W.L. Stefanov, Department of Geological Sciences, Arizona State University, Tempe, AZ 85287 USA

M. Ramsey, Department of Geology and Planetary Science, 200 SRCC Building, University of Pittsburgh, Pittsburgh, PA 15260

P.R. Christensen, Department of Geological Sciences, Arizona State University, Tempe, AZ 85287 USA

Urban centers located along the U.S.-Mexico border represent significant sources of fugitive (airborne) dust. This dust, which can lead to adverse health effects, arises from several factors including, construction activities related to land use conversion (i.e., agricultural to residential), unpaved roadways, agricultural activities, and human disturbance of the soil. Fundamental baseline data needed for modeling and monitoring of particulate generation and transport are accurate regional classification of land cover, degree of disturbance, and a metric of land cover change. Identification and delineation of fugitive dust source regions using a purely field-based approach is time and labor intensive and can lead to errors over time as land use changes. Further, restrictions on access to specific areas (such as private lands and reservations) may impede or prevent site investigations in these areas. Remotely gathered information can be used to circumvent these difficulties and provide rapid dust source region identification with quantitative area measurements required in transport models.

In order to carry out this analysis, Landsat ETM+ data was used to identify and delineate surficial materials that were either potential fugitive dust source regions or were important factors in dust transport and deposition. Using a knowledge-based system, land cover was classified into three generalized types: natural and disturbed soils (dust generation sites); asphalt, concrete, and urban materials (dust transport areas); and vegetated areas (dust deposition sites). Accuracy of the land cover classification was assessed using field verification, comparison of field and image reflectance spectra, and digital aerial orthophotographs. Results of image classification and field verification for Lansat data acquired during the winter of 2000 show a strong correlation, and will be used with data collected during the summer dry season for change detection analysis. The digital format of the classified data is optimal for input into fugitive dust transport models, and is available for use by federal, state, and municipal research and regulatory entities.

------------------
Submitted: American Geophysical Union Spring Meeting
Date: May 29 - June 2, 2001