Miljøprojekt, 1000 – BAM's skæbne i grundvand

The Fate of BAM in Groundwater

Summary and Conclusions

The present project is a follow-up on the project ”Pesticides and Waterworks” (Miljøstyrelsen, 2002). The purpose of the project is to provide further documentation regarding the fate of the metabolite BAM in soil and groundwater in order to evaluate the extent and duration of BAM contamination in aquifers and thereby the consequences for the waterworks. New knowledge regarding the fate of BAM (sorption and degradation) and the applied mass of the mother compound dichlobenil has therefore been obtained. Furthermore, the literature was reviewed, to evaluate the importance of evaporation for the removal of dichlobenil. The new data regarding sorption, degradation, evaporation and applied dichlobenil was included in mass balances to quantify the mass of dichlobenil and BAM in soils and waters.

One of the most important parameters for the duration of the BAM contamination is the degradation of BAM. Therefore the degradation of ¹⁴C-labeled BAM was investigated in both uncontaminated and in contaminated sediments, where microorganisms may have been adapted to BAM. The results show that BAM is microbially degraded in the upper meter below surface (mbs), whereas the degradation potential is limited below 1 mbs (average half-life 61 years). BAM is degraded much more easily in the contaminated sediments than in the uncontaminated sediments. Furthermore, in the upper meter below surface the degradation of BAM is faster in the investigated contaminated sandy sediments (average half-life 0.33 years) than in the contaminated clayey till sediments (average half-life 24 years).

Degradation experiments with different initial concentrations showed no ambiguous relation to degradation rates. In the contaminated clayey till sediments the degradation rates increase with increasing BAM concentrations, which indicates that the degradation of BAM is controlled by growth kinetics. High BAM concentrations may cause growth of specific BAM degraders, which thereby increases the degradation. This indicates that, due to differences in the applied mass and subsequent sediment concentrations at locations where dichlobenil has been applied, the degradation rate may have varied over time. In contrast, the degradation rate is independent of the initial concentration in the sediments from the contaminated sandy location (Eskærhøjvej). Growth kinetics has therefore not been observed in this sediment, maybe because the number of specific degraders was high already before the addition of BAM, due to a long historical duration of dichlobenil and BAM contamination.

The degradation experiments demonstrate that BAM is mineralised to CO₂, primarily in the upper meter of the unsaturated zone. The detection of 2,6-dichlorbenzoic acid in two out of three sediments shows that BAM is degraded to 2,6-dichlorbenzoic acid and further to CO₂. Detections of 2,6-dichlorbenzoic acid in the field may therefore indicate BAM degradation. Survey data of 2,6-dichlorbenzoic acid were reviewed, in order to get a general view over the distribution and frequency of detections in Denmark (and thereby the degradation of BAM) and possible correlation to geology and land use. The review shows that 2,6-dichlorbenzoic acid was detected at BAM-contaminated locations, and the detections can frequently be related to the unsaturated zone, where also the degradation experiments in the laboratory have shown degradation of BAM. This does indeed indicate that detection of 2,6-dichlorbenzoic acid may be evidence of BAM degradation under field conditions. The survey data for 2,6-dichlorbenzoic acid were, however, too limited to correlate the detections to geology and land use.

Desorption of BAM has not been investigated before, and sorption and desorption were investigated in reduced clayey till sediments. The results demonstrate that BAM is relatively strongly adsorbed to the reduced sediments, because more than 40% of the adsorbed BAM did not desorb during the experiment. These results therefore illustrate that desorption of BAM only occurs very slowly. As a consequence, the transport of BAM through the clayey till layers covering the aquifers will be longer than predicted, because slow desorption or irreversible sorption are not yet incorporated in models.

Evaporation of dichlobenil has been proposed to be an important transport process, which will reduce the mass of BAM percolating to the ground water. In this project a review of evaporation of dichlobenil showed that the literature is old and insufficient, but the data do indicate that evaporation of dichlobenil is limited under Danish weather conditions. Although dichlobenil has a relatively high Henry's constant, which may indicate a risk of significant evaporation, a combination of the strong sorption to soil and the product being formulated as granulate, which is the dominant formulate used in Denmark, seems to reduce the evaporation. If dichlobenil evaporated significantly to the atmosphere it would be detectable in precipitation and/or air, because dichlobenil is photochemically stable and hardly degrade to BAM in precipitation and air. The detections of dichlobenil in precipitation is also very limited, which does indeed support the conclusion that the evaporation of dichlobenil is limited and without importance for the removal of dichlobenil under Danish conditions.

