Evaluating continuous-move sprinkler machines using time-series statistics

Abstract

Traditional methods of describing uniformity of water applied by continuous-move sprinkler machines provide a measure of variability but fail to identify sources of variability. Thus, we investigated the potential of time series statistics as a tool for identifying patterns of variability in catch-can data collected during evaluations of three continuous-move sprinkler machines. Transects of catch cans were installed along the laterals, and the can data were analyzed using autocorrelation and spectral analysis. However, for one machine, transects across individual spans with closely spaced cans were also used. The traditional approach showed similar measures of variability among the machines. The time-series analysis, however, revealed very different patterns of variability along the lateral, attributed to a modification made to one machine and random behavior for another. For a third machine, the uniformity along the lateral was similar to that across individual spans, but along the lateral, random variability occurred, whereas, across the spans, inadequate sprinkler overlap was the major source of variability. We conclude that time-series statistics is an evaluation tool to be used in identifying periodic patterns of variability.