Spatial and temporal analysis of pesticides concentrations in surface water: Pesticides atlas

Transcription

1 Journal of Environmental Science and Health Part B (2008) 43, Copyright C Taylor & Francis Group, LLC ISSN: (Print); (Online) DOI: / Spatial and temporal analysis of pesticides concentrations in surface water: Pesticides atlas MARTINA G. VIJVER, MAARTEN VAN T ZELFDE, WIL L.M. TAMIS, KEES J.M. MUSTERS and GEERT R. DE SNOO Leiden University, Institute of Environmental Sciences (CML), Leiden, the Netherlands Dutch water boards have a well-established program for monitoring pesticide contamination of surface waters. These monitoring data have been processed into a graphic format accessible online and designed to provide insight into pesticide presence in Dutch surface waters and trends over time: the Pesticides Atlas ( With this tool one can easily get maps of where a pesticide is being measured and where it might possibly constitute an environmental problem over the years. Presently, results of the periods 1997/1998 until 2005/2006 are available at the level of individual active ingredients. At a national level, the percentage of pesticides concentrations that exceed the maximum tolerable risk has declined 30% to 38% over the years 2003/2004 compared with 1997/1998. This means that surface water quality in the Netherlands has improved with respect to pesticides, however there are still many locations at which the measured concentrations exceed the environmental quality standards. The results on linking land use to pesticides concentrations were shown to assist in optimization of monitoring programs. By developing the present Internet tool, many new opportunities for environmental risk assessment and risk management were identified, e.g. optimization of monitoring strategies and communication to policymakers. Keywords: Pesticides; monitoring; surface waters; geographic presentation; internet tool. Introduction The inappropriate use of pesticides is causing world-wide pollution. In addition, the pollution of aquatic environment, is becoming more complex as more pesticides are used. [1,2] Pesticide pollution usually comes as a result of the legitimate application of pesticides by farmers and not just through illegal use or accidents. [2,3] However, pesticide use is not only confined to agriculture. These substances are also used on roads and rail tracks, in homes and in gardens. [4] They are used as anti-fouling paints on ships and boats, [5] timber treatments and surface biocides and against human disease vectors such as mosquitoes. Pesticides enter in the surface waters with several ways. Pesticides may be washed into ditches and rivers by rainfall; [6,7] surface waters can be contaminated by sprays either intentionally (e.g. in order to control water weeds) or unintentionally, (i.e. when crops are sprayed near ditches [8,9] or via runoff and leaching from Address correspondence to Dr. Martina G Vijver, Leiden University, Institute of Environmental Sciences (CML), P.O. Box 9518, 2300 RA, Leiden, the Netherlands; vijver@cml. leidenuniv.nl This paper was presented at the CEMEPE conference held on June 24 28, 2007 at the Skiathos Island, Greece. Received May 15, agricultural fields. [10] )Pesticides that are sprayed on hard surfaces such as pavements are subject to particularly high runoff into drainage pipes and sewers. In most countries, national and regional water authorities monitor pesticides concentrations in surface waters frequently. These monitoring programs are a tool in policy making, either to keep track of the effectiveness of environmental, water, or pollution policies (e.g European Union Water Framework Directive [11] )orassurveillances indicating where actions should be taken such as clean up actions of locations, restriction of use of certain pesticides in specific areas or re-evaluation of the standards of active ingredients. Also in the Netherlands these surveillances of the surface waters is done under different monitoring programs. The results of these programs, however, are amalgamated and reported periodically by different stakeholders. Due to the lack of transparency of monitoring data, optimal use of these data was difficult especially at a national level. Moreover, the traditional (tabular) presentation of the monitoring results is difficult to interpret and geographic presentation of individual pesticide concentrations in the form of maps has several advantages. [12] Against this background, Leiden University Institute of Environmental Sciences (CML) initiated with solid support of an array of other organizations a study to assess the potential for presenting such maps in the form of a Pesticides Atlas of the

