Some considerations on the use of simple box models of contaminant fate in soils.

A simple box model computing time-averaged concentrations in soil and water of contaminants such as pesticides, following a pulse emission under the assumption of constant removal rates, is presented and evaluated against a benchmark model and some lysimeter experiments representative of different soil and climate settings in Europe. The simple box model allows capturing to some extent the trends of contaminant releases observed from lysimeters or predicted by the benchmark model. This suggests that the correct order of magnitude of environmental concentrations and loads of contaminants can be described as a first approximation by a very simple back-of-the-envelope calculation. In the calculation, the time lag between contaminant application and the start of runoff or leaching should be considered. In the absence of more detailed information, repeating the calculation for the extreme cases of zero and one month lag yields a reasonably realistic range, and the geometric mean of the two is suggested as a first guess estimate. This configures a practical approach for the screening of contaminant losses especially suited for the mapping of predicted environmental concentrations (PECs) for assessment at broad scale (such as continental Europe, a country or a large catchment), where contingent determinants of contaminant fate tend to be averaged out and use of more detailed models usually promises to yield little improvement of PECs.