/AnMtgsAbsts2009.52764 Crop Management Zone Delineation Based On Landscape Position.

Monday, November 2, 2009
Convention Center, Exhibit Hall BC, Second Floor
John Shanahan, USDA-ARS, Lincoln, NE, Richard Ferguson, Agronomy and Horitculture, Univ. of Nebraska, Lincoln, NE, Viacheslav I. Adamchuk, Biological Systems Engineering, Univ. of Nebraska, Lincoln, Lincoln, NE, Luciano Shiratsuchi, Agronomy & Horticulture, Univ. of Nebraska, Lincoln, Lincoln, NE and Larry Hendrickson, John Deere, Urbandale, IA
The objective of this work was to develop a robust process for delineation of management zones based on use of landscape position and other spatial data layers that can be used to more efficiently apply crop inputs such nitrogen (N) fertilizer.  Eight sprinkler irrigated fields from across Nebraska with significant spatial variability of yield and representing a wide range of soil and climatic conditions were selected to address the study objective.  In 2009, four N treatment rates (50, 100, 150, and 200 kg/ha), imposed as either field length replicated strips or as small plots replicated across the landscape, were applied at planting at each location. Landscape variability was assessed by acquiring elevation (using real time kinematics grade GPS) and apparent electrical conductivity (ECa) with Veris EC Surveyor prior to planting.  Other spatial data collected included soil color (aerial photograph after planting) and soil chemical properties (i.e., pH, soil carbon, N, and P) using a grid soil sampling scheme.  Grain yields were obtained with a yield mapping combine.  Economic optimum N rate (EONR) to optimize grain yield was determined as a function of landscape variability. The various spatial data layers were imported into a GIS environment and spatial analyses were conducted. Those included use of algorithms to process elevation data into a landscape position (LSP) attribute, which provides an indirect measure of variation in water availability across the landscape. Then multivariate analyses were used to determine which combinations of landscape attributes (i.e., ECa, soil color, pH, LSP) would be most useful in predicting EONR.  Results from this work will be presented.