Clark volcano of the Kermadec arc, northeast of New Zealand, is a large stratovolcano comprised of two coalescing volcanic cones; an apparently younger, more coherent, twin-peaked edifice to the northwest and a relatively older, more degraded and tectonized cone to the southeast. High-resolution water column surveys show an active hydrothermal system at the summit of the NW cone largely along a ridge spur connecting the two peaks, with activity also noted at the head of scarps related to sector collapse. Clark is the only known cone volcano along the Kermadec arc to host sulfide mineralization.
Volcano-scale gravity and magnetic surveys over Clark show that it is highly magnetized, and that a strong gravity gradient exists between the two edifices. Modeling suggests that a crustal-scale fault lies between these two edifices, with thinner crust beneath the NW cone. Locations of regional earthquake epicenters show a southwest-northeast trend bisecting the two Clark cones, striking northeastward into Tangaroa volcano. Detailed mapping of magnetics above the NW cone summit shows a highly magnetized “ring structure” ~350 m below the summit that is not apparent in the bathymetry; we believe this structure represents the top of a caldera. Oblate zones of low (weak) magnetization caused by hydrothermal fluid upflow, here termed “burn holes,” form a pattern in the regional magnetization resembling Swiss cheese. Presumably older burn holes occupy the inner margin of the ring structure and show no signs of hydrothermal activity, while younger burn holes are coincident with active venting on the summit.
A combination of mineralogy, geochemistry, and seafloor mapping of the NW cone shows that hydrothermal activity today is largely manifest by widespread diffuse venting, with temperatures ranging between 56° and 106°C. Numerous, small (≤30 cm high) chimneys populate the summit area, with one site host to the ~7-m-tall “Twin Towers” chimneys with maximum vent fluid temperatures of 221°C (pH 4.9), consistent with δ34Sanhydrite-pyrite values indicating formation temperatures of ~228° to 249°C. Mineralization is dominated by pyrite-marcasite-barite-anhydrite. Radiometric dating using the 228Ra/226Ra and 226Ra/Ba methods shows active chimneys to be <20 with most <2 years old. However, the chimneys at Clark show evidence for mixing with, and remobilizing of, barite as old as 19,000 years. This is consistent with Nd and Sr isotope compositions of Clark chimney and sulfate crust samples that indicate mixing of ~40% seawater with a vent fluid derived from low K lavas. Similarly, REE data show the hydrothermal fluids have interacted with a plagioclase-rich source rock.
A holistic approach to the study of the Clark hydrothermal system has revealed a two-stage process whereby a caldera-forming volcanic event preceded a later cone-building event. This ensured a protracted (at least 20 ka yrs) history of hydrothermal activity and associated mineral deposition. If we assume at least 200-m-high walls for the postulated (buried) caldera, then hydrothermal fluids would have exited the seafloor 20 ka years ago at least 550 m deeper than they do today, with fluid discharge temperatures potentially much hotter (~350°C). Subsequent to caldera infilling, relatively porous volcaniclastic and other units making up the cone acted as large-scale filters, enabling ascending hydrothermal fluids to boil and mix with seawater subseafloor, effectively removing the metals (including remobilized Cu) in solution before they reached the seafloor. This has implications for estimates for the metal inventory of seafloor hydrothermal systems pertaining to arc hydrothermal systems.