Pesticide leaching at SILSTRUP
Chapter 4, PLAP - Monitoring results May 1999 - June 2002
4.1 Materials and methods
4.1.1 Site description and monitoring design
The test field at Silstrup is located south of Thisted in northwestern Jutland (Figure 1). The cultivated area is 1.69 ha (91 x 185 m) and slopes gently 1-2° to the north. Based on two
profiles excavated in the buffer zone bordering the field the soil was classified as Alfic Argiudoll and Typic Hapludoll (Soil Survey Staff, 1999). The topsoil content of clay in the two profiles was 18.3 and 26.6%, and the organic carbon content was 3.4 and 2.8%. The geological description showed a rather homogeneous clay till rich in chalk and chert, containing 20-35% clay, 20-40% silt and 20-40% sand (Figure 21). In some intervals the till was more sandy, containing only 12-14% clay. Moreover, thin lenses of silt and sand were found in some of the wells. The gravel content was approx. 5%, but could be as high as 20%. A brief description of the sampling procedure is provided in Appendix 2. The monitoring design and test site are described in detail in Lindhardt et al. (2001) and the analysis methods in Kjær et al . (2002).
4.1.2 Agricultural management
Cattle slurry (36.5 tonnes/ha) was spread on 19 April 2000, whereafter the field was ploughed. Fodder beet (cv. Kyros) was sown on 5 May and emerged unevenly across the field within 1 to 3 weeks. The herbicides metamitron, phenmedipham, desmedipham and ethofumesate were applied on 22 May, 15 June and 12 July. Potassium bromide tracer was applied on 22 May. The field was sprayed with fluazifop-p-butyl on 28 June to combat wild oats and with pirimicarb on 5 July to combat aphids. The crop was harvested on 15 November yielding 134.5 hkg/ha of beet roots (100% dry matter) and 26.3 hkg/ha of beet tops. Taken together, the dry matter yield was at the same level as the normal yield recorded in the area that year.
The field was ploughed in spring 2001. Due to ample precipitation, sowing of the spring barley (cv. Otira) was delayed until 9 May. Crop emergence was recorded 11 days later. The herbicides tribenuron methyl and flamprop-M-isopropyl were sprayed on 9 and 21 June, respectively. The fungicides propiconazole and fenpropimorph were applied on 21 June and 4 July. The insecticide dimethoate was sprayed on 6 July. Despite the very late sowing, grain yield at harvest on 5 September was as high as 88.0 hkg/ha (15% water content). Precipitation prevented the straw being pressed until late October, resulting in a low straw yield of 28.6 hkg/ha (dry matter).
Figure 20. Overview of the Silstrup site. The innermost white area indicates the cultivated land, while the grey area indicates the surrounding buffer zone. The positions of the various installations are indicated, as is the direction of groundwater flow (by an arrow).
Figure 21. Geological description of the Silstrup site (Lindhardt et al., 2001).
On 25 October 2001 the field was sprayed with glyphosate in the form of Roundup Bio (4.0 l/ha). The field was ploughed to a depth of 22 cm on 18 December. Maize (cv. Loft) was sown on 25 April 2002 after the field had been fertilized with cattle slurry (40.3 tonnes/ha) on 22 April. When two leaves had unfolded the maize was sprayed with pyridate + terbuthylazine to combat weeds. This was repeated on 3 June. On 19 June the maize was sprayed with clopyralid to combat weeds. The crop was harvested on 23 September yielding 134.3 hkg/ha (100% dry matter), somewhat less than other cultivars in the area that year. Management practice at the site is detailed in Appendix 3 (Table A3.3).
4.1.3 Model set-up and calibration
The MACRO model was applied to the Silstrup site covering the soil profile to a depth of 5 m b.g.s., always including the groundwater table. The model was used to simulate the water flow in the unsaturated zone during the full monitoring period April 2000-June 2002 and to establish an annual water balance.
The model was calibrated to the observed groundwater table measured in the piezometers located in the buffer zone, to time series of soil water content measured at three depths (25, 60 and 110 cm b.g.s.) from two profiles S1 and S2 (see Figure 20) as well as to measured drainage flow. A simple calibration procedure was applied that only involved adjustment of the empirical BGRAD parameter regulating the boundary flow and the drain depth, which was determined by the groundwater level during drainage periods. All remaining parameters were based on measured data or literature/default values. The calibration procedure is briefly described in Section 2.1.3. For a detailed description of data acquisition, model setup and calibration procedures, see Kjær et al. (2002).