2 666 Vijver et al. Netherlands at the Internet. [13] The Pesticides Atlas aims to provide an insight into pesticide contamination of Dutch surface waters and to investigate: (i) where a pesticide is being measured and when it exceed an environmental standard, (ii) whether pesticides concentrations changes over time, (iii) whether it is possible to link pesticide concentrations with land use data and (iv) how to improve the regional monitoring systems. The Internet tool makes information available at the level of active ingredients. [14] Active ingredients are in this paper further described as pesticides. Additionally, the tool provides an executive summary of the main results of active ingredients that is especially made for management goals. These summaries include total number of measurements, results of groups of pesticides classified by mode of action of pesticides, top ten lists of the most polluting categories of land use and trend-analyses of pesticide-related problems over time. The Pesticides Atlas is not only interesting for scientists, but has also been implemented as a tool that is relevant for society. It is an Internet tool that is publicly available and free accessible at In this paper we will focus on the results of the monitoring following the four main goals of the Pesticides Atlas. Material and methods Processing and aggregation of monitoring data Pesticides monitoring data for the periods were derived from databases owned and administered by the 28 Water Boards in the Netherlands with prior checks being made on data quality and quantity. Criteria checks were among other things verification of CAS-numbers and pesticides identification, a check on x-y coordinates, a check on detection limits, aggregation of time series in reported monitoring data, the exclusion of extreme values by using an outliers check. The raw monitoring data are initially processed and aggregated in a stepwise procedure. Spatial aggregation is carried out at the level of: one by one or five by five kilometer-grid cells. Temporal aggregation is carried out firstly over annual periods followed by two-years periods. [14] Environmental quality standards used to compare the pesticides concentrations are the European drinking water standard, the Maximum Tolerable Risk (MTR) and the pesticide authorization standard applied by the Dutch Board for the Authorization of Plant Protection Products and Biocides (CTGB). The drinking water standard has been set at 0.1 µg/l for almost all individual pesticides. The limit for the sum of pesticides according to the drinking water standard is 0.5 µg/l. MTR and CTGB criteria are pesticide-specific and vary, depending on the ecotoxicity of the pesticide. In the case of the MTR and CTGB standard, monitoring data are aggregated by calculating the 90% percentile. In the case of the drinking water standard, data are aggregated in the form of the maximum-recorded value. The percentage of pesticides exceeding environmental quality standards has been calculated by dividing, for each 5 5kmgrid cell, the number of substances exceeding the standard by the number of substances assessed. A substance is regarded as exceeding the standard if the 90th percentile of the measurements exceeds the standard. Within the aggregation, measurements either below or above or equal to the quantitation limit (LOQ) are proceeded parallel to each other. When comparing aggregate values with standards, the following three situations may therefore arise [14] : (i) when there is just one aggregate value and it is above or equal to the LOQ, then this is the only value available. In this case we have an assessable value, which may or may not exceed the standard. (ii) when there is just one aggregate value and it is below the LOQ, again this is the only value available. One of two situations may then arise: a) when the value is below or equal to the standard, we have an assessable value that does not exceed the standard in question, b) when the value is above the standard, we have a non-assessable value, as there is no way of establishing whether or not the actual, unknown value does indeed exceed the standard. (iii) When there are two aggregate values, one below the LOQ, the other above or equal to it, there are three alternatives: a) when the value above or equal to the LOQexceeds the standard, only this value is used: we then have an assessable value that exceeds the standard, b) when both values are below or equal to the standard, we have assessable values that do not exceed the standard, c) when the value above or equal to the LOQ isless than or equal to the standard and the aggregated value below the LOQ exceeds that standard, we have a non-assessable value, asthere is no way of establishing whether or not the actual, unknown value indeed exceeds the standard. For the calculation of different products on environmental target exceeding a minimum of five to fifteen measurements, depending on the kind of product, are used in order to exclude incidental measurements from the national overviews. Mapping monitoring data Aggregated data on individual pesticides are depicted in a geographical presentation using a geographic information system (GIS)-framework. Colors are used to classify the aggregated pesticides concentrations compared to the different environmental quality standards (Table 1). If no measurements are available, the cell is shown in white. Grey cells represent measurements that are nonassessable as there is no way of establishing whether or not the actual, unknown value exceeds the standard.

