Saturday, 15 July 2006
115-17

Satellite and Airborne Remote Sensing for Rapid Assessment of Surface Soil Properties.

Dana Sullivan1, Joey N. Shaw2, Paul Mask2, D. Rickman3, J.C. Luvall3, and JM Wersinger2. (1) USDA-ARS Southeast Watershed Research, PO Box 748, 2316 Rainwater Rd, Tifton, GA 31794, (2) Auburn Univ, Auburn, AL 36849, (3) Global Hydrology and Climate Center, NASA, Huntsville, AL 35806

Effective erosion control, precision management, and soil survey necessitate methods to rapidly assess variability in surface soil properties. Soil attributes such as soil organic carbon (SOC), particle size distribution and mineralogy have been correlated with reflectance spectra. A major goal of this study was to evaluate high-resolution satellite and airborne data as tools for field and landscape scale assessment of soil property variability in conventionally tilled, agricultural fields. Two separate analyses were conducted: 1) evaluation of the relationship between soil properties and spectral response across three physiographic regions in Alabama using the high spectral and spatial resolution Airborne Terrestrial Applications Sensor (ATLAS), and 2) estimation of SOC and clay content using satellite-derived (IKONOS) data for specific sites in the Tennessee Valley and Coastal Plain regions. Soils consisted mostly of fine-loamy, kaolinitic, thermic Arenic and Plinthic Kandiudults in the Coastal Plain to fine, kaolinitic, thermic Rhodic Paleudults in the Tennessee Valley. Surface soil samples were collected for soil water content, surface crusting, SOC, particle size distribution, and iron oxide content determination. Using the ATLAS sensor, SOC was difficult to quantify in these highly weathered systems, where soil organic carbon was generally < 1 %. ATLAS estimates of sand and clay content were best in the Tennessee Valley region, explaining 42 –59 % of the variability. In the Coastal Plain, sandy surfaces prone to crusting limited estimates of sand and clay content variability. Estimates of iron oxide content were best accomplished using specific spectral band ratios, with regression explaining 36-65 % of the variability in the Tennessee Valley and Coastal Plain sites, respectively. Co-kriging with IKONOS imagery improved field scale estimates of SOC and clay content compared to ordinary kriging and multiple linear regression estimates. Root mean square errors of co-kriged estimates of SOC and clay content were as low as 0.14% and 2.2 %, respectively.


Back to 1.0PA Multiscale Mapping of Soil Properties for Environmental Studies, Agriculture, and Decision-Making - Poster
Back to WCSS

Back to The 18th World Congress of Soil Science (July 9-15, 2006)