Geoscience Reference
In-Depth Information
to semi-anchored ones and semi-anchored dunes will be-
come mobile in the dune and steppe areas of China. In con-
trast, the desertification is predicted to reverse in central
and eastern regions of arid and semi-arid China. Overall
the changes are controlled mainly by precipitation regimes
and potential evaporation.
While Thomas, Knight and Wiggs (2005) and Wang
et al.
(2009) are important marker studies of the poten-
tial for dune reactivation under future climate change,
they also draw attention to the importance of monitoring
contemporary dune processes with a view to improving
the empirical indices on which dune mobilisation was
based. This is partly because of the complexity of dune
processes, with a number of recent studies revealing that
active and fixed dunes may coexist under similar climatic
conditions (Tsoar, 2005; Yizhaq, Ashkenazy and Tsoar,
2007). A related issue is that mobility indices are cali-
brated against current climate conditions whereas dunes
response may lag behind climatic change (Hugenholtz and
Wolfe, 2005). An even more fundamental limitation is the
skill with which climate models simulate precipitation and
the worrying differences in this simulation, particularly at
the regional scale, among different climate models. These
stubborn problems cannot be resolved without concerted
effort. The dramatic results of Thomas, Knight and Wiggs
(2005) and Wang
et al.
(2009) show that such an effort is
clearly warranted.
performance (not something that is possible in future cli-
mate change projections), there remains large uncertainty
in the future projections of dust loadings. In the Fourth
Assessment Report of the IPCC, Meehl
et al.
(2007) note
the conflicting outcomes of several studies. Mahowald
and Luo (2003), for example, find that the global atmo-
spheric burden of soil dust aerosols could decrease by
between 20 and 60 % in association with climate change,
a result confirmed by a similar, later study (Mahowald
et al.
, 2006). Dust production in this set of studies was
mainly shown to respond to changes in the source areas,
which result from vegetation changes. Winds or soil mois-
ture changes are argued to be less important. Tegen
et al.
(2004a, 2004b), using model runs by the European Centre
for Medium Range Weather Forecasts/Max Planck Insti-
tute for Meteorology Atmospheric GCM (ECHAM4) and
UKMO-HadCM3 forced by identical greenhouse gas sce-
narios, demonstrate that the future projections of dust
loadings are model-dependent, with the loadings increas-
ing in one model but decreasing in the other. These latter
two simulations included changes to atmospheric condi-
tions and vegetation cover. Tegen
et al.
(2004a, 2004b) are,
however, able to conclude that dust loadings from agricul-
ture are small (less than 10 %) of the total dust loadings,
even with a maximum estimate of increased agricultural
area by 2050.
Using the UKMO-HadAM3 atmospheric model, which
included feedback of vegetation on climate, Woodward,
Roberts and Betts (2005) point to an order of magnitude
increase in atmospheric dust loadings over 100 years up
to 2100 as a result of desertification and climate change.
More recent work by the same group, but based on a
new version of the Met Office Hadley Centre coupled
climate-carbon cycle model, shows a severe drying over
the Amazon, which, through carbon feedbacks on the at-
mosphere following forest loss, leads to the Amazon be-
coming an important global dust source (Betts, Sanderson
and Woodward, 2008).
There are a number of key problems with numerical
models of the dust cycle that contribute to the ambiguity
evident in changes to projected dust loadings as reported
in the Fourth Assessment Report:
24.5
Climate change and dust
The development of numerical models that simulate the
dust cycle has been ongoing since at least the early 1990s,
making this component of aeolian geomorphology among
the most sophisticated and complex methodologically.
Recognition that numerical weather prediction benefits
from inclusion of dust aerosols (e.g. Perez
et al.
, 2006;
Rodwell and Jung, 2008) has led to a number of exper-
imental and operational dust forecast systems, e.g. the
Navy Aerosol Analysis and Prediction System (NAAPS),
the global and regional Earth-System (atmosphere) mon-
itoring using satellite and in situ data (GEMS) at the
European Centre for Medium Range Weather Forecasts
(Morcrette
et al.
, 2007) and the dust regional atmospheric
model (DREAM) (Nickovic
et al.
, 2001). Simultaneously,
inclusion of dust aerosols has been an important step in the
further development of climate and Earth System models.
In the Fourth Assessment Report of the IPCC, dust was
included in several of the 20 or so climate models used in
climate change projections.
While operational dust forecasting schemes provide the
Dust models are effectively unconstrained at dust source
regions since the available data on dust concentrations
needed to calibrate models at source does not exist.
It is widely recognised from analysis of satellite data
that dust emission occurs primarily in a relatively small
number of extremely remote preferential source regions
(Herman
et al.
, 1997; Prospero
et al.
, 2002; Torres
et al.
,
2002; Washington
et al.
, 2003). Current estimates of
