Phytoliths, pollen, and microcharcoal from the Baringo Basin, Kenya reveal savanna dynamics during the Plio-Pleistocene transition

Chad L. Yost, Sarah J. Ivory, Alan L. Deino, Nathan M. Rabideaux, John D. Kingston, Andrew S. Cohen

Research output: Contribution to journalArticlepeer-review

18 Scopus citations

Abstract

As part of the Hominin Sites and Paleolakes Drilling Project (HSPDP), phytoliths, pollen, and microcharcoal were examined from the 228 m (3.29 to 2.56 Ma) Baringo-Tugen Hills-Barsemoi drill core (BTB13). A total of 652 samples were collected at ~10 to 32 cm intervals, corresponding to sub-millennial to millennial scale temporal resolution. Microcharcoal was well-preserved throughout the core and often peaked in abundance ~5 kyr before and after insolation peaks. Phytolith preservation varied between excellent to total dissolution in alternating intervals throughout the core. Pollen was rarely preserved. These combined datasets indicate that prior to ~3.1 Ma, woody cover fluctuated between open savanna (< 40% cover), woodland (40–80% cover), and forest (> 80% cover) at typically precessional (19–23 kyr) periodicities. During the mid-Piacenzian Warm Period (MPWP; 3.26–3.01 Ma), intervals with exceptionally high microcharcoal abundance suggest that regional turnover from wooded to open habitats was driven in part by fire. After ~3.1 Ma, low-elevation woody cover likely never exceeded 40%, with oscillations between mesic tall-grass vs. xeric short-grass savanna at precessional periodicities. Mesic C4 tall-grass (Panicoideae) peaked in abundance during insolation maxima, whereas xeric C4 short-grass (Chloridoideae) peaked during insolation minima. The onset of Northern Hemisphere glaciation (NHG) at ~2.75 Ma coincided with the appearance of deep lake phases and increases in grass density and fire frequency. Spectral analysis and intervals with well-preserved phytoliths indicate that precession and interhemispheric insolation gradients influenced vegetation via their effects on equatorial precipitation and fire. This study fills a crucial gap in Pliocene vegetation reconstructions from the East African Rift Valley and its associated hominin localities. It also provides orbitally resolved regional vegetation data useful in paleodata–model comparisons for the onset of the MPWP (which is often used as an analog for future warming) and NHG.

Original languageEnglish (US)
Article number109779
JournalPalaeogeography, Palaeoclimatology, Palaeoecology
Volume570
DOIs
StateAccepted/In press - 2020

Bibliographical note

Funding Information:
We thank Kaitlynn Walker, Lizbeth Murillo, Murtadha Al Malallah, Sirawitch Nantanoi, and Marie Mathis from the University of Arizona for assisting with the phytolith/microcharcoal and pollen separations. This research was conducted as part of the Hominin Sites and Paleolakes Drilling Program ( HSPDP ). Initial BTB13 core processing and sampling was carried out at the National Lacustrine Core Facility ( LacCore ), University of Minnesota , USA. BTB13 is archived at LacCore. Funding for this work was provided by National Science Foundation (NSF) grants EAR-1123942 , BCS-1241859 , EAR-1338553 , and the International Continental Scientific Drilling Program (ICDP) . We thank Doris Barboni, Jay Quade, and Steve Archer for commenting on an early version of this manuscript, special issue guest editors Jenni Scott and Mark Sier, and two anonymous reviewers for comments that greatly improved the manuscript. This is publication #25 of the Hominin Sites and Paleolakes Drilling Project.

Publisher Copyright:
© 2020 Elsevier B.V.

Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.

Keywords

  • Hominins
  • Microfossils
  • Mid-Piacenzian Warm Period
  • Northern Hemisphere glaciation
  • Paleofire
  • Tugen Hills

Continental Scientific Drilling Facility tags

  • HSPDP-BTB

Fingerprint

Dive into the research topics of 'Phytoliths, pollen, and microcharcoal from the Baringo Basin, Kenya reveal savanna dynamics during the Plio-Pleistocene transition'. Together they form a unique fingerprint.

Cite this