<p><b>This thesis investigates the magmatic processes that operated at Ōkataina volcanic centre prior to and during the rhyolitic events of its most recent eruption sequence, the Rotorua Subgroup (erupted since 25.4 ka), with a specific focus on timescales of human interest. Further context is derived with data from the 52.8 ka Rotoiti (caldera-forming) and Earthquake Flat eruptions, and selected Mangaone Subgroup events that occurred between the 52.8 and 25.4 ka. The regular occurrence of these events provides regular snapshots of the magmatic system over this time period. Whole-rock, glass, and mineral compositional data and timescales modelled from diffusion gradients are compared within geographical and physical volcanological contexts. These data provide a picture of the recent evolution of Ōkataina, shedding light on how quickly the volcano could awaken from its current dormancy to produce a new rhyolitic event, and what that event might be like. </b></p>
<p>Geochemical data from Mangaone Subgroup eruptives show that the post-caldera magmatic system recovered over c. 27 kyr, with increasingly evolved and radiogenic compositions erupted during these events. The post-25.4 ka Rotorua Subgroup shows a more steady evolution, trending towards less radiogenic products towards the present day. Modern Ōkataina comprises three vent clusters, the intra-caldera Mt. Tarawera (formed by the 21.9 ka Ōkareka, 17.5 ka Rerewhakaaitu, 14 ka Waiohau and 1314 AD Kaharoa eruptive episodes) and Haroharo Massif (the 25.2 ka Te Rere, 9.4 ka Rotomā, 7.9 ka Mamaku and 5.5 ka Whakatāne), and the Ōkareka Embayment (the 25.2 ka Te Rere, 15.6 ka Rotorua) on the western edge of the caldera. Most eruptions were geochemically complex, with multiple magmas, some of which were repeated between eruptions. Variations in whole-rock and glass geochemistry are controlled by both the dominant mineral assemblage and external inputs, and relate to the geographical region and timing of when each magma erupted. Each vent region shows evidence for distinct features in its shallower magmatism (i.e. that producing the final-erupted magma compositions), super-imposed on temporal trends controlled by the deeper system. This deeper system beneath Haroharo and Tarawera appears to be more unitary or inter-connected. However, the Ōkareka Embayment is unique in that it appears to involve multiple magmatic systems.</p>
<p>Back-scattered electron imaging of orthopyroxenes from the Rotorua, Waiohau and Whakatāne eruptives (and selected plagioclase from the Rotorua), coupled with major-element analyses of the orthopyroxenes, imply that open-system processes are common at Ōkataina, at least in the early stages of eruptible magma body assembly. These processes are reflected in diverse core compositions and complex zonation patterns, narrowing to more uniform rims that are in equilibrium with the compositions of the host melts. Not all cores are in equilibrium with the associated whole-rock compositions, but are in equilibrium with whole-rock compositions from other magmas erupted nearby, further demonstrating open-system crystal exchange and the repeated tapping of some magma sources from one event to another. Melt segregation and magma body assembly takes place over decades up to a century or two, with magma residence times on the order of centuries to millennia, or (if prematurely triggered) as short as months. Priming or triggering events such as heating by basaltic injection (common in the Ōkataina rhyolite eruptions) occur over decades prior to eruption.</p>
<p>Quartz-hosted, sealed melt inclusions preserve volatile contents of c. 3-6 wt% H2O and c. 15-125 ppm CO2, with values tending to decrease in younger eruptions, although Waiohau samples have relatively low H2O, and the Rotorua samples have relatively low CO2. Overall, the Ōkataina eruptives have remarkably low CO2 contents compared with international datasets. There are weak correlations between H2O and CO2 contents coexisting in magmas within individual vent regions, but little system-wide coherence due to each region having differing trends (a step-wise drop at Tarawera versus a steady decrease in the young Haroharo eruptions). Scatter in data from sealed melt inclusions, and relatively lower H2O contents in open-ended inclusions (embayments) indicate pre-eruptive volatile losses, driven by crystallisation and/or open-system degassing. The H2O contents of the inner parts of embayments suggest an initial slow decompression at the start of the final magma ascent may also play a minor role. Magma ascent rates are low to moderate (0.1-4.4 m.s-1), with the Ōkareka Embayment eruptives having notably faster ascent rates than those from Tarawera or Haroharo. Both the sealed melt inclusion volatile contents and derived magma ascent rates are similar to those observed or inferred from other silicic eruptions worldwide and are notably comparable to those of super-eruptions, indicating that magma volume and overpressure do not influence ascent rate at eruption onset. </p>
<p>Current geophysical observations suggest that a viable magma source currently exists under the southwestern end of Haroharo, and that basaltic diking activity is ongoing in the Ōkataina area. While basaltic eruptions can occur with little warning (e.g. Tarawera, 1886 CE), it is more likely that a future Ōkataina event will be prolonged and rhyolitic. The volcano can rejuvenate into an eruption-ready state within human lifetimes (i.e. decades), and final pre-eruptive warnings from geophysical monitoring could be as short as only a few hours. A new rhyolitic episode could have major impacts, lasting several years and involving both voluminous lavas and widespread fall deposition.</p>
