The role of hydrovolcanism in the formation of the Cenozoic monogenetic volcanic fields of Zealandia

ABSTRACT

Cenozoic geological evolution of New Zealand centres around the formation of Zealandia, a new continent that became detached from the eastern margin of Gondwana around 105 Ma. Spreading opened the Tasman Sea leaving a fragment of continental lithosphere, largely submerged, in the SW Pacific. Throughout the Cenozoic history, volcanism became an integral part of Zealandia. Continental lithosphere provided the basement for the volcanism, both onshore and offshore. Monogenetic volcanism was common throughout the Cenozoic. The availability of water was ubiquitous through surface water bodies (oceans and lakes) and various other terrestrial hydrous systems provided by the humid temperate climate of Zealandia. Hydrovolcanism, both explosive and non-explosive, has played a significant role in Zealandia's volcanic history resulting in volcano mega- architecture involving edifice geology and volcanic hazards. However, hydrovolcanism has commonly been overlooked in Zealandia's monogenetic volcanism context. Cenozoic monogenetic fields of Zealandia provide a unique laboratory and comparative analogy for other volcanic fields on Earth that are associated with low-lying terrestrial settings or shallow marine environments in a humid temperate climate.

KEYWORDS:

• Phreatomagmatic
• volcanic field
• tuff ring
• maar
• tuff cone
• scoria cone
• Molten Fuel-Coolant interaction (MFCI)
• intraplate
• plate boundary

Introduction

Hydrovolcanism refers to a process where magma and water or water saturated sediments interact either explosively or non-explosively during a volcanic event (Sheridan and Wohletz 1983). Recognition of hydrovolcanism is significant as it can be used as a proxy to determine the eruptive environment the magma entered (e.g. deep to shallow subaqueous, subaerial) (Gutmann 2002; Risso et al. 2008; Graettinger et al. 2013; Németh and Kósik 2020). In this regard, hydrovolcanism contains a signature of the environmental conditions at the time of magma emplacement. It thereby links either the broader environmental conditions of a region where the magmatism and volcanism took place or the environment created by the magmatic process itself; for instance upper conduit, vent or crater processes (Németh and White 2003; Kereszturi et al. 2011, 2014; Kereszturi and Németh 2011; Ross et al. 2011; Németh and Kósik 2020). Hydrovolcanism or hydromagmatism are commonly used as synonyms for phreatomagmatism (Valentine and White 2012; Németh and Kósik 2020). In recent years, however, phreatomagmatism has become associated with magma – water interaction driven volcanic eruptions such as those that typically occur in subaerial conditions and driven by Molten Fuel-Coolant Interactions (MFCI) (Lorenz 1987; White 1991; Németh and Kósik 2020). Hydrovolcanism can take place in any type of volcanic setting including subaqueous (e.g. shallow marine, deep marine, sub-lacustrine, subglacial) and subaerial (e.g. coastal plains, on flank of central volcanoes, along fluvial valleys, intramountain basins) environments and associated with small-volume or more complex, large volcanoes such as stratovolcanoes, composite volcanoes or calderas (Self 1983; Walker 1993; Wilson 1994; Giordano 1998; Muraviev et al. 1998; De Rita et al. 2002; Geshi et al. 2003; Mastin et al. 2004; Darragh et al. 2006; Edgar et al. 2007; Németh and Cronin 2008; Wong and Larsen 2010; Nagaoka and Okuno 2011; Swanson et al. 2012; Van Eaton et al. 2012). In recent years the recognition of hydrovolcanism in any volcanic system has evolved hugely. These include direct observations through experiments designed to understand cratering processes (Valentine et al. 2015; Graettinger et al. 2016; Macorps et al. 2016; Graettinger and Valentine 2017; Sonder et al. 2018) to field descriptions of the resulting magmatic-to-volcanic systems (Leat and Thompson 1988; Heiken et al. 2000; Valentine et al. 2017b; Graettinger 2018; Pedrazzi et al. 2018; Németh and Kósik 2020). Phreatomagmatism plays an important role of the style of volcanism of small-volume, dispersed volcanic centres, commonly associated with volcanic fields generated by monogenetic volcanism (Németh and White 2003; Seghedi et al. 2007; Wijbrans et al. 2007; Németh et al. 2010; Kereszturi et al. 2011; Kshirsagar et al. 2011; Mattsson and Tripoli 2011; Stárková et al. 2011; Buechner and Tietz 2012; Herrero-Hernandez et al. 2012; Ramon Avellan et al. 2012; Valentine 2012; van Otterloo et al. 2013; van Otterloo and Cas 2013; Valentine et al. 2017a). Monogenetic volcanism in this context defines volcanism that involves small-volumes of magma producing relatively simple volcanic edifices in a short time (Németh and Kereszturi 2015; Smith and Németh 2017). The significance of hydrovolcanism in the context of monogenetic volcanic field evolution is that the external water can influence the magma fragmentation in the case of explosive processes, hence the resulting eruptive products and volcanic macroforms (a term referring to a volcanic edifice regardless of the geo-environment – subaqueous vs subaerial) can be greatly affected (Kereszturi et al. 2011; Németh et al. 2012b; Agustin-Flores et al. 2014). In the past decade numerous case studies have used New Zealand examples to demonstrate the link between volcanic eruption style, the role of water in explosive volcanism and the interaction between the environment and the volcanism (Németh et al. 2012b; Kereszturi et al. 2014). As monogenetic volcanism involves a small magma volume in a single or a few magma batches, the presence of external water can dominate or influence the volcanic eruptions significantly and that can be recorded in the eruptive products (Agustin-Flores et al. 2014; Németh and Kereszturi 2015; Smith and Németh 2017). This approach has recently been used to define various volcano types in the context of monogenetic volcanism, and the effect of external water has been utilised as a measure of the influence the environment has on the resulting volcanic macroform and the facies architecture of the deposits. Using various configurations of the volume of the interacting magma and the influence of the environment (external water) a framework classifying various small-volume volcanic landforms has been developed (Smith and Németh 2017). This approach has been validated and proven its usefulness in regions where volcanism (magmatism) takes place in a geo-environment ranging from humid climatic conditions, to those located near sea level or entirely subaqueous. Throughout the Cenozoic, suitable conditions existed for preserving dispersed volcanic deposits in the New Zealand region. Consequently, there are excellent examples of hydrovolcanism ranging from fully subaqueous, emergent styles to subaerial phreatomagmatism. However, vegetation cover or incomplete exposure commonly hinder assessment of key sites, many of which are Quaternary and geologically young. In addition, offshore examples are commonly known only from indirect methods (e.g. geophysical measurements) and direct observation or sampling is limited. In this paper we provide an overview of various monogenetic volcanic fields in New Zealand and the evidence used to establish the hydrovolcanic origin of a volcanic succession. The paper will primarily focus on volcanic megaforms and textures (macro and micro) traditionally used to identify hydrovolcanism across various Cenozoic monogenetic volcanic fields in Zealandia.

Geological context of hydrovolcanism of monogenetic volcanic fields of Zealandia

New Zealand is located in an active plate margin defined by two opposing subduction fronts and associated volcanic arcs separated by the 400⁺ km long dextral strike-slip Alpine Fault (Cole and Lewis 1981; Brothers 1984, 1985; Ballance et al. 1985; Gaina et al. 1998; King 2000; Nicol et al. 2007) (Figure 1A). The continental crustal fragments comprising New Zealand were separated from Gondwana in the late Cretaceous, after 105 Ma (Laird 1993; Sutherland et al. 2001; Laird and Bradshaw 2004; Nicol et al. 2007). The continental mass commenced drifting during active Tasman Sea rifting that started at c. 105 Ma, which was followed by breakup and seafloor spreading in the Tasman Sea from 83 Ma (Sutherland et al. 2001, 2017; Laird and Bradshaw 2004). In addition, in the Taranaki and West Coast, a second, later rift event occurred. These tectonic events culminated in the initiation of the convergence associated with the Hikurangi and Puysegur subduction zones and eventually lead to form the present-day Zealandia, including the current landmass of New Zealand (Sutherland et al. 2001; Segev et al. 2012; Mortimer et al. 2014). Dispersed volcanism has commonly occurred since the Late Cretaceous, but most of the rift related volcanoes are buried within sedimentary basins (Bischoff et al. 2020). These well-preserved Cenozoic volcanic successions tend to be restricted to distinct regions at different times (Timm et al. 2010) (Figure 1). Most of the intraplate volcanism was basaltic (Timm et al. 2010).