3 Pesticides concentration in surface water 667 Table 1. Overview of classes to compare concentrations with environmental standards. Color DW classes CTGB classes MTR classes blue <DW <0.01 <NR green >0.01, <CTGB NR >, <MTR yellow <10x DW >CTGB >MTR orange >2x CTGB >2x MTR red >10x DW >5x CTGB >5x MTR DW = Drinking water criteria; CTGB = Authorization criteria; MTR = maximum tolerable risk, NR = negligible risk both having an ecotoxicological basis. Note that most figures contain a version date of September 2006 on the picture; this is the date at which the database with measurements was updated. Frequently results of the pesticides monitoring are updated at the Internet. For the most recent information, go to pesticidesatlas.nl. [13] Trend analyses Trend analyses were performed to get insight into changes of environmental quality of surface waters with respect to pesticides concentrations over time. The analyses over time were depicted both using geographical visualization and using a trend line graph. To compile these maps, for each 5 5kmAtlas cell the value of the numerical category for the selected period is deducted from that for the reference period 1997/1998. If the result is negative, this means that there has been increased exceeding of the standard in the cell in question and thus the situation has deteriorated. If the result is zero, the situation has remained the same, and if it is positive, the situation has improved. The results are then assigned colors for use on the maps: red, orange or yellow for a decline in quality, green for an improvement and blue for unchanged. If, for a given 5 5kmcell, there is no measurement available in one of the two periods or if the measurement is non-assessable in one or both of the periods, the cell is colored grey. If no measurements are available for either of the periods, the 5 5kmcell is colored white. The advantage of this method is that it is simple, in line with the idea of an atlas involving a literal comparison between map codings and likely to be rather insensitive to outliers in the monitoring data. A drawback is that it will also be insensitive to any major changes occurring within an individual numerical category. Thus, if a particular compound changes from exceeding the MTR by a factor 100 to exceeding it by a factor 10 in a given cell, for example, the cell would still be marked on the map as unchanged, as both values are in the category > 5x MTR. When measurements of the pesticides were made in every period, trend analyses was done on the actual (aggregated) data. In case of missing values in the monitoring over the years, data were imputed using TRIM software. [15] The reason to use an imputing technique is because missing data are not randomly distributed over space and time; rather full regional or whole monitoring periods are missing. For these missing values, the anticipated value is then imputed. A number of basic assumptions were involved: (a) The differences between periods are the same in all cells, (b) The only cells that could be assigned to a category other than no data or non-assessable in at least one period, c) Only pesticides having, in all periods, at least one cell with an imputed value greater than 0 or exceeding of the standard, d) Only pesticides for which fewer than 50% of cells have a missing value. In the data presentation, trends in which missing values have been imputed can be recognized by the fact that the trend is then indexed, with the average value for the reference period 1997/1998 being set at 100 and changes expressed relative to that period. Comparing monitoring data with land use data Observed concentrations of the pesticides were analyzed for correlation with acreages of dominant crops in The Netherlands (statistically analyzed at the 1 1kmbase using a non-parametric Spearman test). In this atlas we are concerned solely with cases that give an indication of the possible causes of high pesticide levels, i.e. in the positive correlations. Significant correlations are those that have a P-value <0.05. Twenty-four major crop types were distinguished based on two national sources; the national land use database [16] largely based on satellite images and a digitized database of landscape ecological characteristics. [17] To establish the relations between exceeding of quality standards and crop acreage, all assessable measurements were statistically analyzed (non-parametric Mann-Whitney U-test). If there are significantly larger areas devoted to a particular type of land use in cells where standards are exceeded, the pesticide is deemed to be in exceeding for that category of land use. The final step is to count, for each land use category, how many pesticides exceed limits. Results and discussion The number of active ingredients and the number of locations where measurements have being carried out over the periods of 1997 to 2006 are shown in Table 2. From this table it is clear that for every period data of about Table 2. Overview of the amount of pesticides data in the database. Number of active ingredients # Number of locations 1997/ / / / / # active ingredients and relevant metabolites.

