Dust Observations and Climatology - Mineral Dust: A Key Player in the Earth System - page 163

sidering some limitations of the approach (optically thick cloud). Such new
developments highlight that dust transported above clouds is significant at a
global scale and has to be added to previous cloud-free satellite retrievals of
dust.

7.3.3 Interannual Variability and Trends

Although visibility data collected at meteorological stations present potential biases
that require careful uses to properly evaluate dustiness, they represent the longest
time series available (Mahowald et al. 2007 ). Several studies have used them for
analysis of long-term variability of dust events, that is, over periods of 30-40 years
(Sun et al. 2001 ; Mahowald et al. 2007 ), especially in regions where no other
continuous long-term dust observations are available. As an example in Asia, Shao
and Dong ( 2006 ) examine the evolution of the total number of dust days since
1961, as recorded at 175 meteorological stations located in dust-affected areas of
China. Their analysis (Fig. 7.7 ) shows that a marked decrease occurred in the late
1970s and the 1980s, with a clear negative trend until 1997. In contrast, this trend is
not observed in annual variations of dust-event days in transport-affected regions
such as Japan (Kurosaki and Mikami 2003 ) over the period 1971-2002. These
discrepancies are not well understood and require further study (Shao and Dong
2006 ).
In the northern tropical Atlantic, the Barbados station in operation since 1965
has highlighted large interannual changes in African dust transport (Prospero and
Lamb 2003 ). In the last 15 years, thanks to the availability of multiyear satellite
observations of dust, substantial progress has been made in our understanding of
dust variability on relatively long timescales, that is, over periods on the order of
10 years or more (Hsu et al. 2012 ). Indeed, certain satellite sensors in operation
since the late 1970s (TOMS/Nimbus 7) or early 1980s (Meteosat, AVHRR) and
some more recent products cover periods long enough to support analyses of
dust variability at interannual timescales with attempts to identify their controlling
factors (Engelstaedter et al. 2006 ). Most of these studies focus on the export of
North African dust. Based on Meteosat observations, Moulin et al. ( 1997 )analyse
11 years of dust transport out of North Africa and reveal the control of the North
Atlantic Oscillation (NAO) on interannual variations of dust recorded over the North
Atlantic Ocean and Mediterranean Sea.
Chiapello et al. ( 2005 ) analyse long-term measurements of surface concentra-
tions of dust from Barbados (1966-2000) together with a 22-year satellite dust
record combining TOMS and Meteosat observations over the tropical North Atlantic
(1979-2000; Fig. 7.8 ). Their analysis shows a large regional impact of Sahel
drought conditions on interannual variations of dust transport over the Atlantic, in
both summer and winter seasons. The NAO influence appears to be restricted to
interannual variability in winter. Focusing on the winter season and using a 23-year