Submitted to: Agronomy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/1/1996
Publication Date: N/A
Interpretive Summary: Crop models offer hope as tools to efficiently manage resources, maximize returns to producers, and reduce impacts on water quality. This paper demonstrates how corn yield predictions with the ALMANAC and a revised version of CERES-Maize compare with county yield estimates at nine U.S. counties over ten years. Both models appeared appropriate at predicting individual year's yields for most counties. The data sets developed herein provide a starting point for ALMANAC and CERES users simulating corn yields at other U.S. locations.
Technical Abstract: Crop models can be evaluated based on accuracy in simulating several years' yields for a location or based on accuracy in simulating long-term mean yields for several locations. Our objective was to demonstrate how the ALMANAC (Agricultural Land Management Alternatives with Numerical Assessment Criteria) model and an improved version of CERES-Maize (Crop- Environment Resource Synthesis) simulate maize (Zea mays L.) yield for ten years in nine U.S. counties. We compared simulated yields for 1983 to 1992 with yields for these counties reported by the National Agricultural Statistical Service (NASS). We used a representative soil for each county and measured weather. Mean simulated yield for each county was always within 5% of the mean measured yield for the location. Simulated yields accounted for a significant amount of the variability in measured yields at all but two counties. Both models reasonably simulated the variability around mean yields at most sites. Both models also appeared appropriate for predicting an individual year's yield for most counties. The data sets provide a starting point for ALMANAC and CERES users simulating maize yields at other U.S. locations.