4 668 Vijver et al. 200 active ingredients (and metabolites) are available. The number of measurement locations has been increased over time. Of all active ingredients the number of measurements varies strongly. Totally for every 2-year period there are about 150,000 measurements available. The geographical variation in the number of pesticides measured through the Netherlands varies. Pesticide monitoring data and environmental standards The percentage of locations at which pesticides concentrations exceed the European drinking water standard over the different months in a year, is given in Fig. 1 (upper graph). The results show that at 28 62% of the locations, levels exceeding of drinking water criteria can be found over the Fig. 1. Percentage of locations with pesticides concentrations exceeding an environmental standard during the year (2003/2004): Upper graph: European drinking water standard (0.1 µg/l), lower graph: authorization standard.

5 Pesticides concentration in surface water 669 Fig. 2. Left side: Percentage of pesticides exceeding the European Union (EU) drinking water standard during the period 2003/2004. Right side: Percentage of pesticides exceeding the maximum tolerable risk (MTR) during the period 2003/2004. Figures were made on a5 5kmgrid scale. different months of the monitoring period 2003/2004. The highest percentage of locations that exceed the drinking water criteria can be detected in the months May and June. This period is the growing season for many different crops and therefore a large pesticides use can be expected. It is remarkable that pesticide concentrations were also above this standard during the winter period, because during winter pesticides are hardly used. The percentage of locations where the pesticide authorization standard (CTGB) is exceeded is much lower (right graph, Fig. 1), between 5 and 14%. The highest percentage of locations that exceed the authorization standard occur in May and December. The differences between the numbers of pesticides concentrations exceeding the drinking water standard and these exceeding authorization standard can be explained by the differences in values of the quality targets. Moreover to a minor extent not all pesticides that exceed the EU-drinking water standard have an authorization standard. The percentage of pesticides in the Netherlands that exceed the European drinking water standard (0.1µg/L) and MTR over the periods 2003/2004 is also presented in a geographical way (Fig. 2). Many sites can be found that have at least 10% of the pesticides measured exceeding a target value. Regional differences are obviously seen from these figures. Especially in the western part of the country many pesticides concentrations exceed the standards. With this geographical presentation, the overall environmental quality of surface water can be evaluated. Interpretation of the results at the regional scale is also possible by means of zooming at a certain area (not shown in this paper). With this facility, additional layers of landscape ecological characteristics, boundaries of water body basins and water discharge areas can be evoked. Changes in pesticide concentrations over time Changes in the percentage of pesticides exceeding the MTR standard have been calculated by subtracting the percentage of substances exceeding the standard in the reference period (1997/1998) from the corresponding percentage in the period 2003/2004 (Fig. 3). Percentages are regarded as unchanged if they differ by less than 5% between the two periods. From Fig. 3 locations can be identified where the quality has improved or is deteriorated. The results show that especially in the northern and western part of the country water quality regarding pesticides concentrations has improved. In the southern part of the country increased percentages of exceeding pesticides were

6 670 Vijver et al. Fig. 3. Differences in the percentage of locations exceeding the maximum tolerable risk (MTR)-value between 2003/2004 period and 1997/1998 period. observed, meaning that the quality has deteriorated. At the moment for many locations evaluation of the trend was not possible (grey cells in Fig. 3). The same data can be used to calculated the average numerical category of the percentage of substances that exceed the MTR value per 5 5kmcell. Here the overall change in pesticides exceeding at a national scale is indexed to the average value in the reference period 1997/1998 that is set to 100 (Fig. 4). As can be seen, the percentage of pesticides found in the Dutch surface waters has decreased approximately by 30 38% (Fig. 4) meaning that the overall surface water quality regarding pesticides improved between the period 2003/2004 and 1997/1998 at national scale. Not that