4.2 Results and discussion
4.2.1 Soil water dynamics and water balances
The model simulations were largely consistent with the observed data, thus indicating a reasonable model description of the overall soil water dynamics in the unsaturated zone. The dynamics and level of the groundwater table were captured well by the model except for the initial rise in the autumn 2000, when percolation and drainage flow were initiated. The delayed rise in the groundwater table resulted in a delayed response in the drainage flow in November 2000. Similarly, the modelled drainage flow is delayed compared to the measured drainage in September 2001 due to the simulated groundwater table being too low (Figure 22B and C).
The overall trends in soil water content could be modelled reasonably well, especially in the A horizon (Figure 22D). In the subsoil the model tends to describe a dryer soil during the summer periods than measured by the deeper> TDR probes (Figure 22E and F). Unexpectedly, the measured time series at 60 and 110 cm b.g.s. were not affected by the lower groundwater table during the two summer periods. Measured water saturation ranged from 90 to 110%, with the highest values during the driest periods. According to the measured retention curves the soil water content should have been approximately 80%, as simulated by the model. The quality of the measured time series has been thoroughly analysed. The unexpected pattern could be related to the use of the general relationship between measured primary TDR data and the calculated soil water content. The use of a calibrated, soilspecific relationship would improve the findings. Another explanation may be the limited applicability of TDR in near-saturated soils and in the heavy clayey and swelling soils in which the probes could have been installed at Silstrup (clay content up to 43%, according to Lindhardt et al ., 2001).
Closer examination of measured and modelled drainage flow during the period October 2001 to April 2002 (Figure 23) reveals that the drainage flow pattern at Silstrup is dominated by transient peaks of high flow typically lasting 4-7 days, often separated by no-flow periods. The flow pattern thus seems to be dominated by macropore flow generated during major precipitation events - a finding supported by the rapid occurrence of bromide in the drainage water and suction cups located 1 m b.g.s (Figure 24A and B). The model, which is as yet uncalibrated, could not fully match this flow pattern, but the overall trends and dynamics of the drainage flow are reasonably well simulated.
The resulting water balance is shown in Table 7 (July to June) for the three modelled years. The first and third years were wet years at Silstrup, while the second year was dry. The simulated drainage flow was higher than the measured drainage flow in both of the two monitoring years in which drainage flow was measured, the discrepancy being greatest in the rather wet year of 2001/2002. Despite a difference in precipitation of 138 mm between these two monitoring years the measured drainage flow was similar. Thus the additional precipitation input in the last monitoring year mainly percolated deeper into the soil. Simulated groundwater recharge ranged from 257 to 373 mm/yr. Simulated percolation 1 m b.g.s. is generally continuous from September/October until late spring with precipitation events exceeding approximately 15 mm/d immediately being reflected in the percolation (Figure 22A).
Table 7. Annual water balance for Silstrup (mm/year). Precipitation is corrected to the soil surface according to the method of Allerup and Madsen (1979)
Figure 22. Soil water dynamics at Silstrup: Locally measured precipitation and simulated percolation 1 m b.g.s. (A), simulated and measured groundwater level (B), simulated and measured drainage flow (C), and simulated and measured soil water saturation (SW sat.) at three different soil depths (D, E and F). The measured data in B derive from piezometers located in the buffer zone. The measured data in D, E and F derive from TDR probes installed at S1 and S2 (see Figure 20).
Figure 23. Simulated and measured drainage flow at Silstrup from October 2000 to April 2002.
4.2.2 Bromide leaching
Two large storm events occurred a few days prior to and after the application of the bromide tracer on 22 May 2000. The first event caused the onset of a minor flow of drainage water, while the second resulted in rapid percolation and breakthrough of bromide to the drainage system, with the concentration reaching 5.1 mg/l on 29 May (Figure 24C). At Silstrup the upper macropore zone extends down to 1.3 m b.g.s. (Lindhardt et al ., 2001). The zone is heavily fractured and contains numerous biopores coated with clay and organic matter. When the bromide was applied, the groundwater table was located around 1.25 m b.g.s. (Figure 22B). The presence of macropores and the location of the groundwater at the time of bromide application were reflected in the almost instantaneous occurrence of bromide in the drainage water, suction cups S1 and S2 (Figure 24A, B and C) and in the uppermost filters of all but one of the downstream, vertical wells (Figure 25). The orientation and magnitude of the fractures may also explain why bromide was detected in the lowermost screen of M12, which is located upstream of the test field.
Total bromide recovery during the two-year monitoring period was 2.1 kg/ha, indicating that only 11% of the applied tracer had leached into the drains. The elevated bromide concentration detected in the suction cups and drainage water in 2002 indicate that bromide continued to leach from the unsaturated zone as long as two years after application. In conclusion, the overall distribution of bromide in the test field indicates that most of the bromide is retained in the upper part of the soil profile, probably in the clay matrix. Continuous, slow leaching of bromide can therefore be expected for a long period of time.
