1077 
ROAD DUST AS AN INDICATOR FOR AIR POLLUTION 
TRANSPORT AND DEPOSITION. AN APPLICATION OF 
SPOT IMAGERY 
J.Keller, R.Lamprecht 
Paul Scherrer Institute (PSI), CH-5232 Villigen PSI, Switzerland 
ABSTRACT 
A simulation model for atmospheric diffusion and dry deposition of coarse dust particles developed at the 
Paul Scherrer Institute was recently applied to calculate the dispersion and deposition of road-generated dust 
from the Dalton Highway which is a high-speed gravel road in arctic Alaska travelled mainly by large vehicles. 
The propelled dust is deposited on the adjacent vegetation where it may cause detrimental effects to the 
plants of the highly fragile tundra. During a field experiment in 1991, all meteorological parameters as well 
as the size distribution of the deposited dust particles were measured. This data was used to calculate the 
dust distribution pattern and the amount of dust transferred into the vegetation at specific locations adong 
the Dalton Highway. 
The scope of this paper is to identify and, as far as possible, to quantify this dust deposition pattern 
along the Dalton Highway by multispectral SPOT imagery. The spatial distribution of the dust on both sides 
of the road is distinctly visible in the XS3 channel (0.79 — 0.89 pm) of a SPOT satellite image. On the basis of 
the ground reflectance and the reflectances of pure dust and pure vegetation, the dust load can be calculated. 
The dust load depends on the particle size distribution which can be derived from the size spectra measured 
in the field experiment. 
The spatial dust load obtained from the SPOT data is compared with the distribution computed with 
the simulation model. As the simulation is based on only a limited number of days, the dust load scaling is 
arbitrary. Taking this fact into account, the general shapes of the two distributions agree remarkably well 
within a strip of about 1 km width along the road. In general, the agreement is better across the main wind 
direction than sdong that direction. 
Apart from this application on a local settle, suspended dust may also be detected on larger scales, 
e.g. as volcanic ash or as tropospheric or stratospheric aerosols. 
KEYWORDS: air pollution, dust, SPOT, reflectance, arctic 
1 INTRODUCTION 
In the framework of the research activities in the field of air pollution modeling at the Paul Scherrer Institute 
(PSI), a Lagrangian stochastic air quality model for the simulation of atmospheric diffusion and dry deposition 
of coarse dust particles has been developed (Lamprecht (1989)). Recently, this model has been applied to 
calculate road-generated dust from the Dalton Highway, which is a high-speed gravel road in Alaska. The 
Dalton Highway is heavily travelled mainly by large vehicles which generate considerable dust plumes in their 
wake. This dust is deposited on the adjacent vegetation where it may cause detrimental effects to the highly 
fragile plants of the arctic tundra. During a field experiment in the summer of 1991 near Toolik Lake, the 
meteorological parameters relevant for the model as well as the size distribution of the deposited dust particles 
were measured. This data was used as input to calculate the dust distribution pattern and the amount of dust 
transferred into the vegetation at specific locations along the Dalton Highway (Lamprecht (1992), Lamprecht 
and Graber (1993)). 
As the dust is clearly visible within a strip of about 1 to 2 km width along the road, we expected that 
satellite-borne remote sensing methods are capable of detecting the spatial dust distribution. The scope of 
this paper is to identify and, as far as possible, to quantify the dust load pattern along the Dalton Highway 
by satellite imagery. The image data has to meet at least two requirements: a) the spatial resolution of the 
sensor has to be equal to or finer than the model grid width, b) the multispectral sensor has to be sensitive 
in the visible and in the near infrared radiation spectrum where dust and vegetation are discernible. 
These requirements are met by the multispectral bands of the satellites SPOT (Satellite pour l’Obser- 
vation de la Terre). The output signal of the satellite sensor is strongly correlated with the reflectance of the 
sensed object at ground level. Including the reflectances of pure dust and pure vegetation (e.g. from image 
sections comprising sufficiently large areas of the respective land cover) the proportion of the area covered 
with dust can be derived for each image pixel. As this proportion is a function of the dust load, it is possible 
to derive that load from the signal of the satellite’s sensor. In principle, the evaluation method is independent