Pesticiders udvaskelighed

Summary

Title : Leachability of pesticides - assessment of the variation of DT₅₀- and K_oc-values, used in the GUS index.

By Bo Lindhardt, Inge S. Fomsgaard, Walter Brüsch and Rossana Bossi.

The object of this project is to assess the validity of ranking mobility of pesticides in soil by using the GUS index. The project includes an assessment of the variation in degradation rates in soil, DT₅₀, and the distribution coefficient between organic carbon and water, K_oc, for 12 selected pesticides. The ranking of these selected pesticides, based on the GUS index, was compared with information from lysimeter studies and data from surveys of drainwater and young groundwater.

In collaboration with the Danish Environmental Protection Agency 12 pesticides were selected (atrazine, bentazone, diuron, glyphosate, isoproturon, MCPA, mecoprop, metametrion, metsulfuron methyl, phenmedipham, propiconazol and triasulfuron), for which the variation in degradation rates in soil, DT₅₀, and distribution coefficients, K_oc were assessed. Data were primarily collected from the reports used by the Danish Environmental Protection Agency in their approval procedures for plant protection products.

There was a significant difference in the number of available degradation rate data for these 12 pesticides, from 1 to 62 results. The data derived from laboratory as well as field experiments. The average were estimated. The relative standard deviation for the 11 pesticides, which had more than one result, was between 36 and 112%. Using only results from laboratory experiments, the data sets was expected to be more homogeneous. A consequence of making such a delimitation, was that there would be no data for three of the selected pesticides; diuron, glyphosate and metsulfuron methyl. Within this project it has not been possible to determine to which extent the variation in data only reflected a real variation in the degradation rates. A part of the variation may be assigned to the different condition of the experiments and a fluctuation in quality of the single experiments. Another part of the variation may be assigned to the reduction of results to only one parameter, the DT₅₀ value. In some of the investigations the reduction of the degradation kinetics to a 1.order rate was associated with significant uncertainty and in some cases not correct. In addition to this, the number of processes included in the DT₅₀ values are considerably different. In the laboratory investigations the DT₅₀ only described the microbial degradation, whereas in the field experiments the parameter included volatilisation, leaching, hydrolysis and photolysis together with the microbial degradation.

From the available data it has not been possible to estimate the significance of soil types, content of organic carbon, pH, water content, and microbial activity as the variation in DT₅₀ has been too big in relation to the available data.

It is difficult to form a clear picture of the variation in the distribution coefficient between organic carbon and water, K_oc. The relative standard deviation for the selected pesticides vary between 9 and 174%. The quality and the extent of the data foundation for the single pesticide vary a lot.

For one of the pesticides, glyphosate, it is assessed that a description of sorption by a distribution coefficient between organic carbon and water is not valid as sorption of this pesticide is controlled only to a small extent by the content of organic matter of the soil. Presumably, the sorption is governed by other factors such as the content of clay and oxides.

The majority of the investigations show that the sorption coefficients are dependent on the concentration which is why the estimated distribution coefficients will be valid only within a very limited concentration interval. It has not been possible to point out extreme values for the individual pesticides to make the data sets more homogeneous and permit calculation of meaningful averages and variations.

A possible explanation of the considerable variation in the distribution coefficients might be that the sorption to the soil matrix is dominated not only by interaction between the pesticides and the content of organic matter of the soil, but e.g. also by the content of clay and oxides. This is significant for a number of the pesticides containing different more or less polar groups.

The test soils have been collected all over Europe and North America. Only a few number of the tests have used Danish soils. The variation in DT₅₀ and K_oc for soils which are assessed to be typical for Denmark have been estimated. Selecting results from experimentes where there have been used soils with a content of organic carbon between 0.5 and 3%, and a clay content of less than 20% which are assumed to be typical for Denmark, does not give a significant reduction of the relative standard deviation in the degradation rates DT₅₀, and the distribution coefficients K_oc, for the 12 pesticides.

Therefore, as matters now stand it is estimated to be impossible to state a statistically meaningful variation in DT₅₀ or K_oc for the selected pesticides. It is recommended to use median values and to express the variation in parameters by minimum and maximum values. The median values are less influenced by extreme values than the average values.

Based on the median values determined for DT₅₀ and K_oc the GUS index for the pesticides was calculated. The GUS_med indicate that bentazone, triasulfuron, metsulfuron methyl, and diuron must be expected to be "probable leachers", whereas atrazin, mechlorprop, isoproturon, MCPA, propiconazol, and metamitron must be described as "transition compounds". Phenmedipham is stated as "improbable leacher". The calculated GUS_min and GUS_max show that all 11 pesticides with extreme combinations of degradation and sorption conditions must be assessed to be "probable leachers", and that 8 out of 11 pesticides under other conditions must be assessed to be "improbable leachers". This shows that ranking by using the GUS index is subject to considerable uncertainty.

Lysimeter test results for 7 of the selected pesticides are available in the approval procedure material of the Danish Environmental Protection Agency. Compared to the ranking which appeared from the GUS_med it corresponds reasonably with the ranking of the lysimeter tests. This ranking should be regarded tentatively as a consequence of the very varied test conditions.

The ranking of the 12 selected pesticides after the registered occurrence in groundwater and surface water differs in several respects from the ranking obtained by use of the GUS_med index. From the actual findings the metamitron must be considered to be relatively more mobile than predicted by the GUS_med, whereas diuron should be relatively less mobile. The metamitron application pattern does not explain this difference. For diuron the limited amount of data for DT₅₀ together with the relatively limited use might be a reason for the deviation between the two rankings of this pesticide.

Assessed from the data on DT₅₀ and K_oc available, in the approval procedure material of the Danish Environmental Protection Agency for 12 selected pesticides, it must be concluded that it is not reasonable to use the GUS index when performing an actual ranking of the mobility of these pesticides in soil because of the limited number of data for some of the pesticides and the very large variation in these two parameters for most of the other pesticides.