Date of Award

June 2014

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Earth Sciences


Christopher A. Scholz


Africa, Lakes, Paleoclimate, Turkana

Subject Categories

Physical Sciences and Mathematics


Lake sediments are some of the best archives of continental climate change, particularly in the tropics. This study is focused on three ~10m sediment cores and high-resolution seismic reflection data from Lake Turkana in northern Kenya. Lake Turkana is the world's largest desert lake and the largest lake in the Eastern Branch of the East African Rift System. It is situated at ~2 °N at 360 m elevation and is ~250 km long and ~30 km wide with a mean depth of 35 m. The lake surface receives less than 200 mm yr-1 of rainfall during the twice-annual passing of the Intertropical Convergence Zone via Indian Ocean-derived moisture, and evaporation is >2300 mm yr-1. This study is the first to quantify the climate and deepwater limnologic changes that have occurred in the area during the African Humid Period (AHP) and since the Last Glacial Maximum.

A 20-kyr, multiproxy lake level history was derived from ~1100 km of CHIRP seismic reflection data, in conjunction with gamma ray bulk density, magnetic susceptibility, total organic carbon, total inorganic carbon, core lithology, and scanning XRF data from sediment cores that were chronologically constrained by radiocarbon dates. Two desiccation events occurred at 18.5 and 17 ka when the lake was at least 100 m lower than today, as evidenced by basin-wide, high-amplitude reflections correlated to sand intervals in the sediment cores. Lake level rose abruptly at ~11 ka as interpreted from an increase in organic carbon content and abrupt shift to silt and clay-sized sediment, after which Lake Turkana overflowed into the White Nile River System. AHP highstand conditions lasted until ~5 ka when the lake became a closed basin. The loss of Lake Turkana as a White Nile input likely had significant implications for nascent communities living along the Nile. Lake level has fluctuated but remained at a moderate lowstand since its mid-Holocene closure.

An easterly shift of moisture derived from the Congo Basin has been proposed as an additional moisture source for environments in East Africa during the AHP, as evidenced by an expanding number of hydrogen isotope records from lake basins in East Africa. To determine if Atlantic-derived moisture reached Lake Turkana in the past via the Congo Basin, compound specific (C28 n-alkanoic acids) hydrogen isotopes (δDwax) were measured from the composite sediment core record from Lake Turkana.

The record revealed that δDwax values were depleted by more than -60 / during the AHP. δDwax depletion occurs abruptly at 13.7 ka and is sustained throughout the AHP, and values gradually become more enriched after 7 ka. This depletion suggests that Congo-derived moisture had a significant influence on Turkana precipitation during the AHP, but depletion of that magnitude cannot be explained solely by a change in moisture source. In addition to the source effect, a possible vegetation effect was quantified using stable carbon isotopes of leaf waxes, and the amount effect was estimated using modern Kenyan amount effect coupled with a published basin fill model. It was determined that vegetation can account for up to -17 /, and amount effect estimations range from -17 / to - 24 / of the total AHP depletion. After accounting for these effects, precipitation during the AHP is depleted relative to a 100% increase in precipitation amount, which we suggest is driven by an influx of moisture derived from the Congo Basin. Our calculations suggest that at least 45% of the moisture supplied during the AHP was Congo-derived.

The precipitation-evaporation balance of a lake system is closely related to its heat budget. To quantify changes in temperature through the onset and termination of the AHP, the paleotemperature proxy called TEX86 was used to generate a 14-kyr record of lake surface temperatures for Lake Turkana. This proxy has successfully reconstructed regional and high-latitude paleotemperatures in other large African lakes.

TEX86 temperatures from 14 to 0.4 ka are highly variable and range from 24.3 °C to 28.6 °C with a mean of 25.9 °C. There is a long-term trend within the temperature record that follows mean northern hemisphere peak summer insolation, and a mean temperature of 26.2 °C during the early Holocene decreases to 25.7 °C during the late Holocene. A century-scale fluctuation of ~1 °C persists throughout most of the record which appears to be related to lake mixing processes that overprint the regional climate signal based on fluctuations in metal ratios. While this reduced sensitivity of the TEX86 proxy was an unexpected result from this analysis, it has shown that this proxy may have limitations in its application to well-mixed, tropical, arid-land lake systems. Despite high-frequency variation, TEX86 temperatures remain close to the mean for the duration of the record, which is attributed to the evaporation response of this lake system.


Open Access