7 Pesticides concentration in surface water 671 Fig. 4. Trend analyses of the percentage of pesticides exceeding the maximum tolerable risk (MTR). The accuracy of the calculations is indicated as the standard error: the smaller the error, the more accurate the calculation. although the water quality with respect to pesticides improved, there are still many locations at which the MTR is exceeded (Fig. 2). A list of most serious problem pesticides exceeding the MTR was calculated based on a weighted approach. The weighted number takes into account the level of exceeding per 1 1kmgrid cell and the number of 1 1km grid cells in which the substance has been assessed. The level of exceeding was divided into classes for purpose of the calculation. Substances that have been assessed in fewer than ten 1 1kmgrid cells were omitted. The pesticides giving problem in the Netherlands in the period 2003/2004 were in descending order: (1) imidacloprid, (2) fenamiphos, (3) pyrimiphos-methyl, (4) aldicarbsulfoxide, (5) dithianon, (6) pyridaben. Every two years a new list can be calculated based on the surveillance data of that period. When pesticides seriously exceed limits over the years, measures can be taken for e.g. the effectiveness of policies activities such as banning of certain pesticides, replacement of certain pesticides or restricted uses and application techniques. A specific example is the Dutch Plant Protection policy [18] which is aimed at reaching sustainable agriculture in the Netherlands. Operational goals with respect to the environment based on MTR exceedings are given in this policy document. Bottlenecks in the achievement of goals can be identified using the information of the Pesticides Atlas. Optimizing monitoring effort One of the aims of the Pesticides Atlas is to assist in building an optimal monitoring program. The costs of gathering pesticides concentrations in the various surface waters (think of sampling and analytical costs but also concerning data interpretation) are very high. Yet many Dutch Water Boards can improve their programs by investigating which pesticides are allowed to be used in their areas and in the stream areas that are connected to their districts. The necessity of this is illustrated by measurement data on metribuzin for the period 1999/2000 (Fig. 5, left side) next to land use data from potato fields (Fig. 5, right side). The pesticide metribuzin is used almost exclusively on potatoes. Obviously monitoring in the south-east part of the country can be discussed. In contrast, in the polders in the middle of the country, many potato fields can be found, however there are no measurements of this pesticide in this area (see the arrows in the Fig. 5) because Water Boards did not include metribuzin in their surveillance programs. In the Atlas, statistical correlations have been computed between surface water pesticide concentrations on the one hand and the areas devoted to types of land use (generally types of crop) on the other. Statistical relationships that are scientifically sound could be derived for 249 combinations of active ingredients and land use. By knowing this correlation, the expected pesticides to be used in an area can be calculated based on the presence of a certain land-use type and compared to the actual measured active ingredients (Fig. 6). It appears that at most locations in the Netherlands a low percentage of the expected pesticides is included in monitoring programs. Actually, nowhere is this exceeding 75% and the percentage of measured pesticides varies largely among regions. In the western part of the country, monitoring programs are adjusted to expected pesticides compared to other regions. In the eastern part of the country,

