Magnetic susceptibility, XRF and color reflectance data of DSDP Hole 72-516F, cores 113 and 114

Deep marine successions of early Campanian age from DSDP site 516F drilled at low paleolatitudes in the South Atlantic reveal distinct sub-Milankovitch variability in addition to precession and eccentricity related variations. Elemental abundance ratios point to a similar 5 climatic origin for these variations and exclude a quadripartite structure - as observed in the Mediterranean Neogene - of the precession related cycles as an explanation for the inferred semi-precession cyclicity in MS. However, the semi-precession cycle itself is likely an artifact, reflecting the first harmonic of the precession signal. The sub-Milankovitch variability is best approximated by a ~ 7 kyr cycle as shown by 10 spectral analysis and bandpass filtering. The presence of sub-Milankovitch cycles with a period similar to that of Heinrich events of the last glacial cycle is consistent with linking the latter to low-latitude climate change caused by a non-linear response to precession induced variations in insolation between the tropics.

XRF iron (Fe) counts of ODP Hole 171-1050A and 171-1051A

Upper Paleocene to lower Eocene sediments drilled at Ocean Drilling Program (ODP) Site 1051 (Blake Nose, off Florida) display well-defined orbital cycles, a detailed magnetic stratigraphy, and a suite of planktonic foraminiferal datums. We derived a cyclostratigraphy by using spectral analysis of high-resolution records of elemental concentrations obtained by an X-ray fluorescence (XRF) Core Scanner. XRF counts of iron serve as a proxy for the relative amount of terrestrial material. Sliding-window spectral analysis, bandpass filtering, and direct counting of precession and obliquity cycles yield minimum durations for magnetic polarity chrons C22 to C26 (~49 to ~61 Ma), calculations of sediment accumulation rates, as well as constraints on the timing of biostratigraphic and climatological events in the vicinity of the Initial Eocene Thermal Maximum (IETM). Durations of polarity chrons as represented in sediments drilled at Site 1051 were estimated using a conservative assignment of 41 k.y. for obliquity cycles and 21 k.y. for precession cycles. Combined polarity chrons C26r and C26n span 3.61 m.y., and chron C25r spans 1.07 m.y. Polarity chron C24r is estimated as 2.877 m.y. The interpretation of polarity chron C24n is ambiguous, but its duration is probably <1.23 m.y. Polarity chron C23r spans 0.53 m.y., chron C23n is 0.74 m.y., and chron C22r is 0.9 m.y. Spectral analysis through this interval indicates that spectral peaks shift through time and are related to changes in sedimentation rate in Site 1051. The sedimentation rates dramatically increased ~200 k.y. after the IETM and remained high for most of chron C24r.

Seismic Data from MSM21/4 to Knipovich Ridge on the western Svalbard margin

The seismic data were acquired north of the Knipovich Ridge on the western Svalbard margin during cruise MSM21/4. They were recorded using a Geometrics GeoEel streamer of either 120 channels (profiles p100-p208) or 88 channels (profiles p300-p805) with a group spacing of 1.56 m and a sampling rate of 2 kHz. A GI-Gun (2×1.7 l) with a main frequency of ~150 Hz was used as a source and operated at a shot interval of 6-8 s. Processing of profiles p100-p208 and p600-p805: Positions for each channel were calculated by backtracking along the profiles from the GI-Gun GPS positions. The shot gathers were analyzed for abnormal amplitudes below the seafloor reflection by comparing neighboring traces in different frequency bands within sliding time windows. To suppress surface-generated water noise, a tau-p filter was applied in the shot gather domain. Common mid-point (CMP) profiles were then generated through crooked-line binning with a CMP spacing of 1.5625 m. A zero-phase band-pass filter with corner frequencies of 60 Hz and 360 Hz was applied to the data. Based on regional velocity information from MCS data [Sarkar, 2012], an interpolated and extrapolated 3D interval velocity model was created below the digitized seafloor reflection of the high-resolution streamer data. This velocity model was used to apply a CMP stack and an amplitude-preserving Kirchhoff post-stack time migration. Processing of profiles p400-p500: Data were sampled at 0.5 ms and sorted into common midpoint (CMP) domain with a bin spacing of 5 m. Normal move out correction was carried out with a velocity of 1500 m s-1 and an Ormsby bandpass filter with corner frequencies at 40, 80, 600 and 1000 Hz was applied. The data were time migrated using the water velocity.