The Depth Distribution of Mantle Serpentinisation at Magma Poor Rifted Margins: Geophysical Evidence from the Iberian, Newfoundland and Nova Scotia Margins

Basement sampling studies reveal exhumed serpentinised mantle between thinned continental crust and basaltic oceanic crust in the ocean-continent transition (OCT) of magma-poor rifted margins. We have made estimates of the density and depth extent of serpentinised mantle at the Iberia, Newfoundland and Nova Scotia margins by comparing seismic refraction and gravity inversion Moho depths for these margins. Moho depth has been derived using a 3D spectral domain gravity inversion method which incorporates a lithosphere thermal gravity anomaly correction. The adjusted crustal basement density required by the gravity inversion to give a gravity Moho depth that matches the seismically observed depth of the 8.0 km/s P wave velocity has then been determined. Within the OCT of the rifted margins sampled by IAM9, SCREECH 3 and SMART 1 & 2, where serpentinised exhumed mantle is believed to exist, the reconciliation of seismic (8 km/s depth) and gravity inversion Moho depth requires a mean density of serpentinised mantle of between 3000 and 3100 kg/m3. Proposed density-depth relationships for serpentinisation (Cole et al. 2002; Skelton et al. 2005) on IAM 9 show a surface layer of high serpentinisation of thickness 1 - 1.5 km, beneath which serpentinisation decays exponentially with depth reaching approximately 5% at 5 km depth. Densities determined by reconciliation of gravity and seismic estimates of Moho depth are consistent with seismological estimates of serpentinisation with depth for IAM 9 (Iberian margin) from Cole et al. (2002) and Skelton et al. (2005). The average densities of serpentinised mantle determined from comparison of seismic (8 km/s) and gravity Moho depths for SCREECH 3, SMART 1 and SMART 2 are also consistent with the serpentinisation-depth relationship determined seismically for IAM9 which suggests that a similar distribution of serpentinisation with depth exists for these margins. The best fit between serpentinised mantle density determined by comparison of seismic (8 km/s) and gravity Moho depth, and from a density-depth relationship derived from Skelton et al. (2005), is achieved assuming a magnetite content of 4.5%. Additional tests suggest that the degree of serpentinisation is controlled by the depth beneath sea-bed, not depth within mantle beneath Moho. The proposed serpentinisation-depth relationships and gravity inversion suggest that there is little mantle serpentinisation deeper than 5 km, and in which case the average density of serpentinised mantle is 3000 and 3100 kgm-3. We predict that where serpentinised mantle penetrates deeper than ~2.5 km then its average density exceeds oceanic and continental crustal basement density. This work also suggests that it is unlikely that significant serpentinisation occurs under thinned continent crust at the OCT, and that if it occurs then the average density of serpentinised mantle is greater than that of crustal basement.