The dense root of the Iceland crust O Ł lafur Gudmundsson  Danish Lithosphere Centre, Geocentre Copenhagen, Òster Voldgade 10 L, 1350 Copenhagen K, Denmark Received 5 July 2002; received in revised form 22 November 2002; accepted 25 November 2002 Abstract Bathymetry and topography in the North Atlantic Ocean around Iceland are compared to estimates of crustal thickness in the area. Iceland lies much lower than expected based on crustal thickness. This suggests an anomalously low density contrast between crust and mantle beneath Iceland [W. Menke, Geophys. Res. Lett. 26 (1999) 1215^1218]. The relationship between bathymetry and depth to Moho along ridges adjacent to Iceland suggests a normal density contrast there. Continuity of this relationship leads to the conclusion that most of the change occurs in the crust, i.e. the abnormally low density contrast between crust and mantle within Iceland is primarily due to a heavy crust, not light mantle. Gravity modeling also requires the average density of the crust to be unusually high. I argue that this is in fact not abnormal. The lower crust in Iceland is denser because of a higher degree and/or depth of melting beneath Iceland than at adjacent ridges, because of phase transformations occurring within the thick crust and possibly due to fractionation processes in the crust. ß 2002 Elsevier Science B.V. All rights reserved. Keywords: Iceland; density; isostasy; gravity; melting; phase transformations 1. Introduction Recent seismic studies in Iceland have revealed a very thick crust. Three independent types of seismic constraints from wide-angle Moho re£ec- tions, receiver functions, and surface-wave disper- sion, provide consistent evidence of a crustal thickness of about 40 km where it is thickest in central Iceland [2^4]. In the meantime Iceland has received considerable attention as a potential ex- ample of a ridge-centered plume. Ito and cowork- ers have modeled the interaction of an active plume upwelling and passive spreading [5,6]. By comparing their model results with ¢eld observa- tions in and around Iceland they conclude that, if the mantle is anhydrous, a relatively broad plume conduit underlies Iceland, about 200 km in radius. If, on the other hand, the mantle contains a sig- ni¢cant amount of water, a much more vigorous and narrower plume conduit is required. The ex- istence of a plume conduit is in general agreement with the low-velocity anomaly imaged by [4,7] beneath Iceland at depths between 200 and 400 km. The bathymetric anomaly around Iceland and the Bouguer gravity anomaly centered on Ice- land are reasonably matched along the ridge axis with crustal thickness contributing about 70% of the mass de¢cit to explain the elevation of Iceland 0012-821X/02/$ ^ see front matter ß 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S0012-821X(02)01110-X * Tel.: +45-3814-2651; Fax: +45-3311-0878. E-mail address: og@dlc.ku.dk (O. Gudmundsson). Earth and Planetary Science Letters 206 (2003) 427^440 www.elsevier.com/locate/epsl