Desert dunes - Climatic Change in Deserts - page 114

the specialist texts (e.g., Pye and Tsoar, 1990 ; Pye and Lancaster, 1993 ; Lancaster,
1995 ; Tchakerian, 1995 ; Goudie et al., 1999 ; Warren, 2013 ) and the references listed
at the beginning of this section.

8.2 World distribution of active and stable dunes
Perhaps the most interesting aspect of desert dunes from a paleoclimatic perspective is
the present-day global distribution of active and stable dunes ( Figures 8.3 and 8.4 ). For
example, along the southern margins of the Sahara, a wide belt of now vegetated and
stable dunes extends for at least 500 km to the south of the present-day southern limit
of active dunes in the Sahara, which coincides with the 200 mm isohyet, presumably
indicating greater-than-present aridity when those fixed dunes were active. Likewise,
in the Thar Desert of north-west India, vegetated and now stable dunes extend to the
east far beyond the present limit of mobile dunes into areas that receive well in excess
of 450 mm of rain a year (Goudie et al., 1973 ). The critical questions here are when
the fixed dunes south of the Sahara and the eastern Thar Desert were last active, and
whether they were active at the same time. On a global basis, the area covered by
active dunes during the last major phase of dune activity seems to show that the deserts
were once far more extensive than they are today, at a time when the world climate
was thought to be much drier in the intertropical zone (Sarnthein, 1978 ; Sarnthein
et al., 1981 ). Indeed, Sarnthein ( 1978 ) commented quite explicitly that active sand
dunes presently occupy about 10 per cent of the land area between 30

° N and 30 ° S

±

but that during the Last Glacial Maximum (21 2 ka), the corresponding percentage

of area covered by active dunes between those two latitudes was probably closer to
50 per cent, especially when taking into account the greater land area resulting from
lower sea levels.
In order to test this model with appropriate rigour, two things are necessary. One
is a robust chronology of dune formation, particularly because many linear dunes are
polygenic (Fujioka et al., 2009 ;Cohenetal., 2010a ; Singhvi et al., 2010 ; Fujioka and
Chappell, 2011 ;Yangetal., 2011b ). Difficulties arise in interpreting whether the ages
obtained by luminescence dating of dune sand samples collected from different depths
within the dune reflect peak dune activity or simply the final transitional phase from
maximum movement to waning accumulation (Swezey, 2001 ; Swezey, 2003 ). The
second requirement is a set of well-dated climate proxies that are able to show quite
independently of the dune evidence whether the local or regional climate was indeed
more arid than it is today in that region. Thomas ( 1997 ) has provided a comprehensive
assessment of such evidence, which seems on balance to favour the glacial aridity
model of Sarnthein ( 1978 ), at least for the hot tropical deserts, but the reality on a
more local scale is in fact far more complex.
It is also not always obvious whether the inferred desert expansion involved a
physical migration of dunes into what are now semi-arid and subhumid regions