Fig. 2: Central Arabian speleothems reveal the episodic occurrence of humid intervals over the past 8 Myr.
From: Recurrent humid phases in Arabia over the past 8 million years
a, The δ18O (‰ VPDB) benthic stack records covering 8,000 ka (ref. 15) to present relative to epoch and age (S4 refers to Stage Four). b, Temporal distribution of key events and hominin information based on appearance dates and their associated assemblage uncertainty (grey bars). c,d, Summary of terrestrial environmental change for northeastern Africa including the percentage of C4 plants in eastern Africa from plant wax biomarkers44 (c) and pedogenic carbonate δ13C (‰ VPDB) (d)45. e, The timing of the two major fossil deposits in Arabia: the Baynunah Formation2 in the late Miocene and the Nefud Desert deposits during the middle-to-late Pleistocene associated with stone tools and fauna. f, U–Pb-derived ages (grey-filled circles) from this study, and U–Th-derived ages using solution multi-collector inductively-coupled-plasma mass spectrometry (yellow squares) and laser ablation multi-collector inductively-coupled-plasma mass spectrometry (red diamonds) with the 95% confidence interval uncertainty indicated by the horizontal error bars. One age with an uncertainty greater than >±1.5 Myr is not shown but can be found in Supplementary Data 1. g, Frequency distribution statistics of both U–Pb-derived and U–Th-derived ages from this study using 300-kyr bins (the average uncertainty of the U–Pb ages). As not all speleothem layers were dated, frequencies should not be used as a record of palaeoclimate.
