Wednesday, 9 November 2005
10

Comparison of First Order Soil Surveys to Alternative Approaches for Characterizing Cotton Productivity on Alabama Ultisols.

Joey Shaw, J. Owen, C. Burmester, D.W. Reeves, and P.L. Mask. Auburn University, 5102 Cress Lake Rd, Auburn, AL 36830

Intensive cotton (Gossypium hirsutum L.) production occurs on Tennessee Valley (AL) Ultisols. Increased adoption of site-specific management requires the temporal and spatial characterization of agronomic properties, and soil, crop, and remotely sensed data are often used. We hypothesized that first order soil surveys are equal or superior to other techniques for characterizing cotton productivity over multiple growing seasons. The experiment was conducted during 2001-2003 at the Tennessee Valley Research Center (Belle Mina, AL) on a highly variable 5-ha site where soils ranged from Oxyaquic to Typic Paleudults. An Order 1 soil survey (1:2500) was created, terrain attributes (e.g. slope, CTI) were developed from high resolution DEMs (RTK-GPS), and satellite remote sensing data were collected. Fuzzy k-means clustering of terrain, remote sensing, field-scale electrical conductivity, and yield data were used to develop zones. Eighteen random sampling sites were established, and soil properties (nutrients, volumetric water content), crop properties (leaf temperature, tissue nutrients, node counts), and cotton productivity (yield, fiber quality) were measured at each site. Significant (p<0.05) differences in cotton yield were observed between all delineation techniques, however, fiber length and strength were only different in zones developed using remote sensing data (NDVI). Factor analyses reduced multivariate data into five factors representing 80% of the data variability, and the first three factors represented soil moisture, leaf tissue nutrients, and soil nutrients, respectively. These three factors described 50, 52 and 78% of yield, fiber length, and strength, respectively. In particular, soil moisture was critical to yield and fiber strength. Soil moisture and tissue nutrient factor scores were significantly affected by all zones, but soil nutrients were only different among soil survey delineations. The aggregate of data suggests that first order soil surveys and remote sensing imagery are effective resources for creating cotton management zones in this region.

Back to Innovative Approaches for Soil Mapping
Back to S05 Pedology

Back to The ASA-CSSA-SSSA International Annual Meetings (November 6-10, 2005)