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

Facilitated transport of pesticides

5 Discussion

5.1 Pesticide leaching

The estimated total losses of Pendimethalin to the drains (0.0013 % of the applied amount on average for the two seasons) are in fair agreement with losses reported elsewhere for structured field soils. The quantities of Pendimethalin lost with drain water from tile drains range from 0 to 0.040 % of the applied dose (Czapar et al., 1994; Jones et al., 1995; Traub-Eberhard et al.,1995; Vencill et al., 1999). Traub-Eberhard et al. (1995) investigated the movement of Pendimethalin into tile drains in a one year experiment involving two soil types. The estimated total Pendimethalin loss following autumn application was less than 0.001% of the applied dose on a silt loam soil, whereas no Pendimethalin was detected in drain water collected from a sandy soil. They attributed the absence of Pendimethalin in drain water from the sandy soil to less occurrence of preferential flow due to low stability and lack of continuity of macropores. Between 0.009 and 0.040 % of the applied Pendimethalin dose was lost with drain water in a one year field study on a cracking, intensively drained heavy clay soil in England (Jones et al., 1995). Czapar et al. (1992) studied the movement of Pendimethalin in packed soil columns with and without artificially made continuous macropores. Following heavy irrigation, Pendimethalin was detected in drainage from columns with a continuous macropore but not in columns without macropores. Starret et al. (1996) studied the movement of Pendimethalin in 50 cm long undisturbed soil columns. Traces of the herbicide were found in leachate from the columns following heavy irrigation (four 2.54-cm applications) but not after light irrigation (sixteen 0.64-cm applications). On average, 0.2 % of the applied Pendimethalin was found in leachate from the columns after heavy irrigation.

The maximum Pendimethalin concentration reported by Traub-Eberhard et al. (1995) (0.7 µg l^-1) was more than ten times smaller than the maximum concentration found in the present study. This difference may be caused by natural variations between the soils in the two studies. It could also be caused by different lag periods between application and first rainfall generating drain flow. However, in the present study the leaching of Pendimethalin was first detected 25 days after application in the season 1999-2000 and after 13-14 days in 2000-2001. This and the moderate amounts of precipitation in the two seasons does not make the present study a “worst case scenario”. The difference in maximum concentration of the present study and that of Traub-Eberhard et al. (1995) may also be due to the analytical methods, the results of Traub-Eberhard et al. (1995) being based on filtered water samples alone (undefined pore size), or it may be due to a differing sampling strategy.

The large amount of Pendimethalin remaining in the soil in February 2000 (81 %) indicates that only a small part of the applied Pendimethalin has been degraded. This is in agreement with the results of Traub-Eberhard et al. (1995), who reported an estimated time to 50 % dissipation (DT50 value) for Pendimethalin of more than 300 days following autumn application on a silty loam soil. It also confirms with earlier reported DT50 values of 3-4 months (Walker and Bond, 1977). The investigation for remaining Pendimethalin was based on soil coring in February 2000 to a depth of 35 cm. This depth was chosen mainly in consequence of the findings of Traub-Eberhard et al. (1995). They sampled the soil intensively to a depth of 90 cm but were unable to detect Pendimethalin below ploughing depth.

5.2 Leaching dynamics

5.2.1 Soil particles

Soil particle concentrations (expressed by turbidity) generally peaked shortly after initiation of the drain flow events and before the occurrence of maximum drain flow rates. This is in agreement with observations from similar field experiments (Villholth et al., 2000) and erosion experiments showing that the sediment concentrations peak within the first 3-5 minutes of a rainfall event (Sander et al., 1996).

On unfrozen soil, large peak values for turbidity (T > 50 NTU) always occurred during or shortly after (within 8 hours) relatively strong rain events, coinciding with the occurrence of large moisture contents (θ_L) at 8 and 16 cm depth. Very large turbidity values were observed in break of frost situations where the soil was still partly frozen. These are both conditions (very wet conditions and frozen soil) that favour macropore flow and most likely also reduce reattachment/filtering of particles since pores will be fully wetted and flow will be fast.

Thus, in order to sample these peak values, focus should be given to drain flow events occurring in break of frost situations, and on rain events creating soil moisture contents considerably above field capacity in the entire plow layer.

5.2.2 Pendimethalin and Ioxynil

Pendimethalin concentrations as turbidity generally peaked shortly after initiation of the drain flow events and before the occurrence of maximum drain flow rates.

