Suzanna H.A. van de Lagemaat, Peter J.J. Kamp, Lydian M. Boschman, Douwe J.J. van Hinsbergen 

Following hundreds of millions of years of subduction in all circum-Pacific margins, the Pacific Plate started to share a mid-ocean ridge connection with continental Antarctica during a Late Cretaceous south Pacific plate reorganization. This reorganization was associated with the cessation of subduction of the remnants of the Phoenix Plate along the Zealandia margin of East Gondwana, but estimates for the age of this cessation from global plate reconstructions (~86 Ma) are significantly younger than those based on overriding plate geological records (105-100 Ma). To find where this discrepancy comes from, we first evaluate whether incorporating the latest available marine magnetic anomaly interpretations change the plate kinematic estimate for the end of convergence. We then identify ways to reconcile the outcome of the reconstruction with geological records of subduction along the Gondwana margin of New Zealand and New Caledonia. We focus on the plate kinematic evolution of the Phoenix Plate from 150 Ma onward, from its original spreading relative to the Pacific Plate, through its break-up during emplacement of the Ontong Java Nui Large Igneous Province into four plates (Manihiki, Hikurangi, Chasca, and Aluk), through to the end of their subduction below East Gondwana, to today. Our updated reconstruction is in line with previous compilations in demonstrating that as much as 800-1100 km of convergence occurred between the Pacific Plate and Zealandia after 100 Ma, which was accommodated until 90-85 Ma. Even more convergence occurred at the New Zealand sector owing to spreading of the Hikurangi Plate relative to the Pacific Plate at the Osbourn Trough, with the most recent age constraints suggesting that spreading may have continued until 79 Ma. The end of subduction below most of East Gondwana coincides with a change in relative plate motion between the Pacific Plate and East Gondwana from westerly to northerly, of which the cause remains unknown. In addition, the arrival of the Hikurangi Plateau in the subduction zone occurred independent from, and did not likely cause, the change in Pacific Plate motion. Finally, our plate reconstruction suggests that the previously identified geochemical change in the New Zealand arc around 105-100 Ma that was considered evidence of subduction cessation, may have been caused by Aluk-Hikurangi ridge subduction instead. The final stages of convergence before subduction cessation must have been accommodated by subduction without or with less accretion. This is common in oceanic subduction zones but makes dating the cessation of subduction from geological records alone challenging.