Figure 24. Bromide concentration at Silstrup. A and B refer to suction cups located at S1 and S2. The bromide concentration is also shown for drainage runoff (C), the horizontal monitoring wells H1 and H2 (D) and vertical monitoring well M5 (E). The green vertical line indicates the date of bromide application.
Figure 25. Bromide concentration at Silstrup. The data derive from the vertical monitoring wells (M5-M12). Screen depth is indicated in m b.g.s. The green vertical line indicates the date of bromide application.
4.2.3 Pesticide leaching
Monitoring began at Silstrup in April 2000 and presently encompasses 12 pesticides and 13 degradation products (Table 8 and Figure 26). It should be noted that precipitation in Table 8 is corrected to soil surface according to Allerup and Madsen (1979), whereas percolation (1 m b.g.s.) refers to accumulated percolation as simulated with the MACRO model. It should also be noted that as tribenuron methyl (applied here as Express) and pyridate (applied here as Lido) degrade rapidly, the leaching risk is more associated with their respective degradation products, triazinamin-methyl and PHPC. For the same reasons it is the degradation products and not the parent compounds that are monitored in the PLAP (Table 8).
Table 8. Pesticides analysed at Silstrup with the product used shown in parentheses. Degradation products are in italics. Precipitation and percolation are accumulated from the date of first application (App. date) until 1 July 2002. 1 st month percolation refers to accumulated percolation within the first month after application. Cmean refers to average concentration in the drainage water. The number of pesticide-positive samples is indicated in parentheses.
Figure 26. Pesticide application, precipitation and irrigation (primary axis) and simulated percolation 1 m b.g.s. (secondary axis) at Silstrup. Desm.: desmedipham; Phenm.: phenmedipham; Fenp.: fenpropimorph; Prop.: propiconazole. Pesticides applied later than April 2002 are not included.
With desmedipham, phenmedipham and fluazifop-P-butyl, the leaching risk was found to be negligible at the Silstrup site. Apart from one sample containing less than 0.1 µg/l of desmedipham and fluazifop, none of these compounds nor the degradation products listed in Table 8 were detected during the two-year monitoring period. Metamitron, metamitrondesamino, ethofumesate and pirimicarb did leach from the root zone, but not at unacceptable levels. The findings are briefly summarized below. For a detailed description of the leaching pattern, see Kjær et al . (2002).
The leaching risk of the pesticides applied to the spring barley in 2001 and maize in 2002 will not be evaluated until the 2003 monitoring results become available, i.e. when two years of monitoring data have been collated. Nevertheless, no evidence was found to indicate leaching of fenpropimorph and triazinamin-methyl, which were only detected in one sample (fenpropimorphic acid, 0.019 µg/l; Table 8). In contrast, the other four pesticides applied in 2001 were found to leach from the root zone during the current monitoring period:
- Pirimicarb was detected in several drainage water samples, although always at concen trations below 0.1 µg/l. Pirimicarb was not detected in the suction cups, but was detected at a concentration of 0.01 µg/l in three groundwater samples. Leaching of pirimicarb was confined to a nine-month period ending in April 2001.
- Evidence of ethofumesate leaching was seen at the S2 suction cups located 1 m b.g.s. as well as in the drainage water. Throughout the 2000/2001 leaching period, ethofumesate leached to the drainage system at an average concentration reaching 0.03 µg/l. The concentration only exceeded 0.1 µg/l in one sample. In groundwater, ethofumesate was detected at concentrations ranging from 0.01 to 0.02 µg/l in four samples. Leaching of ethofumesate was confined to a six-month period ending in January 2001.
- Throughout the 2000/2001 leaching period, metamitron and metamitron-desamino leached to the drainage system at average concentrations of 0.05 and 0.06 µg/l respectively. In addition, both compounds were detected in several samples of groundwater and samples from the suction cups located 2 m b.g.s. In total, only four groundwater samples contained concentrations exceeding 0.1 µg/l. The maximum concentration detected was 0.17 µg/l for metamitron and 0.13 µg/l for metamitron-desamino. Leaching of metamitron was confined to a nine-month period ending in April 2001, whereas metamitron-desamino continued to leach to the drains at low concentrations more than one year after application. During the leaching period 2001/2002, metamitron-desamino was detected in nine samples, the latest of which was a drainage water sample from October 2002. In all instances the concentration was below 0.1 µg/l.
