Agronomy and Horticulture, Department of


Document Type


Date of this Version

April 2004


Published in Precision Agriculture, 5, 207–225, 2004. Permission to use. 2004 Kluwer Academic Publishers. Manufactured in The Netherlands.


The goal of this research was to determine the potential for use of site-specific management of corn hybrids and plant densities in dryland landscapes of the Great Plains by determining (1) within-field yield variation, (2) yield response of different hybrids and plant densities to variability, and (3) landscape attributes associated with yield variation. This work was conducted on three adjacent fields in eastern Colorado during the 1997, -98, and -99 seasons. Treatments consisted of a combination of two hybrids (early and late maturity) and four plant densities (24,692, 37,037, 49,382 and 61,727 plants ha-1) seeded in replicated long strips. At maturity, yield was measured with a yield-mapping combine. Nine landscape attributes including elevation, slope, soil brightness (SB) (red, green, and blue bands of image), ECa (shallow and deep readings), pH, and soil organic matter (SOM) were also assessed. An analysis of treatment yields and landscape data, to assess for spatial dependency, along with semi variance analysis, and block kriging were used to produce kriged layers (10 m grids). Linear correlation and multiple linear regression analysis were used to determine associations between kriged average yields and landscape attributes. Yield monitor data revealed considerable variability in the three fields, with average yields ranging from 5.43 to 6.39 mg ha-1 and CVs ranging from 20% to 29%. Hybrids responded similarly to field variation while plant densities responded differentially. Economically optimum plant densities changed by around 5000 plants ha-1 between high and low-yielding field areas, producing a potential savings in seed costs of $6.25 ha-1. Variability in yield across the three landscapes was highly associated with landscape attributes, especially elevation and SB, with various combinations of landscape attributes accounting for 47%, 95%, and 76% of the spatial variability in grain yields for the 1997, -98, and -99 sites, respectively. Our results suggest site-specific management of plant densities may be feasible.