Goldrick 1

Eruptive Cycles within Akaroa Lava Flows and Recognition of 1

Contemporaneous Explosive Eruptions 2

1,2Emma Goldrick and 1Samuel Hampton 3

1Department of Geological Sciences, University of Canterbury, Private Bag 4800, 4

Christchurch, New Zealand 5

2Department of Geology and Geophysics, Yale University, 210 Whitney Ave, New Haven, 6

CT 7

8

ABSTRACT 9

Geochemical and stratigraphic study of the Akaroa Volcanic Complex can 10

reveal pre eruptive processes and inform eruption style. Volcanic deposits record 11

many details from their formational history, including source material, magma 12

chamber dynamics, and eruption style. Geochemistry can be used to show the 13

evolution of stacked lava flow composition as well as the beginning of new volcanic 14

cycles. All data must be taken into account when studying this complex, intraplate 15

system. Previous assumptions of a single vent, shield volcano will continue to be 16

revised as more evidence shows Akaroa as a large complex with many vents 17

erupting from shallow lava chambers sourced by a deep reservoir. This study 18

analyzes lava flow and ash deposits in the southeastern area of the volcano to create 19

an eruptive model recognizing contemporaneous explosive eruptions. 20

INTRODUCTION 21

Volcanic deposits on Banks Peninsula tell a geological history, reflecting the 22

processes that occurred deep below the surface millions of years ago. Banks 23

Peninsula offers a unique view on volcanism due to its intraplate nature, a less well 24

understood tectonic setting for volcanism. Geochemistry, mineral textures, and 25

stratigraphic sequencing of rock samples provide clues to reconstructing the nature 26

of past eruptions. The Akaroa Volcanic Complex on Banks Peninsula, NZ shows 27

clear, cyclic geochemical trends repeating all over the peninsula (Beckham 2015, 28

Johnson 2012). Beckham (2015) studied six eastern bays to create a shallow magma 29

chamber model. This model explains Akaroa’s the geochemical evolution from 30

Goldrick 2

picritic basalt to trachyte and decreasing phenocryst percentage within successive 31

lava flows. 32

This study seeks to add three transects in the southeastern area of the 33

volcano (Figure 1) to the geochemical database, analyse the beginning stages of the 34

batch eruption sequence, and investigate Sewell et al.’s 1992 map of the peninsula. 35

GEOLOGICAL BACKGROUND 36

Banks Peninsula, located on New Zealand’s South Island east coast, consists 37

of two Miocene volcanic complexes. The Lyttelton Volcanic Complex (12.3-10.4 Ma) 38

is older but the younger Akaroa Volcanic Complex (9.4-6.8 Ma) constitutes two 39

thirds of the peninsula’s rocks (Weaver and Smith, 1989). The two complexes are 40

thought to be a result of two large lithospheric detachments, causing asthenospheric 41

upwelling and large intra-plate eruptions (Timm et al., 2009). Previously the 42

volcanoes were thought of as simple composite shield volcanoes fed by intraplate 43

dykes, but more recent studies have revealed a more complex picture (Sewell et al., 44

1992). 45

There is one 1:100000 scale geologic map of the peninsula created in 1992 46

and described by Sewell and Weaver in a series of guides. The map was created on 47

the assumption of a singular vent for both the Lyttleton and Akaroa complexes 48

(Sewell et al., 1992). Recent work undertaken by Hampton and Cole (2009) was able 49

to map 15 separate eruptive vents for the Lyttleton volcanic complex, showing how 50

much revision is needed in the Sewell (1992) map and formation units. 51

The Akaroa Volcanic Complex formation has been studied most recently by 52

Johnson (2012), Patel (2013), and Beckham (2015). The volcanics are 53

predominantly lava flows ranging in composition from picrite basalt to benmorite 54

