Roadmap | The review process | Home
   Cook-Austral chain
Arago Seamount

The Cook-Austral volcanic chain

Alain Bonneville

Institut de Physique du Globe de Paris

The Cook-Austral volcanic chain is located on the southern part of the Pacific plate, in a region of anomalous shallow seafloor known as the South Pacific superswell [McNutt & Fischer, 1987]. This region also corresponds to a broad geochemical anomaly called SOPITA (South Pacific Isotopic and Thermal Anomaly, [Staudigel et al., 1991]). The Cook-Austral chain extends to the northwest for > 2,200 km from Macdonald Seamount, an active submarine volcano, to the island of Aitutaki (Figures 1a & b). The chain comprises 11 islands and two atolls with little area above sea level (the largest island is 70 km2). Although oriented roughly in the direction of present Pacific plate motion (11 cm/yr in a N 115° direction), the pattern of both the aerial and submarine volcanoes is rather complex. Two recent geophysical cruises, one at the southern end of the chain [McNutt et al., 1997], and the other over the central and northern parts [Bonneville et al., 2002] have revealed the complexity of the overlapping volcanism.

Figure 1a. Shaded view of seafloor topography for the Cook-Austral region (see upper inset for location). Map projection is made along direction of present motion of Pacific plate (from right to left). White lines correspond to crustal magnetic anomalies, and their corresponding ages are in white numerals. Black diamonds represent places where K-Ar or Ar/Ar ages are known. These ages, in Ma, are printed in black. Click on image for enlargement.

Figure 1b. Three-dimensional view of the Cook-Austral region. Click on image for enlargement.

The morphology and geometry of the island groups suggest the existence of two distinct volcanic alignments. The Aitutaki-Mauke Islands group (Rimatara, Rurutu, Tubuai, Raivavae, and Président Thiers bank) form the northeast alignment. Rarotonga and Mangaia Islands, Neilson Bank, Rapa, Marotiri, and Macdonald Seamount, the only known active volcano, form the southwest branch.

The age of the oceanic crust along the chain ranges from ca. 35 Ma to 80 Ma [Mayes et al., 1990]. Several good K-Ar or Ar/Ar ages have been measured for almost all the islands, on seamounts in the Taukina and Ngatemato chains and in the northern Austral region (Figure 1a). In the north of the Cook-Austral region, Rarotonga has a younger age of 1.1 Ma [Duncan & McDougall, 1976]. At Aitutaki, a 1.2 Ma age coexists with a 8.5 Ma stage [Turner & Jarrard, 1982]. In Rurutu, two different volcanic stages have been identified, an old one at 12 Ma [Duncan & McDougall, 1976], compatible with the progression in ages along the northeastern volcanic alignment, and a young one at 1.1 Ma.

Figure 2. Three-dimensional view of seafloor in vicinity of Arago Seamount. The location corresponds to the black box in Figure 1a.

The initial construction stage of Rurutu can be linked to the magmatic source that formed Tubuai, as it has the same petrologic and geochemical characteristics (see below) and the distance between the two islands is compatible with the absolute Pacific plate motion at that time. However, the source of the later 1.1 Ma volcanic event on Rurutu must be sought on a seamount 130 km southeast of Rurutu [Bonneville et al., 2002; Figure 2], the Arago Seamount, which was named after the French Navy ship that discovered it in 1993. Notably, numerous cones exist between Rurutu and Arago in this 4,500-m-deep basin, but no clear crustal swell seems to be associated with this axis.

Figure 3. Hotspot-track reconstruction for the three groups of volcanoes identified (see text). Black numbers indicate time (in Ma) along each track. The northernmost track starts at 6.5 Ma for the last volcanic event known on this track. The active Macdonald and recently active Arago Seamounts are represented by black stars. Click on image for enlargement.