At Silstrup, pesticide leaching appears to be associated with pronounced macropore transport resulting in very rapid movement of pesticides through the unsaturated zone. According to the hydrological modelling, flow was dominated by macropore flow generated during major storm events (Section 4.2.1.). These findings are in concert with the observed pattern of pesticide flow, where the leaching of pesticides to the drains was completely governed by the individual storm/flow events. Thus sudden storm events accounted for 92% of all the metamitron leached, 89% of the metamitron-desamino, 97% of the ethofumesate, 88% of the glyphosate and 80% of the AMPA.
- Shortly after dimethoate had been applied in July 2001, a large precipitation event caused rapid leaching through the unsaturated zone (Figure 26). On 24 July, just eight days after application, dimethoate was detected at a concentration of 1.42 µg/l in the drainage water. This one sample is the reason why the average drainage water concentration of dimethoate is 0.02 µg/l (Table 8).
- Minor leaching of flamprop-M-isopropyl, flamprop (free acid) and propiconazole was also observed. All three substances were detected in several drainwater samples. Apart from one sample containing 0.11 µg/l of flamprop-M-isopropyl, all concentrations were below 0.1 µg/l (Figure 27).
- On 25 October 2001 the field was sprayed with glyphosate (1.44 kg/ha) in the form of Roundup Bio (4.0 l/ha). Prior to application there had been 4 major storm events yielding up to 11 mm/day of drainage runoff (Figure 28). The day before the field was sprayed there was 9 mm of precipitation. The preceding 13 days were practically precipitation-free, however. Five days after spraying, 12 mm of precipitation caused approximately 2 mm of runoff in which the flow-proportional concentration of glyphosate was 4.7 µg/l, and the time-proportional concentration was 1.9 µg/l. The corresponding AMPA concentrations were 0.06 and 0.14 µg/l, respectively (Figure 28B). The glyphosate concentration constantly decreased during the remainder of the leaching period 2001/2002. The AMPA concentration was lower, but more stable during the leaching period (Figure 28C; Appendix 6). Glyphosate and AMPA were detected in all drainage water samples except one. The weighted average concentration of glyphosate in the drainage water was 0.13 µg/l, while that of AMPA was 0.06 µg/l. The concentrations might have been even higher had not November and December been so much dryer than usual (Appendix 4), resulting in considerably greater drainage runoff than in the preceding and following periods (Figure 28). It should be noted that drainage runoff commenced about one month prior to the application of glyphosate, and that the weighted average concentration refers to the period from the date of application until 1 July 2002. In addition glyphosate was detected in 3 groundwater samples and AMPA in 7, in each case at concentrations below 0.1 µg/l (Appendix 6).
It should be noted that in Figure 27 and Figure 28, time-proportional sampling refers to continuous drainage runoff occurring throughout the whole drainage season, whereas the flow-proportional sampling refers to the drainage runoff induced by the sudden storm events occurring several times during the drainage season.
Figure 27. Precipitation (A) together with concentration of flamprop-M-isopropyl (B), flamprop (free acid) (C) and propiconazole (D) in the drainage runoff at Silstrup. The green vertical lines indicate the date of application.
Figure 28. Precipitation (A) together with concentration of glyphosate (B), AMPA (C) and bromide (D) in the drainage runoff at Silstrup. The green vertical lines indicate the date of application. Bromide was applied in May 2000.
At Silstrup, the leaching risk of pesticides applied in 2000 can be summarized as follows:
The leaching risk of pesticides applied in 2001 and 2002 cannot be fully evaluated at present as the potential leaching period extends beyond the current monitoring period. The preliminary findings are that:
- With desmedipham, phenmedipham and fluazifop-P-butyl, the leaching risk was negligible.
- Metamitron, metamitron-desamino, ethofumesate and pirimicarb leached from the root zone during the current monitoring period, but not at unacceptable levels. Although the concentration exceeded 0.1 µg/l in several samples, the average concentration did not. Metamitron-desamino continued to leach from the root zone more than one year after application, whereas leaching of pirimicarb and metamitron/ethofumesate was confined to a six-month and nine-month period, respectively.
- No evidence was found to indicate leaching of fenpropimorph and triazinamin-methyl (degradation product of tribenuron methyl), which were only detected in one sample (fenpropimorphic acid, 0.019 µg/l).
- Minor leaching of flamprop-M-isopropyl, flamprop (free acid), propiconazole and di-methoate, but not in unacceptable levels. Apart from two samples, all concentrations were below 0.1 µg/l.
- Glyphosate leached from the root zone at concentrations exceeding 0.1 µg/l, the average concentration in the drainage water being 0.13 mg/l. AMPA was also detected in the drainage water, but the average concentration was only 0.06 µg/l. In addition, glyphosate was detected in 3 groundwater samples and AMPA in 7, in each case at concentrations below 0.1 µg/l.