A multi-model approach to the spatial and temporal characterization of the African Humid Period

[Data availability] Original datasets are avaliable via the Neotoma Palaeoecological Database (www.neotoma.org). Code used in the analyses presented here are avaliable via: https://github.com/Squirrelmonkey93/Bachelor_project_M.Holewijn.git, and http://www.doi.org/10.6084/m9.figshare.29608400.

During the last c. 20,000 years, African climates experienced temperature shifts related to the last period of global deglaciation and moisture availability changes that defined what has become known as the African Humid Period (AHP) c. 14,800–5500 years ago. However, spatio-temporal variations in climate, and the mechanisms behind them, across Africa remain poorly defined and debated. Here, we first characterize climatic change across tropical Africa (15°N to 15°S) over the last c. 20,000 years based on two independent site-specific modelling approaches at nine locations, using: (i) probability-based reconstructions based on pollen data from nine previously published pollen records, and (ii) climate simulation-based reconstructions based on mechanistic models of the Earth system. Trends in past climate change per site were found to be similar between the two modelling approaches; however, estimates of precipitation were higher in the pollen-based reconstructions when compared with those from the mechanistic model. Given the overall similarity between the two modelling approaches at the sites, we then used the mechanistic model to produce maps of past climate across Africa at 1000-year time slices. Interrogation of the model supports previous suggestions that the AHP in the west was driven by increasing precipitation (c. 13,000 years ago). In the eastern and southern portions of the study region, the AHP signal is more complex, likely driven by the interplay between different climate mechanisms; with the onset of the wettest AHP conditions in eastern Africa not commencing until c. 9000 years ago.

The authors would like to acknowledge the contributors to the Neotoma and GBIF databases whose data were extracted and analyzed in this manuscript. We also thank the data stewards and all those involved in constructing and maintaining these valuable databases. We give special thanks to Brice Gijsbertsen, (Discipline of Geography, University of KwaZulu-Natal) for cartographic support, and two anonymous reviewers whose comments and suggestions help to develop the manuscript. WDG, MC, MF, AH, TH, ML and SK-B were variously supported through the INQUA multi-year project “Mapping Ancient Africa” (#2103). MC was supported by the German Federal Ministry of Education and Research (BMBF) as a Research for Sustainability initiative (FONA; https://www.fona.de/en) through the PalMod Phase II and Phase III project (grant nos. 01LP 1926D, 01LP2308A (MC)). MLF has received financial support from Germany Research Foundation (DFG) grant TR 419/21. MH conducted the pollen data compilation and initial CREST analysis during her research thesis (BSc Informatiekunde, University of Amsterdam, 2022–2023) under the supervision of MC and WDG. ML and AM were funded by the Leverhulme Research Grant RPG-2020-317. SK-B acknowledges funding from Deutsche Forschungsgemeinschaft (DFG) grant KA 4757/10-1.

Peer reviewed