Geoscience Reference
In-Depth Information
intervals (Stone and Thomas,
2008
). The greater abundance of OSL ages revealed
that the dunes of the south-west Kalahari had been partly active throughout the last
120 ka and suggested that in this region, the dunes had been close to their threshold of
reactivation throughout much of the late Quaternary, so earlier work invoking discrete
phases of dune sand accumulation needed to be reassessed (Stone and Thomas,
2008
).
Furthermore, aridity may not have been the main cause of sand movement and could
be outweighed by changes in wind velocity (Chase,
2009
), which means that dunes
are unreliable as indicators of past climate (Chase and Brewer,
2009
). Perhaps what
is now needed to resolve this impasse is a greater focus on the fossil soils within
the dunes and any microfossils within them. Dunes will always be hard to interpret,
because they reflect a number of different controlling agents, notably wind speed,
plant cover, sand supply and effective precipitation, all of which could affect renewed
dune movement.
Given the difficulties involved in using desert dunes to reconstruct past climatic
changes, greater attention is now focussed on other lines of evidence, including
lake deposits, speleothems and tufas, and pollen. The Makgadikgadi Lake complex
at the distal end of the Okavango River in semi-arid Botswana has its headwaters in
the equatorial highlands of Angola. This lake system is perhaps the best-dated set
of lakes in semi-arid southern Africa, with more than 140 OSL ages now reported
(Huntsman-Mapila et al.,
2006
; Burrough et al.,
2009a
; Burrough et al.,
2009b
). The
Okavango high shorelines have OSL ages of 104, 92, 64, 39, 27, 17 and 8 ka, with
small error terms of only a few ka. Older shorelines of 288, 267 and 131 ka have error
terms between 25 and 16 ka (Burrough et al.,
2009a
). Hydrologic models suggest that
above a certain size, the large lakes could have an influence on both local and regional
climate (Burrough et al.,
2009b
). At present, when the Angolan highlands are wet,
Botswana tends to be dry, and conversely. Huntsman-Mapila et al. (
2006
) considered
that this anti-phase relationship between late Quaternary rainfall in southern and
equatorial Africa also obtained during the late Quaternary. When Lake Ngami, which
was fed by the Okavango, was high from 19 to 17 ka, there were signs of increased
aridity in Botswana, which means that the LGM was arid in Botswana but wet in the
Angolan headwaters of the Okavango.
Brook et al. (
1997
) compiled the
14
C and U-series ages of speleothems and tufas
collected from the summer rainfall zone of southern Africa (Namibia, Botswana,
northern Cape and the Transvaal) and compared them with the ages obtained for
similar deposits across Somalia. Southern Africa was apparently wetter at 202-186,
50-43, 38-35, 31-29, 26-21 and 19-14 ka. Conditions there were wet during late
glacial times and dry during the early Holocene. In Somalia, speleothem, tufa and
rock-shelter sediments indicated wetter conditions in this presently arid region at
260-250, 176-160, 116-113, 87-75, 13, 10, 7.5 and 1.5 ka. From their survey, they
concluded that over the last 35 ka at least, when it was wet in southern Africa it was
dry in Somalia, and vice versa.