Figure 1. Topographic and tectonic settings of the broader environment of New Zealand (Zealandia) with respect to the locations of Cenozoic monogenetic volcanism; Auckland Volcanic Field (AVF) (inset A), South Auckland Volcanic Field (SAVF) and Waikato coastline (W) (inset B), Taupo Volcanic Zone (TVZ) (inset C) and Central and East Otago with Waipiata Volcanic Field (WVF) and Dunedin Volcanic Complex (inset D). Bathymetric information of offshore Zealandia was derived from the 250 m resolution gridded bathymetric data (2016), National Institute of Water and Atmospheric Research (NIWA) (Mitchell et al. 2012). Tectonic and structural information from Mortimer et al. (2017). Regional maps of the TVZ, South Auckland and Waikato region and the central Otago region were plotted to the 8 m NZ DEM (LINZ – Land Information New Zealand 2012), whereas the Auckland Volcanic Field was mapped on a 1 m LiDAR DEM (LINZ – Land Information New Zealand 2013).

Some small volume, monogenetic felsic volcanoes also occurred close to plate boundaries; the most recent of these are associated with rifting in the Taupo Volcanic Zone (Wilson et al. 1995; Wilson and Rowland 2016; Kósik et al. 2018, 2019, 2020) (Figure 1). Within the Taupo Volcanic Zone basaltic as well as more silicic, dacitic to rhyolitic volcanism, commonly resulted in the formation of small-volume, and likely short-lived, lava dome eruptions (Rooney and Deering 2014; Kósik et al. 2016; Wilson and Rowland 2016; Kósik et al. 2019). The explosive silicic volcanism commonly associated with Taupo also generated numerous silicic tuff rings that filled with lava domes (Brooker et al. 1993; Kósik et al. 2019).

Today New Zealand is part of a recently defined continent Zealandia [https://www.gns.cri.nz/Home/News-and-Events/What-s-new/The-origin-and-meaning-of-the-name-Te-Riu-a-Maui-Zealandia]. This continental mass, which is largely submerged today, fulfils the geophysical criteria for a continent. However, not all geoscientists agree with this interpretation and strong debate continues (Dowding and Ebach 2017). The stratigraphic framework of Zealandia can be grouped into two stratigraphical mega-units (Austral Superprovince and Zealandia Megasequence); the older Austral Megaunit is related to the geological evolution of the eastern Gondwana margin and includes the Eastern and Western terrain slices, defined as ‘provinces’ (Mortimer et al. 2014). The younger Zealandia Megasequence consists of the late Cretaceous to Cenozoic succession including the early Cenozoic terrestrial sedimentary units that infilled the rift basins created during the earlier rifting phase (Mortimer et al. 2014). These units were gradually covered by marine deposits as water depth increased in reponse to the gradual submergence of the continent through the early part of the Cenozoic (Mortimer et al. 2014). A large portion of Zealandia was submerged by late Oligocene coincident with the onset of extensive subaqueous volcanism forming shallow marine plateaus. Volcanism in a largely marine setting meant subaqueous to emergent volcanic successions inter-fingered with marine sedimentary rocks (Stilwell and Consoli 2012; Thompson et al. 2014).

Intraplate volcanism of various styles has taken place in Zealandia periodically throughout the Cenozoic. With the development of the modern New Zealand plate boundary in the early Miocene reuplift led to an archipelago developing into an ever-increasing landmass with shallowing of the marine conditions in the surrounding offshore regions. The resulting geo-environment included abundant Miocene to Pliocene volcanic centres dispersed from the Chatham Rise and Campbell Plateau to the present day Otago region in the South Island (Gamble et al. 1986; Cas 1989; Briggs and McDonough 1990; Kear 1994; Hoernle et al. 2006; Timm et al. 2010; Scott et al. 2013; Scott and Turnbull 2019) (Figure 1). Many of these are well preserved. Direct sea floor dredging and geophysical surveys confirms the presence of dispersed volcanism to the east of the plate boundary in the south where a large number of mafic volcanoes occur offshore, on land (Timm et al. 2010) and out into the deep-submarine farther offshore (Uruski 2020). In eastern Otago, Paleogene submarine to emergent volcanism occurred in a shallow continental shelf setting, producing Surtseyan volcanic mounds and tuff cones the deposits of which are exposed along the present day coastline, (Price and Coombs 1975; Coombs et al. 1986, 2008; Cas et al. 1989; Németh et al. 2003; Price et al. 2003; Sorrentino et al. 2011; Fox et al. 2015; Moorhouse et al. 2015; Moorhouse and White 2016; Jones et al. 2017; McLeod and White 2018; Scott et al. 2020). The most extensive group of volcanoes formed during the Cenozoic in the Otago region are the continental alkaline basaltic volcanoes centred on the Dunedin area. Known as the Dunedin Volcanic Group (Scott et al. 2020) it includes, the Waipiata Volcanic Field (WVF) and the Dunedin Volcanic Complex (DVC) (Coombs et al. 1986). In recent work (Scott et al. 2020) it is proposed that the Waipiata Volcanics/Waipiata Volcanic Field be replaced with Dunedin Volcanic Group, since it is actually not easy to subdivide Waipiata from the Dunedin Volcano as they overlap in age, chemistry, isotopic properties and were only distinguished by Coombs et al. (1986) as ‘where the main flows end’, which is ambiguous.

There are several onshore Cenozoic volcanic fields in the southern part of the South Island. One of the largest and best preserved is the Miocene Waipiata Volcanic Field (WVF), an eroded intracontinental volcanic field (Németh and White 2003) adjacent to the Dunedin Volcano (or Dunedin Volcanic Complex after Coombs et al. 1986).

The pre-volcanic Cenozoic units consist of non-marine and marine clastic sediments deposited on an early Cretaceous erosional surface cut into schist basement (Bishop and Turnbull 1996) and successively overlain by thick Oligocene transgressive deposits (Bishop and Turnbull 1996). These marine units consisting of generally fine-grained glauconitic sandstones and siltstones are well preserved in the northern part of the Waipiata Volcanic Field (Németh and White 2003). The contacts of the volcanic conduits exposed today display cross-cutting relationships with the pre-volcanic marine units. The area re-emerged in the early Miocene in response to transpressional tectonics associated with the inception of the Alpine Fault and marked by the first non-marine (terrestrial fluvio-lacustrine) deposits (Mortimer et al. 2014). This geological setting for intraplate volcanism in the South Island is the perfect geo-environment in which to have rising magma interact with external water, either en-route at shallow crustal level or with surface waters such as rivers, lakes or shallow marine bays. The known pre-volcanic lithologies, sand, silt or gravelly beds, are good aquifers and are, in places, closely associated with carbonate rocks, some with extensive fracture networks. Vent alignments have also been recognised suggesting the presence of strong structural control on the distribution of volcanism. This structural fabric likely manifested in the formation of elongated fluvial networks providing broad water-rich environments. Later on, farther north intraplate volcanism developed west of the plate boundary forming a laterally extensive zone of dispersed volcanic fields including the monogenetic volcanic fields from south Waikato region through to Northland (Briggs and McDonough 1990) (Figure 1A). This volcanism developed in a low-lying terrestrial to shallow subaqueous environment forming volcanoes typical of emergent Surtseyan style to more terrestrial type of volcanoes including tuff rings, maars and scoria cones commonly associated with extensive outpourings of lava (Goles et al. 1996; Briggs et al. 1997; Cook et al. 2003, 2005; Smith and Briggs 2008; Ilanko et al. 2009; Gibson et al. 2010).