8 672 Vijver et al. Fig. 5. Example on measurements performed on metribuzin, a pesticide allowed to be used only in potatoes growing. On the left side: measurements on metribuzin, on the right side: potato cultivation in the Netherlands. monitoring programs cover only 3 25% of the expected pesticides. In general in the whole country, monitoring programs can be improved substantially by adjusting the list of included pesticides to the calculated expected pesticides based on local land use. Correlations used for environmental risk assessment purposes Another use of the correlations based on exceedings is to find out in which type of land use the largest amount & levels exceeding of environmental quality standards are calculated. For instance the most problematic pesticide of the period 2003/2004 was imidacloprid, for which levels have strong significant (P-value <0.001) relationship with greenhouse, floriculture and potatoes growing. For fenamiphos and pyridaben the contribution of land use categories to the exceedance of the MTR could not be calculated due to insufficient monitoring data. Levels of pirimiphos-methyl concentrations to the MTR could be correlated significantly to greenhouse and floriculture (P-value <0.001). MTR exceedings of aldicarbsulfoxide could be correlated to growing cabbage (P-value >0.01 and <0.05). The exceedings of dithianon could not be correlated to a specific land use category. From these examples it is shown that, the correlations may assist in prioritizing effective policy measures to diminish emission of problematic pesticides by actions such as restricted use of substances in certain crops or stimulation of technical developments in those agricultural businesses. The pros and cons of using measurements Measuring pesticides concentrations in surface waters is time-consuming and rather expensive. Another disadvantage is that not all pesticides that exist can be covered in the surveillance programs. Inherent to the use of surveillance data is the fact that detailed spatial and temporal information is lost. As a result the exact distance to the crop of the surface water sample can only be estimated. Also the application time compared to the time of sampling is unknown. To partially overcome these disadvantages and to make it possible to work with surveillance data, statistical techniques such as data aggregation and impute of missing data as described above are used in order to make general interpretations. Obviously not all disadvantages can be dealt with. Nevertheless, the advantages of the use of monitoring data are substantial since actual detected concentrations are being used. These result from all emissions, including these from illegal use, from pesticides used in nonagricultural practices and from pesticides from neighboring countries. Furthermore, surveillance data provide an actual view of a specific moment in time. These concentrations reflect the state of affairs in the dynamic agricultural business which can hardly be incorporated in models that

9 Pesticides concentration in surface water 673 Fig. 6. Measuring map of pesticides potentially present according to land use. predict pesticides concentrations. Agricultural business react differently each year, e.g. due to weather changes that have impact on crop diseases and subsequently the specific pesticides that need to be used to control these diseases, change and progression in the different application techniques and annually rotating crops. Therefore it is possible to monitor the actual field situation, in this case the pesticides concentrations in the surface waters. Since the Pesticides Atlas is easy to use and freely available, communication can be stimulated. This can result in the optimization of a cost-effective monitoring program. It is also giving risk assessors a powerful decision-making tool for prioritizing of problem areas, problem crops, or problematic active ingredients. Future perspective is that the

10 674 Vijver et al. Pesticides Atlas enables environmental policies to be compared across different areas and different water managers, so-called benchmarking. Conclusions The unique way of mapping pesticides concentrations as done in the Pesticides Atlas allows for transparent communication to a broad user group such as national and local water managers. With the actual tool available on the Internet, a wide range of actors such as policy makers, regulators, farmers, chemical industry, food industry and nongovernmental organizations can have a good visual impression of the geographic spread of pesticide concentrations in surface water in the Netherlands. The pesticides maps give relevant information about where and when a pesticide is measured, found and where