An age-depth model for Lake Bosumtwi (Ghana) to reconstruct one million years of West African climate and environmental change

Situated within a 1.07 million-year-old meteorite crater, Lake Bosumtwi in Ghana stands as a pivotal location for comprehending climatic, ecological and environmental fluctuations within the sub-Saharan region of West Africa. The region's susceptibility to seasonal environmental shifts and climate oscillations is heightened by the annual movements of the tropical rain belt driven by atmospheric circulation. Yet, there is no satisfying age-depth model available for the entire sedimentary sequence strongly limiting our understanding of changes in this circulation pattern and associated (broad-scale) environmental responses during the last million years in the local to regional context of Lake Bosumtwi. To overcome this, we statistically examine the cyclicity in total natural gamma ray (NGR) data on a core from the lake's centre and create a cyclostratigraphic age-depth model. The calculated maximum age of 946 ka agrees well with the meteorite impact age (∼10 % offset). In order to refine this purely statistical approach, we also perform a correlative age-depth model using 33 tie points accounting for the complexity of climatic and environmental imprints to the NGR record that may exceed direct insolation related effects. Special attention is paid to the core's robustly dated (¹⁴C, OSL, U/Th) uppermost part covering the last 200 ka. Here, high NGR and co-varying K counts coincide with warm periods (except of the water-saturated and unconsolidated Holocene part) and the inverse for glacials and stadials. Based on this, we define tie points for correlating our NGR data to the age-depth model of a NE Atlantic SST record. Comparing our results to the correlation target, other global climate records and Sahara dust flux data reveals striking similarities and supports a proxy understanding with increased in wash of K-enriched terrigenous material from the crater rims in warm and moist periods (high NGR) and K-depleted dust input in stadials possibly contributing to low NGR values in addition to reduced input of K-enriched sediments from the crater rims. Our correlative age model results in precession amplitudes matching eccentricity well, providing further support especially because an over-tuning is unlikely with the used 33 tie points. Overall we provide crucial chronological context to numerous datasets along with environmental constrains that can be used to study the potential habitat availability of early anatomically modern humans in West Africa.