Large-scale statistical analysis of crop rotation effects
Historically, field trials have served as the predominant method to assess yield impacts of farm management practices. As the amount of data produced by farm operations has increased over the past few years; however, the statistical analysis of large, less structured datasets from farm operations has emerged as an alternate method of undertaking this important research.
In a recent article published in Agronomy Journal, researchers used a dataset containing nearly three-quarters of a million yield records joined with weather, soil and crop rotation information not only to replicate a key finding of the agronomic literature (that crop rotations increase yield) but also to expand on that finding – mapping where and under what conditions these effects are the strongest.
The team found that the yield penalty for continuous corn was the largest in dry years and years of poor harvests while the penalty for continuous soybean was larger in high-moisture areas. The continuous soybean penalty increased monotonically with the number of years continuously cropped while the continuous corn penalty did not.
While these techniques cannot unambiguously identify causation, the results support many of the findings from field trial studies, importantly showing that similar effects surface in data from actual farms.