(Beckham, 2015). Johnson (2012) correlated cycles of repeating geochemical 55

sequences spatially around different parts of the volcano, suggesting a model of 56

many eruptive vents similar to Hampton and Cole (2009). These batches of lavas are 57

sourced from multiple shallow magma chambers beneath the peninsula, fed by a 58

much deeper reservoir linked to the lithospheric detachment (Johnson, 2012, Timm 59

et al., 2009). This gives the whole complex similar geochemical trends while having 60

Goldrick 3

varying evolutions depending on the individual vent and shallow chamber. 61

Replenishing events begin the eruptive cycle with picrite basalt containing altered 62

crystal inclusions and over time to become more aphyric evolved lavas (Beckham, 63

2015). These altered crystals are the remnants from the past cycle’s unerupted 64

magma, essentially having new cycles start off by clearing out the shallow magma 65

chamber (Beckham, 2015). 66

METHODS 67

Five transects were taken in southeastern bays in Banks Peninsula 68

(Haylock’s, Damon’s and Flea Bay South, North and North East) resulting in 49 69

samples. Fieldwork was conducted in July 2015 and overseen by Frontiers Abroad. 70

Samples were taken in stratigraphic order, working either from the top to the 71

bottom of ridgelines or from the bottom to top. 72

Whole-rock geochemical data obtained from X-Ray Fluorescence and 73

Diffraction (XRF) was organized based on stratigraphy and plotted in IgPet2012 to 74

identify rock type, magma batches, and trace element variation within picritic 75

basalts. Rock type was determined through the Total Alkali-Silica Diagram. Element 76

and compound trends were plotted to define the boundaries of individual batches 77

and spatially mapped onto the study area (Figures 2-4). Thin sections were 78

photographed under normal and cross polarized light to highlight textural features. 79

Mineral textures were analysed according to the Beckham (2015) Petrographic 80

Guide (Figure 6) based off of Viccaro et al. 2010. 81

Shankle (2015) sampled 17 ash beds across Banks Peninsula and obtained 82

XRF/XRD data and conducted thorough thin section microscopy analysis. 83

RESULTS 84

Substantial, stratigraphic geochemical data was obtained for three of the five 85

transects (Haylock’s Bay, Damon’s Bay and Flea Bay South) taken in southeastern 86

Banks Peninsula. 37 data points were organized vertically by compounds and 87

elements to show geochemical variations within batches (Figures 2-4). Thin sections 88

were compared under normal and cross polarized light. Batches with three or more 89

sequential samples were analysed under thin section microscopy, looking for trends 90

Goldrick 4

and patterns within the matrix and mineral composition as the batch evolves 91

(Figures 7-9). 92

DISCUSSION 93

UNIT DISTINCTION IN AKAROA COMPLEX 94

In 1992 the New Zealand Geological Survey created the first 1:100000 scale 95

map of Banks Peninsula. In this map there is a clear distinction between two 96

formations in the study area as seen in Figure 10. These two formations are the 97

French Hill Formation (light green) and the Te Oka Formation (dark green). As 98

listed in “Geology of the Akaroa West Area”, the Te Oka Formation unconformably 99

overlays the French Hill formation and is marked by a shift from porphyritic varying 100

basalts to steeply dipping aphyric hawaiite flows. Two transects collected by 101

Frontiers Abroad students crossed over the dividing line between the two 102

formations and two stayed completely within the French Hill Formation, as shown 103

by red lines in Figure 10. 104

A spider plot using Pearce 1983 of trace elements in picritic basalts from the 105

four transects was created in IgPet2012 (Figure 11). There appears to be no 106

significant variation between the trace elements in varying transects or within the 107

two transects that cross the formation’s boundary (Damon’s Bay and Flea Bay 108

South). Field observations recorded by Frontiers Abroad students showed no sign of 109

an unconformity within the Damon’s Bay or Flea Bay South transects. In addition, 110

Sewell & Weaver (1990) state Te Oka Formation rock type is exclusively hawaiite. 111

