On the mechanisms of climate, vegetation and lake change in Late Quaternary Africa

Late Quaternary African palaeoclimate records imply humid to arid shifts on varying geographical and temporal scales. Regional responses to climate forcings are not uniform, and the mechanisms responsible for geographically varying sensitivity to orbital insolation forcing, internal climate feedbacks and teleconnections with North Atlantic climate events remain unclear. Here a data synthesis and suite of model simulations are presented which help to quantify the magnitude and direction of environment change across Late Quaternary Northern and Eastern Africa.
During the deglaciation and at several times prior to the LGM numerous palaeoclimate archives record rapid, millennial scale, climate oscillations in time with Heinrich events, which caused complex patterns of humidity change. Mid-Holocene orbital conditions strengthened the North African monsoon generating large lakes and significant vegetation in what is now the Sahara desert.
To investigate these climate oscillations idealised Heinrich event simulations are analysed over Africa, these simulations generate drying in Northern Africa and wetting in the south. In East Africa complex high amplitude shifts from Heinrich induced drying to wetting appear to explain regional inconsistencies in the palaeoclimate record.
A suite of experiments with dynamic vegetation performed across the Holocene capture the enhanced monsoon and vegetation change. The contributions of SST and vegetation are shown to substantially improve GCM and BIOME4 estimates of the mid-Holocene precipitation and vegetation conditions, accounting for approximately 2 degrees north shifts in the Sahara's southern boundary. The largest contribution from vegetation occurs at 9 kyr.
To reduce error in GCM simulations of mid-Holocene North Africa a dynamic atmosphere-vegetation-lake model is proposed. The HYDRA lake model was limited by poor GCM estimates of its input forcing fields reducing its applicability without further research. Nevertheless climate-lake simulations suggest a significant feedback effect from the hydrosphere, and we show that vegetation feedbacks stimulate significant lake development.