Extended Data Fig. 4: Age-distance relationships of the Rurutu-Arago, Samoan, and Louisville plumes.

A) Rurutu-Arago hotspot track. B) Samoa hotspot track. C) Louisville hotspot track. See Background Data Sources For Figures in Methods for data sources. Data shown here are only samples with age determinations or published age estimates (where well constrained by nearby volcanoes^14,29 and Sr-Pb-Nd data to permit color coding. APM models are also included for reference (WK08⁴¹ = Wessel and Kroenke, 2008; D12⁶ = Doubrovine et al., 2012 without plume drift correction; K01¹⁸ = Koppers et al., 2001; K01m = modified Koppers et al., 2001 from this study). The oldest portions of the Hotspot Highway are in good agreement with K01m model predictions; some scatter occurs as a function of plume drift (e.g., Cretaceous portion of the Samoa hotspot; see Extended Data Fig. 8). For the hotspots shown here, the data are consistent with age progressions that can be traced back into the Cretaceous, including the successful 91.3 Ma age determination from our new sample set. Rurutu-Arago has HIMU to FOZO-like compositions, while Samoan volcanoes are EM-type to FOZO in composition. The Rurutu-Arago age progression can be clearly traced into the Wakes and back to ~ 120 Ma. The Samoan plume was active during the Cretaceous, forming the Magellan chain in the West Pacific where EM2 and EM1 compositions consistent with those found in Samoan shield and rejuvenated volcanoes, respectively (see Extended Data Fig. 6 for details). The FOZO Louisville hotspot track and older Ellice Basin Seamounts as well as Seka Seamount; which are likely FOZO-to-DMMlike with Pb isotopes partly overprinted by seawater U ingrowth, are also age-progressive.