| Front page | | Contents | | Previous | | Next |

Vurdering af koncentration og varighed af BAM-forurening i grundvand

Summary and conclusions

Preconditions

This report evaluates the extent and durability of contamination of groundwater and waterworks with the pesticide metabolite BAM (2,6-dichlorbenzamide) for representative main aquifer types in Denmark.

The evaluation is carried out by means of mathematical modeling using the groundwater model codes FRAC3Dvs and MODFLOW/MT3D.

Using a 21 km² hypothetical groundwater catchment as overall model frame, the evaluation includes the following main geological conditions and aquifers,

1. thick (90 m) sandy aquifer without protection from overlaying aquitards (clay layers)
2. thick sandy aquifer covered by fractured aquitards of different thickness (6, 16 and 30 m)
3. thin (20 m) limestone aquifer covered by fractured aquitards of different thickness (6, 16 and 30 m)
4. thick sandy aquifer covered by terminal moraines or otherwise heterogeneous clay layers
5. thick sandy aquifer underneath inter-layered clay layers with and without "sand windows"

Model input data for the contaminants represents laboratory measurements for sorption and degradation of the parent pesticide dichlobenile and its metabolite BAM, together with the following assumptions and estimates for the application of the diclobenile,

1. pesticides (Prefix and Casoron) containing dichlobenile have exclusively been applied in urban and agricultural land use types (74% of area in Denmark, 1982-figures)
2. the actually treated area is 8 times large in the urban land use type than in the agricultural type
3. the total treatment in the two land use types is equal to the total sale of the pesticide in Denmark (29 tons/year, 1966-97)
4. half of the amount sold has been flushed through surface drainage, while the remaining pesticide mass has been infiltrated into the soil

Model in-put data furthermore include representative values for the overall water balance, and representative soil data and geological conditions for the aquifer types, including the influence of fractures and minor sandy layers in clay aquitards.

Local deviation and uncertainty for the applied modeling parameters and conditions is inevitable. However, it is expected that the model evaluation provides an overall picture and a useful analysis of the mean extent and durability of the BAM pollution of the groundwater and waterworks in Denmark.

Modeling results

The modeling shows that approximately 99% of the infiltrated dichlobenile has been transformed to its metabolite BAM, and that currently (year 2001) approximately 100% of this compound has been leached to greater depth than 2 – 3 m. This result is in agreement with the findings of residuals of dichlobenile and BAM in soil samples from the upper few meters of investigated soil profiles.

Under the crucial precondition that BAM is not degraded under the root zone, the modeling shows that BAM will appear widely in the groundwater and/or in the above aquitards and in surface water.

The modeling shows that the BAM pollution will appear with a high frequency in water extraction wells in extensive parts of Denmark during the following 20 years to more than 100 years

The highest BAM concentrations appear in groundwater and wells under the urban area of the model catchment, while the BAM concentrations in the agricultural area in general are under, or near, the drinking water standard (0.1 µg/l). The low BAM model values in the agricultural model area indicates that the frequency of BAM findings in actual water wells under this type of land use will be very sensitive to local variations in the amount and distribution of dichlobenile application.

The modeling moreover show that the extent and future evolution of the BAM pollution is a strong function of local geological and hydrological conditions, which suggests the existence of a potential for minimizing problems for the water supply through planning and management.

The modeling indicates that the highest current BAM concentrations will occur in the groundwater of aquifers without or with thin protecting clay aquitards. However, these are also the situations where the contamination has the shortest durability. The shortest durability of the BAM aquifer pollution will occur in areas which additionally are located down stream in the groundwater catchment and are not influenced by upstream BAM sources.

The modeling shows that the lowest current BAM concentrations in the groundwater should be expected in areas, where the contamination occur through thick clay layers. These areas, however, are also the areas where the contaminations will have the longest durability. The longest durability is indicated to occur in areas where BAM contaminant sources are situated in large areas upstream to the water extraction.

For the model area with 16 m clayey till above the aquifer, the modeling indicate current BAM concentrations in the water extraction wells, which corresponds to typical concentrations (0.01- 0.1 µg/l) found in the groundwater monitoring. For the time until year 2080 the modeling indicate that the concentrations will increase to exceed the drinking water standard in the urban model area, but only up to the same standard in the agricultural area. The frequency of actual BAM findings in aquifer underneath the latter type of land use, therefore will be uncertain and sensitive to the local pattern of dichlobenile application .

The modeling moreover indicate that aquifers underneath BAM-contaminated surface layer will be less vulnerable in areas with high run-of in the surface layers and small groundwater recharge. For specific areas the actual BAM values in the groundwater will, hence, be an additional function of the area distributed water balance in ground water catchments.

The modeling moreover indicate that BAM contamination of groundwater underneath thick clay layers (30 m or more) will not break through to the aquifers before approximately year 2050 to 2070, however, hereafter the BAM concentration will increase until approximately year 2150. Hence, the modeling indicates that water extraction may occur without BAM content from aquifers underneath thick clay aquitards until the middle or late in this century. A similar result is indicated to occur for aquifers underneath inter-layered clay layers in sandy deposits and underneath clay aquitards without fractures.

It should be noticed, however, that the aquitard simulations are highly sensitive and uncertain to the appearance of fractures in deep clay layers. Modeling of sensitivity indicate that there will be current BAM breakthrough (however, with much higher concentrations than observed in the monitoring program) in clay areas with wider spacing between deep fractures (5 – 10 m) than simulated for the above main aquifer types.

The modeling moreover indicate that the groundwater will be more exposed to BAM pollution underneath aquitards of glacially disturbed clay, e.g. terminal moraines and other heterogeneous clay types containing abundant thin sand layers. In such areas the model indicate that, everything else equal, there will be current breakthrough of BAM to aquifers and following long duration of contamination in water extraction wells.

The modeling indicates that the lowest BAM concentrations, everything else equal, will occur in water extractions wells of high yield (365.000 m³/year) from thick aquifers. This is because of dilution due to vertical mixing with water free from BAM. At the same token, however, there should in many cases be expected a higher frequence of BAM findings in such wells, because they will mobilize BAM contamination from more distant sources than wells with smaller yield (<50.000 m³/year).

The dilution mentioned above will be absent or smaller, in the case of water extraction from thin aquifers. This has particular relevance to limestone aquifers, because water extraction from these often occur from a relatively thin glacially crushed upper zone in the limestone, which allows for only limited vertical mixing. For the same reason are the modeled BAM concentrations and frequency of findings in the limestone aquifer model situation generally higher relative to the other aquifer types. This is in agreement with the general distribution of BAM observed in the groundwater monitoring program.

In comparison with the observed current BAM concentrations in the ground water monitoring the model apparently predicts too high concentrations in the actual model set-up for the aquifer types with no clay layer or with only a thin clay layer (6 m) above the aquifer.

On the basis of such local data, it will be feasible to use the model-setup developed in the current study, to identify sub-areas within specific groundwater catchments where groundwater supply can be maintained without or with minimum influence of BAM pollution.