Drift risk assessments (RA) in most countries rely on either Tier 1 drift tables (e.g. the EU) or simple statistical models derived from data (e.g. AgDisp in the USA, Australia, Brazil, etc.) resulting in buffer distances that can be excessively large and unrealistic. A major missing component of a tiered drift RA is an absence of publicly available and regulatory acceptable higher tier mathematical models that can provide conservative yet realistic estimates of spray drift – and here we expressly mean ground-based applications. Thus, a refined mechanistic arable crop spray drift model – the Casanova Drift Model (CDM) – is presented. The model is a mixture of ballistic tracking of droplets (Lagrangian transport of water droplets through air in time) combined with equations dealing with droplet evaporation, a representation of the horizontal wind profile, an approximation of canopy effects, a droplet size distribution, and a simplified description of spray sheet breakup. A cloud-based computationally-efficient interface has been developed using the R-Shiny platform – a single run takes less than a minute – and the entire coding is in C++ and R. The model has been calibrated and validated against not only Bayer (internal) field data but also European research data from four countries. Preliminary work assessing relative performance (i.e. relative changes in drift due to operational variables such as wind speed) and absolute performance (i.e. modelled drift curves that match results on the ground) are presented. The model includes modules that may help address issues raised during the SETAC DRAW discussions (e.g. sampler capture efficiency; locale). Plans are being made to have this model reviewed by key drift model experts. Afterwards, the source code, interface, and a library of droplet spectra for the standard flat fan nozzles used in EU drift trials will be made freely available. In this upload, we provide presentations and papers presented at various conference as well as a first draft of the user manual.
Casanova et al. (Mon,) studied this question.