it is exceeding an environmental standard. Based on this type of information not only monitoring systems of water authorities can be improved (measurements at relevant sites etc.), but the Pesticides Atlas might also be used for evaluating environmental policy over time as well. At the national scale, surface water quality in the Netherlands has been improved by approximately 30 38%. Hence, the surface water quality can differ over the regions which are obviously seen in the geographical representation of the results. Since it is shown that in many cases pesticides concentrations and exceedings of an environmental standard can be statistically linked with land use data, the maps can also be used in the registration process, and decision making for land use planning. Acknowledgment We thank Dennis Kalf (Waterdienst) and his colleagues for their assistance in the development of the Pesticides Atlas. We thank Roel Knoben and his colleagues (Royal Haskoning) for hosting and designing our Internet tool. We thank Ton van der Linden, National Institute for Public Health and the Environment (RIVM), for his comment on the manuscript and the three anonymous reviewers. References [1] Stoate, C.; Boatman, N.D.; Borralho, R.J.; Rio Carvalho, C.; De Snoo, G.R.; Eden, P. Ecological impacts of arable intensification in Europe. Journal of Environ Manag. 2001, 63, [2] De Snoo, G.R.; De Jong, F.M.W. Eds. Bestrijdingsmiddelen en milieu; Uitgeverij Jan van Arkel: Utrecht. 1999, (in Dutch). [3] De Snoo, G.R. Variation in agricultural practice and environmental care. In Pesticides: Problems, Improvements, Alternatives; Den Hond, F. Groenewegen., P. Van Straalen, N.M., Eds.; Sustainable Pest Manag. Series Blackwell Science: Oxford, UK, 2003, [4] Kempenaar, C.; Spijker, J.H. Weed control on hard surfaces in the Netherlands. Pest Manag Sci. 2004, 60, [5] Konstantinou, I.K.; Albanis, T.A. Worldwide occurrence and effects of antifouling paint booster biocides in the aquatic environment: a review. Env. Intern, 2004, [6] Van Dijk, H.F.G.; Van Pul, W.A.J.; De Voogt P., Eds. Fate of Pesticides in the Atmosphere. Implications for environmental risk assessment. Kluwer Academic Publishers: Amsterdam, Netherlands, 1999, [7] De Jong, F.M.W.; Van der Voet, E.; Canters, K.J. Possible sideeffects of airborne pesticides on fungi and vascular plants in the Netherlands. Ecotox. Environ. Saf. 1995, 30, [8] De Snoo, G.R.; De Wit, P.J. Buffer zones for reducing pesticide drift to ditches and risks to aquatic organisms. Ecotox. Environ. Saf. 1998, 41, [9] Snoo, G.R. De Wit, P.J. Pesticide drift from knapsack sprayers to ditches and ditch banks. Proc. Brighton Crop Protection Conferencec Weeds, Brighton, UK, November 22 25, 1993, 2, [10] De Snoo, G.R.; Wegener Sleeswijk, A. Use of pesticides along ditch banks and field margins. Med. Fac. Landbouww. Univ. Gent. 1993, 58, [11] European Union. Directive 2000/60/EC of the European parliament and of the council of October 23, 2000 establishing a framework for Community action in the field of water policy. [12] Van t Zelfde, M.; De Snoo, G.R. Atlas of pesticide concentrations in Dutch surface waters: A pilot study. Comm. Appl. Biol. Sci. Ghent University 2003, 68, [13] De Snoo, G.R.; Tamis, W.L.M.; Vijver, M.G.; Musters, C.J.M.; Van t Zelfde, M. Risk mapping of pesticides: the Dutch atlas of pesticide concentrations in surface waters; Comm. Appl. Biol. Sci. Ghent University. 2006, 71, [14] TRIM software. Netherlands Statistics, CBS [15] Landelijk grondgebruiksbestand Nederland (LGN-1). Wageningen Universiteit: Wageningen, the Netherlands. (In Dutch) [16] Bolsius, E.C.A. Een digital bestand voor de lanschapsecologie van Nederland: eindrapport van het LKN-project. Rijksplanologische dienst: The Hague, the Netherlands. (In Dutch) [17] Van der Linden, A.M.A.; Van Beelen, P.; Van den Berg, G.A.; De Boer, M.; Van der Gaag, D.J.; Groenwold, J.G.; Huijsmans, J.F.M.; Kalf, D.F.; De Kool, S.A.M.; Kruijne, R.; Merkelbach, R.C.M.; De Snoo, G.R.; Vijftigschild, R.A.N.; Vijver, M.G.; Van der Wal, A.J. Midterm evaluation of the plant protection policy of the Netherlands; Environment RIVM-report. [18] Institute of Environmental Science (CML), University Leiden. Available at Pesticide Atlas, download database version 12 March This Internet tool is financed by Dictoraat Generaal Water, Ministry of VenW in cooperation with the Ministries of LNV, VROM and UvW and CTgB.