Seismic data from Lake Tana, Ethiopia, has been analysed and the key reflection events have been

picked. These have then been dated using an age depth model from previously obtained cores from

Lake Tana. The airgun data from Lake Tana aided the identification of the bedrock below the

accumulated sediment in the Lake Tana basin. The bedrock was dated to 410 k yr BP ± 22 k yr BP,

and this indicates that there is a possible 410 k yr record of lake level fluctuations within the Lake

Tana sediment. The Seistec boomer data allowed the identification of many sub-bottom features

form 98 k yr BP to present day. The main features are the five P layers which indicate that since 98 k

yr BP there have been five major lowstands in Lake Tana. The features that appear between these P

layers include less intense desiccation layers, downcutting events, large sand ripples and prograding

wedges. These findings are in agreement with those of Bates et al. (2007) and Lamb Bates et al.

(2007) who have also used other seismic records to identify lake level fluctuations in Lake Tana.

The lake level fluctuations identified in the seismic records from Lake Tana correlate well with many

of the lake level fluctuations of other large African lakes, such as Lake Victoria (Johnson et al, 1996),

Lake Challa (Moernaut et al., 2010) and Lakes Tanganyika and Malawi (Scholz and Rosendahl, 1988).

From 98 k yr BP to 91 k yr BP there are only a few other records of the fluctuations that are observed

in Lake Tana. Between 91 k yr BP and 72 k yr BP there is no evidence of any lake level fluctuations,

but from 72 k yr BP to present there are numerous other lake level records that recognise the lake

lowstands observed in Lake Tana. The lowstands in Lake Bosumtwi are isolated as being in very good

agreement with the lake level fluctuations in Lake Tana. This is due to the presence and timing of

lowstands in Lake Bosumtwi (Talbot and Johannessen, 1992; Shanahan et al., 2006; Brooks et al.,

2007) being synchronous with the lowstands of Lake Tana. This may be due to the fact that the SST

of the Atlantic Ocean is a more dominant control over the amount of rainfall over the Lake Tana

catchment area than the SST of the Indian Ocean, despite the proximity of the Indian Ocean to the

Ethiopian highlands.

The features between the P layers display repeating patterns, suggesting that there were two

different depositional environments during the past 98 k yr. From 98 k yr BP k yr to 72.4 k yr BP k yr

BP the deposition environment is exemplified by prograding wedges and from 72.4 k yr BP k yr to

the present day the depositional environment is exemplified by downcutting events. The main

control over the deposition environment appears to come from the amount of variation within the

insolation at the top of the atmosphere at 12 °N, the SST of the Atlantic Ocean and the ice volume of

the northern high latitudes. From 98 k yr BP k yr to 72.4 k yr BP k yr BP there appears to be a great

deal of rapid and large variation within the elements that control rainfall over Lake Tana, which may

have led to greater variations in the level of rainfall over Lake Tana. The rapid nature of these

variations may have meant that at times when rainfall was low; it was not low for long enough for

Lake Tana to desiccate completely, which permitted the prograding wedges to form. This variation

decreases between 72.4 k yr BP k and 15.2 k yr BP during the period where there were many

lowstands and the downcutting features. This decrease in the variation and the reduction in the

speed of these variations of the controlling elements may have resulted in rainfall decreases that

were substantial in length to cause considerable reductions in rainfall and resulted in the repeated

lowstands that appear in the seismic record from Lake Tana. The increase in variability after 15.3 k yr

BP to the present day appears to not have had the significant effect that the previous high levels of

variability appear to have had.

Up to 38 k yr BP, ice volume in the high latitudes appears to exert little control over the quantity of

rainfall over the Lake Tana catchment area. From 38 k yr BP to the present day it appears that

northern high latitude ice volume does exert a control over the amount of rainfall over Lake Tana. It

is also evident that ice volume in the southern high latitudes has no control over the climate over

the Lake Tana region. δ18Oatm from Vostok (Petit et al., 1999) shows that the increases and decreases

in ice volume in Antarctica are not in sync with the lake level fluctuations of Lake Tana. Many of the

lowstands during this time period are synchronous with increases in the ice volume of northern high

latitudes, as exemplified by the δ18Oice values from NGRIP (2004). Heinrich Events also appear to have

an impact on the level of rainfall over the Lake Tana region as the many of the large and small

desiccation events from 38 k yr BP in Lake Tana appear to be associated with Heinrich Events. The

reason that Heinrich Events and ice volume increases in the northern high latitudes has such an

impact on rainfall over the Lake Tana catchment area is that the ice rafting that results from rapid

increases in continental ice volume that causes mass ice rafting (Heinrich, 1988). This ice then melts

and the excess of freshwater input to the north Atlantic alters the SSS and will slowdown the AMOC.

The slowdown of the AMOC will result in an interhemispheric gradient in SST that results in lower

SSTs in the northern hemisphere (Carto et al., 2009). This effect again supports the conclusion that

changes in the Atlantic Ocean have a large impact on the climate over east Africa, despite the large

distance between the Atlantic Ocean and Lake Tana.