Buried late Miocene volcanoes are inferred to exist east of Dunedin and the Campbell Plateau (Adams 1981; Gamble et al. 1986; Gamble and Thomson 1990; Hoernle et al. 2006). Large shield volcanoes similar to the Dunedin Volcano exist in the south at the Auckland Islands (Carnley and Ross volcanoes) (Gamble and Adams 1985; Gamble et al. 1986) and Campbell Island (Morris 1984). Intraplate volcanism is reported from the Antipodes Islands (Scott et al. 2013) and the Chatham Islands (Eocene - Oligocene/40 - 35 Ma and Miocene – Pliocene/6 - 2.6 Ma) (Hoernle et al. 2006). Recent work on the intraplate volcanoes of the South Island, Chatham Rise and Campbell Plateau suggests volcanism occurred intermittently over a large area throughout the Cenozoic, with no recognisable lineament array (Hoernle et al. 2006; Timm et al. 2010) or melting of metasomatised lithospheric mantle (Panter et al. 2006; van der Meer et al. 2017).

Currently there are two concurrent models to explain the origin of Zealandia intra-plate volcanism. Some authors relate it to the presence of a mantle plume that initiated during the final stage of eastern Gondwana breakup (Panter et al. 2006; van der Meer et al. 2017) while others ascribe volcanism to asthenospheric upwelling induced by removal of parts of the subcontinental lithosphere throughout the Cenozoic (Hoernle et al. 2006; Timm et al. 2010).

Scales of observation used to recognise hydrovolcanism in the New Zealand geological record

The determination of whether or not magma fragmentation was influenced or directly triggered by magma and water interaction is challenging as there are no tools available to directly test it. Evidence of the action of hydrovolcanism during magma fragmentation is usually indirect often from a common set of micro- and macro-textural features consistent with magma and water interaction including micro and macro textural features of on the chilled margins of pyroclasts, presence of accretionary lapilli and the dominance of pyroclastic density current deposits rich in accidental lithic clasts derived from conduit walls at the crater rim (Németh and Kósik 2020). The problem is perfectly highlighted in a recent review pointing it out that ‘absence of evidence is not evidence of absence’ (White and Valentine 2016). This statement highlights the problem that even if we cannot see direct evidence that hydromagmatic processes took place during a volcanic eruption that is not evidence to say it didn't happen (White and Valentine 2016). In this paper we will restrict ourselves to a list ‘common wisdoms’ used to identify features generated by hydrovolcanism-associated with eruptions of small volume volcanoes in the New Zealand context. There are at least three levels of scale of observations used in New Zealand volcanology to establish magma-water influenced eruption styles in the Cenozoic. These are based on landscape level recognition of volcanic features and macroforms that can be linked to hydrovolcanism, outcrop scale lithofacies and facies relationships in both a vertical and lateral sense, and microtextural works focusing on juvenile pyroclast textures that can be associated with MFCI processes. Unfortunately none of the above mentioned observations can provide unique and decisive arguments to establish hydrovolcanic origin of an eruption sequence of a volcano, volcanic erosion remnant or volcanic successions in the geological record (Lorenz 1987; Graettinger et al. 2015; White and Valentine 2016; Amin and Valentine 2017; Sonder et al. 2018; Németh and Kósik 2020). Recent research is heavily focused on finding a link between eruption styles and resultant volcanic macroforms and deposits. Here we provide some common methods applied in New Zealand in recent years to establish phreatomagmatic origin of volcanoes and volcanic successions.

Volcanic macroform-scale evidence

The macroforms of small volume volcanoes such as maar craters (Figure 2A), tuff rings (Figure 2B) or tuff cones (Figure 2C) are commonly used to infer hydrovolcanic processes responsible for the growth of those macroforms (Morrissey et al. 2000; Brand and Brož 2015; Brož and Németh 2015; De Hon 2015; de Silva and Lindsay 2015; Valentine and Connor 2015; White et al. 2015). Recent research, however, has demonstrated that the process of cratering, for instance the formation of a maar, should be looked at as process creating the volcanic macroform that may or may not be linked to an explosive magma – water interaction hence hydrovolcanism, or more specifically to phreatomagmatism (Graettinger et al. 2015, 2016; Graettinger and Valentine 2017). Here we use maar as a volcanic landform that consists of a flat floored crater cut into the syn-volcanic landscape (even if it is rugged) creating a ‘hole-in-the-ground’ morphology (Christenson et al. 2015; Graettinger 2018; Németh and Kósik 2020), without linking it directly to explosive phreatomagmatic eruptions. Some reports suggest that dry maar craters may form in response to a gas-rich magma subterranean explosion (Mattsson 2012; Berghuijs and Mattsson 2013; Balashova et al. 2016). However, the preserved tephra rings that surround a maar crater show many microtextural features indicative of magma and water interaction, hence phreatomagmatism (Graettinger 2018; Ort et al. 2018).

Figure 2. A – Pukekohe East maar (PEM) from the air (South Auckland Volcanic Field), Note the flat crater floor and the smooth surface of the tephra ring surround the maar crater; B – Overview of the Motukorea/Brown Island (Auckland Volcanic Field) complex tuff ring (tr). Note the scoria cone (sc) complex filling the crater of the tuff ring.; C – Overview of the Maungauika/North Head tuff cone (Auckland Volcanic Field) from the east. The core of the tuff cone (tc) is preserved, its outer edifice has been shoaled away. The centre of the tuff cone is filled with a scoria cone (sc); D – Overview or the preserved part of the Foulden Maar (Waipiata Volcanic Field). The image was taken from the north. Small quarry operation of diatomite is in the middle of the depression (red arrow). In the left side of the view an erosional remnant of a scoria cone (Bald Hill) sitting on the tuff ring.

Erosional remnants of scoria cones, tuff rings and maars of the Waipiata Volcanic Field (WVF) form a characteristic landscape comprising volcanic buttes and knobs (Figure 2D). Cenozoic sedimentary units are mainly preserved along the northern margin of the WVF and not in the central, uplifted fault/fold block areas due to the presence of thick lava flows (Bishop and Turnbull 1996). The only direct evidence of their previous existence or absence comes from studying the pyroclastic deposits preserved in various volcanic pipes, diatremes, in volcanic fields like the Waipiata Volcanic Field (Németh and White 2003).

The WVF is an example of an eroded monogenetic volcanic field where shallow surface water and multiple groundwater tables facilitate water-influenced small-volume volcanism including explosive phreatomagmatic eruptions (Németh and White 2003). This field includes erosional remnants of at least 55 volcanoes in an area recently revised to 7800 km² (Scott et al. 2020). The original small-volume volcanoes, and their associated lava fields, are preserved in the various vent-filling, intra-crater and proximal pyroclastic successions within tuff rings and suspected maar craters (Németh 2003).

Eroded volcanic edifices are also common in younger volcanic regions such as the Taupo Volcanic Zone (Figure 1C). Here entire small-volume volcanoes are commonly covered by thick, younger ignimbrite or reworked volcaniclastic successions. These have been exposed by river downcutting or quarrying. A section of a suspected diatreme is exposed in the interior of the K-trig basalt quarry near Taupo (Figure 1C) (Brown et al. 1994), where a dish-like feature composed of accidental-lithic-rich basaltic tuff overlying coarse breccia infilling a conduit is exhumed (Figure 3A). Similarly, a previously unknown volcanic centre, Te Hukui diatreme, has been discovered near Taupo. This basaltic lapilli tuff and tuff succession forms an odd hill-side associated with a younger silicic tuff ring and maar chain and is interpreted as a remnant of a diatreme (Kósik et al. 2017) (Figure 3B). In both cases the pyroclastic rocks of the preserved successions are dominated by glassy juvenile lapilli and ash particles hosted in a fine-grained accidental-lithic rich matrix, suggesting intensive excavation of country rocks most probably by explosive magma-water interaction (Figure 3C & D).

