A Not So-Random Walk with Wind: Evaluating Wind Velocity Update Methods in Ground Based Spray Deposition Models

Abstract

The notion that wind speed and direction can be approximated by adding a random fluctuation to the previous value was investigated. The data were recorded at one meter above a field to simulate conditions that are present at a ground sprayer's boom. Variance ratio tests were carried out to test the null hypothesis that wind possesses similar properties to a random walk versus the alternative that wind does not. More specifically, are the random fluctuations auto correlated with one another in time? This process was done to a 10Hz sample and averages of the measured wind data at 0.5, 1, 5, 10, 30, 60, 300, and 600 seconds. It was found that for all tests, except for the 300 and 600 second data samples, the null hypothesis was rejected at greater than 99.9% certainty. This indicates that there is evidence of autocorrelation (rather than randomness) in the measurements of wind speed and direction, associated with each other in time.