Description: Little is known about the effects that subducting an oceanic large igneous province (LIP) has on the petrogenesis of submarine arc volcanoes and their geochemical composition. The southern Kermadec arc represents a rare example where an LIP—the Hikurangi Plateau—is currently subducting and where its effect on mantle composition, element recycling and arc volcanism can be studied. We present mineral chemistry and whole-rock major and trace element, and Sr–Nd–Pb isotope data from samples recovered from the southern Kermadec arc volcanoes Rumble II East and Rumble II West, together with shipboard gravity and magnetic measurements. The Rumble II volcanoes (including a volcanic cone ∼10 km further west) form an ∼23 km long arc–backarc transect located ∼250 km north of New Zealand above the subducting Hikurangi Plateau. Although only a short distance apart, rocks from the two volcanoes have different mineral and whole-rock geochemical compositions. Lavas from Rumble II East are predominantly basaltic and contain primitive olivine phenocrysts (≤Fo91), high-Mg# clinopyroxene (≤96) and anorthitic plagioclase (≤An97). Geochemically these lavas are very diverse and cover a spectrum from low Th/Yb (〈0·15) at high Ba/Th (〉1014) to higher Th/Yb (〉0·15) at lower Ba/Th (〈844). This spectrum, together with 206Pb/204Pb and 143Nd/144Nd in the range of 18·74–18·83 and 0·51309–0·51298 respectively (at similar to slightly elevated 87Sr/86Sr), suggests a mantle wedge that has undergone previous melt extraction and significant fluid addition from the subducting Pacific Plate and that contains sediment and HIMU-type Hikurangi Plateau components. The geochemistry of the sediment–HIMU-type components is exemplified in an olivine pyroxenite (e.g. 206Pb/204Pb = 20·02; 87Sr/86Sr = 0·70516; 143Nd/144Nd = 0·5126). We propose that the olivine pyroxenite formed through melt or fluid–rock metasomatism and represents the first direct evidence of a near Moho arc mantle rock that shows the imprint from a subducting HIMU-type (Hikurangi) seamount. Conversely, lavas from Rumble II West and the cone ∼10 km to the west are generally more silica rich than Rumble II East lavas and mainly contain plagioclase with less ortho- and clinopyroxene + olivine phenocrysts. The low Ba/Th (〈470) and 206Pb/204Pb (〈18·74), a range of 143Nd/144Nd (0·51297–0·51307) and elevated Th/Yb (0·13–0·39) in these lavas can best be explained by minor sediment input into a less depleted mantle wedge. In addition, the geochemical composition of the Rumble II West lavas does not require involvement of a Hikurangi component, placing a spatial limit on Hikurangi material influencing regional melt generation beneath the backarc. Supported by a gravity model requiring two distinct magma chambers, the different geochemical compositions of Rumble II East and West lavas are inconsistent with a shared magma plumbing system. The different geochemical compositions of lavas from the two Rumble II volcanoes furthermore demonstrate that across-arc geochemical heterogeneities can occur within a few kilometres and may originate from both a geochemically heterogeneous mantle wedge and Moho transition layer, recording inherited geochemical heterogeneities beneath the volcanoes.

Description: The age and style of opening of the Havre Trough back-arc system is uncertain due to a lack of geochronologic constraints for the region. 40Ar/39Ar dating of 19 volcanic rocks from across the southern Havre Trough and Kermadec Arc was conducted in three laboratories to provide age constraints on the system. The results are integrated and interpreted as suggesting that this subduction system is young (〈2 Ma) and coeval with opening of the continental Taupo Volcanic Zone of New Zealand. Arc magmatism was broadly concurrent across the breadth of the Havre Trough.