Figure 3. A – A small diatreme is exhumed as the base of the Punatekahi volcanic complex (Taupo Volcanic Zone). Note the v-shaped architecture of pyroclastic rocks in the base of the section (white dashed lines), marking an exposed diatreme (d).; B – Te Hukui Basalt formation exposing typical phreatomagmatic pyroclastic successions in a well-defined small area inferred to be a diatreme near the Puketerata Volcanic Complex, Taupo Volcanic Zone (note people for scale); C – pyroclastic rocks recovered from the Punatekahi diatreme is composed of glassy juvenile ash and lapilli and abundant accidental lithic clasts forming a bedded, cross-bedded succession inferred to accumulated by base surges; D – Accretionary lapilli-bearing basaltic lapilli tuff and tuff succession from the middle section of the Te Hukui diatreme.

A basaltic diatreme, Ngatatura diatreme (Heming 1980), exposed on the Waikato coastline, (Figure 4A) has been K-Ar dated at 1.83–1.54 Ma (Briggs et al. 1989; Briggs et al. 1994, 1997). However, new K-Ar dating reporting ages of 3.34 ± 0.06–1 Ma are well outside of the commonly accepted Ngatatura ages and similar to those known from Okete Formation and Alexandra Volcanics (Van Niekerk 2016). The diatreme itself is dominated by pyroclastic breccias (Figure 4B) and cannot be linked to any surface expression of a volcano nearby (van Niekerk et al. 2015b). A recent study has revealed the rock defined as Ngatatura diatreme is part of a complex volcano with a history of multiple events potentially spanning a few hundreds of thousands of years and contributing to a record of dramatic changes in the coastal region where near-sea level, coastal conditions have dominated and controlled small volume volcanism over time (van Niekerk et al. 2014, 2015a). As a result, the Ngatatura diatreme and enclosing deposits reveal a history of shallow, subaqueous volcanism that produced hyaloclastite (Figure 4C) and pillow lavas subsequent to the diatreme formation (Van Niekerk 2016).

Figure 4. A – Overview of the Ngatatura diatreme (d) from the NW located along the Waikato coastline (Figure 1D). Note the steep diatreme wall (white dashed lines) that slightly enlarged upsection. The cliff face is about 60 m high; B – Close up view of the pyroclastic breccia of the Ngatatura diatreme rich in angular glassy juvenile pyroclasts (j) and variable rounded sedimentary (s) rocks from the pre-volcanic deposits; C – Over the Ngatatura diatreme shallow marine sediments (ms) intercalated with a hyaloclastite breccia (hy) suggesting a younger eruptive phase in the same area where the diatreme formed previously.

Evidence of magma-water interaction-driven explosive eruptions that can be likened to Surtseyan style eruptions (e.g. explosion taking place within a water body, while the eruption cloud pierces the water surface) are abundant in New Zealand. Along the Waikato Coastline, near Ngatatura diatreme, a complete half section of a tuff cone is exposed with its monotonous pyroclastic succession including abundant dunes and soft sediment deformation (Figure 5A). In the offshore Gannet Island/Karewa a half section of a nephelinitic tuff ring has been reported as a hydrovolcanic edifice (Briggs et al. 1997). Similar half sections are known from East Otago, such as those on the Boatmans Harbour in Oamaru (Figure 5B), or in Cape Young at Chatham Islands (Figure 5C) and interpreted to be result of shallow subaqueous eruptions that gradually built emergent hydrovolcanic edifices through Surtseyan style eruptions (Cas et al. 1989; Maicher 2003; Corcoran and Moore 2008; Sorrentino et al. 2011; Stilwell and Consoli 2012; Sorrentino et al. 2014; Agustin-Flores et al. 2015; Moorhouse et al. 2015). Another interesting example documented for its characteristic facies architecture, is Northhead in the Auckland Volcanic Field which is defined as a Surtseyan volcano formed in very shallow water (30 metres maximum) (Agustin-Flores et al. 2015) (Figure 5D). Kaiapo volcano, a basaltic tuff cone in the Taupo Volcanic Zone, has been reconstructed on the basis of characteristic blocky pyroclasts (Figure 6B) in a 500 m long, albeit dissected exposure consisting of a 50 m thick monotonous, palagonite tuff succession (Brown et al. 1994) (Figure 6A) exposed along the Kaiapo Fault near Taupo. Not far from Kaiapo, poorly exposed basaltic tuffs occur in shoreline cliffs at Acacia Bay, Lake Taupo. These are interpreted as an erosional remnant of a phreatomagmatic tuff ring that formed in subaerial conditions (Wilson and Smith 1985). These tuff ring deposits are dominated by pumiceous lake sediments mixed with 15–20 volume % juvenile glassy basaltic ash and lapilli deposited by pyroclastic fall and surge (Wilson and Smith 1985). None of these volcanoes are well preserved nor do they show their original landform geometry. Their hydromagmatic origin and potential Surtseyan type edifices are inferred on the basis of their pyroclastic texture and facies architecture as outlined in the next section. Ohakune Volcanic Complex, on the southern Ruapehu ring-plain, is a small volcanic field composed of at least four volcanic craters. One of them is being actively quarried and since the first documentation of this complex tuff ring and multiple scoria cones (Houghton and Hackett 1984), new exposures provide insight into the volcano's evolution (Kósik et al. 2016). During its emplacement this scoria cone was influenced by shallow water entering the erupting vents that changed their locations as the edifice grew. However, a complete ‘wet’ eruption history is not preserved. This is a common scenario in large scoria cones with initial phreatomagmatic tuff ring basement (Murcia et al. 2015). The tuff ring at Ohakune shows evidence of magmatic fragmentation generating lava spatter dominated dark horizons (Figure 6C) that are commonly interbedded with fine ash often accompanied by dewatering pipes and plastering effects (Figure 6D).

Figure 5. A – Half section of a Surtseyan tuff cone along the Waikato coastline, just south of Karioi Mountain. Note the strong erosional features, u-shaped channels (u), and the dune-bedded pyroclastic succession typical for an emergent volcano, commonly defined as a result of Surtseyan style eruptions (hammer is for scale); B – Various distal beds of eruption fed density currents at Boatman Harbour in Oamaru as part of a complex emergent volcano. Soft sediment deformation (ss) features are apparent; C – Half section of a Surtseyan volcano with thick pyroclastic succession (pyx) at Cape Yoiung, Chatham Island intruded by feeder dykes with peperitic margin (d), section is about 70 m high; D – Lower sequence of the pyroclastic succession of the Maungaika/North Head Surtseyan volcano at the Auckland Volcanic Field.

Figure 6. A – Overview of the half section pyroclastic succession of an emergent volcano (Kaiapo) near the city of Taupo, Taupo Volcanic Zone (hammer is for scale). Arrows indicate coarse-grained juvenile pyroclast-rich layers; B – Lapilli tuff from the middle section of the pyroclastic succession of the Kaiapo volcano. Larger angular juvenile, low-vesicularity pyroclasts arrowed with white. Some orange-brown coloured pyroclasts are juvenile palagonitized pyroclasts (yellow arrows); C – Pyroclastic succession of the Ohakune Volcanic Complex exposes alternating magmatic explosive (m) and phreatomagmatic explosive (ph) pyroclastic deposits (Taupo Volcanic Zone); D – Fine grained tuff of phreatomagmatic origin at Ohakune Volcanic Complex basal section with vertical pipes (arrows).

Outcrop-scale evidence

At outcrop scale pyroclastic deposits that can be inferred to have originated from pyroclastic density currents (PDC) include base surges which are considered indirect indicators for phreatomagmatism (Valentine et al. 2017b). The recognition of abundant PDC deposits is another indirect line of evidence for hydrovolcanic explosive eruptions. PDC-generation is a common process documented with violent, explosive hydrovolcanic eruptions in young sites or directly observed and recorded from known eruptions (Self et al. 1980; Geshi et al. 2003; Doronzo et al. 2017; Ort et al. 2018; Sweeney et al. 2018). However, in recent explosive eruptions identifying the pyroclast textures that may reflect magma-water explosive interaction is compounded by the difficulty identifying juvenile pyroclasts (Pardo et al. 2014).

