Spread Pattern Analysis Tool (SPAT): I. Development and theoretical examples

A custom Windows-based program called SPAT was written in MatLab (1999) to theoretically investigate the overall quality of spread patterns. It computes overlapped patterns, spread pattern characteristics, and produces cv-swath width and rate-swath width graphs. Pattern shapes that are often seen in practice such as Triangles, Trapezoids, M- and Crown shapes were analyzed. The effect of skewness, a pattern distorting factor in aerial as well as ground-based application, was studied as well. Traditionally, the quality of a spread pattern is defined in terms of the measured application rate (% deviation from the target rate) and the uniformity (expressed as the statistical coefficient of variation, cv). Both factors are not constants, but are functions of the swath width which makes their practical value disputable. In this study a new descriptor called "robustness" was defined. Robustness is a value based on the average cv-values over a range of potential swath-widths. A robust pattern allows the applicator to change the swath width (and the application rate), and be confident that no damage to the crop will occur. The most robust pattern found was a Gaussian shape, also known as a 'bell curve'. The second most promising pattern found was a symmetric triangle with the target rate value in the center. Trapezoids were found to be acceptable and even more efficient than triangles, although too much emphasis on large swath widths can make the pattern non-robust. Skewness was found to be a major factor in producing non-robust patterns; it tends to increase the cv over the whole range of useful swath widths. Although SPAT was written for analysis of aerial spread patterns, the overall results of this study can be applied to ground based application.