Towards a robust and consistent middle Eocene astronomical timescale

International audienceUntil now, the middle Eocene has remained a poorly constrained interval of efforts to produce an astrochronological timescale for the entire Cenozoic. This has given rise to a so-called "Eocene astronomical timescale gap" (Vandenberghe et al., 2012). A high-resolution astrochronological calibration for this interval has proven to be difficult to realize, mainly because carbonate-rich deep-marine sequences of this age are scarce. In this paper, we present records from middle Eocene carbonate-rich sequences from the North Atlantic Southeast Newfoundland Ridge (IODP Exp. 342, Sites U1408 and U1410), of which the cyclical sedimentary patterns allow for an orbital calibration of the geologic timescale between ∼38 and ∼48 Ma. These carbonate-rich cyclic sediments at Sites U1408 and U1410 were deposited as drift deposits and exhibit prominent lithological alternations (couplets) between greenish nannofossil-rich clay and white nannofossil ooze. The principal lithological couplet is driven by the obliquity of Earth's axial tilt, and the intensity of their expression is modulated by a cyclicity of about 173 kyr. This cyclicity corresponds to the interference of secular frequencies s3 and s6 (related to the precession of nodes of the Earth and Saturn, respectively). This 173-kyr obliquity amplitude modulation cycle is exceptionally well recorded in the XRF (X-ray fluorescence)-derived Ca/Fe ratio. In this work, we first demonstrate the stability of the (s3-s6) cycles using the latest astronomical solutions. Results show that this orbital component is stable back to at least 50 Ma, and can thus serve as a powerful geochronometer in the mid-Eocene portion of the Cenozoic timescale. We then exploit this potential by calibrating the geochronology of the recovered middle Eocene timescale between magnetic polarity Chrons C18n.1n and C21n. Comparison with previous timescales shows similarities, but also notable differences in durations of certain magnetic polarity chrons. We present a revision of previous astronomical timescales from the Equatorial and South Atlantic, to overcome the differences between different mid-Eocene astrochronologies. Using our new records from the North Atlantic, combined with existing records from the South Atlantic (ODP Site 1263 and Hole 702B) and Equatorial Atlantic (ODP Site 1260), we revise the durations of magnetic polarity Chrons C18n.1n to C21n, thereby arriving at a robust and self-consistent closure of the middle Eocene astronomical timescale gap