<p>Detailed mapping studies of Quaternary stratovolcanoes provide critical frameworks for examining the long-term evolution of magmatic systems and volcanic behaviour. For stratovolcanoes that have experienced glaciation, edifice-forming products also act as climate-proxies from which ice thicknesses can be inferred at specific points in time. One such volcano is Tongariro, which is located in the southern Taupō Volcanic Zone of New Zealand’s North Island. This study presents the results of new detailed mapping, geochronological and geochemical investigations on edifice-forming materials to reconstruct Tongariro’s volcanic and magmatic history which address the following questions: (1) Does ice coverage on stratovolcanoes influence eruptive rates and behaviour (or record completeness)? (2) What is the relationship between magmatism, its expression (i.e. volcanism) and external but related processes such as tectonics? (3) How are intermediate-composition magmas assembled and what controls their diversity? (4) What are the relative proportions of mantle-derived and crust-derived materials in intermediate composition arc magmas? (5) Do genetic relationships exist between andesite and rhyolite magmas in arc settings? Samples from 250 new field localities in under-examined areas of Tongariro were analysed for major oxide, trace element and Sr-Nd-Pb isotope compositions. Analyses were performed on whole-rock, groundmass and xenolith samples. The stratigraphic framework for these geochemical data was established from field observations and 29 new 40Ar/39Ar age determinations, which were synthesised with volume estimates and petrographic observations for all Tongariro map units. Mapping results divide Tongariro into 36 distinct map units (at their greatest level of subdivision) which were organised into formations and constituent members. New 40Ar/39Ar age determinations reveal continuous eruptive activity at Tongariro from at least 230 ka to present, including during glacial periods. This adds to the discovery of an inlier of old basaltic-andesite (512 ± 59 ka) on Tongariro’s NW sector that has an unclear source vent. Hornblende-phyric andesite boulders, mapped into the Tupuna Formation (new), yield the oldest 40Ar/39Ar age determination (304 ± 11 ka) for materials confidently attributed to Tongariro. Tupuna Formation andesites are correlated with Turakina Formation debris flows that were deposited between 349 to 309 ka in the Wanganui Basin, ~100 km south of Tongariro, which indicates that Ruapehu did not exist at this time, at least not in its current form. Tongariro has a total edifice volume of ~90 km3, 19 km3 of which is represented by exposed mapped units. The total ringplain volume immediately adjacent to Tongariro contains ~60 km3 of material. The volume of exposed glacial deposits are no more than 1 km3. During periods of major ice coverage, edifice-building rates on Tongariro were only 17-21 % of edifice-building rates during warmer climatic periods. Because shifts in edifice-building rates do not coincide with changes in erupted compositions, differences in edifice-building rates reflect a preservation bias. Materials erupted during glacial periods were emplaced onto ice masses and conveyed to the ringplain as debris, which explains reduced preservation rates at these times. MgO concentrations in Tongariro stratigraphic units with ages between 230 and 0 ka display successive and irregular cyclicity that occurs over ~10-70 kyr intervals, which reflect episodes of enhanced mafic magma replenishment. During these cycles, more rapid (≤10 kyr) increases in MgO concentrations to ≥5-9 wt% are followed by gradual declines to ~2-5 wt%, with maxima at ~230, ~160, ~117, ~88, ~56, ~35, ~17.5 ka and during the Holocene. Contemporaneous variations in Tongariro and Ruapehu magma compositions (e.g. MgO, Rb/Sr, Sr-Nd-Pb isotope ratios) for the 200-0 ka period coincide with reported zircon growth model-ages in Taupō magmas. This contemporaneity reflects regional tectonic processes that have externally regulated and synchronised the timings of elevated mafic replenishment episodes versus periods of prolonged crustal residence at each of these volcanoes. Isotopic Sr-Nd-Pb data from metasedimentary xenoliths, groundmass separates and whole-rock samples indicate that two or three separate metasedimentary terranes (in the upper 15 km of the crust) were assimilated into Tongariro magmas. These are the Kaweka terrane and the Waipapa or Pahau terranes (or both). Subhorizontal juxtapositioning of these terranes is indicated by the coexistence of multiple terranes in the same eruptive units. Paired whole-rock and groundmass (interstitial melt) samples have effectively equal Sr-Nd-Pb isotope ratios for the complete range of Tongariro compositions. Despite intra-crystal isotopic heterogeneities that are likely widespread, the new data show that crystal fractionation and assimilation occur in approximately equal balance for essentially all Tongariro eruptives. Assimilated country rock accounts for 22-31 wt% of the average Tongariro magma. Initial evolution from a Kakuki basalt-type to a Tongariro Te Rongo Member basaltic-andesite reflects the addition of 17 % assimilated metasedimentary basement with a mass assimilation rate/mass crystal fractionation rate ratio—a.k.a. ‘r value’ of 1.8-3.5. Subsequent evolution from a Te Rongo Member basaltic-andesite to other Tongariro eruptive compositions represents 5-14 % more assimilated crust (r values of ~0.1-1.0). Magma evolution from high (>1) to lower (0.1-1.0) r values can explain the dearth of andesitic melt inclusions in (bulk) andesite magmas observed globally. High relative assimilation rates characterise rapid evolution from basalt to basaltic-andesite bulk compositions which contain andesitic interstitial melts. Thus, andesitic melt inclusions have a reduced chance of being preserved in crystals which can explain their low representation in global datasets.</p>
