Submitted to: International Conference on Precision Agriculture Abstracts & Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 7/17/2002
Publication Date: 7/17/2002
Citation: JAYNES, D.B., KASPAR, T.C., COLVIN, T.S. COMPARISON OF TECHNIQUES FOR DEFINING YIELD POTENTIAL ZONES IN AN IOWA FIELD. INTERNATIONAL CONFERENCE ON PRECISION AGRICULTURE ABSTRACTS & PROCEEDINGS. 2002. CD-ROM. MADISON, WI: ASA-CSSA-SSA. Interpretive Summary:
Technical Abstract: A first step in determining field management zones can be to determine the areas of similar yield potential within a field. This can be a major challenge however, because yields typically vary across space and also vary among years as well. In this study, we investigated three methods for determining corn (Zea mays L.) yield potential zones based on 6-yr of yield information. The methods included cluster analysis of multiple year yield data, cluster analysis of easily-measured field attributes believed to be important in determining corn yield, and multiple regression of these field attributes on corn yield. All methods were applied to corn yield data collected from 224 plots arranged along eight transects spanning a 16-ha field. Each method divided the 6-yr of yield data into years having wetter than average and drier than average growing season precipitation. This reflects the importance of water availability in determining corn yield in this rain fed field. All three methods illustrated the importance of landscape position on corn yield, with yield generally decreasing from footslope to summit positions in dry years. All three methods also illustrated that excess precipitation caused increased yield variability especially in lower landscape positions. Clustering was particularly effective in separating a pothole area from the remainder of the field. Multiple regression and yield clustering were superior in reducing the residual yield variance. Overall, each method was effective in delineating areas of similar yield potential within the field.