pesticides and groundwater - GEUS

6 downloads 0 Views 198KB Size Report
The selected pesticides were: Mecoprop, dichlorprop, 2,4,5-T, metsulfuron-methyl, isoproturon, and atrazine at concentrations of 250 µg/l (HPLC) and 25 µg/l ...
The Danish Environmental Research Programme

PESTICIDES AND GROUNDWATER

Annual Report

The Groundwater Group December 1998

1. INTRODUCTION The Groundwater Group has now been working for 2½ years within the Danish Environmental Research Programme, ´Pesticides and Groundwater´ (SMP 96). The time has been characterised by increased findings of pesticides in Danish groundwater and, due to that, an intense public debate in both the Parliament and in public media. Pesticides and groundwater constitute a very complex field of research, mainly due to the chemistry of the compounds and the fact that pesticides with new chemistries are introduced all the time by the agrochemical industry. In Denmark, it has been decided to establish protection zones where pesticide use are either banned or restricted in certain vulnerable catchment areas, despite that our knowledge of e.g. transport mechanisms, sorption, and degradation is still too limited to accurately define the size of such zones needed to protect the drinking water resource. Although such problem areas may appear to complex to expect final and definite answers of the present efforts, the ongoing research will make a strong contribution of the many questions related to transport and fate of pesticides in aquifers. The research carried out by the Groundwater Group generally follows the approved programme. Within project 1, “Preferential transport of pesticides to groundwater” all field sites have been established and studies of pesticide transport are in progress. A thorough characterisation of the clayey tills has been carried out and it is concluded that preferential transport in macropores predominantly controls the migration of pesticides to the groundwater. Heterogeneous structures were also observed at sandy soils and the importance of these structures for transport and degradation of pesticides is now being investigated. Within project 2, “Fate of pesticides in aquifers” sorption and degradation of several pesticides in aquifers and wetlands are studied. One pesticide field injection study has been finished and another is in progress. Until now microbial degradation of the pesticides was only seen at aerobic aquifer conditions and compared to topsoil the degradation proceeds much more slowly. However, due to the long pesticide resistance time in aquifers microbial degradation is expected to control the fate of these compounds at least at aerobic conditions. Pesticide degrading pure cultures of bacteria have been isolated and used in studies of factors limiting degradation in aquifers. Within project 3, “Large-scale modelling of pesticide transport” appropriate structures for both the catchment and regional models have been identified. Furthermore, field sites to be used for validation of models have been selected, and programmes for data collection and field investigations are in progress. An evaluation of the Groundwater Group activities was initiated at a meeting at Sonnerupgård Gods, September 11-12. At the meeting the content and basic ideas of each project were discussed with a scientific evaluation panel headed by Prof. Poul Harremoës, and comprising Prof. Joseph M. Suflita, Prof. Wolfgang Kinzelbach and Docent John Stenström. In general, the outcome of the meeting was positive, although the panel emphasised some weaknesses e.g. in defining the pesticide input function needed for modelling pesticide transport in soil and groundwater (project 3). Based on comments from the evaluation panel the Project Board decided to distribute most of the reserved funds to specific projects of the programme. Since the research has to be finished by the end of next year, money was allocated only to projects supporting ongoing activities. The following activities were funded (in DKK) : (1) Supplementary studies of pesticide transport at the Havdrup test site (200,000.-), (2) Classification of fractures in clayey tills (150,000.-), (3) Quantification of matrix and fracture fluxes trough a fractured till (100,000.-), (4) CFC-age dating of groundwater from the field sites (100,000.-), (5) Supplementary tracer tests to identify pathways of pesticide transport in the unsaturated zone (50,000.-), (6) An additional pesticide field injection study at the Vejen test site (470,000.-) (7) Identification of bottlenecks for pesticide degradation in aquifers using pure cultures of pesticide degrading bacteria (380,000.-), (8) Tracer test at the Volby Bæk Wetlands (60,000.-), (9) Collection of data to be used for validation of the catchment model (260,000.-), (10) Mapping of redox conditions at the Eggeslevmagle test site (40,000.-), and (11) Purchase of radiolabelled mecoprop (100,000.-).

2

Until now the Groundwater Group research activities within pesticides and groundwater have resulted in 11 international articles in peer reviewed journals, 28 contributions to conferences, workshops, and symposia, and several national articles. At present 9 Ph.D. students are financed by the programme. The Groundwater Group wishes to transfer approx. DKK 2,500,000.- to 1999. The main reasons for this transfer are the same as described in the annual report 1997, namely at the initial phase of the project period it was difficult to find qualified Ph.D. students and a general delay occurred in the selection of field sites. It should be mentioned that the amount of money to be transferred to 1999 is much smaller than the money transferred from 1997 to 1998. The Project Board of the Groundwater Group expects that most of the money originally allocated to the programme will be used during 1999, except for what is needed to finish the late initiated Ph.D. studies.

3

STATUS FOR PROJECTS

Project 1: Preferential flow of pesticides towards the groundwater In 1998, the establishment of the three field sites was completed (Fig.1). At these sites experiments are carried out on transport and degradation of pesticides in the unsaturated zone. At Flakkebjerg, fracture transport in clayey till from the soil surface to the groundwater table has been studied under different topsoil treatment conditions. At Havdrup, transport studies have been carried out in deep-seated fractures in clayey till. At Fladerne Bæk, transport experiments in the shallow meltwater-sand aquifer are carried out. In addition to transport studies in the field and in lysimeters in subproject 1.1, laboratory experiments on degradation and sorption have been carried out in subprojects 1.2 and 1.3. Activities under subproject 1.4 have comprised parameter estimations and parameter selection for the Flakkebjerg site model and review of models for flow and transport in a partly saturated fractured medium. The project as a whole follows the time schedule laid down in the proposal although some activities have been delayed due to problems at the beginning of the project. The two field experiments at Havdrup have been accomplished while some clarifying studies have been planned for next year. At Flakkebjerg, the field studies with pesticides, bromide tracer and dye tracer are proceeding according to the plans and also the laboratory experiments with Flakkebjerg samples are well on the way. At Fladerne Bæk, the field studies will be commenced early next year, while the laboratory experiments on samples from Fladerne Bæk are proceeding as planned. The modelling has been commenced on the Havdrup data while the modelling for Flakkebjerg and Fladerne Bæk is still awaiting data.

Figure 1. Location of the three field sites included in Project 1.

4

Subproject 1.1: Transport Flakkebjerg: Activities and results Field investigations: Based on the field studies in the two excavations a macropore model for the unsaturated zone in the clayey till has been established and described: A top zone with burrows, root channels and seasonal desiccation fractures, a transition zone with desiccation fractures and horisontal and vertical tectonic fractures and a bottom zone with vertical tectonic fractures. The main part of the hydraulic activities is anticipated to be restricted to the macropores and fractures. The areas of the four field plots have been cultivated: Two plots with normal cultivation and two plots with reduced cultivation. After the harvest of winter wheat in September 1998 glyphosate was applied to the plot and in December isoproturon together with bromide tracer were applied. In the corners of the field plots dye tracer (Rhodamine) experiments were added in small restricted areas. The sampling of water from the plot zero tension samplers was started in September and the samples analysed for base values of pesticides. The collection of water samples for measuring of isoproturon and bromide has proceeded. At two plots 200 mm of water was added to the normal precipitation. Dye tracer tests with Rhodamine and Brilliant Blue are carried out and demonstrate the transport pathways from the ground surface down to approximately 2 m below ground surface. The data from the experiments are still being evaluated and analysed, but the first results show that the transport occurs in selected matrix areas and macropore ”channels” while other areas and macropores have not taken part in the transport. From field hydraulics tests in multisampler boreholes close to the plot area the bulk permeability is estimated and varies over 4 orders of magnitude. It decreases with depth and falls into three groups corresponding to the top, transition and bottom macropore zones. Continuous groundwater level monitoring, pumping test and bromide tracer test data revealed that the till is fractured throughout its entire sequence in accordance with the macropore model. Locally there is fracture connectivity from the shallow till to an underlying sand aquifer situated 10 m below ground surface. In winter and spring groundwater levels are at maximum and a water table is present within the shallow till and fracture flux occurs in the horisontal fractures. During summer the upper till dewaters and the fracture flux is then controlled by the lower permeability till zones. Flow is now from matrix into the fractures and vertical transport dominates. Laboratory experiments: A total of three large undisturbed columns (LUC) were collected from 3.45 to 5.65 m depth along the most prominent fracture observed in the excavation. The LUC sampling was coordinated with fracture mapping in the excavations. Preliminary bulk hydraulic conductivities of the columns are 6.10-9 m/s to 3.10-7 m/s. Shelby tube matrix samples have conductivities of approximately 2.10-9 m/s. Flow in the fractures is to a high extent correlated with the occurence of channels. These structures appear to close with depth. Dye tracer distributions are consistent with the channelling observed prior to the experiments. Lysimeter studies: Application of the lysimeters with 14C-labelled glyphosate took place in September 1997 after winter wheat was sown, and in October 1997 14C-labelled isoproturon was applied to the lysimeters. Until now the amount of 14C in the leachate has been determined using liquid scintillation counting. Development of analytical methods - LC/MS - for analyses of isoproturon and glyphosate and some primary metabolites in water are nearly finished. According to the liquid scintillation counting still no leaching has been measured for glyphosate, the detection limit being about 0.03 µg/l. For isoproturon approximately155 l of water have percolated the lysimeters since application, the average concentration measured as 14C-equ. Being 0.36 µg/l. Clay colloidal studies: Flocculation-dispersion phenomena of clay colloids play an important role in colloid mobility in soils and groundwater aquifers. The repulsive forces in the electrical double layer, responsible for clay dispersion, are affected mainly by the charge characteristics of the clay surfaces. It is commonly observed that natural clay colloids exhibit different charge characteristics and stability

5

behaviour compared to their pure reference clay counterparts. These observations have been attributed mainly to the presence of adsorbed humic substances and coatings of iron and aluminum oxides on the surfaces of clay colloids. However, the lack of information on the composition and physicochemical properties of natural mobile soil colloids hampers our ability to model and understand the role these colloids may play in transporting contaminants. The aim of the present study was (i) to determine the mineralogy and chemical composition of natural water-dispersible clay colloids (WDC) collected from a clayey till situated at Flakkebjerg, (ii) to examine the effects of pH and ionic strength on the charge characteristics and stability behaviour of the WDC, with special emphasis on the role of surface adsorbed natural organic matter and coatings of iron and aluminum oxides. In order to achieve this, water-dispersible colloids were fractionated from bulk samples, collected from several horizons throughout the soil profile, as well as from special samples, collected around macropores in the lower soil profile, containing translocated and illuviated clay. The extraction of the WDC fractions was accomplished by mixing soil samples with deionised H2O (without addition of dispersing agent) with a soil:water ratio of 1:8, shaking overnight and separating the WDC by particle size centrifugation, at 2 and 0,2 µm, yielding the fractions 0,2-2 and < 0,2 µm. Subsamples of the colloid suspension were treated with sodium peroxodisulphate and sodium citrate-bicarbonate-dithionite for selective removal of organic matter and iron-aluminum oxides, respectively. The mineralogical composition of WDC was determined by XRD analysis and used for semi-quantitative mineral estimation. The content of FeCBDand AlCBD in all WDC fractions has been determined by the procedures outlined above. Content of carbon, the C:N ratio and the surface area of the WDC fractions are to be determined. Measurements of electrophoretic mobility and particle size distribution, as a function of pH and ionic strength, of the natural and treated WDC suspensions, as well as on reference clay minerals, are being analysed by laser Doppler velocimetry-photon correlation spectroscopy using a Zetasizer 3000 (Malvern Instruments). The dispersion-flocculation phenomena of the WDC suspensions are studied turbidimetrically using a HACH turbidimeter. The experiments will be published in the beginning of next year. Havdrup, activities and results: 1997 and 98: Performance of the Havdrup field experiment no. 1. 1997: Sampling of large undisturbed columns (LUC) at Flakkebjerg and Havdrup. 1997 and 98: Laboratory experiments with LUC and small columns from Flakkebjerg and Havdrup. 1998: Initiation of field experiment no. 2 at Havdrup. 1998: 6 hydrogeology master students are educated at the University of Copenhagen/Geological Institute with their field work integrated in the experiments funded by SMP 96. Field experiment no. 1: The vertical bulk hydraulic conductivities of two hydraulically separated blocks (each 40 m2 and 2 m deep taken from 3 to 5 m depth) in the unoxidised clay were 7.6 10-9 to 1.1 10-8 m/s. Excavation of the plot after infiltration of a dye tracer showed that the majority of the stained fractures in the plot was not conductive or had very low conductivity. More than 99% of flow occured as preferential flow in widely spaced root channels or other channelling structures. These structures all followed the Fe-oxide stained fractures in the till. Breaktrough concentrations of the applied nonreactive tracer (bromide), and the mobile pesticides (metsulfuron-methyl and MCPP), and the immobile pesticides (prochloraze, and colloide transported cypermetrine) were monitored at the base of the till. Breakthrough of the pesticides was spatially heterogeneous and consistent with breakthrough of the bromide and the traces of the dye. Breakthrough concentrations of mobile pesticides and the nonreactive tracer were approximately the same and range from the detection limit to approximately 0.3 with no apparent retardation of the pesticides relative to bromide. Prochloraze occurred with the same spatial pattern as the mobile pesticides, however, in very low concentration (0.01-0.03ì g/l) appearing erratically over time. The influence of colloide transported cypermetrine has not yet been analysed. Field experiment no. 2: Natural infiltration and 3D solute transport were investigated in a sloping area of 40 x 30 m. The investigations are based on a network of 80 monitoring wells installed with screens