Total Alkali- Silica Diagrams created in IgPet2012 list rock types from these 112

transects as varying from picrite basalt to hawaiite. 113

These results indicate that a large-scale revision must take place in analysing 114

the formation units of the Akaroa Volcanic Complex. The origin of the distinction 115

between the Te Oka and the French Hill formation must be questioned. With new 116

research indicating the complex’s formation came from multiple vents it would be 117

wise to create a new map of the whole volcano, as local formations may appear 118

depending on their spatial relation to varying vents. This map has already begun 119

Goldrick 5

through the collaboration of many generations of Frontiers Abroad students, but 120

should be opened to all geologists wishing to join. 121

ERUPTIVE MODEL 122

Batches of evolving geochemistry within bay transects are consistent with 123

Beckham's (2015) shallow magma chamber model designed for eastern bays in 124

Banks Peninsula. Transects in Figures 2-4 show clear, cyclic geochemical evolution 125

where increasing SiO2 corresponds to increasing K2O, MgO, P2O5 and Zr and 126

decreasing Al2O3, FeO, V, and Sr. The consistency within these transect's 127

geochemistry implies the southeastern bays went through magma recharge events 128

(Figure 12). It is possible one magma chamber was not shared by all the samples, 129

but it is likely the same primitive source underlay the volcano. 130

Thin sections show evolution of melt through mineral percentages 131

and textures. As magma is recharged into the shallow chamber it mixes 132

leftoverphenocrysts, partially melting or refractionating crystals, leading to a 133

plethora of textures and higher mineral percentages seen in the early picrite 134

basalts (Figures 7-9). Sieving, patchy cores, and resorbed rims all support 135

thehypothesis of an intrusive, hot melt signalling the beginning of a new 136

batch and altering existing crystal cumulate mush. As the source for the 137

injectiondepletes, mixing stops and flows become more aphyric as well as evolve 138

intohigher silica lavas. In the southeastern bays the lavas evolved from picrite 139

basalt to hawaiite to mugearite. However, before these lavas even erupted, another 140

important stage in the batch cycle occurs. 141

Shankle (2015) found that ash deposits on Banks Peninsula are almost 142

exclusively sourced from early stage picritic magma. This would indicate that 143

explosive, ash producing eruptions only occurred early in the eruption sequence 144

when the newly injected magma was still primitive. The ash deposits sampled were 145

almost all found stratigraphically below picrite basalt flows (Figure 13). As a result a 146

collaborative model has been developed to explain the evolving nature of the 147

Akaroa Volcano going from explosive to effusive in eruption style. 148

Eruptive Model Stages 149

Goldrick 6

1) A capped vent traps volatiles within the vent and the attached shallow magma 150

chamber, crystal cumulates sit at the bottom of the chamber, leftover from the last 151

magma batch 152

2) A new injection of magma into the chamber begins to mix the crystals, partially 153

melting and recrystallizing them, as well as increasing the pressure within the 154

chamber and conduit as volatile bubbles grow and coalesce 155

3) The cap fails at a critical pressure resulting in an explosive eruption driven by the 156

rapid expansion of volatiles, depositing layers of ash – immediately followed by 157

continued explosive activity, erupting picritic basalt containing the sieved and 158

resorbed minerals phenocrysts 159

4) The magma injection is over and explosive eruptions become less frequent, 160

magma within the chamber begins to fractionate new crystals which sink to the 161

bottom without the mixing the injection had provided 162

5) Lava erupted becomes more evolved and aphyric over time, from picrite basalt to 163

hawaiite to mugearite, and fractionated crystals build up on the bottom of the 164

magma chamber 165

6) In the final stages the evolved, aphyric lava is erupted effusively and the conduit 166

closes signifying the end of the injected lava batch and returning to stage 1 167

