Viscosity structure of the oceanic lithosphere inferred from the differential late Quaternary sea-level variations for the southern Cook Islands

Late Quaternary sea-level variations for the southern Cook Islands such as Rarotonga and Mangaia provide information on the time-dependent crustal movement due to viscoelastic arching in response to loading by the Pleistocene volcanic island of Rarotonga, The lithospheric responses to both external and internal loads have been investigated to estimate the viscosity of the lower part of the lithosphere and to examine the initial stage of swell formation. Detailed observations of sea-level variations for the past 125 kyr indicate that the crustal uplift for Mangaia is greater than 10 m, while Rarotonga was apparently stable for this period. The following geophysical implications for the lithospheric rheology and loading model are derived from these observations, The observed differential crustal movement implies that the viscous relaxation associated with this volcanic loading is still proceeding in the lithosphere. The layer supporting stresses has therefore been migrating with time from weaker lower zones into the stronger upper zones for a lithosphere with a depth-dependent viscosity structure. This fact provides an important constraint on the viscosity of the lower part of lithosphere. The observation that Rarotonga has been apparently stable for this period is indicative of a local buoyant internal load in the upper mantle. This load may be related to small-scale and secondary convection in the asthenosphere. Surface uplift due to an internal load is therefore required to cancel the subsidence by volcanic loading. This problem has been examined for two simplified background density models. One is a model in which the density of the lithosphere is equal to that of the asthenosphere. For this model, very large mass anomalies which are 10 times larger than the external load are required beneath the lithosphere in order to explain the observed differential crustal movement of the islands. For an earth model for which the density of the lithosphere is greater than that of the asthenosphere, which is possible for mature oceanic lithosphere, the observed differential crustal movement is explained for an internal-load model with density anomalies of less than 20 kg m(-3). The volume of the internal load is at most twice the volume of the external load. A high-viscosity layer with an effective viscosity of 10(24) pa s and with a thickness of greater than 60 km is required beneath the top elastic layer with a thickness of 10-15 km. The thickness of thermal lithosphere estimated by the plate age of this region is approximately 80-90 km, regardless of the age-thickness relationship adopted. It is therefore suggested that the major part of the thermal lithosphere is composed of a viscoelastic layer with an effective viscosity of 10(24) pa s and with a relaxation time of 1 Myr.