Modelling the effects of CO2 on C3 and C4 grass competition during the mid-Pleistocene transition in South Africa

Michaela Ecker,Douglas Kelley,Hiromitsu Sato

doi:10.1038/s41598-020-72614-2

Michaela Ecker, Douglas Kelley + Show 1 more

Open Access

https://doi.org/10.1038/s41598-020-72614-2

Copy DOI

Abstract

Palaeoenvironmental reconstructions of the interior of South Africa show a wetter environment than today and a non-analogous vegetation structure in the Early Pleistocene. This includes the presence of grasses following both C3 and C4 photosynthetic pathways, whereas C3 grasses decline after the mid-Pleistocene transition (MPT, c. 1.2–0.8 Ma). However, the local terrestrial proxy record cannot distinguish between the potential drivers of these vegetation changes. In this study we show that low glacial CO2 levels, similar to those at the MPT, lead to the local decline of C3 grasses under conditions of decreased water availability, using a vegetation model (LPX) driven by Atmosphere–Ocean coupled General Climate Model climate reconstructions. We modelled vegetation for glacial climates under different levels of CO2 and fire regimes and find evidence that a combination of low CO2 and changed seasonality is driving the changes in grass cover, whereas fire has little influence on the ratio of C3:C4 grasses. Our results suggest the prevalence of a less vegetated landscape with limited, seasonal water availability, which could potentially explain the much sparser mid-Pleistocene archaeological record in the southern Kalahari.

Highlights

Palaeoenvironmental reconstructions of the interior of South Africa show a wetter environment than today and a non-analogous vegetation structure in the Early Pleistocene
South Africa can be divided into three seasonal rainfall zones, which in turn strongly influence the distribution of v egetation[1] (Supplementary Fig. S3)
Palaeoenvironmental studies hypothesized that areas that are in the summer rainfall zone of the savanna biome might have been under the influence of an expanding winter rainfall zone during Pleistocene glacial conditions[2,3,4]

Summary

Introduction

We use the LPX-DGVM21 driven by Atmosphere–Ocean coupled General Climate Model (AOGCM) climate output[22], to test three possible drivers forcing vegetation change that have been suggested in the Wonderwerk Cave palaeoenvironmental study[18]: (1) atmospheric CO2 levels, (2) rainfall seasonality, and (3) disturbance through fire. Under identical glacial climate reconstructions from four different AOGCM outputs, we modulated C O2 levels to determine its effect on vegetation cover, the expansion or contraction of C 3 and C 4 grasses.

Results

Conclusion