A dynamic model for the Iceland Plume and the North Atlantic based on tomography and gravity data

Abstract

The North Atlantic around Iceland is characterized by a large geoid high and an anomalous shallow ocean floor; both observations are presumably due to upwelling of hot material in the mantle. We extracted this region from a global tomography data set to study the structure of the mantle in more detail. The tomography data reveal a confined low-velocity structure from the core-mantle boundary (CMB) to the upper mantle, stretching from a location at the CMB below southwest Greenland towards the upper mantle in a strongly eastward inclination. In the present study we compare the observed gravity potential field in the North Atlantic with a modelled field based on mantle temperature variations estimated from tomography. Seismic traveltime residuals are converted to temperature variations assuming a linear relation between seismic velocity and density and a pressure-dependent thermal expansivity. We found a maximum excess temperature in the plume conduit at the CMB of ∼250 °C, weakly decreasing towards the phase transition zone (PTZ). In the PTZ a temperature rise of ∼50-70 °C is found, which is in agreement with the latent heat release by the olivine phase transitions. The 3-D temperature field is then used as the driving force for viscous flow in a Cartesian convection model. The model dimensions are chosen four times as large as the tomographic section to allow resolution for long wavelengths and convective return flow. For a number of constant viscosity and temperature- and depth-dependent viscosity cases the dynamic topography, gravity anomaly and geoid undulations are calculated and compared with the EGM96 potential field coefficients in the wavelength range of 400 to 4000 km. The observation data were also corrected for ocean lithosphere cooling and isostatic compensation of continental crust. The best agreement between observation and modelled data (78 per cent fit for geoid and 47 per cent for gravity) is obtained for a temperature-dependent viscosity of about one order of magnitude for 500 °C temperature variation and an increase of viscosity with depth by no more than a factor of 50 from the upper to the lower mantle. The generally good spatial agreement supports the tomographic model, at least for the upper mantle, and indicates that the East Greenland margin as well as the outer Faroer Ridge are dynamically supported. Low-density material west of the Kolbeinsey Ridge might be linked to low-density anomalies below the Greenland Shield. The presence of lower mantle anomalies causes a large-scale geoid high of ∼3 to 5 m in agreement with observations, but our approach cannot further constrain the spatial distribution of anomalies in the lower mantle. © 2004 RAS.