|
Constraining the Iceland Low-velocity Anomaly to Test Causal Hypotheses
R. M. Allen University of Wisconsin-Madison J. Tromp California Institute of Technology AGU 2002 fall meeting For several decades the mantle plume hypothesis has been the most prevalent model cited as the cause of the geophysical and geochemical anomalies around Iceland. Recently the hypothesis has come under increasing pressure as various workers argue that the apparent anomalies are not particularly anomalous, and alternative models, operating entirely within the upper mantle, are presented as the causal mechanism. Seismic tomography provides the only method of "imaging" 3D mantle structure in situ, and three seismograph networks have been deployed across Iceland to collect the necessary data. Several velocity images of the Icelandic mantle using traveltime delays recorded by these regional networks have been published; all use ray-theoretical tomographic inversion techniques. To first-order they are consistent, showing a low velocity anomaly with a horizontal width of a few hundred kilometers, and extending from the surface to the maximum depth of resolution around ~400 km. However, small variations in the structure imaged, and inherent distortions associated with the inversion techniques, have provided for a range of interpretations. Here we present constraints on the geometry and amplitude of the low-velocity anomaly beneath Iceland. They are the results of tests using both ray-theoretical and full 3D wave propagation methods designed to test the extent to which the anomaly can be bent and squeezed. Ray-theoretical tests to squeeze the low-velocity anomaly both horizontally and vertically show that low-velocities are required to at least 350 km depth. They also suggest that the traveltime dataset could be satisfied by a narrow low velocity column, 100 km in diameter. Using the Spectral Element Method (SEM) we calculate synthetic waveforms and traveltime delays for stations across Iceland given various anomaly geometries. The SEM delay maps show a much broader delay footprint than ray-theoretical calculations would predict, implying that the Iceland anomaly could be about half the width of the ray-theoretical tomography results. However, the amplitude of the delays is also significantly reduced for narrow anomalies. We conclude that the Iceland low-velocity anomaly must extent to at least 350 km depth, is 100 to 200 km wide and does not extend laterally along the North Atlantic Ridge. |