CONCLUSIONS 168

The shallow magma chamber model for Akaroa developed by Johnson (2012) 169

and refined by Beckham (2015) applies to the sampled southeastern bays: 170

repeating geochemical and thin section textural evolutionary trends matched 171

those described in Beckham (2015) 172

Formation of Akaroa complex in the southeastern area of the volcano 173

requires revision: no significant geochemical or morphological differences 174

were found to distinguish between the French Hill Formation and the Te Oka 175

Formation, no boundary between the two was found in transects 176

Ash deposits can inform on early eruption cycle dynamics: in conjunction 177

with Shankle (2015), a model was created to integrate an early explosive 178

Goldrick 7

eruption stage, supported by intermittent ash layers with picrite basalt 179

origins, into the already existing shallow magma chamber model 180

Future study should focus on pre eruptive magmatic processes revealed 181

through thin section crystal alteration, in addition the field area should be 182

revisited to find more ash deposits overlain by lava flows to further the 183

model created here 184

ACKNOWLEDGMENTS 185

A massive thank you to Frontiers Abroad and the University of Canterbury 186

for the opportunity to conduct this research. I would like to acknowledge the help 187

and guidance given by Dr. Sam Hampton, Mr. Max Borella, Dr. Darren Gravley, Dr. 188

Travis Horton, Ms. Elisabeth Bertolett, and Mr. Rob Spiers. I would also like to thank 189

my fellow FA students for inspiring new questions and especially Maddie for her 190

collaboration. Lastly, thank you to Zoe for supplementing me with technology while 191

in New Zealand. 192

FIGURES 193

194

Figure 1 195

Goldrick 8

196

Figure 2 197

198

Figure 3 199

Goldrick 9

200

Figure 4 201

Goldrick 10

202

Goldrick 11

Figure 5203

Goldrick 12

204

Figure 6 205

206

Figure 7 207

208

Goldrick 13

209

Figure 8 210

Goldrick 14

211

Figure 9 212

Goldrick 15

213

Figure 10 214

Goldrick 16

215

Figure 11 216

217

Goldrick 17

218

Figure 1219

Goldrick 18

220

221

Figure 13222

Goldrick 19

FIGURE CAPTIONS 223

Figure 1. Highlighted study area on a map of Banks Peninsula taken and outlined on 224

Google Earth 225

226

Figure 2. Haylock’s Bay stratigraphic geochemical plot. Horizontal dashed lines 227

represent batch dividers, vertical solid lines represent related flows within single 228

batches, and vertical dashed lines represent unrelated flows between two batches. 229

Geochemical data trends shown for MgO, SiO2, Al2O5, K2O, P2O5, FeO, V, Zr, and Sr. 230

231

Figure 3. Damon’s Bay stratigraphic geochemical plot. Horizontal dashed lines 232

represent batch dividers, vertical solid lines represent related flows within single 233

batches, and vertical dashed lines represent unrelated flows between two batches. 234

Geochemical data trends shown for MgO, SiO2, Al2O5, K2O, P2O5, FeO, V, Zr, and Sr. 235

236

Figure 4. Flea Bay South stratigraphic geochemical plot. Horizontal dashed lines 237

represent batch dividers, vertical solid lines represent related flows within single 238

batches, and vertical dashed lines represent unrelated flows between two batches. 239

Geochemical data trends shown for MgO, SiO2, Al2O5, K2O, P2O5, FeO, V, Zr, and Sr. 240

241

Figure 5. Geochemical plots mapped spatially onto the study area. Red lines indicate 242

transects and arrows within the red lines indicate stratigraphic up. 243

244

Figure 6. Beckham’s (2015) petrographic guide created using Viccaro et al. (2010) 245

as a model. 246

247

Figure 7. Batch 1 from the Haylock’s Bay transect, resorbed rims and melt inclusions 248

at the beginning of the batch indicating a magma recharge event with stages of rapid 249

crystal regrowth. 250

251

Goldrick 20

Figure 8. Batch 2 from the Damon’s Bay transect, large phenocrysts with varying 252

textural indicators begin this batch, steadily decreasing in size and percentage as 253

well as plagioclase microlites in the matrix becoming more aligned. 254

255

Figure 9. Batches 1 and 10 from the Flea Bay South transect, batch 1 evolving from a 256

highly altered, crystal rich picrite basalt to an almost aphyric mugearite, batch 10 257

evolving from a porphyritic flow with clear resorbed rims to an aphyric, aligned 258

flow. 259

260

Figure 10. Map area taken from Sewell et al. (1992) showing the four collected 261

transects as red lines. 262

263

Figure 11. Spider diagram showing picrite basalt flows from Damon’s Bay, Flea Bay 264

