Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Comparison of different methodologies for selecting PELMO input parameters for groundwater modelling of plant protection products including current EU guidance (SANC0/12117/2014 - final, 2014)

: Klein, Michael

Fulltext (PDF; )

Berlin: UBA, 2019, 53 pp.
Texte - Umweltbundesamt, 147/2019
Report, Electronic Publication
Fraunhofer IME ()

For registration of plant protection products in Germany the groundwater risk assessment at lower tier is based on modelling results using simulated leaching concentrations of the Hamburg scenario from the FOCUS PELMO model. Measured leachate concentrations for active compounds and transformation products from outdoor lysimeter studies are accepted as higher tier endpoints and could over-write modelling results. The objective of this project was to determine, whether lower tier calculations are still more conservative than higher tier results from lysimeter studies. For this purpose three different methodologies for selecting input parameters for modelling were considered: two variations of the EU methodology (previous: Kfoc value based on the arithmetic mean and new: Kfoc value based on the geometric mean according to EFSA 2014) and the national input parameter selection procedure (Holdt et al. 2011).Overall, the German methodology was found to be slightly more conservative than the EU methodology based on the geometric mean Kfoc value (DE: 82%, EU: 79% simulations above lysimeter results). The results did not depend on the type of substance (active compound, transformation product). In contrast, results based on the previous EU methodology using the arithmetic mean Kfoc value showed less situations where PELMO simulations were above respective lysimeter results (74%). Furthermore, results were dependent whether active compounds (79%) or transformation products (72%) were analysed. When evaluating the regulatory impact of the different methodologies for selecting in-put parameters no significant differences were found when focusing on false negative results (lysimeter showed a problem which PELMO could not determine): independent on the input parameter selection procedure 3% of the active compounds and 6% to 7% of the transformation products belonged to that class. Comparing the three methodologies with regard to false positive comparisons (regulatory necessity for additional higher tier studies) the results were dependent whether active compounds or transformation products were simulated. For active compounds both EU methodologies led to 2 to 3 three times lower false positive results than the national input parameter selection. For transformation products the situation was different, since the new EU methodology (geometric mean Kfoc value) resulted in similar percentages as the national parameter input selection procedure. In contrast, the previous EU methodology (arithmetic mean Kfoc value) was characterized by lower deviations of these false positive results.