6

at depths from 1 to 10 m. The experimental set up was accomplished in autumn 98 and hydraulic measurements and solute infiltration experiments have been initiated. Laboratory experiments: A total of 4 LUC were collected from the field plot after the hydraulic and pesticide transport experiments. The columns were collected along stained fractures from appoximately 3.5 to 4.5 m depth in the unoxidised clays. Preliminary bulk hydraulic conductivities of the columns are < 10-11 m/s to 4 10-7 m/s, which are anticipated to reflect fractures without hydraulic conductivity and fractures with high hydraulic conductivity, respectively. Bromide and dye tracer experiments are planned for comparison with field experiments. Colloidal transport studies: In this project, the mobility of pesticides and artificial microspheres are investigated in a large-scale field injection test in lower, fractured clayey till. Field activities have now been finalised and the results are being interpreted. Preliminary conclusions: - Preferential pesticide transport studies are being carried out at 4 scales, providing detailed data of fracture/matrix interactions (laboratory studies) and natural contaminant 3D transport behaviour at the field scale. - At Havdrup preferential flow presents a very high percentage of total flow to at least 5-6 m depth in the till. This result is in agreement with measurements in tills at other places in Denmark. - Deep roots preferably penetrate vertically along existing stained fractures to form highly conductive channelling structures in the fracture network. - Preferential transport was exclusively observed along such stained structures in the till and no flow was observed in "invisible fractures" as commonly inferred. - Apart from the channel structures mentioned above clearly stained fractures appear to be closed in the unoxidised clays. This result is correspond to results from other Danish till localities. - Mobile pesticides (mesulfuron and MCPP) migrated rapidly (0.5 m day) at natural vertical gradients with only a small decline in concentrations by preferential flow in the channel structures of the tills. Fladerne Bæk, activities and results Field investigations: The purpose of the field site investigation is to study the role of “macropores” and other structures in sand deposits (Fladerne Bæk) for the degradation of pesticides. During 1998 installation of suction cells and TDR probes in four plots north of the previous installations at Fladerne Bæk has been carried out. While installing the equipment a geological description of the pits (100 m walls) was made. The meltwater sand showed a pronounced structure with bleached sand in vertical tubes of approx. 20-40 cm in diameter and reaching the groundwater table at 1.9-2.2 m depth. Adjacent to the bleached sand concretion of precipitated iron-organic matter, manganese was present aligning the bleached sand. The origin of these formations is unknown but it is suggested that it could be remnants of the former vegetation as the structures have an equidistance of around 3-5 m. An origin as a periglacial structure could also be considered. The hydraulic implications of both the topology of the fields (spring cereals or potatoes) and the sand structures will be investigated by tracer experiments on selected plots. Samples will be taken in highly conductive as well as in minor conductive sections in order to determine differences in microbial properties. Implementation of comments by the Evaluation Panel. The Fladerne Bæk site investigations have been evaluated. The relationship between the field- and laboratory scale experiments will be described. The field transport studies include pesticide and bromide leaching and dye tracer expriments. The laboratory experiments include microbiological degradation experiments, mineralisation and transport adsorption experiments. The potential use of research results on preferential flow and transport from the Flakkebjerg and Havdrup sites in the implementation of regional and national groundwater protection strategies in Denmark will be described in a separate report.

7

Activities for next period The field experiments with pesticides, bromide and dye tracer will be finished and the collected data modelled. The lysimeter studies on isoproturon and glyphosate will be continued and completed. Laboratory experiments on mobility of the natural colloids in undisturbed soil columns will be carried out. Subproject 1.2: Degradation Flakkebjerg and Fladerne Bæk, activities and results The studies have included microbiological characterisation of the macropore structures in the clayey till (Flakkebjerg) as well as in the meltwater sand (Fladerne Bæk), using traditional methods for enumeration of the bacterial populations (plate counts, most probable number method, acridine orange direct count) and Biolog EcoPlates for estimating how the populations differ with respect to carbon utility. Finally, incubation experiments are conducted to study whether the macropore structures are differing in their ability to degrade pesticides using 14C-labelled pesticides. The preliminary results indicate differences between macropores and sediment matrix in the clayey till as observed earlier in other clayey deposits. The sand deposit seems to be more heterogeneous than previously assumed. The results of the Biolog EcoPlates measurements showed significant differences between the different structures in sandy soil subsoil profile as indicated in the Fig. 2. Mineralisation of four different pesticides showed quite a variation in rate constants mostly depending on available organic matter for cometabolic degradation. Additionally, adsorption experiments showed that strong adsorption will deminish or even stop degradation of some pesticides e.g. glyphosate. Inocculation of the sediment batches with mecoprop degraders showed that the sediment composition or state may inhibit the degradation. A new developed method for transport adsorption studies on undisturbed soil columns has significantly improved the understanding of transport corridors in sandy as well as clayey sediments. The same procedure will be modified to make it possible to study transport and adhesion of bacteria through soil columns.

8 6

PC 2

4 2 0 -2 -4 -6 - 1 0 -8

-6

-4

-2

0

2

4

6

8

10

PC 1

Figure 2: Principal component analysis of Biolog Ecoplate data from Fladerne Bæk. The symbols represent different structures visualised as different colours in the soil profile. The degradation studies in clayey till have been continued. The development of the analytical method for analysis of isoproturon and its primary metabolite in soil is now finalised. With the developed

8

method, analysis of water extractable as well as solvent extractable isoproturon and its primary metabolite can be performed, which will be useful for modelling the fate of the compounds in varying compartments. With use of the vacuum system, described in last years report, the degradation studies which focus on the disappearance of the parent compound and the formation of metabolites have been started. Initial trials have shown that the degradation of isoproturon is relatively slow, and no results on degradation kinetics and degradation rates have yet been presented. Activities for the next period The microbiological degradation experiments will be continued. Subproject 1.3: Sorption Flakkebjerg, activities and results Geochemistry, clay mineralogy and sorption of pesticides: At the study site at Flakkebjerg the soil formation was described in excavation 2, and matrix and micromorphological features from the different horisons were studied in the laboratory. Also sediment samples from the excavation were collected from a depth between 0.5 m and 4 m to below the redox interface between the oxidised and the reduced till. A sampling intensity of 1 sample per 10 cm was used. Subsequently the samples have been described and analysed in the laboratory for redox and sorption properties. The soil at the Flakkebjerg site was classified as a typic Agrudalf. Samples from the A, B, and C horisons have been analysed for physical, chemical and mineralogical parameters. Macropores in the horisons were mapped and described with respect to occurrence of clay skins, accumulations of organic matter, iron- and manganese oxides, and mineralogical composition. The development of different geochemical environments (pH and Eh) after the last glaciation has been studied in the clayey deposits. Measurements of the content of calcium carbonate and pH in suspensions of sediment were used to characterise the leaching of inherited calcium carbonate (CaCO3) after the glaciation and to define the different pH zones. Redox zones were described by the colour of the sediments using the Munsell Soil Colour Charts, and by analyses of redox-sensitive elements (e.g., organic matter (TOC), nitrate, iron and manganese oxides) present in the sediments. Preliminary results indicate the presence of 3 major redox environments; an upper oxidised and CaCO3-free zone down to about 2 m, and oxidised and CaCO3-rich zone down to about 3 m, and a deeper reduced and Ca-rich zone. The content of 17 pesticides within the different zones was analysed by GC/MS after the pesticides were extracted from the sediments using selective solvents. In all samples the contents of pesticides were below the detection limit of 0,05 mg/kg and the methods here used indicated no differentiation of the pesticides according to pH-Eh zones. More studies on sorption of pesticides in samples of different geochemical environments have been planned. So far sorption of pesticides to clay minerals separated from typical Danish clayey sediments of the oxidised and reduced zones has been studied by screening experiments using atrazine, MCCP, isoproturon, bentazone, and 2,4 D. Also, isotherms and sorption kinetics of two clay samples of oxidised zones and atrazine, MCCP, and isoproturon have been studied. The results indicate a greater sorption of atrazine than of other chosen pesticides. Also desorption experiments with the same two clay samples indicate differences for the here chosen pesticides, but more studies are still needed. Until now the experiments have included a number of well-known clay samples (< 2 µm) of different redox status. For some of the pesticides, e.g., atrazine, this property seems very important for the sorption. The adsorption of glyphosate was determined by following the principles in the OECD testguideline 106, revised October 98. 14C-labelled glyphosate was used for the experiments. Radiochemical purity was > 99% and the specific activity 23.56 mCi mmol-1. The adsorption was determined in 50 cm3 plastic centrifuge tubes after 24 and 48 hours, and there was only a small difference observed. Desorption was determined by removing the supernatant after centrifugation and applying fresh CaCl2solution and shaking for another 48 hours. The table shows the adsorption percent and the calculated Kd- and Koc-values. The table shows that glyphosate is strongly adsorbed to this soil type with a Kd value of 326. The adsorption was supposed to be high in this soil due to a relatively high

9

content of clay (14,3%). In previous experiments the adsorption of MCPP, isoproturon and atrazine has been determined in the plough layer in this soil. The Kd -values were 0.5 for MCPP, 1.7 for isoproturon and 1.6 for atrazine. The adsorption of glyphosate is much higher than for the other compounds. It is therefore expected that, if glyphosate is leached out into this soil, it will be through macropores and carried by transport of soil colloids. Table 1. Adsorption, desorption and mass-balance for 14C from glyphosate in soil sampled in the plough layer from Flakkebjerg. Soil Adsorption DesorpKd Koc MassClay Total C 3 3 % tion, 48 h cm /g Cm /g Balance % % Flakkebjerg 97.0 1.8 326 29655 97.7 14.3 1.1 Activities for next period The adsorption of glyphosate has until now only been determined in the plough layer. The experiments in 1999 will concentrate on sorption to clayey till sampled in the subsurface. More sorption experiments with reduced clays are planned. Also sorption experiments with sediment samples of the three major redox environments and redoximorphic features collected at Flakkebjerg will be included in the ongoing sorption experiments. The importance of the mineralogical composition of the clay minerals for the sorption of pesticides will be considered in coming sorption experiments using subfractions of clay and involving a wide range of the commonly occurring clay constituents. Subproject 1.4: Modelling Existing numerical models for flow and transport in partly saturated fractured medium have been reviewed. Evaluation of numerical modelling approaches to assess pesticide flow indicates that the influence of preferential flow is so significant that assessment of pesticide risks using conventional total porosity piston flow models (e.g. PELMO, PESTLA) is underestimating the observed pesticide transport rates and concentrations with orders of magnitude. Significantly improved results are achived using preferential transport models (e.g. FRACTRAN and 3DFRACvs). Activities for the next period The modelling of transport of pesticides and tracers based on the data from the three field sites will be started and intensified during the period. One Ph.D student will have this modelling work as a central task.