In the WVF, preserved and exposed pyroclastic successions along with the common volcanic landform elements were used to group volcanic rocks to specific vent types (Németh and White 2003). Detailed facies architecture of the preserved volcanic erosion remnants and the type of volcanic successions preserved in association with the vents were classified into 3 types (Németh and White 2003). Type 1 vents are inferred to be the remnants of scoria cones and are dominated by scoriaceous juvenile lapilli (Németh and White 2003). Occasionally small outcrops of chaotic lapilli tuffs with glassy juvenile pyroclasts and accidental lithics are preserved along dykes indicating that while Type 1 vents likely represent former scoria cones, we cannot rule out that phreatomagmatism was, at times, involved in the eruptions during edifice growth (Németh and White 2003). Recent studies in young Quaternary volcanic fields elsewhere have found similar features on scoria cones and interpreted them as showing the scoria cones had a short phreatomagmatic phase during their initial eruption phase (Murcia et al. 2015). Vent successions dominated by pyroclastic infill are classified as Type 2 vents and are inferred to have been the substructures of phreatomagmatic tuff ring and/or maar volcanoes. Typical pyroclastic facies are non-volcanic-lithic rich, grain- or matrix-supported, massive tuff breccias and lapilli tuffs that consist predominantly of accidental lithic blocks and lapilli (Figure 7A). These facies are often interbedded with unsorted, matrix-supported, diffusely stratified tuff breccia and lapilli tuff beds. Tuff breccias and lapilli tuff rocks with abundant accidental lithic content are volumetrically dominant in these vent remnants and are interpreted to be vent-filling pyroclastic units formed by phreatomagmatic explosive eruptions. Thinly bedded, locally scour-fill, cross stratified, accidental lithic-rich lapilli tuff and tuff beds form decimetre scale domains with uniform bedding dipping toward the centre of the vent. These domains are enclosed within other pyroclastic rocks and indicate syn-eruptive collapse and/or subsidence; these are commonly claimed processes in the formation of maar-diatremes (Lorenz 1986, 1987, 2007; White 1991; Suhr et al. 2004; Suhr et al. 2006; Lorenz and Kurszlaukis 2007; Valentine et al. 2015; Kurszlaukis and Lorenz 2016; Lorenz et al. 2017). The sedimentary textures of the facies suggest deposition from base surges. Many of the pyroclastic rocks recovered from Type 2 vents show evidence of sudden cooling of pyroclasts at every scale suggesting that they formed due to explosive magma – water interaction. The low vesicularity and the blocky shape of the fine lapilli to ash fractions also support fragmentation driven by external forces. Type 3 vent complexes are groups of closely spaced or overlapping vents, many share similar pyroclastic textures to those preserved in Type 2 vents suggesting phreatomagmatic fragmentation. Many of the Type 3 vents are associated with thick and extensive lava flow fields that commonly show intra-crater facies characteristics. Type 3 vents, such along the Pigroot Hill along the Palmerston – Kyeburn SH85, are interpreted to be the remnants of volcanoes comprising adjoining to coalescing maars and tuff rings with explosive and effusive magmatic products (Figure 7B). These volcanic centres often had simultaneous eruptions from more than one of their vents, producing intercalated, mostly proximal beds of explosive and effusive products. Hence, a Type 3 vent complex is a group of coalesced Type 2 and/or Type 1 vents. Type 3 vent complexes locally include deposits formed on the paleo-ground surface adjacent to the volcano and thus represent the best preserved, least eroded volcanic remnants in the field. Some of these (for instance tuff units characterised by dune-bedding and/or accretionary lapilli and/or vesiculated tuff) are typical of medial to distal deposits of phreatomagmatic tuff rings. Different types of scoriaceous, structureless to weakly bedded, tuff breccia, lapilli tuff, and tuff beds are preserved in thick piles (>10 m) as capping units, indicating a ‘dry’ explosive magmatic phase preceded or accompanied the effusion of the lava. Clastogenic lava flows are commonly preserved in dish-like structures. Spatter-rich pyroclastic beds are intercalated with thinly bedded sideromelane-rich pyroclastic beds, suggesting simultaneous deposition from magmatic and phreatomagmatic activity at closely spaced vents (Houghton and Hackett 1984; Houghton and Schmincke 1986; Houghton et al. 1996; Kósik et al. 2016). The pyroclastic deposits of most of the Waipiata vents record initial explosive phreatomagmatic activity fuelled by groundwater, followed by Strombolian-style eruptions (Németh et al. 2012a).

Figure 7. A – Pyroclastic breccia rich in accidental lithic fragments from an exhumed diatreme near Longland Station, Swinburn Plateau (Waipiata Volcanic Field). View is about 20 cm across; B – Base surge units of the Pigroot Hill Volcanic Complex along the Red Cutting Summit (Waipiata Volcanic Field). Camera bag is for scale (12 cm long); C – Ballistic bomb (arrow shows impact direction) causing impact sag on the phreatomagmatic pyroclastic successions of Motukorea/Browns Island (Auckland Volcanic Field). Pen is 14 cm long; D – Cauliflower-shaped bombs (arrows show impact directions) from the basal section of the Orakei Basin maar (Auckland Volcanic Field). Arrows point to the impact direction (50c coin is for scale).

Younger monogenetic fields such as those located in part of the South Auckland, Auckland and Northland Volcanic Fields contain several well-preserved but normally covered and poorly exposed volcanic landforms typical for phreatomagmatic explosive eruptions. These include shallow maars and tuff rings of unusual scoria cones with intermittent phreatomagmatic explosive phases (Hopkins et al. 2020). Younger Cenozoic volcanic fields, mostly in the northern part of the present-day North Island provide opportunity for more detailed geological work to establish the important role of hydrovolcanism and phreatomagmatism in the formation of their volcanoes (Kereszturi et al. 2014). The Quaternary Auckland Volcanic Field is considered to be active and studies to date recognise that explosive phreatomagmatic eruption styles have occurred in nearly all of its known vents (Kereszturi et al. 2014). This has formed a strong basis for current volcanic hazard studies as Auckland is New Zealand's largest city and the economic powerhouse of New Zealand which is located directly over of this active volcanic field (Lindsay et al. 2011; Kereszturi et al. 2013, 2014; Le Corvec et al. 2013; Blake et al. 2017; Leonard et al. 2017; Hayes et al. 2018; Hopkins et al. 2020). Pyroclastic successions from at least 45 of the currently known 55 vents starts with a typical lapilli tuff and tuff deposits considered to be the product of explosive magma – water interaction (Kereszturi et al. 2014). In some cases, such as the Brown Island (Motukorea), Orakei Basin, Waitomokia or Mangataketake the initial pyroclastic successions are typically rich in accidental lithic bombs and blocks from the country rock immediately underlying the region (Figure 7C & D). The initial pyroclastic units consist of plastically deformed, fine-grained tuff beds consistent with deposition in a water-saturated environment. These are ballistic bombs which impacted the underlying pyroclast horizons and are commonly large accidental lithics. In addition, complex cauliflower textured bombs, that commonly enclose xenoliths of sedimentary country rock are also more common in the basal part of the pyroclastic succession, however, more work is required to understand this relationship. The basal layers are generally less than a metre thick and change up-section into alternating coarse-fine-coarse lapilli tuff and tuff beds rich in lapilli strings, scour fill, low angle cross stratification or dune bedding all indicating transportation involving water; base surges are most likely to be responsible for their development (Figure 8A & B).

