Magnetic studies of magma-supply and sea-floor metamorphism: Troodos ophiolite dikes

Dikes of the eastern Troodos ophiolite of Cyprus intruded at slow ocean-spreading axes with dips ranging up to 15° from vertical and with bimodal strikes (now NE–SW and N–S due to post-88 Ma sinistral microplate rotation). Varied dike orientations may represent local stress fields during dike-crack propagation but do not influence the spatial-distributions or orientation-distributions of dikes' magnetic fabrics, nor of their palaeomagnetic signals. Anisotropy of magnetic susceptibility (AMS) integrates mineral orientation-distributions from each of 1289 specimens sampled from dikes at 356 sites over ∼400 km2 in the eastern Troodos ophiolite of Cyprus. In 90% of dikes, AMS fabrics define a foliation (kMAX–kINT) parallel to dike walls and a lineation (kMAX) that varies regionally and systematically. Magma-flow alignment of accessory magnetite controls the AMS with a subordinate contribution from the mafic silicate matrix that is reduced in anisotropy by sea-floor metamorphism. Titanomagnetite has less influence on anisotropy. Occasionally, intermediate and minimum susceptibility axes are switched so as to be incompatible with the kinematically reasonable flow plane but maximum susceptibility (kMAX) still defines the magmatic flow axis. Such blended subfabrics of kinematically compatible mafic-silicate and misaligned multidomain magnetite subfabrics; are rare. Areas of steep magma flow (kMAX plunge ≥ 70°) and of shallow magma-flow alternate in a systematic and gradual spatial pattern. Foci of steep flow were spaced ∼4 km parallel to the spreading axes and ∼6 km perpendicular to the spreading axes. Ridge-parallel separation of steep flow suggest the spacing of magma-feeders to the dikes whereas ridge-perpendicular spacing of 6 km at a spreading rate of 50 mm/a implies the magma sources may have been active for ∼240 Ka. The magma feeders feeding dikes may have been ≤ 2 km in diameter. Stable paleomagnetic vectors, in some cases verified by reversal tests, are retained by magnetite and titanomagnetite. In all specimens, the stable components were isolated by three cycles of low-temperature demagnetization (LTD) followed by ≥ 10 steps of incremental thermal demagnetization (TD). 47% of primary A-components [338.2 /+ 57.2 n = 207, α95 = 3.9; mean TUB = 397 ± 8 °C] are overprinted by a B-component [341.4 /+ 63.5, n = 96, α95 = 8.7; mean TUB = 182 ± 11 °C]. A- and B-components are ubiquitous and shared equally by the N–S and NE–SW striking dikes. A-component unblocking temperatures (TUB) are zoned subparallel to the fossil spreading axis. Their spatial pattern is consistent with chemical remagnetization at some certain off-axis distance determined by sea-floor spreading. A-components indicate less microplate rotation and more northerly palaeolatitudes that are consistent with metamorphic remagnetization after some spreading from the ridge-axis. Thus, their magnetizations are younger than those of the overlying volcanic sequence for which ChRMs are commonly reported as ∼274 /+ 33 (88 Ma).