Project 2: Fate of pesticides in aquifers The overall scope of this project is to investigate the transport and the fate of pesticides in aquifers and to evaluate, if there is a potential for ‘selfpurification’ of pesticide contaminated aquifers. The project is in good progress. After a minor delay in the beginning of the project period used for development and implementation of methods for analysing the pesticides etc., selection of field sites and sampling locations, most of the planned samples are now collected, the laboratory experiments are running, and the field experiments have been started. Figure 3 shows the different sampling locations and field sites used. Since many of the experiments have to run for long periods, only relatively limited results have yet been obtained and interpreted, and the last year of the project period is therefore dedicated to finalise the experiments and complete all the analyses.

10

Drastrup

Fladerne Bæk Mølgårde (Gjern Å)

Asserbo Nykøbing Ballerup

Grindsted

Vejen

Bromme Frankerup

Drengsted

Tyske Mose Køge Ås Sigerslev Fakse

Figure 3. Overview over the different sampling locations and field sites included in Project 2. The content and the basic ideas of the project were presented to and discussed with the evaluation panel during a meeting at Sonnerupgård Gods the 10th – 11th September where each subproject presented the experiments carried out and the results obtained so far. Generally the comments from the evaluation panel were very positive, and the project was characterised as well outlined and well executed. The following discussion resulted in fruitful suggestions to new experiments, which now are included in the future activities. To ensure a good co-ordination between and within the subprojects several smaller meetings have been arranged. Furthermore, a progress seminar for Project 2 was held at DTU on March 6th 1998 with a full programme for the whole day. Apart from all the participants in project 2, the co-ordinators from the other Projects and the Project Board were invited. Such a progress seminar is also planned for January 1999. Subproject 2.1: Sorption/desorption of pesticides in aquifers. The aim of the project is to determine the sorption of selected pesticides onto aquifer sediments with low organic carbon content, and to investigate the controlling factors for the sorption (i.e. TOC, mineralogy, available surface area, and pH). The sorption has been measured onto 12 aquifer sediments, all with low TOC. The investigated pesticides were atrazine, bentazone, 2,4-D, glyphosate, isoproturon, mecoprop, metamitrone, and metsulfuron-methyl. The sediments were collected in Jutland as well as in Zealand. The sorption was investigated at a concentration of about 250 µg/l for each combination of pesticide and sediment. A significant sorption of glyphosate, metamitrone, atrazine, isoproturon, mecoprop, and 2,4-D was observed, whereas the sorption of bentazone and metsulfuron-methyl was insignificant. The variation in Kd between the sediments was significant. Based on these results sorption processes in Danish aquifers are expected to reduce the mobility of glyphosate, metamitrone, atrazine, and isoproturon and, to a smaller extent, the mobility of mecoprop and 2,4-D. The observed sorption was higher compared to Kdvalues calculated from the actual TOC content and literature data on top soil Koc-values. Consequently, TOC is not the only factor controlling the sorption in low TOC sediments; also the mineralogy, available surface area, and pH affect the sorption. Currently the sorption data are being further interpreted with respect to the different sediment and groundwater characteristics, using the knowledge from the mineral study, and the results from the image analysis. These investigations include statistical analysis, e.g. partial least square calculations, of the sediment and groundwater

11

characteristics and the Kd-values for glyphosate, metamitrone, atrazine, isoproturon, mecoprop, and 2,4-D. Furthermore, the sorption isoterms and kinetics for selected pesticides are currently being investigated for the 12 sediments. The experiments will be finished by the end of the year. In order to investigate the kinetics of the sorption/desorption for the different sediments under more realistic conditions, column experiments will be carried out in the future. The purpose is to investigate the significance of sorption in a dynamic experiment with realistic flow conditions. To improve the understanding of the difference in sorption behaviour between the sediments, BSE image analysis will be carried out in order to quantify the mineralogy of surfaces available for sorption in the sediment. In an associated Ph.D.-project, sorption of pesticides on pure minerals is studied in order to quantify how much the different constituents of aquifer solids sorb the pesticides, and to compare sorption characteristics of pesticides. Five pesticides are investigated: Atrazine, bentazone, 2,4-D, isoproturon, and mecoprop. Experiments with calcite, quartz, aluminium oxide, kaolinite, and three iron-oxides (2line ferrihydrite, goethite, and lepidocrocite) have been carried out. The results demonstrate the importance of the type of mineral phase and the surface charge of these aquifer minerals. The future work will focus on the effects of pH and ionic strength.

Subproject 2.2. Degradation of pesticides under different redox conditions in Danish aquifers. The aim of the project is to investigate the potential of degradation of pesticides under various redox conditions present in Danish aquifers. The experiments include chemical analysis of the groundwater and sediment, a microbial characterisation of the sediment, cell number and rate reduction of electron acceptors, and degradation of added pesticides, both as initial degradation (HPLC) and mineralisation of radioactively labelled pesticide. One part of the subproject focuses on the investigation of the spatial variation of the potential for pesticide degradation. The sampling and investigation of a number of locations initiated last year have in 1998 been extended with four more locations, where sediment cores with adjacent water samples were collected. These four new locations are: Drastrup kildeplads, northern Jutland (2 depths), Frankerup, western Zealand (3 depths), Bromme, central Zealand (4 depths) and Grindsted, central Jutland (4 depths). Two sampling depths from Grindsted and the four sampling depths from Bromme represented aerobic aquifer conditions, the rest were anaerobic. Thus a total of eight locations each with 2-4 sampling depths have been included, and sediment cores and water samples have been collected from a total of 25-26 different sampling depths. For each sampling depth experiments have been set up to investigate the degradation of the pesticides MCPP (Mecoprop), isoproturon and metsulfuron-methyl as sole pesticides; a mixture of MCPP (Mecoprop), isoproturon, MCPA, DNOC, atrazine, 2,4,5-T and dichlobenil; and a mixture of metsulfuron-methyl, 2,6dichlorbenzamid (dichlobenil-degradation product, BAM) bentazone, 2,4-D and dichlorprop. The characterisation of the samples with respect to the redox conditions is in progress, and the results show that the collected samples represent various redox conditions, from aerobic to methanogenic conditions, with the major part within Fe(III), Mn(IV) and sulphate reducing conditions. The general microbial activity was characterised in terms of turnover of benzoic acid. These experiments showed evident microbial activity, even in the anaerobic samples, but with substantial variation between the samples including a few samples which showed very low turnover of the benzoic acids. These data will be related to the results from the investigations of the ongoing microbial redox processes. Currently all the samples are being characterised with respect to hydro- and geochemistry and ongoing microbial redox processes (denitrification, Mn(IV), Fe(III), sulphate reduction and methane production). Since the degradation experiments for each sample are planned to continue for at least one full year, no experiments have yet been analysed completely. Therefore no complete pattern regarding the degradation of the pesticides is present. However, the preliminary results indicate only slow or no degradation for most pesticides under anaerobic conditions, however, e.g DNOC seems to be transformed most rapidly under some anaerobic conditions. The major activities in the next period for this part of the project will be to finalise the pesticide analysis.

12

The aerobic degradation of several pesticides was investigated in an associated master thesis project. The investigated sediments were collected from two different locations in the aquifer at the Vejen site and the pesticides were investigated at a concentration of 1 or 25 µg/l. Mecoprop (MCPP), dichlorprop and DNOC were degraded, whereas bentazone, isoproturon and BAM were not degraded during the investigation period of 230 days. The influence on the degradation of several environmental parameters was investigated in a factorial experiment. Oxygen (saturated conditions compared with a content of approximately 2 mg O2/l) stimulated the degradation of MCPP and dichlorprop, but the combination of a high content of oxygen and elevated pH-level (pH 7.5-8.5 compared to the background level of pH 4.3-6.0) reduced the degradation rate. Addition of a primary substrate in terms of benzoic acid seemed to stimulate the degradation of dichlorprop and MCPP. Addition of inorganic nutrient salts reduced the degradation of dichlorprop and MCPP, but stimulated the degradation of DNOC. To investigate some locations in more detail, experiments were carried out as slurry experiments with sediment and groundwater from either Fladerne Bæk (aerobic and denitrifying) or Asserbo Plantage (sulphate reducing and methanogenic) kept either aerobic or anaerobic and added the electron acceptor of relevance. The selected pesticides were: Mecoprop, dichlorprop, 2,4,5-T, metsulfuron-methyl, isoproturon, and atrazine at concentrations of 250 µg/l (HPLC) and 25 µg/l (mineralisation). The degree of mineralisation measured as 14CO2 during 150 days was very low under all conditions, below 1% for all pesticides except mecoprop which reached 2,5-5%. Samples from the HPLC experiment have yet to be analysed. The low rates of mineralisation were discussed at the evaluation in September 1998, and it was suggested that we apply pure cultures of pesticide-degrading bacteria in the investigation of bottlenecks in the degradation. In accordance with this input a new series of experiments are planned. The aim is to uncover the limiting factor in the degradation of atrazine and mecoprop. Sediment from Fladerne Bæk is amended with pesticide and either a pesticide-degrading culture, carbon source, electron acceptor, or a combination of these. The pesticide-mineralising strains used will be (1) an atrazine-mineralising Pseudomonas sp. ADP (Mandelbaum et al. 1995), which degrades the atrazine both under aerobic and denitrifying conditions, (2) transconjugants of pseudomonads isolated from Fladerne Bæk which contain a plasmid that encodes for the initial steps in the metabolism of atrazine, and (3) a mecoprop-degrading bacterium isolated from the Fladerne Bæk aquifer. This bacterium degrades both chiral forms of mecoprop and it has been tentatively identified as a Burcholderia cepacia. At the moment a more thorough characterisation based on 16-S-RNA is initiated. Furthermore, attempts to develop sulphate reducing atrazine-degrading transconjugants are in progress. A new method to investigate in situ degradation of pesticides in aquifers by a diffusion cell is under development and being tested in the Asserbo aquifer. The groundwater chemistry in the Asserbo aquifer (the new field site in Tisvilde Hegn) has been characterised, and the results have shown distinct zones of iron reduction, sulphate reduction, and methanogenesis. Diffusive emitters for release of 2,4,5Trichlorophenoxy acetic acid (2,4,5-T) and 2,4-Dichlorophenoxy acetic acid (2,4-D) have been established in the methanogenic zone, and measurements inside the releasing wells have demonstrated that LDPE (Low Density PolyEthylene) emitters can release relatively steady concentrations of the 2 phenoxy acids as well as tritiated water (tracer). However, until now it has not been possible to locate the plume in the aquifer. Probably this is due to insufficient development of the wells combined with seasonal fluctuations in the groundwater flow direction. Laboratory degradation experiments with sediment and groundwater from the methanogenic zone have shown no degradation of 2,4,5-T within 12 months. Therefore, we are currently considering to use another pesticide which is more likely to be transformed under methanogenic conditions. Abiotic degradation experiments have shown that removal of 2,4,5-T can be coupled to the presence of iron(III)hydroxide (ferrihydrite) together with iron(II). The reaction is found to be pH dependent. In order to investigate this process under field conditions, 10 more wells screened in the iron-reducing zone as well as the upper part of