Figure 8. A – Dune-bedded base surge deposits in the mid-section of the Motukorea/Browns Island tuff ring (Auckland Volcanic Field). Arrow shows transport direction, hammer is for scale; B – Climbing dunes of a base surge succession in the upper section of the tuff ring of the Lake Pupuke maar pyroclastic deposits, Auckland Volcanic Field (Arrow shows transport direction); C – Weathered, palagonitized dune bedded phreatomagmatic tuff from a tuff ring of Pukewairiki/Highbrook Park (Auckland Volcanic Field) (Arrow shows transport direction); D – Weathered accretionary lapilli-rich phreatomagmatic tuff from the southern crater rim of Pukekohe East maar (South Auckland Volcanic Field)

In well-preserved sections accretionary lapilli beds are recognised and their sedimentology records the involvement of magma-water interaction. In the Auckland and South Auckland Volcanic Fields, fine tuff beds are commonly, almost exclusively, composed of accretionary lapilli. These units display thin, parallel bedding dominating the upper part of the pyroclastic successions. In some cases, despite the young ages of most of the Auckland volcanoes, the tephra rings are covered and poorly exposed. Once exposed, the pyroclastic units weather quickly and intensely. However, on freshly exposed surfaces, most common in the Auckland and South Auckland regions, weathering enhances the texture of dune-bedded tuff successions (Figure 8C & D). Weathering of tuffs can make it difficult to establish the eruption styles responsible for the formation of the pyroclastic successions.

The upper sections of the exposed pyroclastic units of Auckland's volcanoes are composed of fine-grained matrix-rich tuff with common clusters of ballistic impact bombs. These beds are almost exclusively composed of fine-grained matrix derived from the underlying fine-grained siliciclastic sediments. From the Auckland and South Auckland region, where thick (10 m +) pyroclastic sections are preserved and accessible, the main part of the succession clearly shows repetitive fine- and coarse-grained pyroclastic bed alterations without significant visible juvenile and accidental lithic content variations. Most of the sections have some characteristically different types of pyroclastic interbedding such as thick scoriaceous units (e.g. Brown Island/Motukorea) (Figure 9A). The top sections of all the known volcanoes across Auckland and South Auckland have a magmatic cap. This gradually develops from the underlying tuff and lapilli tuff dominated successions leaving a transitional zone only a few metres thick (Figure 9B & C). In places the magmatic caps culminate in a coherent lava succession with a clastogenic character in the base (Figure 9D). Across the Auckland and South Auckland region to date no volcano where the magmatic cap is covered by a new pyroclastic succession similar in texture, componentry and appearance to those forming the basal successions of the same volcanoes have been observed.

Figure 9. A – Nearly a metre thick scoriaceous bed (s) in the middle of the phreatomagmatic succession of Motukorea/Browns Island tuff ring indicates intermittent changes in the eruption style during the edifice growth of the tuff ring (Auckland Volcanic Field). Note the normal fault that dissect the section (arrows); B – The Motukorea/Browns Island tuff ring is capped by a thick scoriaceous lapilli and agglomeratic unit (white dashed line) indicating the complete eruption style changes to magmatic explosive in the end of the volcanic activity. Note the scoria cone (sc) in the background filling the crater of the tuff ring (tr); C – Well-exposed section at Te Manurewa o Tamapahore/Wiri Mt tuff ring (Auckland Volcanic Field) showing a capping magmatic succession (m) over the phreatomagmatic pyroclastic units (ph); D – Complex magmatic capping unit at Te Manurewa o Tamapahore/Wiri Mt showing gradual eruption style changes and the growth of a complex lava spatter and scoria cone intercalated with clastogenic lava flows (s1-s2-s3) and the entire section covered by lava flows (lf).

Micro-scale evidence

Undoubtfully the strongest supporting evidence for phreatomagmatic fragmentation are microtextures preserved in the pyroclasts. Juvenile pyroclast vesicularity, vesicle shape and bubble density are the strongest evidence supporting explosive magma-water interaction. The textural characteristics, such as microtextures in fine grained pyroclasts can be used as strong evidence of rapid cooling and brittle fragmentation (Zimanowski et al. 1997a, 1997b; Buettner et al. 1999, 2002). These clasts are typically a few tens of microns across and are inferred to have been derived from the interaction zone where the MFCI took place. Such clasts are commonly referred to as active particles from the MFCI perspective and supports explosive magma-water interaction during eruptions (Heiken 1972, 1974; Aramaki et al. 1986; Zimanowski et al. 1997b; Buettner et al. 1999, 2002; Morrissey et al. 2000; Mastin 2007; Németh 2010; Németh and Cronin 2011; Duerig and Zimanowski 2012; Jordan et al. 2014). The MFCI processes are associated with the finest fraction of juvenile pyroclasts inferred to be particles derived from the interaction zone between molten fuel (magma) and external coolant (water or water saturated sediment) (Zimanowski et al. 1997b) (Figure 10A & B). Textural and sedimentary indicators, such as ash aggregates, accretionary lapilli or vesicular tuffs, are thought to form during phreatomagmatic explosive eruptions (Figure 10C). Their presence and abundance in a tephra layer, however, indicates moist conditions in the transporting agent (either fall or PDC) suggesting presence of water which transforms to steam in the heat of the eruption and then forms ash aggregates. Similarly, the presence of abundant cauliflower shape lapilli and bombs in the pyroclastic deposits supports the presence of water prior to the fragmentation of magma forming rapidly cooled and contracted pyroclasts (Lorenz 1986, 1987) (Figure 10D). Most of the rocks of the facies contain sideromelane with various degrees of palagonite around the rim, low vesicularity and microlite content all indicative for fast cooling rate of the pyroclasts due to hydrovoclanism (Heiken and Wohletz 1986) (Figure 11A). Most of the pyroclastic products of hydrovolcanic eruptions contain a small proportion of opaque volcanic glass called tachylite that cooled slower than sideromelane glass, hence indicating variable chilling and cooling conditions in a complex eruption cloud (Figure 11B). The vent remnants are commonly topped with pyroclastic units rich in lava spatter. These are inferred to indicate exhaustion of the water supply to the explosion sites and hence cessation of phreatomagmatic fragmentation. A sudden increase in mantle-derived xenoliths in the top of the pyroclastic succession prior to the appearance of the magmatic capping units is documented in the WVF suggesting relatively rapid upward movement of the magma in the later stages of the eruptions. Many of the pyroclastic successions of the Cenozoic monogenetic volcanic fields across New Zealand are dominated by pyroclastic rocks rich in accidental lithic fragments (Figure 12A). On micro-scale tuffs are commonly high in minerals or rock fragments typically derived from the underlying country rocks (Németh et al. 2007; Hencz et al. 2017) (Figure 12B). Large juvenile lapilli or bombs often enclose blocky accidental lithic fragments that were entrained from the conduit wall as observed from the Waipiata Volcanic Field, South Auckland and Auckland Volcanic Fields (Figure 12C). In special cases exotic mineral phases such as glauconite (e.g. in diatremes of the Waipiata Volcanic Fields) have been used to infer that glauconitic sandstone cover must have been widespread in areas in the past and been ripped from the conduit wall during the later phase of phreatomagmatic volcanism (Németh 2001) (Figure 12D).

Figure 10. A – SEM image of an interactive particle, a glassy pyroclasts with fractured texture and blocky, non-vesicular appearance indicating MFCI process to be responsible for its formation; B – Close up view of a glassy pyroclasts from Orakei Basin maar showing surface textural features (arrows) suggest brittle fragmentation caused by magma-water explosive interaction; C – Thin section view (plane parallel light) of an accretionary lapilli (ac) from a fine grained tuff of the Crater Hill tuff ring succession. The image is about 1 mm across; D – A peperitic domain as a lapilli in the basal tuff ring sequence of the Mangere Mountain (Auckland Volcanic Field)

Figure 11. A – Slightly palagonitized sideromelane pyroclast in thin section (plane parallel light) from the ejecta ring of the Pukaki Lagoon maar, Auckland Volcanic Field. Note that the pyroclasts has low vesicularity, it is blocky and has abundant microlites. The core of the pyroclasts (bounded by white dashed line) is still fresh while the outer part (o) became dark yellow indicating gradual palagonitization. The image is about 0.5 mm across; B – Tachylite (black) and sideromelane (yellow) glass in thin section (plane paralell light) from the tuff ring of Crater Hill (Auckland Volcanic Field). Note the palagonitization on the sideromelane glass shard in form of colour changes as well as the thin rim around large pyroclasts forming armoured pyroclasts (arrows). The image is about 2 mm across.

