Combining Spatial and Temporal Corn Silage Yield Variability for Management Zone Development

Abstract

Precision agriculture requires an understanding of yield variability. The objectives of this study were to (i) document the temporal and spatial variability of corn ( Zea mays L.) silage yields on dairy farms in New York, and (ii) derive farm‐based management zones that account for both types of variability. Silage yield data from 847 fields (9084 ha; six farms) were collected by yield monitoring systems between 2015 and 2017. Raw yield data were cleaned of errors via a standardized postharvest data cleaning protocol. The whole‐farm area‐weighted average yield across years and the temporal SD of yield across years for fields with 3 yr of data were used to divide each field into 10‐ by 10‐m grid‐cells. Each grid‐cell was assigned a quadrant (Q), with Q1 and Q4 having consistently higher and lower yield than the farm average yield, respectively; Q2 having variable but higher yield than the farm average; and Q3 having variable and lower yield than the farm average. The evaluation showed variability in average yield per farm, yield per field, and within‐field yield, in addition to variability across years. Spatial and temporal variability were uncorrelated, suggesting that management zones need to consider both spatial and temporal variability. The area per farm classified as variable (Q2 and Q3) ranged from 30 to 44%, illustrating the importance of implementing precision agriculture technologies and in‐season management adjustments. Research is needed to determine the optimum number of zones per farm and the number of crop years to include in developing yield stability zones. Core Ideas Corn silage yield monitors collect relevant yield data for dairy farmers. Management zones can be developed from yield stability maps. Both temporal and spatial variability are important factors to consider. A yield‐stability‐based approach can generate precision management zones.

Abstract from the paper’s OpenAlex record, shown for indexing and research. Copyright remains with the authors or the publisher. Read at the source ↗

The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.