13

the sulphidogenic zone have been installed. In April-July 1998, 2 different laboratory experiments investigating diffusion of tetrachloroethylene (PCE) through nylon tubing were conducted in order to improve the knowledge on design and performance of diffusive emitters. This work resulted in development of a field releasing system at the C.F.B. Borden field site, which turned out to perform very well. Further laboratory experiments have been conducted in order to optimise the emitter system used in the Asserbo aquifer, using a stronger and vastly more efficient type of LDPE. The future plans are to develop new LDPE emitters for release of 2,4,5-T and other pesticides, probably the nitroaromatic DNOC and to install them in the iron reducing/sulphidogenic and the methanogenic zone. The wells screened in the methanogenic zone will be developed again before the new emitters are installed. More multilevel piezometers will be installed to locate the plumes. Further laboratory work will be directed towards uncovering the mechanism of abiotic 2,4,5-T removal coupled to the presence of iron(III)hydroxide (ferrihydrite) together with iron(II). Similar degradation experiments with DNOC will be conducted. With regard to the laboratory experiments we expect to demonstrate the reductive dechlorination of 2,4,5-T in the presence of iron(III)-iron(II) surface complexes as well as the reduction of nitro groups on the DNOC molecule under similar conditions. The aim is also to be able to demonstrate these processes occurring under field conditions. Furthermore we hope to see reduction of DNOC-nitro groups under methanogenic conditions. Using the LDPE emitter system, we expect to be able to generate steady and uniform pesticide plumes in the different redox environments. Sub-project 2.3. Kinetics of pesticide degradation at different redox and nutrient conditions Studies on kinetics of degradation under aerobic conditions of 2,4-D, MCPP and 2,4,5-T have been performed with aquifer sediment sampled at Fladerne Bæk, Jutland or at Grindsted, Jutland. The degradation has been followed by means of mineralisation measuring 14CO2 evolved after addition of different concentrations (0.1 - 10.000 µg/l) of the three herbicides. The results of this study showed (i) mineralisation of 2,4-D, although with different degradation kinetics depending on the substrate concentration added, (ii) mineralisation of MCPP above a certain threshold value, and (iii) no mineralisation of 2,4,5-T. It appears that within a certain range of 2,4-D-concentrations or concentration ratios the kinetics shift from growth-linked kinetics with an accelerating rate of degradation to a slow first order degradation with negligible growth. For typical groundwater low in degradable organic material threshold concentrations for non-growth kinetics have been found to be more than an order of magnitude below those found in eutrophic surface water. Examples of observed threshold concentrations are 5-10 µg/l for degradation of the pesticide 2,4-D in river water and in some cases below 0.2 µg/l in aerobic groundwater. To investigate the presence of threshold concentrations in more details new experiments with a very low pesticide concentration (0.02 - 1.00 µg/l) are currently in progress. A highly sensitive method has been developed for such low concentrations (as low as actually occurring pesticide concentrations in groundwater). The method is based upon trapping CO2 from a large test volume into an external absorber. The experiments are being conducted with groundwater with and without sediment fines to elucidate the degradation kinetics and the role of bacterial attachment at low pesticide concentrations. The effects of inorganic nutrients and organic carbon sources on the degradation of 2,4-D were studied also with sediment from Fladerne Bæk. At low 2,4-D concentrations (1µg/l) no stimulation of the mineralisation was seen by addition of either casamino acid, inorganic minerals or both. In contrast, all nutrient additions stimulated mineralisation of 2,4-D at high concentrations (10.000 µg/l). Apparently, at this high 2,4-D concentration, growth of specific 2,4-D degrading bacteria is limited by growth factors present in the nutrient solutions. Similar studies with MCPP and 2,4,5-T are in progress. Together with activities within subproject 2.2 attempts to establish enrichment cultures degrading isoproturon have been carried out. No enhanced mineralisation of isoproturon was seen in enrichment cultures with high pesticide concentrations (25 mg/l; figure 4). In contrast, a rapid mineralisation of the demethylated metabolite monodesmethyl-isoproturon (DMI) was observed and this capability could easily be transferred to fresh DMI containing media. It is hypothesised that the initial

14

demethylation of isoproturon limits the complete mineralisation of the herbicide in soil and groundwater. Experiments investigating whether this initial step proceeds by biotic or abiotic processes are in progress. Recently, Dr. Alan Walker, Horticulture Research International, United Kingdom provided us with a British soil (Deep Slade) with a high isoproturon degrading ability. From this soil enrichment cultures degrading both isoproturon and monodesmethyl-isoproturon have been established.

60%

CH3 H

50%

C

H 14 C

N

CH3

H C O

N C H3

14

C Mineralized (%)

Monodesmethyl-isoproturon 40%

30%

20%

C H3 H

C

H 14C

N

C H3

C H3 C O

N C H3

Isoproturon

10%

Autoclaved soil 0% 0

5

10

15

20

25

30

35

Time (days)

Figure 4. Mineralisation of isoproturon and monodesmethyl-isoproturon in soil from Græse (top soil).

Mean values of triplicate samples ± standard deviation.

The research next year will focus on mechanisms of adaptation to pesticide degradation. At the evaluation in September 1998 Dr. Stenström argued that adaptation, defined as an increased degradation following pre-exposure to a compound, often is a result of growth of specific degrader organisms. The importance of such growth-related ´adaptation´ in sediments exposed to low pesticide concentrations will be examined. Transfer of catabolic plasmids may also lead to adaptation. At the moment effects of such transfer on the kinetics and lag period for degradation of 2,4-D in groundwater sediments are under investigation. In this research the effect on gene transfer at groundwater conditions by the presence of a selective pressure as well as the sizes of bacterial donor and recipient populations will be demonstrated. Subproject 2.4. Field investigations of transport and fate of pesticides in a sandy aquifer The aim of the project is: (1) to identify the processes governing the fate of pesticides in a natural flow system; (2) to determine actual field-scale degradation rate constants; (3) to develop and test a reactive solute transport model for pesticides; and (4) to compare the fate of pesticides in different experimental systems (lab. batch and column, field injection). This is accomplished by conducting a continuous field injection experiment at ambient flow gradients in an aerobic sandy aquifer which has a high potential for pesticide degradation. Supplementary laboratory experiments studying sorption and degradation will be carried out as a part of this subproject or subprojects 2.1, 2.2 and 2.3. The experimental focus area is a shallow, aerobic aquifer (Quarternary sand deposits) situated near Vejen in the western part of Denmark. The injection experiment was designed as a continuous injection of pesticides in two parallel experiments. The project activities in winter and spring 1998 were dedicated the planning and preparation of the field injection experiment. A summary report

15

describing the field site, the background, and aim of the experiment and the experimental set up was submitted to Ribe Amt and Vejen Kommune. This report and a meeting were used as background for obtaining permission from the local authorities to perform the experiment. At the same time an agreement with the landowner was obtained. The field injection experiment was designed as two parallel experiments in order to cover a broader range of pesticides and study interactions between different pesticides. Two injection areas have been instrumented (injection wells, multilevel samplers, dataloggers) in the aerobic part of the aquifer. The injection was performed by a fully automatic system into six injection wells perpendicular to the flow direction. In June 1998 the injection experiment was initiated, however, due to mechanical failure of the pumping system at both sites, the injection was ceased after 7-9 days. At site A, MCPP and 2,4-D were injected, while bentazone, BAM (degradation product of dichlobenil), DNOC, bentazone, MCPP, dichlorprop and isoproturon were injected at site B (see figure 5). Bromide was injected as a tracer at both sites. The aimed concentrations of the pesticides were 50 µg/l. The injected pulse was sampled up to 20 m from the injection wells for a period of two months. The samples were analysed and the results were evaluated. The pore flow velocity was in the order of 0.2-0.3 m/day. The spatial distribution indicated geological heterogeneities in the aquifer. This in combination with a slight change in flow direction with time made evaluation of breakthrough curves beyond 10 m from the injection wells difficult. The sorption in the sandy aquifer was insignificant for most compounds injected, however, a slight retardation of isoproturon and pronounced retardation of DNOC were observed from the breakthrough curves. DNOC is probably also subject to degradation. In addition, retardation is markedly different for breakthrough curves sampled at different depths. All other pesticides did not show any significant degradation within the experimental period. Analytical work focusing on degradation products is still ongoing. 3DADE, a computer programme for threedimensional solute transport during steady unidirectional water flow in porous media with uniform transport and flow properties, is currently applied in inverse mode for estimating transport parameters from the field data.

Figure 5. Field injection experiment: Breakthrough curves for bromide (tracer), DNOC and isoproturone at a distance of 1 m from the injection wells. Based on the experience from the first experiment, a modified sampling network was designed in order to start a second injection in site B. This experiment will continue for a period of 4-6 months to overcome possible lag phases. Furthermore, the effect of pre-adaptation may be evaluated as the same pesticides are injected at the site within a period of a few months. The concentration of the tracer is reduced with a factor of two in the injection solution to reduce density problems. Currently, the supplementary multilevel samplers are being installed and an improved pumping system developed. The monitoring network will after modification contain 500 sampling points, which will be sampled on

16

a weekly basis in the beginning. Later on more emphasis will be put on obtaining mass balances for the pesticides in fences of multilevel samplers at different distances. Parallel to the field activities, column experiments have been carried out in a master thesis project closely related to the project. Six columns were packed with aquifer material sampled from two different locations (three columns of each) in the aerobic part of the Vejen site. Groundwater from the site spiked with a mixture of pesticides (similar to the pesticides in the field injection experiment) was applied to each of the columns. The pore flow velocities were 20 m/day corresponding to a retention time of 20 days in the column. The results in terms of breakthrough curves were interpreted by use of onedimensional reactive solute transport models. Sorption was insignificant for the MCPP, 2,4-D and dichlorprop, BAM and bentazone, while isoproturon and DNOC significantly sorbed to the aquifer material (Kd-values in the range of 0.06-0.32 l/kg). MCPP, 2,4-D and dichlorprop were all degraded after a lag phase (16-33 days) in the column experiments. The degradation rates (zero order) were in the range of 1.0-2.6 µg/l/day. DNOC was also degraded, however, the lag phase was substantial (up to 80 days). BAM, bentazone and isoproturon did not show any degradation within the experimental period of 140 days. An analytical method based on the solid phase micro extraction principle (SPME) for determination of phenoxy acids and their degradation products has been developed. The method and the analyses are carried out on a GC-MS. The advantage of the method is that only a small volume of sample is needed and the use of extraction chemicals is avoided. The method seems promising, however, more practical experience is needed to evaluate the potential compared to other methods. The model development is based on three-dimensional flow and finite element codes using Galerkin's technique. The flow code is developed at University of Waterloo, Canada, whereas the transport code is developed as part of this study. At present, the transport code is capable of calculating transport subject to advection-dispersion, linear equilibrium sorption, and first-order decay assuming spatially homogeneous parameters. Future enhancement of the transport code will include refined description of sorption and degradation processes and allow for spatial variability. The sorption and degradation processes will be identified by analysing the field injection experiment and laboratory experiments. New activities will be coupled to the transport code by operator splitting technique. The purpose of developing a new numerical transport code is to include reliable descriptions both sorption and degradation processes to enhance the knowledge of the coupled effect of these processes. A new field injection experiment will be started November 1998 and continue for a period of 4-6 months depending on the results. The monitoring network will be sampled, samples analysed, and preliminary interpretation of the results will take place during the same period. Hydrogeological observations and modelling of the results in terms of breakthrough curves and spatial distribution of the tracer and the pesticides will be used for planning of the sampling strategy. The site will be further characterised with respect to geology and hydraulic properties after sampling of the injected cloud has stopped. The need for characterisation of other properties (microbial, geochemical or sorption characteristica) and for supplementary laboratory experiments in order to support the interpretation will be considered. Final interpretation of the data including detailed modelling and reporting of the results will take place autumn and winter 1999. Subproject 2.5. Transport, degradation and sorption of pesticides in wetland areas The aim of this subproject is to investigate the potential of degradation of various pesticides under the redox conditions present in the riparian area of Mølgårde. The experiments include chemical analysis of the groundwater, a microbial characterisation of the sediment, cell number and rate of reduction of electron acceptors, and degradation of added pesticides, both as initial degradation (HPLC) and mineralisation of radioactively labelled pesticide. The experiments are carried out as slurry experiments with sediment and groundwater sampled from aerobic, denitrifying, sulphate reducing, and methanogenic spots in the area. The slurries are kept either aerobic or anaerobic according to the original conditions and added the electron acceptor of relevance. The sediments have been

