Evaluation and improvement of mathematical models of pesticide mobility id soils and assessment of their potential to predict contamination of water systems.

摘要

In this project with participants from Germany, the United Kingdom, Italy and France, the simulation models LEACHP, PRZM2 and VARLEACH were examined with regard to their efficiency in predicting the contamination of water (including groundwater). Forthis purpose, about 80 data sets from the participating countries were stored in a database, which also provides most of the input data necessary for modelling, e.g. soil and climatic data. This comprehensive pool of data from contrasting climates offersthe possibility to investigate the suitability of the models under varied conditions. A user-friendly interface (MARVEL - Model Attribute Relationship Validation and Entry Layout) was created to facilitate use of the three main models LEACHP, PRZM2 andVARLEACH. It allows data input a sequence of standardized screens, from which model input files (e.g. input or weather data files) are created automatically. Additionally a routine for graphical output of different results is included. Measured and simulated results (from LEACHP, PRZM2 and VARLEACH) using 26 selected data sets with comprehensive input data were compared with respect to distribution of residues in soil and, where possible, leachate concentrations. No general conclusions can be drawn regarding the suitability of the models. Even within time-phase samples from a single experiment, there was considerable variability in model performance, with no obvious explanation. Using several models on one data set, each might give the best fit at a different sampling date. Overall, redistribution and persistence in soil were described better than breakthrough and concentration in leachate.