South, Flea Bay Northeast, and Haylock’s Bay transects. 265

266

Figure 12. Goldrick and Shankle (2015) eruptive model that take contemporaneous 267

explosive eruptions and lava flows into account as explained by the captions below: 268

1) A capped vent traps volatiles within the vent and the attached shallow magma 269

chamber, crystal cumulates sit at the bottom of the chamber, leftover from the last 270

magma batch 271

2) A new injection of magma into the chamber begins to mix the crystals, partially 272

melting and recrystallizing them, as well as increasing the pressure within the 273

chamber and conduit as volatile bubbles grow and coalesce 274

3) The cap fails at a critical pressure resulting in an explosive eruption driven by the 275

rapid expansion of volatiles, depositing layers of ash – immediately followed by 276

continued explosive activity, erupting picritic basalt containing the sieved and 277

resorbed minerals phenocrysts 278

4) The magma injection is over and explosive eruptions become less frequent, 279

magma within the chamber begins to fractionate new crystals which sink to the 280

bottom without the mixing the injection had provided 281

Goldrick 21

5) Lava erupted becomes more evolved and aphyric over time, from picrite basalt to 282

hawaiite to mugearite, and fractionated crystals build up on the bottom of the 283

magma chamber 284

6) In the final stages the evolved, aphyric lava is erupted effusively and the conduit 285

closes signifying the end of the injected lava batch and returning to stage 1 286

287

Figure 13. Picrite basalt overlying a thick ash deposit in Haylock’s Bay 288

289

References 290

Beckham, E. (2015). Magmatic evolution of The Akaroa Volcanic Complex, Eastern 291

Banks Peninsula. University of Canterbury. 292

Hampton, S. J., & Cole, J. W. (2009). Lyttelton Volcano, Banks Peninsula, New 293

Zealand: Primary Volcanic Landforms and Eruptive Centre Identification. 294

Geomorphology, 104(3-4), 284-298. 295

Johnson, J. (2012). Insights into the magmatic evolution of Akaroa Volcano from the 296

geochemistry of volcanic deposits in Okains Bay, New Zealand. University of 297

Canterbury. 298

Patel, S. (2013). On the Subsurface Evolution of Batch Lavas at Stony Bay, Banks 299

Peninsula, New Zealand: A Study in Plagioclase Phenocryst Development. 300

University of Canterbury. 301

Sewell, R.J. and Weaver, S.D. (1990). Geology of the Akaroa West Area. New Zealand 302

Geological Survey, Department of Scientific and Industrial Research. 14-19. 303

Sewell et al. (1992) Geological Map of New Zealand: Banks Peninsula. 1:50 000. New 304

Zealand: New Zealand Geological Survey. 305

Shankle, M. (2015). Ash Homogeneity Informing Eruption Dynamics on Banks 306

Peninsula. University of Canterbury. 307

Timm, C. Hoernle, K. Van Den Bogaard, P., Bindeman, I., & Weaver, S. (2009) 308

Geochemical evolution of intraplate volcanism at Banks Peninsual, New 309

Zealand: interaction between astehnospherica and lithospheric melts. Journal 310

of Petrology, 40(6), 989-1023. 311

Goldrick 22

Viccaro, M., Giacomoni, P. P., Ferlito, C., & Cristofolini, R. (2010). Dynamics of magma 312

supply at Mt. Etna volcano (Southern Italy) as revealed by textural and 313

compositional features of plagioclase phenocrysts. Lithos, 116(1), 77-91. 314