17

characterised chemically and microbiologically incl. determinations of number and activity of aerobic, denitrifying, sulphate reducing and methanogenic bacteria. The pesticides selected for the experiments are: Mecoprop, metsulfuron-methyl, isoproturon, and atrazine at concentrations of 25 µg/l. Cultures are set up both for HPLC and 14CO2 measurements and they have incubated now for 365 days. The HPLC experiment is not yet terminated and analysed. The experiments with radioactively labelled pesticide show mineralisation under aerobic conditions for mecoprop (33%), metsulfuron-methyl (7-20%), and isoproturon (12-15%). Under anoxic conditions only mecoprop is mineralised, 4-6% under denitrifying conditions and 11-12% under sulphate reducing or methanogenic conditions. The experiments will soon be terminated. In 1997-1998 the spatial variation in hydraulic and transport characteristics of the discharge zone in the wetland study area were initially investigated. Hydrologic measurements during 1997-98 are compiled in hydrographs. Hydraulic tests (slug test) conducted in 1997 were analysed and results from a previous natural-gradient conservative tracer test were re-evaluated to obtain both bulk and spatially distributed hydraulic properties. The flow pattern was time variant and 3D in the foot of the hill slope and constant along the transects in the wetland. Objectives and investigation strategy or concept have been presented to the evaluation panel. The panel was generally positive concerning the project, however, the only concern was that it might be of significant difficulty to separate the controlling hydraulic and transport mechanisms (dispersion, diffusion, sorption, and degradation) in the breakthrough curves of the planned combined natural gradient pesticide infiltration experiment. In light of these comments an additional conservative tracer test is included to obtain high spatial resolution of flow paths and residence time in the hydraulic active horizons of the discharge area and to obtain a satisfying determination of the water balance and solute mass recovery. The influence of pesticides on biogeochemical processes in the wetland - especially denitrification - is studied by use of microcosms (continuous flow soil columns). Four continuous flow soil columns have been set up in the laboratory under in situ conditions. 14C-labelled pesticides, i.e. either isoproturon, bentazone or MCPP were added to three of the columns, while the fourth column is used as a control. So far all columns showed significant denitrification rates before addition of the pesticides. The experiment is still running. Adsorption studies with wetland soil from the Mølgårde site have been initiated. The soil samples are incubated under aerobic and anaerobic conditions, respectively. Further, from the above-mentioned continuous flow soil columns water samples are taken through the cylinder wall for every 3 cm through septa in the core wall and analysed for 14C activity. At the end of the experiment the columns will be cut into sections and analysed for 14C activity. During the next period the main activities will be an additional instrumentation of the monitoring system (seepage meters) and execution of the additional conservative tracer test. If the results of the additional conservative tracer test show satisfying mass recovery, a modelling of the transport, sorption, and degradation processes of the pesticides in the discharge zone will be carried out. A natural gradient pesticide infiltration test is planned, and the steady state saturated groundwater and surface water flow will be modelled in the wetland area.

Project 3: Large scale modelling of pesticide transport The project is divided into three subprojects: 3.1 Development of methods for analysing up-scaling, 3.2 Catchment modelling, 3.3 Regional modelling. A significant part of the project activities in 1998 has been devoted to (1) identification of the appropriate model structure and complexity for both the catchment model and the regional model and (2) identification of suitable test areas for the two models. Identification of the appropriate model structures is a balance between complexity i.e.

18

incorporation of the most relevant physical, geochemical, and microbiological processes that influence the fate of pesticides in the subsurface at the scale of interest; data availability; and expectation and credibility of the model predictions. Different model formulations are proposed for the different scales as a compromise between the above factors. For all scales under consideration the models developed as part of the project are to be considered as research tools that will be tested and evaluated according to specified performance criteria. Given the complexity of the behaviour of pesticides in the subsurface and that many of the involved processes are not known in details and therefore cannot be formulated in mechanistic terms, the project does not claim that the models being developed will be of general validity. Consequently the models cannot directly be transferred to regulating agencies and used as planning or regulating tools. Another major activity in 1998 has been identification of field sites for testing and validation of the models for both catchment and regional scales. It is not possible to identify sites that fulfil all requirements to data due to practical and budget constraints and consequently a number of compromises need to be taken. A cause for delay in 1998 has been that the Flakkebjerg site first selected as one of the test sites for the catchment model for several practical reasons turned out to be unsuitable. In view of this finding a new survey of candidate sites was carried out before the Eggeslevmagle site was ultimately selected. Subsequently a new field investigation and monitoring programme was defined. Two test sites for the catchment scale and two for the regional scale are now identified and the field investigation and data collection programmes are in progress in all of them. Subproject 3.1: Up-scaling Objectives The research in this project is targeted at supplying general methods to analyse the upscaling problem and to facilitate hydrologic modelling in the face of it. The up-scaling problem is particularly difficult because many hydraulic properties are not constants, but functions of state variables. On the other hand, because hydrologic models must be manageably simple, the hydraulic functions are commonly represented in a condensed manner via a small number of parameters. Also fundamental mathematical theories of up-scaling operate on the parameters rather than the full functions. For practical applications, up-scaling will have to involve the estimation of effective parameters, and hence this project emphasises the estimation aspect of the suggested models. Based on the above, the research in this project has two goals: -To examine whether a simple parametric representation of hydraulic functions is adequate (otherwise, existing up-scaling theories do not pass the most fundamental requirement) -To develop new models that can represent heterogeneity in another form, with distributions of parameters (but still allowing a concise representation of reality) and determine the uncertainty that this simplification introduces. Different parts of the pesticide transport cycle seem most relevant with respect to achieving each goal: -Unsaturated zone flow and thus implicitly transport -Reactive transport in the saturated zone. Activities and results (1) Laboratory experiments have been conducted to examine stability and reproducibility of the generally accepted procedure to obtain unsaturated hydraulic parameters. The results showed that effects not captured in traditional models of unsaturated flow cause effectively random behaviour, such that no unique hydraulic parameters are valid under all flow conditions.

19

A maximum likelihood estimation method for estimating the parameters of the parametric functions for unsaturated hydraulic functions has been developed. The statistical methodology provides a basis for quantifying the uncertainty in parameter estimates. As expected from the observed behaviour, estimated parameters are significantly different for the same soil sample, but different flow conditions. (2) A “multirate” model has been developed for rate-limited reactive transport, where kinetic properties are represented as a distribution of rate constants. A corresponding statistical methodology has been developed to examine uncertainty in the thus obtained estimates relative to the time scales of available data. The “multirate” model is well suited to examine the data time scale vs. desired estimate time scale problem. Reactive heterogeneity can be represented in relatively concise form. The model could fit data better than several existing ones for reactive heterogeneity suggested recently. Further it has been shown that the traditional first-order rate law cannot be extended to cometabolic degradation of multiple substrates and an alternative rate law with multiple rate parameters has been formulated that also expresses interaction. Some data could be described well based on the new rate law, but the estimation problem appear too difficult to solve in practice. Implementation of comments by the evaluation panel It was suggested to take advantage of the data on fracture spacing obtained in Project 1. Mass transfer in some of the column experiments shows definite signs of rate limitation, but could not be modelled with a single rate parameter. The multirate model thus appears well suited to re-analyse the data. As the FRACTRAN model which has been used for interpretation of the data is based on similar mathematical algorithms as the multirate model, coupling the two should be feasible and is intended for the near future. Activities for next period A generic robust estimation algorithm will be developed. Criticism met in connection with publication of our results has mostly focused on the need to make assumptions for the errors in the data, a necessity for applying maximum-likelihood estimators. An improved robust estimation methodology will not rely on such assumptions. Multirate modelling of tracer experiments in columns of fractured till will be attempted. Subproject 3.2: Catchment modelling Objectives (1) To develop a model code based on the best available knowledge on process descriptions for the considered scale including new findings from Projects 1and 2 (2) To address the up-scaling problems (3) To test the process descriptions included in the model code on field data from Projects 1and 2 (4) To validate the model on two small-scale catchment areas (5) To gain insight into the problems and possibilities of using such code operationally for catchment modelling in Denmark (6) To assess the uncertainty and the reliability of using the developed code for catchment modelling in Denmark

20

(7) To develop a methodology for applying a catchment model operationally on other Danish catchments, including procedures for assessing model parameters and input data on the basis of normal data availability, and procedures for model calibration and validation on specific sites.

Activities and results The activities and results achieved for each of the above seven objectives are briefly described in the sections below. 1. Development of model code The model concept has been discussed and agreed upon in 1997 and minor adjustments have been introduced in 1998. Basically the code will consist of an integration of model codes already existing at the participating institutions together with development of additional pesticide related modules. The following progress has been achieved: • The coupling of MIKE SHE and DAISY is now functioning • A new surface module is being implemented. The surface module includes a new management action in DAISY for spraying and surface pesticide processes. The latter comprises retardation and dissipation (e.g. photolysis, volatization) of pesticides in the crop canopy • A new module for describing the biodegradation of pesticides has been developed and will soon be integrated in the coupled MIKE SHE/DAISY code Thus, the code describing the pesticide movement and transformation on a catchment scale is almost completed. 2. Up-scaling methodologies The adopted methodology has formed the basis for the design of the field monitoring programme and the model validation test scheme. 3. Test of process descriptions in new model code This activity will be carried out based on data and findings obtained in Projects 1 and 2. This activity has not yet started. 4. Model validation on two small catchments The model validation involves four activities, namely identification of test catchments, data collection in combination with extended field monitoring, definition of model validation test schemes, and the actual modelling. 4.1 Identification of test catchments At project start, two different locations were originally selected as test sites for the catchment modelling, a location in Jutland (Fladerne Bæk) dominated by sandy conditions and a location situated on Zealand (Flakkebjerg) dominated by moraine clay conditions. An important reason for selecting these two locations was that activities related to projects 1 and 2 were located here ensuring consistency between process verification and model validation at catchment scale. These sites were also selected so that it was possible to identify sub-catchments, where the result of the pesticide transport and transformation processes, including surface, root zone and groundwater processes, could be tested against field data integrating the responses over the areas, reflecting that the focus of this sub-project is on catchment areas rather than on processes at point scale. However, as data collection and field monitoring were initiated at the Flakkebjerg site three major problems appeared. While data were available for calibration of an overall water balance covering a larger area (26 km2), it became clear that the smaller drainage catchment chosen for pesticide modelling and equipped with a drainage gauge was not suitable for two reasons. First, initial pesticide measurements in the drain system revealed that the drainage system appeared to have direct contact with a site at the research station used for filling and washing spray equipment. The pesticide leaching from this point source was by far the dominant source for pesticide concentrations in the drain water instead of the non-point source, which is the focus of the overall project. Second, the drain gauge became flooded during the autumn 1997 making further flow and pesticide monitoring impossible.

