LAUSR

As agricultural data and information becomes more abundant, diagnostics are needed to quickly and efficiently interrogate these data. Indices exist to identify sensor data with structured spatial variation, conducive to site-specific management. However, these indices do not indicate if this spatial variation is driven by managerial or environmental effects. A new index is proposed to identify perennial (or ordered row) fields that are likely or highly likely to have management effects within the spatial pattern of sensor data. This is determined by investigating differences in anisotropic (directional) variograms parallel and perpendicular to the direction of management (row orientation). Small differences are indicative of isotropic (environmental-driven) variation. Large differences indicate row and management effects. The index is derived, run on a database of 1080 simulated fields and applied to yield data from 124 vineyard blocks to assess index performance and response to different levels of variation. Simulations showed that the index is non-responsive to the magnitude of variation but responds strongly to anisotropy in the data. The stochastic variance in the data was observed to have an effect on index response and may be problematic when applied to noisy data sets. The index scores for the simulated and real-world data showed a similar pattern of response and the index was able to identify vineyard blocks where differential row management had generated differing yield responses. The index scores are continuous and some general guidelines for use of the index are proposed.