A similar pattern of pesticide concentrations in flow from tile drains have been reported elsewhere (Brown et al., 1995; Kladivko et al., 1991; Villholth et al., 2000). The early breakthrough of Pendimethalin indicates, that the primary means of transport of Pendimethalin from the surface to the drains is through macropore transport. The hydraulic conductivity of macropores is mostly decades higher than the hydraulic conductivity of the matrix. The very fast transport through the macropores could cause equilibrium sorption conditions to not be reached, hence allowing for a relatively greater portion of the hydrophobic pesticide to pass through the soil. In the matrix, transport is slower and sorption equilibrium can be reached. Also the filtering capacity to soil particles is significantly (infinitely) higher in matrix than in macropores, which would also explain the early findings of maximum concentrations of particles in drain water. Because of the limited water transport capacity of the macropores (macropores only make up a small percentage of the overall porosity), maximum concentrations of hydrophilic compounds will only occur after breakthrough of matrix transport has been reached (see e.g. Addiscott et al., 1978). Any drain flow originating from the soil matrix is expected to reduce rather than increase turbidity and Pendimethalin concentrations. For Ioxynil, the main transport is expected to occur through the matrix. Breakthrough of Ioxynil could be to some extent controlled by earlier events, in the sense that Ioxynil could be stored in the pore water of the unsaturated zone and in the saturated zone, until a new rainfall strong enough to generate drain flow occurs.

5.3 Particle-facilitated transport

5.3.1 Correlation between turbidity and Pendimethalin

Results showed a strong correlation between turbidity and Pendimethalin concentration during the first most important drain flow events after Pendimethalin application in both seasons. The correlation declined somewhat over time (with event no.), see Table 3.2. This decline is most likely due to the fact that the accessibility of Pendimethalin available for transportation on the soil surface declines with event no., whereas the supply of soil particles is more or less infinite. Gharidi and Rose (1991) by peeling of soil aggregates found, as would be expected, that outer fractions of soil aggregates contained considerably more pesticides than inner fractions. They also found that the enrichment (with pesticides) of the particles detached by rainfall decreased with time (30 and 50 minutes) over a rainfall event, in accordance with the outer fractions of the aggregates being detached before inner fractions.

Even though the correlation is fairly convincing, this does not prove that Pendimethalin is transported associated to colloidal particles. The correlation could be an indirect one, in the sense that Pendimethalin and particle concentrations could both be positively correlated to another phenomenon. This phenomenon could be the flow rate and the mechanism would be that a high flow rate would lead to a rinsing effect of drain tiles, in which particles and Pendimethalin were deposited in a porous deposit on the bottom of the tiles. However, since particle content, and hence Pendimethalin concentrations were inversely correlated to flow rate, this was not the case.

The underlying phenomenon could also be macropore transport. As argued in the preceding section, there is a physically plausible reason why macropore transport would correlate positively with both particle and Pendimethalin concentrations. The decreasing correlation factor with event no. could then be further explained by the fact that in the areas surrounding the ’openings’ of the macropores in the top soil, Pendimethalin would be washed out to the maximum extent during the first events, leaving behind only more strongly bound Pendimethalin.

However, since Pendimethalin sorbs so strongly to bulk soil and soil particles, it is also very likely that some of the soil particles observed to be washed out through the drains carry Pendimethalin with them. Also, the modelling showed that colloid-facilitated transport through macropores was the totally dominant transport mechanism for a strongly sorbing pesticide, such as Pendimethalin, and even modelling Pendimethalin as a conservative tracer, the observed concentration levels were not achieved, again indicating that a supplementing transport mechanism was acting. For Ioxynil there was no correlation between particle content (turbidity) and Ioxynil concentration, indicating that particle-associated transport was not an important mechanism for Ioxynil. The modelling exercise, employing different K_d-values, showed however, that even for lightly sorbing compounds such as Ioxynil (Koc ~100), particle-facilitated transport may be an important additional transport mechanism together with dissolved transport (see Table 4.2). In this case this was not supported by the experimental evidence, that was collected in the current study.

5.3.2 Fractionation analysis

The fractionation analysis for Pendimethalin performed in the present study indicated that particle-facilitated transport of Pendimethalin takes place, but is unclear in quantifying it.

The total mass of Pendimethalin found in the fraction above 0,7 µm filter size was 6 % in the season 1999-2000 and 15 % in the season 2000-2001 at Rørrendegaard. For the single 5 liter sample the same fraction was 19 % and no Pendimethalin appeared to be separated out of the water phase through the subsequent centrifugation (see Figure 3.7). In the samples taken from the model areas, between 30 and 60 % of the total Pendimethalin content was found in the solid phase extracted via the centrifugation. This significant difference in particle associated Pendimethalin could possibly be explained by differences in the composition of the soils at the different places. According to the sorption experiments (Table 3.6), the soil from the model areas should then be richer in organic carbon or clay content. The average clay content for the top soil of the three localities is Lillebæk: ~16%; Rørrendegaard: ~12%; Odderbæk: ~6% and the organic C content is Odderbæk: ~2,3-2,5%, Lillebæk: ~1-1,8%; Rørrendegaard: ~1,2%. The topsoil of the model areas are somewhat richer in organic carbon, but the same hypothesis can not explain the variation between the two model areas. The difference in results could also be caused by the experimental procedure. The samples were stored for periods of days to weeks before filtration and centrifugation. The five liter sample from Rørrendegaard was taken out on December 18, 2000 and centrifuged almost 7 months later on July 12, 2001. The samples from the model areas were centrifuged the day after the sampling.

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