It is interesting to reconstruct the apparent path followed by a hotspot on the seafloor, moving its present location back in time. We used the set of stage poles proposed by Wessel & Kroenke [1997] completed by the pole previously proposed by [Yan & Kroenke, 1993] for the period 0-3 Ma. If we use a 100-km-wide track as representative of the zone of influence of a given hotspot source, we clearly see in Figure 3 that the Macdonald hotspot could not have generated the northern Austral Islands. Furthermore, its track fits well with the 19 Ma age of Mangaia and with the K-Ar age of 9 Ma obtained on a seamount during the same cruise. On the other hand, the northern Austral Islands can be explained well by a hotspot source that probably stopped producing magma at Raivavae ca. 6.5 Ma. The Arago track fits quite well with Rurutu’s age and Cook Island ages. The Cook-Austral archipelago comprises at least 3 volcanic chains from south to north – the Macdonald track (in yellow on Figure 3) from Macdonald seamount (0 Ma) to Mangaia (19 Ma), the Arago track (in pink on Figure 3) from Arago seamount (0.2 Ma) to Atiu Island (8 Ma) and the Raivavae track (in light blue on Figure 3) from Raivavae Island (6.5 Ma) to Rurutu Island and nearby seamounts (12 Ma). Rarotonga and Aitutaki islands are also located at the beginning of a new volcanic track but more data are needed in order to to propose a valid reconstruction for this northwestern part of the Cook-Austral archipelago. Note that volcanoes are often not located exactly on track, which suggests lithospheric control rather than a change in the location of the source.

Figure 4. 208Pb/204Pb vs. 206Pb/204Pb isotope diagram for published data from Austral Islands basalts. NHRL – Northern Hemisphere reference line; EMI and EMII – enriched-mantle sources I and II, respectively; HIMU – high-µ [(U + Th)/Pb] mantle source.

The geochemistry data (Figure 4) are in good agreement with the geodynamical reconstruction proposed above. Although great heterogeneity is observed on the scale of the superswell, relative homogeneity occurs on the scale of each hotspot track. The Arago track appears to be the most homogeneous and could be related to the young Rurutu volcanic episode (1 Ma). Rimatara Island samples, although dated at 27 Ma, clearly belong to the Young Rurutu episode. Rimatara Island thus probably belongs to this track and the existing radiometric age, which has already been questioned by several authors, is likely an error.


Arago Seamount is the most recent surface expression of the hotspot responsible for the recent volcanic activity at Rurutu and probably for other volcanoes in the Cook Islands chain. The only possible track for Macdonald hotspot is along a southeastern path and it could not have supplied the magma for the northern Austral Islands. For these latter islands an extinct magmatic source close to the Austral Fracture Zone is required. With this new scheme added to the already complicated geological history of this region, two common characteristics of all the hotspot tracks identified so far may be highlighted:

  • a short lifespan (< 20 My);
  • several hotspots are active simultaneously on a region of the seafloor < 2,000 km in diameter.

The future goal is to identify the mechanism that could cause this apparent rythmicity of the volcanism in both space and time. These short-lived and closely spaced hotspots are not consistent with discrete mantle plumes of deep origin. A preliminary hypothesis favors more local upwelling in the upper mantle strongly influenced by weaknesses in the lithosphere.


  • Bonneville, A., R. Le Suavé, L. Audin, V. Clouard, L. Dosso, P.-Y. Gillot, P. Janney, A. Jordahl Kelsey, and K. Maamaatuaiahutapu, Arago Seamount : The missing hot spot found in the Austral Islands, Geology, 30 (11), 1023-1026, 2002.
  • Duncan, R.A., and I. McDougall, Linear volcanism in French Polynesia, J. Volc. Geotherm. Res., 1, 197-227, 1976.
  • Mayes, C.L., L.A. Lawver, and D.T. Sandwell, Tectonic history and new isochron chart of the South Pacific, J. Geophys. Res., 95 (B6), 8543-8567, 1990.
  • McNutt, M.K., D.W. Caress, J. Reynolds, K.A. Jordahl, and R.A. Duncan, Failure of plume theory to explain midplate volcanism in the Southern Austral islands, Nature, 389, 479-482, 1997.
  • McNutt, M.K., and K.M. Fischer, The south Pacific Superswell, in Seamounts, Islands and Atolls, edited by B. Keating, P. Fryer, R. Batiza, and G. Boethlert, pp. 25-34, AGU, Washington D.C., 1987.
  • Staudigel, H., K.-H. Park, M. Pringle, J.L. Rubenstone, W.H.F. Smith, and A. Zindler, The longevity of the South Pacific isotopic and thermal anomaly, Earth Planet. Sci. Lett., 102, 24-44, 1991.
  • Turner, D.L., and R.D. Jarrard, K/Ar dating of the Cook Austral island chain; a test of the hot spot hypothesis., J. Volc. Geotherm. Res., 12, 187-220, 1982.
  • Wessel, P., and L. Kroenke, A geometric technique for relocating hotspots and refining absolute plate motions, Nature, 387, 365-369, 1997.
  • Yan, C.Y., and L.W. Kroenke, A plate tectonic reconstruction of the Southwest Pacific, 0-100 Ma, in Proc. Ocean Drilling Project, Scientific Results, pp. 697-709, 1993.