21

Finally, the third problem with the Flakkebjerg site was the nature of the field experiments on the land areas belonging to the research station. If the validation approach (described below) should be possible, it is crucial that a realistic input function can be established. At project start it was expected that such a function could be established back in time on the basis of information from the research station. This has, however, not been possible at larger scale, and as the pesticide use at the research station is not comparable to common agricultural practice, it would not be possible to use statistical data to establish a realistic input function. It was therefore necessary to find another moraine location for the model validation exercise. Two other possibilities, viz. the Lillebæk land monitoring catchment (LOOP) on Funen and the Eggeslevmagle groundwater monitoring site (GRUMO) near Flakkebjerg, were investigated. As both of these sites are part of a national monitoring programme focusing on surface- and groundwater quality, respectively, some data on pesticide findings are available. The two sites, however, differ considerably with regard to their main purpose. Thus, the Lillebæk catchment would provide most data from the near surface (monitoring of agricultural practice together with pesticide measurements in shallow groundwater (3-5 m depth), drain water, and stream water), but almost no data from deep aquifers. The Eggeslevmagle area would, on the other hand, provide data from the deeper aquifers only. The nature and possible outcome of the model validation task would therefore differ depending on the site chosen. Selecting the Lillebæk site would provide validation data for a catchment study including water and mass fluxes in the surface-near layers, whereas selecting the Eggeslevmagle site would provide validation data for a groundwater study over a similar area. Thus, it was not possible fully to combine the catchment approach enabling mass balance control integrated over an area with the objective also to focus on aquifer processes. As the main focus of the entire project is groundwater and as the Eggeslevmagle site is located close to the originally selected Flakkebjerg site and offers a much better possibility for sharing of information and co-ordination with the other SMP96 projects, it was finally decided to proceed with the Eggeslevmagle site. 4.2 Field monitoring at the two test catchments To supplement the existing data the following field programme has been planned and is under implementation for the two test catchments: Fladerne Bæk At the Fladerne Bæk site no continuous measurements of the streamflow have been made in the past. With respect to the calibration and validation of the model, it is of major importance to establish a streamflow gauging station near the outlet of the Fladerne Bæk to sample flow continuously. A gauging station has just been established at an appropriate location a few hundred metres upstream from the outlet. In addition, the flow will be measured as single measurements every month, right at the outlet, in order to relate the outflow to the continuous flow. A measurement campaign, conducted this summer at multiple stream locations, revealed that more than half of the flow at the outlet originates from contributions downstream from the five established wells. With respect to monitoring of the groundwater at the downstream end of the catchment it has been decided to establish six more wells. These wells will be installed in pairs on each side of the stream at different distances from the outlet, during the next few months. Samples for pesticide analyses will be taken from the stream at the outlet every month. The samples will be analysed for Bentazone and MCPP. The first sample has just been collected and this sample is presently being screened for 40 different pesticides. To map concentrations of pesticide in groundwater, 48 samples will be collected from the five existing wells and from the six boreholes to be established. The samples from the existing wells will be taken from various depths whereas the new boreholes will contain only one filter. Eggeslevmagle – general At the Eggeslevmagle site measurement campaigns, conducted this summer at multiple stream locations, revealed a remarkable increase in the flow at summertime in the

22

Lungrende over a short distance between Frankerup and the outlet into the Øllemose Rende. This increase in flow is most likely mainly due to a contribution from a near-surface aquifer in the area to the west of the Lungrende. A detailed geological and geochemical mapping from boreholes conducted in Project 2 confirms the existence of this local aquifer. Also at the Eggeslevmagle site no continuous measurements of the streamflow have been made in the past. Thus, a streamflow gauging station will be established at the outlet of the Lungrende into the Øllemose Rende. In addition, the flow will be measured as single measurements a few times through the year at various locations of the stream. It must be expected that the concentration of pesticide in the stream is highly varying over time. Thus, it has been decided to measure the concentration of pesticide in the stream only at one location but continuously and proportionally to the flow. The sampling will take place at the outlet of the Lungrende in the Øllemose Rende. The first sample has just been collected and this sample is presently being screened for 40 different pesticides. Pesticide sampling from 10 boreholes in the near-surface aquifer (3-4 m depth) is planned. The boreholes will be constructed using a ramming technique. The reason why no unsaturated zone sampling will take place as originally proposed is that it is very difficult to develop a reliable sampling technique in the unsaturated till. Also the mentioned finding of a local near-surface aquifer is part of the reason. Analyses of pesticides of the near-surface groundwater will be conducted from a network of boreholes established between fields. At least two campaigns of sampling will be conducted. Depending on the outcome, several samples (time series) from selected boreholes or additional samples from all boreholes will be taken and analysed. Eggeslevmagle – redox environment In the Frankerup area, located within the Eggeslevmagle area, sediment samples from 11 sites were collected to depths of maximum 9 m, and stored at 5 °C until analysis. Matrix colour and redoximorphic features were described following the Munsell Soil Colour Chart System. Distribution of free CaCO3 and exchangeable Fe2+ in the profiles was analyzed. Sorption of atrazine and mecoprop to sediment samples of different geochemical environments (pH and Eh) was obtained using the batch equilibration procedure with radiolabelled and non-radioactive solutions of atrazine and mecoprop. In the Frankerup area more than 95% of the area are used for intensive crop production. A moraine landscape mainly with clayey soils, extramarginal stream valley with sandy and organic soils and an esker make up the 3 major morphological units within the area. Deposits from the last two glacial advances have been recognised. Often melt-water sand, gravel or melt-water clay/silt indicate the transition between the upper and lower till unit. Leaching and consumption of inherited CaCO3 have formed a CaCO3-free zone to depths of maximum 3 m. After the last glaciation oxidation-processes have formed a yellowish brown or brown zone, followed by the unchanged gray coloured sediments. Also, in the deeper sandy layer an oxidised zone has formed. Three major geochemical environments were identified; an upper CaCO3-free and oxidised zone, a CaCO3-rich and oxidised zone, and a CaCO3-rich reduced zone. The distribution of exchangeable Fe2+ is a very good indication of the redox environment, often the highest concentrations occur in the gray, reduced zone. Sorption of atrazine reflects the development in redox environments. Mecoprop demonstrates a close relation with the presence of CaCO3 (and pH) but seems only little influenced by the redox status of the sediments. The study has demonstrated mapping of different pH-Eh geochemical environments within morphological units as a valuable tool in the description and modelling of pesticides in clayey till areas. Also, a close correlation between sorption of different pesticides and pH or redox status makes already existing profile records with information on calcium carbonate (pH) and colours (Eh) applicable in an assessment of leaching of pesticides at a regional scale.

23

4.3 Model validation tests Model validation test scheme The model validation tests will include tests for the three steps in the modelling: • • •

Flow modelling, where data on streamflow and groundwater heads will be used Transport modelling (conservative solutes), where CFC data on groundwater age will be used Pesticide modelling (reactive solutes), where pesticide concentrations will be used

The main idea in the validation test methodology is to conduct a series of stepwise tests, where more and more field data are included in the model calibration. In this way the predictive capability of the model will be tested for different levels of data availability and conclusions can be derived on the data requirements and accuracy in model predictions. The methodology will be as follows: a) The performance criteria for the flow, transport and pesticide modelling will be based on the types of criteria described below. The numerical acceptance values for each criterion will be specified on the basis of an analysis of the existing data b) Validation test of flow model. This implies a traditional split-sample test with calibration using part of the available data and subsequent validation test using independent data and comparing the performance criteria with the achieved model results c) On the basis of b) validation tests for the transport and pesticide models are performed in a ‘blind test mode’ implying simulation without prior calibration d) Calibration of the transport model on the basis of the available data on groundwater age e) On the basis of d) a second validation test for the pesticide model is performed in a ‘blind test mode’ implying simulation without prior calibration f) Calibration and a third validation test of the pesticide model. This implies, if possible, a split of the pesticide data in two parts, of which one part is used for calibration and the other for the validation tests. Performance criteria for model validation Due to incomplete knowledge on the pesticide input function (temporal and spatial distribution of pesticide application) attempts will be made to achieve as good as possible a statistical representation (in terms of spatial mean and standard deviation) of this input rather than a correct point by point representation. The up-scaling procedure for the root zone processes, including the pesticide transformation and leaching in the root zone, will be based on an aggregation of results to subcatchment areas representing so many computational grid points that the statistical properties of the pesticide input can be preserved although the point by point (georeferenced) input cannot be expected to be correct. As the major part of the pesticides (> 95%) is expected to be degraded in the root zone before reaching the groundwater system, it is important to assess the capability of the model to predict these loss terms. After having reached the groundwater table a significant part of the water (and hence a similar part of the pesticides) moves horizontally towards drains and river systems, where it appears as catchment runoff. Therefore a test of the model’s capability to predict the total flux of pesticide out of the catchment is an important indirect check of the input function to the aquifer system. Although the focus in the present project is on pesticide concentrations in the aquifer system, this catchment flux is considered an important measure and is hence included in the performance criteria. Most observations indicate very low pesticide concentrations as compared both to the analysis method and to the process understanding. As both the representativeness (in time and space) of the monitored samples and the simulated concentrations are rather uncertain, and definitely more uncertain than in most other groundwater pollution cases, the model cannot be expected to be able to reproduce concentration levels very accurately. Therefore, a criterion focusing on the extent of areas with

24

concentrations above/below a certain critical value is included. This critical value may be chosen either as the detection limit, the maximum allowable concentrations, or a value in between. On this basis the performance criteria for the pesticide model will be as follows: 1. Annual flux of pesticide out of catchment simulated within given accuracy (accuracy within a factor of ?). This criterion can maybe only be used for Fladerne Bæk. 2. Groundwater concentrations in top layer (upper 1-3 m). Statistical distribution within catchment compared to measurements: • •

% of catchment with concentrations lower than the critical level (accuracy within +/- ?) average concentration (accuracy within a factor of ?)

3) Groundwater concentrations in deeper layer (to be defined, so that it includes anoxic zones). Statistical distribution within catchment compared to measurements: • •

% of catchment with concentrations lower than the critical level (accuracy within +/- ?) average concentration (accuracy within a factor of ?)

The performance criteria for the transport model will be: 1. Groundwater age • distribution curve over catchment • age along flow lines The exact numerical figures to be used as success criteria will be specified before conduction of the actual model validation tests. The above criteria all have the full catchment/area as a basis. In addition it will be attempted to divide the aquifer area in sub-areas with different geological/redox conditions and try to carry out the tests within each of the sub-areas. 4.4 Modelling work Model constructions for the two test catchments are in progress. Initial model runs focusing on calibration of the flow models have been made. 5. Post-validation analyses of model performance This activity has not yet been initiated. 6. Assessment of uncertainties in model predictions This activity has not yet been initiated. 7. Methodology for applying pesticide catchment models operationally This activity has not yet been initiated. Implementation of comments from the evaluation panel We have noted the critical comments and the concerns expressed by the evaluation panel with regard to: • • • • •

Difficulty to test the model completely because the data from Projects 1and 2 will be insufficient Difficulty to assess the input function pesticide modelling at catchment scale Feasibility of “statistical approach” for pesticide modelling Need for rigorous and transparent model parameterisation Need for rigorous validation test schemes

25

We acknowledge in general that pesticide modelling at catchment scale is a non-trivial exercise with a considerable risk of failure. The main ambition in the project is to test whether a model comprising the main knowledge on processes and parameters from Projects 1 and 2 has any predictive capabilities at catchment scale. We should like to emphasise that we do not claim so beforehand, but we should like to test it during the project. We believe that an answer to this question is of strategic interest, amongst others in the light of the more and more widespread use of models for registration purposes. We agree with the evaluation panel that we cannot test the model completely and can hence not claim that the model is correct with regard to description of the main processes. Therefore the main objective of the modelling exercise is limited to testing the predictive capability of the model with respect to simulation of groundwater age and concentration in the upper aquifer layer of two specific catchments. The adopted validation test scheme is tailored to this objective. Thus we do neither make any claims of general validity of the model, nor of validity with regard to simulation of other variables than those included in the validation tests. We acknowledge the difficulty in assessing the input function, i.e. the actual history of pesticide application and agricultural management practices. In our opinion the only feasible way of assessing the input function, except for small research catchments, is a “statistical approach” where information on cropping pattern, agricultural practice, and pesticide application, derived from available agricultural statistics and the agricultural extension service, is applied on a catchment scale. This implies that the exact locations of the different crops and pesticide applications are not known, but that their statistical distributions within the catchment are preserved. Such approach has previously been applied successfully in nitrogen modelling (e.g. Styczen and Storm, 1993) and is now also being used in pesticide modelling (Loague et al., 1998). Thus, contrary to the opinion of the panel, we believe that such approach is feasible and worth pursuing. As a consequence of the comments from the panel we have, however, put more emphasis on the validation of the model against field data from the two test catchments. Therefore the field monitoring programme has been expanded. Activities for next period The key activities for 1999 will be: • Finalisation of model code development • Tests of process descriptions against field data from Projects 1 and 2 • Completion of field monitoring on flows, groundwater heads, and pesticide concentrations for Eggeslevmagle and Fladerne Bæk • More drillings near Frankerup with collection of sediment samples and subsequent assessment of a regional distribution of geochemical environments and related sorption properties • Conduction of model validation tests for the two catchments • Further analysis of model performance at the two catchments • Assessments of uncertainty and reliability of model predictions • Preparation of a methodology for applying the model operationally on other Danish catchments. Subproject 3.3: Regional modelling Objectives To obtain a better understanding of the influence of geology on pesticide occurrence in groundwater on a regional scale.

