Skip to main content
ARS Home » Southeast Area » Auburn, Alabama » Soil Dynamics Research » Research » Publications at this Location » Publication #212406
Title: Multivariate crop productivity zones in the Alabama coastal plain

Author
item STONE, H - AUBURN UNIVERSITY
item SHAW, J - AUBURN UNIVERSITY
item RODEKOHR, K - AUBURN UNIVERSITY
item Balkcom, Kipling
item RAPER, RANDY
item REEVES, DONALD

Submitted to: Southern Conservation Agricultural Systems Conference
Publication Type: Proceedings
Publication Acceptance Date: 6/25/2007
Publication Date: 6/25/2007
Citation: Stone, H.D., Shaw, J.N., Rodekohr, K.S., Balkcom, K.S., Raper, R.L., Reeves, D.W. 2007. Multivariate crop productivity zones in the Alabama coastal plain. In: Wright, D.L., Marois, J.J., Scanlon, K., editors. Proceedings of the 29th Southern Conservation Agricultural Systems Conference, June 25-27, 2007, Quincy, Florida. Available at: http://www.ag.auburn.edu/auxiliary/nsdl/scasc/.

Interpretive Summary: Various data are used to develop management zones for site-specific crop production. Most evidence indicates that the technique used for zone development is crop and management dependent. Researchers from Auburn Univ. and USDA-ARS locations in Auburn, AL and Watkinsville, GA evaluated which field-scale data are most appropriate for developing management zones for characterizing crop productivity and variability over multiple growing seasons and managements. Specifically, we are evaluating: 1) if field-scale crop yield variability is better described in zones derived from temporal or static data, and 2) the relationships between zone development approach and soil management system. This study was conducted on a field-scale (20 acre) experiment evaluating the interaction of soil management systems (conventional versus conservation) with soil landscapes on a site in the Alabama Coastal Plain. Field-scale data include satellite remote sensing imagery, terrain attributes generated from a LiDAR derived digital elevation model (DEM), field-scale electrical conductivity, and a first-order soil survey (1:5000). Management zones were developed with clustering techniques. Our results indicate that all evaluated data were generally suitable for characterizing crop productivity and variability using a clustering approach. As expected, satellite remote sensing data collected in season were more highly related to yield compared to terrain and soil variables. This study illustrates the relative effectiveness of these data for describing yield variability is most dependent on crop, and somewhat dependent on management.

Technical Abstract: Various data are used to develop management zones for site-specific crop production. Most evidence indicates that the technique used for zone development is crop and management dependent. The objective of this research is to evaluate which field-scale data are most appropriate for developing management zones for characterizing crop productivity and variability over multiple growing seasons and managements. Specifically, we are evaluating: 1) if field-scale crop yield variability is better described in zones derived from temporal or static data, and 2) the relationships between zone development approach and soil management system. This study was conducted on a field-scale (20 acre) experiment evaluating the interaction of soil management systems (conventional versus conservation) with soil landscapes on a site in the Alabama Coastal Plain. Six replications in a cotton (Gossypium hirsutum L.) - corn (Zea mays L.) rotation traverse the landscape. Soil landscapes range from Typic Paleudults on well drained uplands to imperfectly drained Oxyaquic Paleudults in drainageways. Field-scale data include satellite remote sensing imagery, terrain attributes generated from a LiDAR derived digital elevation model (DEM), field-scale electrical conductivity, and a first-order soil survey (1:5000). Management zones were developed using fuzzy k-means clustering, and geo-referenced crop yield data have been collected (2001-2006). Our results indicate that all evaluated data were generally suitable for characterizing crop productivity and variability using a clustering approach. As expected, satellite remote sensing data collected in season were more highly related to yield compared to terrain and soil variables. The relative effectiveness of these data for describing yield variability is most dependent on crop, and somewhat dependent on management.