Figure 12. A – Typical proximal pyroclastic succession of a phreatomagmatic volcano in the Auckland Volcanic Field. The image was taken from the Waitomokia/Mt Gabriel tuff ring basal section. It is rich in shallow excavated accidental lithic clasts commonly forming distinct bomb horizons (arrows) that can be interpreted as ballistic curtain deposits; B – A typical sieved sample from the fine tuff beds of the Panmure Basin maar ejecta ring succession. The image shows the 125–250 μm size fractions. The grains are mineral phases (mostly quartz) derived from various siliciclastic country rocks that have been ‘recycled’ by the phreatomagmatic explosions and crater excavation. Note the dark volcanic glass shards; C – Sedimentary rock fragment (s) in a cored basaltic bomb from the basal succession of the ejecta ring of the Orakei Basin maar (Auckland Volcanic Field); D – Thin section image of a pyroclastic rock from the Nenthorn Diatreme (Waipiata Volcanic Field). The samples contain abundant glauconite (g) from pre-volcanic marine sediments that were eroded away in the region. The identification of such grains used as proof to state that marine sedimentary cover existed in the region in the time the maar-diatreme formed.

Evidence of hydrovolcanism associated with small volume silicic volcanism along the Taupo Volcanic Zone

The volcanic geology of the Taupo Volcanic Zone (TVZ) indicates that phreatomagmatic, or in broader sense hydrovolcanic activity, were involved in silicic, small to medium-sized eruptions since the 25.4 ka Oruanui caldera-forming eruption of Taupo volcano (Wilson 2001). In contrast, there are only a few known silicic examples, where magma and water interaction was proven prior to the Oruanui eruption for the onshore part of TVZ, even though the paleoenvironmental settings were very similar to the present day (Kósik et al. 2018). A lake side outcrop on Motuoapa Peninsula, southeast shore of Lake Taupo, exposes more than 100 m of coarse PDC and fall dominated pyroclastic deposits (Figure 13A & B) that formed about 80 ka (Kósik 2018; Kósik et al. 2020). The architecture of the remnant pyroclastic cone and sedimentological features of the pyroclastic deposits indicate magma-water interaction occurred during explosive phases of activity. The deposits include accidental lithic rich lapilli tuffs resulting from PDC transportation, and the low vesicularity and micro-texture of juvenile fragments such as micro-fracturing, pitting on glass shard surfaces or fine dust adhering on fine ash grains (Kósik 2018). Paleoenvironmental reconstruction indicates an extensive lacustrine environment existed here between 200–25 ka relating to the ancient Lake Huka (Manville 2001). This is in keeping with the inferred subaqueous to emergent nature of Motuoapa activity. Tuff/pumice cone edifices similar to Motuoapa were also reported from the Haroharo Dome Complex in the Okataina Volcanic Centre such as Pukerimu tuff cone (Nairn 2002) and from the peralkaline Mayor Island volcano located offshore in the Bay of Plenty (Houghton and Wilson 1986).

Figure 13. A – View of the exposed, more than 100 m thick pyroclastic succession of Motuoapa Peninsula from Lake Taupo; B – Coarse PDC and fall-dominated sequence of the middle upper succession of Motuoapa pyroclastic deposits; C – Larger dome of Puketerata Volcanic Complex, which surrounded by an atypical steep-sided tuff ring; D – Wet surge-dominated cohesive layers with a bomb sag of large rhyolitic clast and thin better sorted shower-bedded layers at the base of the Puketerata sequence; E – Chain of maar craters relating to the NE section of the Puketerata fissure eruption. The arrow indicates the location of the Te Hukui diatreme (Figure 3B); F – Block-and-ash-flow deposits found near the top of the western side of the ejecta ring that surrounds the larger dome.

Post-Oruanui activity of Taupo volcano was mostly characterised by small-volume (<0.5km³) sub-Plinian to Plinian rhyolitic eruptions that occurred through Lake Taupo, where some degree of magma-water interaction was present (Wilson 1993). It is the same for the young activity of Okataina, where the initial phases of eruption were usually phreatomagmatic. Their activity was usually characterised by fissure-fed eruptions lasting several months to years (Smith and Houghton 1995; Cole et al. 2010; Houghton et al. 2010), where the erupted volumes of single eruptive episodes ranged between 4–17 km³ and were almost equally shared by effusive dome-forming and sub-Plinian to Plinian explosive activity (Kósik 2018). Phreatomagmatic activity was frequently present during the activity of Okataina Volcanic Centre in the past 45 ka indicated by the deposits of wet pyroclastic surges often containing accretionary lapilli, although their products are volumetrically insignificant.

Outside of the two most active areas of the silicic volcanism-dominated central TVZ (Okataina and Taupo), only a few eruptions were phreatomagmatic. Knowledge about the explosive activity of Ngangiho and Hipaua domes is quite restricted, however, the Puketarata (or Puketerata) Volcanic Complex is extensively studied (Brooker et al. 1993; Kósik et al. 2019), due to the early recognition of a dome surrounded by a well-preserved tuff ring (Figure 13C) (Kósik et al. 2019). Sedimentological observations reveal that initial activity produced fine-grained cohesive surge deposits (Figure 13D) along a 2.5 km long fissure vent. The southwest part of the fissure later produced two lava domes and a composite ejecta ring, whereas the northeast part exhibits the architecture of embryonic activity with a chain of maar craters (Figure 13E). Similar structures with deep excavation are preserved at Okataina such as the Te Whekau crater (Nairn 2002) and at the Whakamaru dome at North Western Dome Complex of Whakamaru caldera (Leonard 2003). At Puketarata, the micro textural observations indicate that fracturing of groundmass glass post-dated the vesiculation and was most likely driven by contraction associated to fast cooling, melt quenching and glass transition, which was estimated to have happened about 620 m below the surface (Kósik et al. 2019). Fragmentation of quenched silicic magma allows the acceleration of pre-mixing with groundwater, essential for MFCI sensu stricto phreatomagmatic eruption (Austin-Erickson et al. 2008). The later stages of activity were dominated by effusive activity, which was frequently interrupted by explosive phases. The juvenile fragments of the fall deposits from these destructive events exhibit low vesicularity (ranges between 0% and 45% with an average 25%), indicating the explosive activity was triggered externally. Destruction of the lava dome usually triggered short-running block-and-ash flows that were mostly deposited in the vicinity of dome (Figure 13F).

Discussion

Hydromagmatism as outlined in the previous sections is an important geological process recognised in various scales of preserved volcanic successions across Zealandia. Hydromagmatism in the eruption styles of many monogenetic volcanoes of Zealandia is a direct consequence of the availability of surface water, abundant ground water, and a general wet climate through the Cenozoic.

At macro, meso and micro-scale there are numerous features commonly associated with magma and water interaction (or water influence upon rising magma), but none of them alone are diagnostic. The micro and macro textural features (juvenile pyroclast vesicularity, shape and pyroclastic deposit/rock componentry alongside with various bedding features such as cross bedding, dune bedding or ballistic curtain deposits), together with the 3D facies architecture and the geological context, provide evidence that hydrovolcanic processes have taken place.