26

Methodology The modelling approach focuses on determining travel path ways and age distribution including residence time in different redox environments of pesticides arriving in aquifers and at well fields under different interpretations of hydrogeology and redox environments on regional scale. Two catchments have been chosen for regional modelling: the clayey till Vårby creek catchment and the sandy Karup River catchment. For the two catchments conceptual hydrogeology models are determined based on • • • • • • • •

geomorphological analysis well data geological maps the pre-Quarternary surface map hydraulic-head maps and time series baseflow observations streamflow time series well screen interval

Steady-state saturated groundwater flow is modelled for the conceptual hydrogeology models. The models are calibrated against existing hydraulic-head and flow data, and if possible an inverse modelling approach is used. Results from Project 2 indicate that redox environments control degradation of most pesticides. Therefore, conceptual models of redox environments for the developed hydrogeology models are determined based on • • •

sediment samples in some of the geomorphological units (based on results from Subproject 3.2) knowledge from existing geological data conceptual hydrogeological models

Transport modelling using particle tracking is performed for the hydrogeology/redox models to determine travel pathways and age distribution including residence time in different redox environments of pesticides arriving in aquifers and at well fields under different interpretations of hydrogeology and redox environments at regional scale. The particles are released continuously in time over the total area. Each particle knows: • • • •

when it was released where it came from years spent under aerobic conditions years spent under anaerobic conditions

When a simulation is completed, the age distribution of particles for aerobic and anaerobic conditions at given locations are studied. Given a specific pesticide, degradation kinetics in the redox environments can be estimated (results from Project 2) and the particles that represent degraded pesticides can be excluded. The remaining particles represent pesticides in the studied aquifer or well field. Using different models of hydrogeology and redox environments this approach makes it possible to analyse the protective capability of geology against pesticide pollution of groundwater at regional scale for the sandy location dominated by matrix flow. On the clayey till location with expected fracture flow the approach has been changed to include preferential flow. Conceptual sub-models are determined for fracture flow based on local studies at

27

the Vårby Creek catchment (based on results from Project 1 and Subproject 3.2) and a parametric study is carried out to obtain estimates of ratios matrix to fracture flow and travel-time distribution. This information is used in the regional models by introducing a ratio of the particles at the bottom of the geological layer with expected fracture flow assuming that the travel time through this layer is diminishing. This approach makes it possible to analyse the protective capability of geology against pesticide pollution of groundwater at regional scale for the clayey till location with matrix and fracture flow. At both locations “relative pesticide concentrations” in a given volume are calculated on the basis of the age information on particles representing pesticides in the aquifer or well field and time series of pesticide leaching. Each particle is assigned a unit mass and is released according to calculated time series of pesticide loads determined from simplification of the modelling approach developed in Subproject 3.2. In a given aquifer or well field the age distribution of particles for aerobic and anaerobic conditions is determined and exclusion of particles is performed as discussed above. Adding the mass of remaining particles and dividing with the volume of the aquifer or well field result is an indication of the pesticide concentration referred to as the “relative pesticide concentration” in this study. This approach makes it possible to relate “relative concentrations” to the results obtained above. Activities and results: field and modelling At the end of the year geological models of Vårby Creek catchment will be completed. Implementation of comments from evaluation panel Objectives, strategy, and modelling concepts have been presented to the evaluation panel. The panel had a very positive attitude toward this project, the only concern was the use of pesticide loads estimated by a similar modelling approach as the one developed in Subproject 3.2. Comments and problems regarding this approach are discussed in Subproject 3.2. It was suggested to use available CFC age-dating results to verify the proposed redox environments for the catchments. Some of the CFC compounds are nearly conservative to redox environments while others are degraded under anaerobic conditions. Thus, comparison of dating results of different CFC compounds can give an indication of the present redox environment. In light of these comments and the implementation of comments in Subproject 3.2 we have decided not to change the present strategy and concepts. Inclusion of age-dating results for redox environment verification will be carried out for available data. Activities for next period Vårby Creek (clayey till): • • • • • • • •

Determination of conceptual models of hydrogeology Modelling of steady-state saturated groundwater flow Calibration using inverse modelling Parametric studies in till to obtain estimates of ratios matrix to fracture flow and travel-time distribution Determination of conceptual models of redox environments Transport modelling incorporating matrix/fracture flow using particle tracking Determination of travel path ways and age distribution of pesticides arriving in aquifers and well fields Analysis of relative concentrations

Karup Creek (sandy):

28

• • • • • • • •

Determination of geological models Determination of conceptual models of hydrogeology Modelling of steady-state saturated groundwater flow Calibration using inverse modelling Determination of conceptual models of redox environments Transport modelling using particle tracking Determination of travel path ways and age distribution of pesticides arriving in aquifers and well fields Analysis of relative concentrations

PRESENTATIONS International scientific papers in peer reviewed journals Fomsgaard, I.S., G. Felding and P. Schjørring (in press). Sampling and substrate application methods for pesticide mineralization experiments in undisturbed soil samples. Int. J. Environ. Anal. Chem. Helweg, A., I. Fomsgaard, T. K. Refstrup and H. Sørensen (in press). Degradation of mecoprop ans isoproturon in soil.- influence of initial concentration. Int. J. Environ. Anal. Chem. Hollenbeck, K. J. and K. H. Jensen 1998. Experimental evidence of randomness and non-uniqueness in unsaturated outflow experiments designed for hydraulic parameter estimation. Water Resour. Res., 34 (4), 595-602. Hollenbeck, K. J. and K. H. Jensen 1998. Maximum-likelihood estimation of unsaturated hydraulic parameters. J. Hydrol., 210 (1-4), 192-205. Hollenbeck, K. J., C.F. Harvey, R. Haggerty and C.J. Werth (in press). A method for estimating distributions of mass transfer rate coefficients with application to purge and batch experiments. J. Cont. Hydrol. Jørgensen, P.R., A. Hoff and N.O. Jørgensen (submitted). A clay aquitard conceptual fluid flow model constructed on the basis of observed fracture scale and porous media parameters. Ground Water. Jørgensen, P.R., L.D. McKay and N.H. Spliid 1998. Evaluation of chloride and pesticide transport in a fractured clayey till using large undisturbed columns and numerical modeling. Water Resour. Res., 34 no.4, 539-553. Jørgensen, P.R. and N.O. Jørgensen (submitted). Assessment of pesticide leaching in fractured clayey till using a discrete fracture numerical model. Nordic Hydrology. Klint, K.E.S. and J. Fredericia (submitted). Calculation of Quantitative Fracture properties in Clayey Diamict Sediments for use in Hydrogeological Investigations. Nordic Hydrology. Klint, K.E.S. and P. Gravesen (submitted). Fractures and Biopores in Weichselian Clayey Till Aquitards at Flakkebjerg, Denmark. Nordic Hydrology. Loague, K, D.A. Lloyd, A. Nguyen, S.N. Davies and R.H. Abrams 1998. A case study simulation of DBCP groundwater contamination in Fresno County, California. 1. Leaching through the unsaturated subsurface. J. Cont. Hydrol. 29, 109-136.

29

Murdock, L., B. Harrar, B. Nilsson, W. Slack and R. Siegrist (submitted). In situ Measurements in Fractured till using sidewall sensors. Nordic Hydrology. Shapir, N., R. T. Mandelbaum and C.S. Jacobsen (accepted). Atrazine mineralization in aquifer sediments under denitrifying conditions by Pseudomonas sp. strain ADP. Environ. Sci. Technol. Styczen, M. and B. Storm1993. Modelling of N-movements on catchment scale – a tool for analysis and decision making. 1. Model description and 2. A case Study. Fertilizer Research 36, 1-17. Thorsen, M, P.R. Jørgensen, G. Felding, O.H. Jacobsen, N.H. Spliid and J.C. Refsgaard (in press). Evaluation of a stepwise procedure for comparative validation of pesticide leaching models. J. of Environ. Qual. Vinther, F. P., A.-M. Lind, F. Eiland and L. Elsgaard (in press). Microbial biomass and numbers of denitrifiers related to macropore channels in agricultural and forest soils. Soil Biol. Biochem.

Other scientific papers Christensen, T.G., N.L. Schouw, H.-J. Albrechtsen and K. Rügge 1998. Kontrollerende faktorer for aerob pesticidnedbrydning i grundvandsmagasiner. In: Pesticider i grundvand og drikkevand, ATV-komiteen vedrørende Grundvandsforurening, Hotel Marselis, Århus 28. oktober, pp. 4756. Akademiet for de Tekniske Videnskaber, Lyngby. Hollenbeck, K. J. (in press). Multiple substrate cometabolism: Modeling and estimating kinetics. Progress Report 78, Department of Hydrodynamics and Water Resources (ISVA), Technical University of Denmark. Tüchsen, P.L., N., Tuxen, K. Rügge, H.-J. Albrechtsen and P.L. Bjerg 1998. Aerob nedbrydning af pesticider i søjleforsøg med grundvandsmateriale. In: Pesticider i grundvand og drikkevand, ATV-komiteen vedrørende grundvandsforurening, Hotel Marselis, Århus, 28. oktober 1998, pp. 57-66. Akademiet for de Tekniske Videnskaber. Lyngby. Tüchsen, P.L., N. Tuxen, K. Rügge, H.-J. Albrechtsen and P.L. Bjerg (submitted). Pesticider i grundvand – sorption i søjleforsøg. Vand & Jord. Tuxen, N., P.L. Tüchsen, K. Rügge, H.-J. Albrechtsen and P.L. Bjerg (submitted). Pesticider i grundvand – nedbrydning i søjleforsøg. Vand & Jord. Oral presentation Albrechtsen, H.-J. 1998. Det Strategiske Miljøforskningsprogram, delprogrammet 'Grundvand & Pesticider’, præsentation af programmet - hvordan kan den ny viden anvendes i miljøforvaltningen. Amternes Videncenter for Jordforurening, Temadag om Punktkildeforureninger med pesticider, Nyborg Strand. 8. oktober 1998. Christensen, T.G., N.L. Schouw, H.-J.Albrechtsen and K.Rügge 1998. Kontrollerende faktorer for aerob pesticidnedbrydning i grundvandsmagasiner. ATV-komiteen vedrørende Grundvandsforurening, møde om Pesticider i grundvand og drikkevand. Hotel Marselis, 28. Oktober 1998.

