Student Research: Ming-Yi Tsai
, Environmental and Occupational Health (EOH), 2007
Faculty Advisor: Michael G. Yost
The Washington Spray Drift Study: Understanding the Broader Mechanisms of Pesticide Spray Drift
The Washington aerial and airblast spray drift studies were conducted over the summers of 2002 & 2004 respectively. This dissertation focuses on spray drift from orchard airblast spraying; however, in chapter 2, we start with modeling results from an aerial application. These results revealed the episodic nature of deposition from drfit and demonstrated the importance of considering both changing meteorology and source location in characterizing spray drift.
Chapter 3 describes the orchard spray drift study and analyzes our deposition results using the geostatistical method of ordinary kriging. Compared with along-wind transects, perpendicular transects (as per U.S. EPA's spray drift test guidelines) underestimated deposition with distance by 22 to 94%, and AgDRIFT underestimated deposition by 88 to 100%. Our orchard spray drift study showed that: 1) following EPA's spray drift test guidelines will underestimate deposition with distance, 2) transects perpendicular to tree rows overestimate near-field and across field deposition, 3) overestimation in the near-field results in underestimation of mass available for longer range drift.
Chapter 4 presents our orchard spray drift model (OSDM). OSDM is based on U.S. EPA's fugitive dust mdoel and adds a complex multi-component source term. The calibrated model accurately estimated actual deposittion in three of four spray events with concordance correlations ranging from 0.950 to 0.996. It also showed the overwhelming contribution to deposition of spray escaping from the top of the canopy, and the importance of chaniging wind direction and source orientation on the total deposition pattern of a spray event.
Chapter 5 compares optical remote sensing (Lidar) data to our OSDM output. Overall, it validates the general trends predicted by the model, but the smoothness of the data are quite different. These differences are a result of: monitoring an actual spray, the intermittent quality of our Lidar data, and the idealized definition of the OSDM source term. The Lidar data confirms the importance of changing wind direction and source position (as shown in time-resolved OSDM output), but suggests that for worker/bystander exposure, a Lagrangian model would be better suited for estimating the probability of being 'drifted' on.