Jogvan Hansen, Dougal Jerram, Christopher Ottley, Mike Widdowson

Abstract: The Paleocene lava succession of the Faroe Islands Basalt Group (FIBG), which is a part of the North Atlantic Igneous Province (NAIP), is intruded by numerous basaltic sills. These can be grouped into three main categories according to their geochemical characteristics: A low-TiO2 sill category (TiO2 = 0.7-0.9), a relatively high-TiO2 sill category (TiO2 = 1.95-2.6) and an intermediate-TiO2 sill that displays major element compositions laying between the other two categories. Mantle normalised plots for the high-TiO2 and low-TiO2 sills display relatively uniform flat LREE trends, with (Ce/Sm)N ratios ranging from 1.11 to 1.27. Mantle normalised HREE trends representing the low-TiO2 sills show relatively low (Sm/Yb)N ratios ranging from 1.09 to 1.26, as compared to ratios of 1.59 to 2.38 for the high-TiO2 sills. The intermediate-TiO2 Morskranes Sill is LREE depleted with an average (Ce/Sm)N ratio of ~0.6, but displays a flat HREE trend with an average (Sm/Yb)N ratio of ~1.
Mantle normalised trace elements of the low-TiO2 sill samples define positive Eu and Sr anomalies, whereas trace elements representing high-TiO2 sill samples display negative anomalies, thus probably indicating the involvement of plagioclase at some stage(s) during magma genesis. Different Nb and Ta anomalies (positive versus negative) in many high-TiO2 versus low-TiO2 sill samples indicate that their respective mantle sources were affected by metasomatism prior to partial melting. Two main isotope discriminators can be detected amongst the actual sill samples: The intermediate-TiO2 sill displays noticeably lower 87Sr/86Sr, 206Pb/204Pb and 208Pb/204Pb ratios relative to both the high-TiO2 and the low-TiO2 sill samples. Pb isotope compositions displayed by local contaminated basaltic lavas imply that some of these assimilated distinct crustal material from E Greenland or basement from NW Britain, while others probably assimilated only distinct E Greenland type of crustal material. However, a third distinct crustal source of E Greenland or Rockall-type basement may be required in order to explain the range in compositions of some lead isotopes of the intermediate-TiO2 Morskranes Sill, unless these were caused by hitherto undetected isotopic heterogeneities in the mantle.
Geochemical modelling indicates that primary melts, which subsequently evolved to Faroese high-TiO2 sills, could have formed by a range of ~4 to 7.5 % batch melting of moderately fertile lherzolitic sourches, while a range of 16 to 21 % batch melting of likewise fertile sources seems to be required in order to produce Faroese low-TiO2 sills. The moderately fertile source inferred for Faroese low-TiO2 sill samples in particular are at odds with the depleted sources envisaged for their low-TiO2 basaltic host-rocks. Primary melts that gave rise to the intermediate-TiO2 sill samples could have formed by a range of 6 to 7 % batch melting of a depleted mantle source, probably with a composition comparable to sources that gave rise to local low-TiO2 and intermediate-TiO2 host-rocks.
The modelling points to garnet-free residues during mantle melting in order to produce the primary melts that subsequently evolved to most of the Faroese sills. These are envisaged to have formed by batch melting of mantle materials comparable in composition to materials reported for the sub-continental lithospheric mantle (SCLM) at depths of  85 km. Their relative enrichments in LREE (and LILE in general), as well as their varying Nb and Ta anomalies, may well point to sources affected by metasomatism. These could have been caused by earlier Cenozoic local basaltic magmatism or by earlier events linked to the complex geologic history of the N Atlantic area.