Through most of the Cenozoic the microcontinent of Zealandia experienced a warm – temperate climate with humid conditions providing abundant surface waters and water-rich subsurface conditions (Buening et al. 1998; Nelson and Cooke 2001; Lindqvist and Lee 2009; Land et al. 2010; Mildenhall et al. 2014; Prebble et al. 2017). While there is no complete climate record through the Cenozoic to link precisely the type of volcanism with the climatic conditions at the time, we have a fairly good climate and paleohydrogeology record for the Miocene when a large number of volcanic fields formed, mostly onshore in the proto-South Island and Chatham Rise regions (Field et al. 1989; Schiøler et al. 2011, 2014; Arnot and Bland 2016). Based on the terrestrial fossil record during the Miocene, New Zealand experienced significant shifts in average annual temperature but stayed within the conditions facilitating hydrovolcanism. Well-preserved floral assemblages recovered from the Dunedin region revealed that during the mid/late Miocene climate, during the Dunedin volcano episode, was warm-temperate to subtropical (mean annual temperature [MAT]: 17–19 degrees C) at Double Hill, similar to modern coastal Queensland, and cool- to warm-temperate (MAT: similar to 12–14 degrees C) at Kaikorai Valley, similar to modern northern New Zealand (Reichgelt et al. 2016). These conditions provide the perfect hydrogeological setting that can heavily influence small-volume, monogenetic volcanism, and provide the emergence of abundant volcanoes formed by hydrovolcanic processes.

Farther inland, in the area of the WVF for instance, paleobotany and paleoclimatic research has revealed a diverse floral assemblage of sub-tropical species in a water-rich environment that likely affected the style of volcanism in that area to be more hydrovolcanic (Lindqvist and Lee 2009; Kaulfuss et al. 2014; Mildenhall et al. 2014; Fox et al. 2015; Kaulfuss and Dlussky 2015; Lee et al. 2016; Jones et al. 2017).

During the Cenozoic, at the maximum submergence of Zealandia, the regions with the thickest continental crust, were covered by a shelf-like, shallow sea into which small volumes of mafic magma were erupted and formed small-volume shallow marine to emergent volcanoes such as those in the Chatham Islands’ Red Bluff Tuff (Sorrentino et al. 2011; Stilwell and Consoli 2012) or the erosional remnant of East Otago's Surtseyan tuff cones and subaqueous pyroclastic mounds (Cas et al. 1989; Moorhouse et al. 2015).

Using the modern analogy, Cenozoic monogenetic volcanism in the Zealandia context shows similarities to other well-studied volcanic regions formed in or near sea level coastal plains and shallow shelfs showing a complete spectrum from shallow marine, emergent to terrestrial volcanism that is influenced heavily by external water (Németh et al. 2010). Cenozoic monogenetic volcanism in Zealandia shows strong physical volcanology similarities to volcanic regions formed in the Miocene to Pleistocene time in the Pannonian Basin in Central Europe where the geodynamics change from a subduction-dominated, narrow marine basin to a broad coastal region that includes intracontinental mafic volcanism through a near-plate boundary setting in low-lying, well-drained fluvio-lacustrine terrestrial environment (Wijbrans et al. 2007; Lexa et al. 2010; Kereszturi et al. 2011; Kovács et al. 2020).

Analogies have recently been used for identifying analogous volcanic fields where the eruption scenarios of individual volcano growth can be viewed as similar to those in the most recent monogenetic volcanic fields in the northern part of the North Island of New Zealand. Comparison of the New Zealand's monogenetic volcanic fields and their volcanic eruption styles with those in Jeju Island, Korea (Brenna et al. 2010, 2011; Sohn et al. 2012), in western Saudi Arabia (Moufti et al. 2015), Japan (Geshi et al. 2011; Geshi et al. 2019) or Argentina (Risso et al. 2008) provides vital insight to the full spectrum of eruption styles that a typical low-lying, well-drained mafic-dominated volcanic system can produce. Terrestrial hydrovolcanic edifices, particularly those where the original volcanic landforms are still recognisable are good sites that can be used to understand cratering (e.g. maar-diatreme growth) processes, or pyroclastic density current transportation and deposition. So far, little research have been conducted to define the relative ratios of various lithofacies preserved in a tuff ring succession surrounding a broad crater (e.g. maar) to help understand the relative explosion depth, the estimated explosion energy release and the processes of crater growth (Graettinger et al. 2015, 2016; Valentine et al. 2015; Macorps et al. 2016; Graettinger and Valentine 2017; Graettinger 2018; Sonder et al. 2018). Most tuff ring successions of the youngest hydrovolcanic edifices formed by phreatomagmatic explosive eruptions resemble natural examples abundant in bedded, dune-bedded and stratified tuffs and lapilli tuffs. Most of the New Zealand examples have initial tuff breccia horizons (Németh et al. 2010, 2012b). In this respect the New Zealand examples resemble the facies architecture similar of Ubehebe maars (California) (Champion et al. 2018), Zuni Salt Lake maar (Utah) (Shoemaker 1957; Onken and Forman 2017), Tihany maars (Hungary) (Németh et al. 2001) or Suwolbong tuff ring (Jeju Island, Korea) (Sohn and Chough 1989). Similar facies architecture is recorded in large scale experiments where the explosion locus was deeper than the optimal scaled depth (Valentine et al. 2015; Graettinger and Valentine 2017). No such systematic study has been undertaken on the tephra rings of New Zealand phreatomagmatic volcanoes. This could be a new and valuable direction for future research because of its relevance to understanding the volcanic hazard such eruptions can pose.

Conclusion

In this overview we provide an up-to-date critical review of hydrovolcanism and its relevance to the volcanic evolution of Zealandia during the Cenozoic. The region has been geographically isolated during these 65 million years and was comprised of a gradually submerging continental landmass that at maximum submergence consisted of an archipelago of island refugia, surrounded by a massive water body and located mostly in a temperate climatic zone. Abundant surface water fed the resultant hydrovolcanism of the monogenetic volcanic field evolution. In near-sea level coastal regions, maars, tuff rings and associated scoria cones formed as ascending dykes interacted with the available water. In contrast emergent to fully subaqueous volcanism dominated those offshore regions, where magma found its way to the surface. If we are looking in a broader Zealandia context and adding the buried or offshore volcanic remnants in the argument, we can say that magma-water influenced volcanism is widespread in the region (Herzer 1995; Bischoff et al. 2019a, 2019b, 2019c). As increasing new data become available from various sedimentary basins worldwide, we can see that in a low-lying coastal environment typical hydrovolcanic landforms can grow, quickly be buried and preserved in a shallow subaqueous sedimentary basin succession along with subaqueous hydromagmatic macroforms developed over the lifespan of the volcanic fields in such geo-environments (Panisova et al. 2018). Hence New Zealand provides a natural laboratory for studying the full spectrum of eruption styles, from fully subaqueous to terrestrial, in time and space. However, the preservation potential of many of New Zealand's Cenozoic monogenetic volcanoes on land is not great due intense weathering and erosion that accompanies the same environmental conditions. In contrast hydrovolcanic macroforms in offshore regions are commonly preserved as volcanic mounds in the sedimentary record of the post-Cretaceous rock record across Zealandia (Bischoff et al. 2020). In pre-Pliocene volcanoes, commonly only the core of the crater or upper conduit zones are preserved as exposed diatremes or core of emergent to fully subaqueous volcanoes. There are a few exceptions such as the Oamaru Surtseyan volcanoes, which still have recognisable crater and near-vent pyroclastic successions exposed in the present day landscape (Moorhouse et al. 2015). In addition, young volcanic landforms are commonly covered by vegetation or are undergoing deep chemical weathering of juvenile pyroclasts thus reducing their usefulness as global reference volcanoes. In this respect we conclude that while the hydrovolcanic landforms in New Zealand are important researching them needs to be undertaken in a global perspective using overseas analogues if we are to better understand the geological information they hold.

Acknowledgements

We grateful for the invitation of the NZJGG to contribute to the IAVCEI 2022 special issue to provide an overview of Cenozoic hydrovolcanism in the Zealandia context. We thankful for all the suggestions and ideas colleagues provided in the past decades to be able to provide a concise overview on this topic. Many thanks for Julie Palmer (Massey University) for reading the original draft and make good suggestions on presentation style and context. The excellent formal Journal reviews by Rosie Cole, Alan Bischoff and an Anonymous reviewer as well as the editorial work of James Scott are greatly appreciated.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.