30

Ernstsen, V. 1998. Atrazine in different redox environments of clayey till. 1998 Annual meeting abstract. America Society of Agronomy, Crop Science Society of America, Soil Science Society of America, Baltimore, Maryland. October 18-22 1998. p. 354 Fomsgaard, I.S. and G. Felding 1997. Comparison of sampling and incubation methods used for pesticide mineralization experiments in undisturbed soil samples. Proceedings of 6th Symposium on Chemistry and Fate of Modern Pesticides, Amsterdam. (Abstract and poster). Harrar, B., B. Nilsson and L. Murdock 1998. Seasonal Variations in Fracture Flux in a Surficial Glacial Till. Abstract for GSA Annual Meeting, Toronto, Ontario, Canada, October 26-29, 1998. Hoff, A., P.R. Jørgensen and C.E. Andersen 1998. The biopore model: A semi-analytical model of mass and solute transport in soil containing cylindrical biopores. Abstract. Conference om Mass Transport in Fractured Aquifers and Aquitards. May 14 to 16., 1998. Geological Institute, University of Copenhagen. Copenhagen, Denmark. Hoffman, M., J. Kistrup, C. Bryde, N.O. Jørgensen, R. Bossi and K, Villholth 1998. In situ pesticide transport experiment in saturated fractured till. Poster at Conference om Mass Transport in Fractured Aquifers and Aquitards. May 14 to 16., 1998. Geological Institute, University of Copenhagen. Copenhagen, Denmark. Hollenbeck, K. J. and K. H. Jensen 1997. Dependence of apparent unsaturated parameters on experiment type and estimation method. International Workshop on Characterization and Measurement of the Hydraulic Properties of Unsaturated Porous Media, Riverside, California, October 22-24, 1997. : Jørgensen, N. O., L. Skjernaa, S. Stipp, P.R. Jørgensen and O. Larsen (editors) 1998. Abstract Volumen to: Conference on Mass Transport in Fractured Aquifers and Aquitards. May 14 to 16., 1998. Geological Institute, University of Copenhagen. Copenhagen, Denmark. Geological Institute, University of Copenhagen. Copenhagen, Denmark, pp. 105. Jørgensen, P.R. 1998. Large undisturbed columns (LUC) for flow and contaminant transport experiments. In "Excursion Guide" to: Conference on Mass Transport in Fractured Aquifers and Aquitards. May 14 to 16., 1998, 44-49. Geological Institute, University of Copenhagen. Copenhagen, Denmark. Jørgensen P.R. and J. Baumann 1998. Macropore scale modeling parameters from clay-rich till aquitards in Denmark. Abstract. Conference on Mass Transport in Fractured Aquifers and Aquitards. May 14 to 16, 1998. Geological Institute, University of Copenhagen. Copenhagen, Denmark. Jørgensen P.R. and N.O. Jørgensen 1998. Numerical prediction of pesticide leaching through clay rich till using laboratory and field measured fracture scale modeling parameters. Abstract. Conference om Mass Transport in Fractured Aquifers and Aquitards. May 14 to 16., 1998. Geological Institute, University of Copenhagen. Copenhagen, Denmark. Klint, K.E.S. 1998. Fractures in Clayey Tills at Flakkebjerg, Denmark "Origin and distribution" Abstract Nordic Geological Vinter-meeting, Århus Jan. 13-16 1998.

31

Klint, K.E.S. and J. Fredericia 1998. Quantitative Fracture Characterisation in Clayey Diamict Sediments.“A Recipe” Extended abstract in: Mass Transport in Fractured Aquifers and Aquitards. Conference at Geoscience center Copenhagen, May 14-16 1998. pp 4. Malinovski, D. and K.E.S. Klint 1997. Geotekniske undersøgelser af moræneler ved Flakkebjerg. Foreløbige resultater af geotekniske undersøgelser ved Flakkebjerg. SMP 96. - Geological Survey of Denmark and Greenland, Report 1997/127, 21 pp. + appendices. Murdoch, L., W. Slack, B. Harrar and T. Siegrist 1998. Sidewall Sensors for Measuring in situ Properties. Remediation of Clorinated and Recalcitrant Compounds Conference Proceedings; Monterey, CA, May 18-21, 1998. Toräng, L. and N. Nyholm 1998. Biodegradation of 2,4-D in groundwater and surface water at ppt-ppb range concentrations". SETAC Europe 98, 14-18 April, 1998. Platform presentation. Tüchsen, P.L., N. Tuxen, K. Rügge, H.-J. Albrechtsen and P.L. Bjerg 1998. Aerob nedbrydning af pesticider i søjleforsøg med grundvandsmateriale. ATV-komiteen vedrørende Grundvandsforurening, møde om Pesticider i grundvand og drikkevand. Hotel Marselis, 28. Oktober 1998. Sørensen, S. R. and J. Aamand 1998. Mikrobiel nedbrydning af isoproturon i jord og grundvandssediment. ATV-komiteen vedrørende Grundvandsforurening, møde om Pesticider i grundvand og drikkevand. Hotel Marselis, 28. Oktober 1998. Posters Clausen, L. and I. Lind 1998. Sorption of pesticides to mineral surfaces. Book of abstracts: 9th International Congress of Pesticide Chemistry, The Food-Environment Challenge, p 6D-037. Published by: The Royal Society of Chemistry & The International Union of Pure and Applied Chemistry. Fomsgaard, I.S. and G. Felding 1997. Comparison of sampling and incubation methods used for pesticide mineralization experiments in undisturbed soil samples. - Proceedings of 6th Symposium on Chemistry and Fate of Modern Pesticides, Amsterdam. (Abstract and poster). Hoffman, M., J. Kistrup, C. Bryde, N. O. Jørgensen, R. Bossi and K. Villholth 1998. Pesticid transport i moræneler. Geologisk Nyt. Nr. 2, 16-19 (in Danish). Jacobsen, O.S. and F. P. Vinther 1997. Semi-automatic field station for monitoring agricultural leaching of nutrients and pesticides. - Proceedings of the First International Conference on Strategies and Techniques for the Investigation and Monitoring of Contaminated Sites. Field Screening Europe. Kluwer Academic Publishers, pp. 37-38. (Abstract and poster). Jørgensen, P.R., J. Baumann, T. Helstrup, J. Urup and K. Butzbach 1998. Hydraulik i sprækket ler. Geologisk Nyt. Nr. 2, 13-16 (in Danish). Klint, K.E.S. and P. Gravesen 1997. Macropores in clayey till at Flakkebjerg, Denmark. NJF Seminar: Preferential flow processes in soil and their importance for water flow distribution and transport of nutrients and pesticides. - The Royal Veterinary and Agricultural University, p. 23. (Abstract and poster).

32

Klint. K.E.S. and P. Gravesen 1998. Fractures and Biopores in Weichselian clayey till Aquitards at Flakkebjerg, Denmark Abstract in: Mass Transport in Fractured Aquifers and Aquitards. Conference at Geoscience center Copenhagen, May 14-16 1998. Madsen, L., and B. Lindhardt 1998. Sorption of pesticides onto aquifer sediments with low organic carbon content. Book of abstracts: 9th International Congress of Pesticide Chemistry, The Food-Environment Challenge, Aug. 1998, p. 7C-041. Published by: The Royal Society of Chemistry & The International Union of Pure and Applied Chemistry. Mortensen, A. P., R. J. Glass, K. J. Hollenbeck, and K. H. Jensen 1998. Visualization of quasi-2D unsaturated flow during dynamic outflow experiments. American Geophysical Union Fall Meeting, San Fransisco, December, 1998. Sørensen, S.R. and J. Aamand 1998. Degradation of the herbicide isoproturon in soil and groundwater. Poster presented at the 42nd Oholo conference: Novel approaches for bioremediation of organic pollution. Eilat, Israel, May 3-7, 1998.

General ’popular’ communication Albrechtsen, H.-J. 1998. Præsentation af Miljøforskningstemaet ’Peticider og grundvand’ på IMT ved forskningsminister Jan Trøjborgs besøg på DTU, den 6. Nov. 1998 Albrechtsen, H.-J. 1998. Grundvandsforureningens betydning for den fremtidige vandforsyningsstruktur i Danmark. Radiointerview 'Miljørapporten' DR1 v. Inger Anneberg, den 13. Sep. 1998 Albrechtsen, H.-J. 1998. Hvordan undersøges pesticidnedbrydning i grundvand? Radiointerview 'Miljørapporten' DR1 v. Inger Anneberg, den 9. Sep. 1998

Master Theses Albrecht, M. and P. Nielsen 1997. Nedbrydning af pesticider i grundvand. Master thesis. Department of Environmental Science and Engineering, Technical University of Denmark, Lyngby. Christensen, F.D. 1997. Transport og nedbrydning af pesticider i grundvand. Master thesis. Department of Hydrodynamics and Water Resources (ISVA), Technical University of Denmark. Christensen, T. G. and N. L. Schouw 1998. Miljøets betydning for mikrobiel omsætning af herbicider under grundvandsforhold - Effekt af ilt, pH, næringssalte, primær-substrat, koncentrationsniveau og deres interaktion. Master Thesis. Department of Environmental Science and Engineering, Technical University of Denmark, Lyngby. Guillaumie, F. 1998. Degradation and degradation products of MCPP, MCPP-P and BAM in groundwater under aerobic conditions. Master Thesis. Department of Environmental Engineering, Technical University of Denmark, Lyngby. Mai, P. 1996. Mikrobiel nedbrydning af MCPP og andre phenoxysyrer. Master thesis, Department of General Microbiology, University of Copenhagen.

33

Rasmussen, L. H. 1998. Mineralogical characteristic of macropore environments and soil matrix of an east Danish clayey till in relation to sorption of dissolved components in leaching soil water. (in Danish). Master thesis. The Royal Veterinary- and Agricultural University, Copenhagen. 113 pp + appendix. Sørensen S. R. 1998. Mikrobiel nedbrydning af isoproturon i jord og grundvandssediment. Master thesis, Department of General Microbiology, University of Copenhagen. Tuxen, N and P. L. Tüchsen 1998. Pesticiders nedbrydning i aerobe akviferer undersøgt ved søjleforsøg. Master Thesis. Department of Environmental Science and Engineering, Technical University of Denmark, Lyngby.

PH.D. PROJECTS FINANCED BY THE PROGRAMME Arildskov, N. P. Department of geology and geotechnical engineering. Title: Degradation of pesticides in aquifers. To be finished 1.8.1999. Hollenbeck, K. J. Department of Hydrodynamics and Water Resources, Technical University of Denmark. Title: Opscaling- and estimation perspective. To be finished 1.4. 1999. Klint, K. H. S. Geological Survey of Denmark and Greenland. Title: Nature and Origin of fractures in diamict deposits. To be finished 1.10.1999. Larsen, L., Geological Survey of Denmark and Greenland. Title: Degradation of pesticides in aerobic and anaerobic aquifers. To be finished 15.11.1999. Lipthay, J. R., Geological Survey of Denmark and Greenland. Title: Adaptation to degradation of pesticides in aquifers. To be finished 1.03.2001. Mortensen, M., Department of Hydrodynamics and Water Resources, Technical University of Denmark. Title: Preferential transport of pesticides to groundwater. To be finished 30.11.2000. Pedersen, P. G., Department of Environmental Science and Engineering, Technical University of Denmark, Lyngby. Title: Degradation of pesticides in aquifers with different redoxconditions. To be finished 1.8.1998. Toräng, L., Department of Environmental Science and Engineering, Technical University of Denmark, Lyngby. Title: Degradation of chemicals in environments at natural concentrations - kinetics and mechanisms. To be finished 31.8.2001. Thomsen, A., Department of Hydrodynamics and Water Resources, Technical University of Denmark. Title: Transport and degradation of pesticides - a field injection experiment. To be finished 1.4.2000.

34