Submitted to: Agronomy Journal
Publication Type: Peer reviewed journal
Publication Acceptance Date: 11/7/2002
Publication Date: 5/1/2003
Citation: KITCHEN, N.R., DRUMMOND, S.T., LUND, E.D., SUDDUTH, K.A., BUCHLEITER, G.W. SOIL ELECTRICAL CONDUCTIVITY AND OTHER SOIL AND LANDSCAPE PROPERTIES RELATED TO YIELD FOR THREE CONTRASTING SOIL AND CROP SYSTEMS. AGRONOMY JOURNAL. 2003. V. 95. P. 483-495. Interpretive Summary: Grain-yield mapping has demonstrated to farmers that much of the yield variability within fields is associated with variability in soil and landscape properties. Yet farmers and consultants have expressed uncertainty about which properties and data analysis methods should be used to best predict yield. Farmers are interested in inexpensive and accurate methods for measuring soil properties that do not rely on intensive soil sampling. This research was conducted to: 1) evaluate the relationship between grain yield and soil electrical conductivity (EC) on 3 contrasting soil and crop systems, and 2) provide a comparison of several data analysis procedures for relating yield with soil and landscape properties. Yield and soil data were obtained from research fields in CO, KS, and MO. While simple correlation analysis that relates grain yield to a single variable is the most often used analysis procedure, we found that a procedure that relates grain yield to several variables and their interactions explained yield variability the best. This result shows that field-scale yield variability is influenced by many factors and that single-factor analysis is inadequate. Visually examining the data in scatter plots also demonstrated the effects of multiple factors causing yield variation. Regardless of the data analysis procedure, soil EC was the most important property for explaining yield variability. Soil EC provides an estimate of within-field differences in soil water storage and availability. Our results will help farmers and crop consultants better understand yield variation and develop more efficient site-specific management plans.
Technical Abstract: Many producers who map yield want to know how soil and landscape information can be used to help explain yield variability and provide insight into improving production. This study was conducted to investigate the relationship of profile soil electrical conductivity (EC), terrain measures, and soil-sampled properties with grain yield for three contrasting soil-crop systems. Yield data was collected with combine yield monitoring systems on three fields [Colorado (Ustic Haplargids), Kansas (Cumuic Haplustoll), and Missouri (Udollic Ochraqualfs)] during 1997-1999. Crops included four site-years of corn, three site-years of soybean, and one site-year each of grain sorghum and winter wheat. Soil EC was obtained using a rolling coulter device. Elevation, obtained by either conventional surveying techniques or RTK GPS, was used to determine slope, curvature, and aspect. Soil organic matter, pH, and CEC came from georeferenced soil samples. Four analysis procedures were employed to investigate the relationship of these variables to yield: correlation, multivariate stepwise regression, nonlinear neural networks, and boundary line analysis. Correlation results, while often statistically significant, were generally not very useful in explaining yield variation. Using either regression or neural networks analysis, soil EC alone explained yield variability better than the other soil and landscape variables in six out of the nine site- years. Combining soil EC and topography measures together usually improved model R**2 values. Boundary lines generally showed yield decreasing with increasing EC for Kansas and Missouri fields. Regardless of the analytical procedure used, soil EC was the most important parameter for explaining yield variability.