In order to determine the distribution pattern of applied dichlobenil, areas where dichlobenil possibly has been applied were mapped in detail at three sites: 1) The clayey site: the town of Søndersø, 2) the clayey site: non-urban areas of Søndersø (both Funen), and 3) the sandy site Bedsted which is a moorland plain located in the south of Jutland. The mapping is based on ortho photos from the period where dichlobenil was used (1967-1997). The applied mass of dichlobenil is subsequently evaluated from expected applied dose, expected frequency of application, and expected application area (percentage of the area where dichlobenil has possibly been applied, and which then actually constitutes a source). For the site the town of Søndersø the applied mass was calculated to be between 1.3 and 5.3 kg/year/km², whereas the calculated mass at the two non-urban sites (Søndersø and Bedsted) was between 0.2 and 1 kg/year/km². The calculated mass was scaled up to the total amount of dichlobenil applied in Denmark from calculations of the total treated area in Denmark, and this estimate was compared to the amount actually sold (29 t/year). From this comparison, the calculated applied mass at the sites seems realistic.

A mass balance was established based on the new data regarding the fate of BAM and dichlobenil and the calculated applied mass of dichlobenil, in order to elucidate the distribution patterns and the mass flows of dichlobenil and BAM in top layers, in the clayey till layers covering the aquifers, and in the aquifers. Another purpose of establishing the mass balance was to estimate the duration of the BAM contamination in the aquifers (however, this is not part of the present project).

The mass balance was calculated using two different methods: 1) a conceptual mass balance, where the mass was calculated based on typical survey concentrations in soils and groundwater, and 2) a detailed mass balance, where the mass was calculated based on model simulations.

The conceptual mass balance is a rough estimate of the actual mass of dichlobenil and BAM in Denmark in the unsaturated and the saturated zone, as well as an estimate of the mass of dichlobenil and BAM which historically has been removed from the groundwater by run-off and pumping. The mass balance was calculated based on the total mass of dichlobenil sold in Denmark and Danish survey data regarding concentrations of dichlobenil and BAM in sediments, pore and groundwater, and surface water. Degradation of BAM was not considered in the conceptual mass balance. The conceptual mass balance estimated the actual mass of dichlobenil and BAM in soil and water and the historical mass removed by run-off and pumping to be between 8 and 51% (73 – 462 tons) of the total mass of dichlobenil sold. Therefore the conceptual mass balance indicates that not all removal processes for BAM and dichlobenil have been considered, since the mass balance maximum only accounts for 51% of the mass of dichlobenil sold. However, the conceptual mass balance shows that a significant mass of dichlobenil and BAM (up till 328 tons) is present in the layers above the aquifers, since more than 95% of the present mass is found in these layers. In comparison only 1-17 tons are in the aquifers at present.

On the basis of the conceptual mass balance for Denmark, both a conceptual mass balance and a model-based mass balance have been calculated for the three sites, where the distribution and mass of dichlobenil applied have been mapped (the town of Søndersø, non-urban areas of Søndersø and Bedsted). In contrast to the conceptual mass balance, the model-based mass balance does not use the actual concentrations of dichlobenil and BAM measured in sediments and waters. Instead the mass balance is solely based on model simulations, which combine the application of dichlobenil with a dynamic description of the transport of dichlobenil and BAM from the key processes regarding the fate of the compounds. The model-based mass balance includes new data regarding the application of dichlobenil distributed by area and the degradation of BAM. On this basis it is evaluated whether the new data on degradation of BAM can explain the mass deficit indicated in the conceptual mass balance.

In general, the conclusions from the mass balances for the three sites, the town of Søndersø, non-urban Søndersø and Bedsted are:

Disregarding BAM-degradation, the conceptual estimated and the model-based mass balance show concurrently that there is a significant mass deficit of dichlobenil and BAM in sediments and waters, compared to the mass applied at the three sites. The deficit is less in the model-based mass balance, because this mass balance includes calculation of the higher concentrations seen historically in surface run-off, whereas the conceptual mass balance is based on concentrations in run-off waters measured after the period where dichlobenil was used.

By including BAM-degradation, the model-based mass balance shows good agreement between the mass of dichlobenil applied at the three sites and masses calculated. When including the degradation rates from the degradation experiments, there is no mass deficit. The model-simulated degradation of BAM accounts for 30-72% of dichlobenil applied.

By including BAM-degradation, the masses in the unsaturated zone and the aquifer at the town of Søndersø and non-urban areas of Søndersø (clayey till locations) calculated on the basis of the model-based mass balance are in good agreement with the masses calculated on the basis of the conceptual mass balance based on survey data in sediments and waters. This indicates that the model-based mass balance gives a realistic description of the transport of masses in sediment and groundwater. The calculation shows that there is a significant mass of dichlobenil and BAM in the layers covering the aquifer (14-18% of the total amount of dichlobenil applied). The mass in the aquifer is less than 3% of the dichlobenil applied. The results are consistent with the conceptual mass balance for Denmark.

