi.org/10.1016/j.energy.2017.02.173. 770