Since 1991, the European Union has been promoting the use of numerical models to assess the environmental fate of pesticides (directive 91/414 EC). In 2000, FOCUS (FOrum for the Co-ordination of pesticide fate models and their USe) issued a set of standard, realistic scenarios and associated software packages for use in assessing the potential movement of pesticides to groundwater in the EU registration process, thus permitting a harmonised and homogeneous evaluation.
To ensure the groundwater safety, the European authority established a quality end point of 0.1 μg/l for the pesticide concentration in groundwater. During the registration procedure, specific mathematical models are used to estimate the mean annual pesticide concentration in leaching at 1 m depth and to ensure that the predicted value is lower than 0.1 μg/l. Because weather conditions change every year, a fixed number of years is simulated (20, 40, or 60), and the most probable 80th percentile of the mean annual concentration of pesticide in groundwater is calculated. European rules do not mention about distribution functions, so the 80th percentile value is obtained by taking, for example, the 16th higher value from 20 years of simulation output.
Recently a group of scientists from the Catholic University and the Marche Polytechnic University (Italy) in association with Informatica Ambientale, the Milan-based research and computer science company, developed a tool that integrates one of the pesticide fate models with GIS software. The paper Balderacchi M., Di Guardo A., Vischetti C., and Trevisan M. 2008, The Effect of Crop Rotation on Pesticide Leaching in a Regional Pesticide Risk Assessment Environ. Sci. Technol. reports a regional scale pesticide risk assessment in which a large number of simulations were performed and a probabilistic approach has been introduced.
Several software products were tested by the research group to introduce distribution functions in the risk assessment study, and EasyFit was selected as the most appropriate tool for analyzing annual mean concentration and estimating the most suitable distributions. Although EasyFit allows to fit many distributions, only a few of them have meaning in the field of groundwater pesticide contamination, according to the suggestions of the Environmental Protection Agency.
The goodness of fit was assessed using the Kolmogorov-Smirnov and Anderson-Darling tests available in EasyFit. Once the Lognormal distribution was selected as the best fitting model, the 80th percentile of mean annual leaching was estimated according to the probabilistic approach. EasyFit was also used to calculate the exceedance probability to overcome the 0.1 μg/l in groundwater, which is another indicator of probabilistic groundwater contamination. This new indicator allows a better communication to stakeholder about pesticide contamination risk.
More information about the study is available at DOI URL: http://dx.doi.org/10.1021/es801287w