The distribution of the masses was different in the mass balance for the sandy site (Bedsted), where the major part of the applied mass is run-off to streams. According to the model-based mass balance 60% of the applied mass is drained this way. Consequently, the mass remaining in the unsaturated zone is only 1% according to the model-based mass balance. For the site Bedsted the model-based mass balance and the conceptual-based mass balance is less consistent than at the two other sites. However, the differences between the mass balances are estimated to be within the deviations of the concentrations used in the conceptual estimated mass balance and the key parameters used in the model based mass balance.

A sensitivity analysis of the model-based mass balance was performed to evaluate the deviation on the results. The sensitivity in the model is evaluated regarding variations in:

Frequency in application of dichlobenil
Draining
Degradation rates for BAM
Thickness of the upper reactive zone regarding degradation.
Surface run-off (through sewer)
Thickness of the unsaturated zone for the moor land location

The sensitivity analysis shows that changes in the selected parameters within a realistic interval cause deviations in the mass balance ranging between a factor 0.4 and 3. The most important parameter is, as expected, the degradation rate for BAM. By using slower degradation rates, corresponding to the observed minimum rates, the mass of BAM degraded reduces significantly to 2% of the calculated mass. Similarly, the sensitivity analysis shows that by reducing the thickness of the reactive zone from 1m to only 0.5 m or 0.25 m, the mass of BAM degraded is reduced 30 and 60%, respectively. In this case a larger mass of BAM is drained to the clayey till layers covering the aquifers and further transported to the groundwater system.

When the deviations on all the parameters in the calculation of both the conceptual mass balance and the model-based mass balance are taken into account, the variations in the model based on the sensitivity analysis seem to be relatively limited. The model-based mass balance is therefore relatively robust due to the large deviations on the input parameters.

In general, the mass balances at the three sites show good agreement between the model-based mass balance and the conceptual-based mass balance calculated based on the survey data for sediment and groundwater in the areas. Therefore the results demonstrate that the simulated model used and the basis data for the fate of dichlobenil and BAM are consistent and useful for performing a validation and a calculation of the duration of the BAM contamination in the ground water, as performed in the previous project “Pesticides and Waterworks” (Miljøstyrelsen, 2002).

Compared to the assumptions made regarding the calculations of the duration of BAM contaminations in the project “Pesticides and Waterworks” (Miljøstyrelsen, 2002) this project contributes new knowledge regarding evaporation, sorption, degradation, surface run-off and the distribution pattern of applied dichlobenil and hence more accurate mass flow estimates. Further, the type of model applied has been improved with regard to unsaturated zone conditions and the use of data for the water balance and thickness of the clayey till layer distributed by area, which not was the case with respect to the model used in “Pesticides and Waterworks”. In “Pesticides and Waterworks” degradation of BAM was not accounted for and to adjust the mass balance, it was assumed that 50% of the amount of dichlobenil applied ran off from terrain surface to sewer drain. In comparisons the model-based mass balance in this project shows that 40% (Bedsted, sandy site) and 77-84% (the town of Søndersø and non-urban areas of Søndersø, clay sites) of the amount of dichlobenil applied is removed by degradation and surface run-off. For the sandy site (Bedsted) baseflow also accounted for a large amount (61 %) of the dichlobenil applied. Baseflow was not included in the model calculations for sandy aquifers in “Pesticides and Waterworks”. Compared to the previous calculations these new assumptions (degradation, run-off and baseflow) will probably result in lower groundwater concentrations and shorter duration of BAM contaminations for both clay and sandy sites.

With regard to the waterworks the renewed estimates of the modelled mass balance covering degradation in unsaturated zones imply that the duration of the BAM contamination in primary aquifers will be shorter than previously found by modelling. The results of the mass balance show that a significant mass of dichlobenil and BAM is present in the unsaturated zone – mainly on locations with clay. As the fast degradation (with a half-life less than one year) seems solely associated with the upper part of the unsaturated zone (within the top one meter), the mass of BAM located in the deeper part of the cover layers – also due to the substantial sorption in these layers – will only be removed very slowly or be transported to the underlying groundwater aquifer. Because this project (and other projects) have not identified degradation of BAM in the deeper cover layers and groundwater aquifers, it is not likely that this mass will be removed by degradation. Hence, the future development in concentrations of BAM is then a question of whether the mass of BAM applied from unsaturated zones is more or less than the corresponding removal of mass by pumping or run-off.