Geoscience Reference
In-Depth Information
and water vapour clouds in the 11 and 12
µ
m wavelength bands. The dif-
ference in the two bands creates a “split-window” image from which areas
of dust are identified. The main limitation of the technique is that it only
works for (semi-) transparent clouds, relying on emitted thermal energy
from the earth's surface passing through them. Verification of the model
predicted dust concentrations is mainly by the use of the satellite image
enhancements described above and quantitatively, from available synoptic
observations of reduced visibility estimated by the Bureau of Meteorology
and observers. The problem in verifying by visibility observations is the
scarcity of reports in inland central and eastern Australia (see Fig. 1) and
once dust becomes suciently elevated and dispersed well away from the
dust source, reductions in visibility due to suspended dust might not be
reported.
5.1.
Case study 1: October 23, 2002
Figures 2(a), (b) are infrared images at 1800 UTC 22 October 2002 (04:00
local, 23 October) and 06:00 UTC 23 October (16:00 local). The model
predicted dust concentrations at 00:00 UTC on 23 October are contoured
from grid values and are defined in area and amount in Fig. 2(c). After
06:00 UTC the area and concentration rapidly diminished (not shown). In
the infrared images the intensity of the dust concentration is indicated by
the color.
Each color step represents a half-degree difference in T 11
µ
m-T 12
µ
m
below a particular threshold. Each color step can be thought of as a change
in the strength of absorption, which approximates the thickness of the dust
cloud, and is also affected by particle size and amount of surface radia-
tion. The dust concentrations are be expected to increase from the morning
through the afternoon due to the increasing depth of the mixed layer and
stronger winds at the surface lifting more dust. This increase is evident in
the two images. Figure 2(d) shows synoptic observations that recorded dust
including the lowest visibility for 23 October. High visibility values amongst
low values mostly indicates stations with few reporting times per day such
that a dust storm with very low visibility may have affected the station but
not at or even close to any reporting time during the day. In Fig. 2(d) the
observations that recorded the lowest visibility tend to be in southern and
central QLD which compares well with the area of highest model predicted
average dust concentration in Fig. 2(c).
