1083 
lu» 
ic radius r e (x, y) 
Dust Load 
300 
•d dust load u (right). 
rows are plotted. The 
e south-north oriented 
sd branch are given as 
maximum dust load is 
.ic profiles is observed, 
ion. The agreement is 
to be recalled that the 
:e, these values should 
rather simple and may 
1 to be deposited on a 
canopy. Moreover, for 
icrical particles which, 
tiple reflexions at dust 
ture of the vegetation 
canopy requires a spatially varying reflectance of pure vegetation. As already mentioned in the previous 
section, the canopy is never clear of dust, hence the retrieval of this information from satellite data is, by 
principle, impossible. 
Du»t Load W-E Profile» 
CTTT? I i i i i I i i i 
' ¡.pbr'diu 1 1 1 1 ' ' 
—— stochastic data 
■— road 
row at 1700 m _ 
row at 1600 m 
----- - 
row at 1900 m 
- : 
row at 1100 m 
A- 
row at 900 m 
row at 700 m 
JL- 
row at 600 m 
A 
row at 300 m 
row at lpo m 
eoo 
1000 - 1600 2000 
Vast-East [m] 
Du»t Load S-N Profile» 
' IsWr'dJtJ 1 1 ' 1 1 ' 1 ' ' 1 
—stochastic data 
- — ro * d A 
column ati 2060 m jF 1 
1 1 ' 1 1 ' 1 
column at 1960 m 
.-A., 
oolumn at 1660 m 
c<j|umnjil 1760 jm _ 
column at 1660 ip 
column at 1660 ip 
column at 1460 m 
coiumnjat 1360 m 
-A . 
- 
columr^jit _1250 rp 
, i 1 , 
o eoo looo leoo «eoo seoo 
South-North [m] 
Figure 5: West-east profiles of dust load in arbitrary units, averaged over 11 image rows (left), south-north 
profiles of dust load in arbitrary units, averaged over 11 image columns. SPOT-based data (solid), stochastic 
data after Lamprecht and Graber (1993) (dashed-dotted) and approximate road position (bar) 
5 SUMMARY AND CONCLUSIONS 
A stochastic model developed at the Paul Scherrer Institute was recently applied to simulate the deposition 
of road-generated dust from the Dalton Highway which is a high-speed gravel road in Alaska travelled mainly 
by large vehicles. The propelled dust is deposited on the adjacent vegetation where it may cause detrimental 
effects to the highly fragile plants of the tundra. In a field experiment in 1991, the relevant model input 
quantities (the meteorological parameters and the size distribution of the deposited dust particles) were 
measured. 
On the basis of the near-infrared band XS3 of the High Resolution Visible (HRV) instrument onboard 
the satellite SPOT, the dust load u(x, y) along the Dalton Highway was derived. The results presented in 
Sec. 3.2 allow to draw the following conclusions: 
• For the band XS3 as well as for XS1, the dust reflectance values corrected for atmospheric effects were 
found to agree with laboratory measurements within the accuracy of the SPOT calibration coefficients. 
For XS2, however, there is a significant discrepancy, which is probably due to an erroneous instrumental 
gain of the satellite sensor. The calibration coefficient can be actuated to a more realistic value by 
assuming that the onboard gain is actually one step higher than the value reported in SPOT (1987). 
• The near-infrared band XS3 is best suited for detecting reflectance differences between vegetation and 
dust. This band is located at wavelengths where the reflectance of vegetation is substantially higher 
than the dust reflectance. In addition, a higher reflectance values cause a higher radiance, thus higher 
digital counts leading to lower errors.