LINKING LAND USE WITH PESTICIDES IN DUTCH SUR- FACE WATERS

Comm. Appl. Biol. Sci, Ghent University, 76/2, 2012 1 LINKING LAND USE WITH PESTICIDES IN DUTCH SUR- FACE WATERS M. van t Zelfde, W.L.M. Tamis, M.G. Vijver & G.R. de Snoo Department of Conservation Biology, Institute of Environmental Sciences, Leiden University (CML), P.O. Box 9518, NL-2300 RA Leiden, The Netherlands Corresponding author, e-mail: zelfde@cml.leidenuniv.nl ABSTRACT Compared with other European countries the Netherlands has a relatively high level of pesticide consumption, particularly in agriculture. Many of the compounds concerned end up in surface waters. Surface water quality is routinely monitored and numerous pesticides are found to be present in high concentrations, with various standards being regularly exceeded. Many standards-breaching pesticides exhibit regional patterns that can be traced back to land use. These patterns have been statistically analysed by correlating surface area per land use category with standards exceedance per pesticide, thereby identifying numerous significant correlations with respect to breaches of both the ecotoxicological standard (Maximum Tolerable Risk, MTR) and the drinking water standard. In the case of the MTR, greenhouse horticulture, floriculture and bulb-growing have the highest number as well as percentage of standard-breaching pesticides, despite these market segments being relatively small in terms of area cropped. Cereals, onions, vegetables, perennial border plants and pulses are also associated with many pesticides that exceed the drinking water standard. When a correction is made for cropped acreage, cereals and potatoes also prove to be a major contributor to monitoring sites where the MTR standard is exceeded. Over the period 1998-2006 the land-use categories with the most and highest percentage of standards-exceeding pesticides (greenhouse horticulture, bulb-growing and flower cultivation) showed an increase in the percentage of standards-exceeding compounds. Keywords: pesticides, surface water, Netherland, standards exceedance, land use INTRODUCTION Use and monitoring of pesticides in the Netherlands In the Netherlands around 800,000 hectares are devoted to agriculture and horticulture, with a further approx. 1,000,000 ha as pasture (CBS 2010). Cultivation of many of the crops concerned is extremely intensive, and exports account for some 2% of the country s Gross National Product (GNP). Maintaining this kind of intensive cropping requires substantial use of fertilizers and pesticides, and compared with most other European countries pesticide use in the Netherlands (and also Belgium and Italy) is relatively high in various sectors (OECD 2009). Over 760 pesticides have been approved for use in the Netherlands, containing around 240 active ingredients (Ctgb, 2011). An impression of Dutch pesticide use can be gained from both sales data (Nefyto) and usage data (Netherlands Statistics; farmer surveys). The former show that total pesticide sales have declined by 56% since 1985 (as per 2010). Sales

2 were lowest around 2001/2002, but with an increase in subsequent years. In 2010 pesticide sales totalled 9.3 million kg active ingredient. As part of water quality monitoring efforts, regional water boards and the Directorate-General for Public Works & Water Management (Rijkswaterstaat) have been monitoring surface water pesticide levels in the Netherlands for many years. These monitoring data are collected at the national level and processed in the Pesticides Atlas (www.pesticidesatlas.nl), with measured concentrations being assessed against relevant quality standards. Monitoring data are available for over 700 Dutch monitoring sites for the years 1997-2010 (Van t Zelfde et al., 2010) These monitoring data show that quality standards have been breached throughout the entire period. Figure 1 shows the percentage of measurements exceeding the Maximum Tolerable Risk (MTR) standard during the period 1997 2010. As can be seen, the percentage of standard-breaching measurements has declined substantially from around 4% in 1997, but the figure has remained more or less unchanged since 2001, at around 1%. 6 4 Netherlands Rhine Meuse Eems Scheldt 2 0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Figure 1: Percentage of measurements exceeding the MTR for the Netherlands as a whole and for the four major river basins in the period 1997-2010. Possible causes of pesticides in surface waters There are two main potential routes by which pesticides can find their way into Dutch surface waters. There may be brought in from neighbouring countries via rivers and other watercourses as well as via atmospheric deposition (Van t Zelfde et al. 2011) and of course pesticide use in the Netherlands itself may also lead to surface water contamination. Although this may be traceable to applications in a range of sectors, pesticide use is highest in agriculture and this article is concerned solely with the contribution from that source.

Comm. Appl. Biol. Sci, Ghent University, 76/2, 2012 3 Surface water monitoring results show that standards exceedances for the various pesticides often exhibit a regional pattern. These regional patterns may be due to various factors. One possible cause is use of a particular pesticide on a particular crop and the regional distribution of that crop. Figure 2 provides an illustrative example, for potatoes and the active ingredient metribuzine, which is used almost exclusively with this crop. Figure 2: Monitoring sites with exceedance of MTR standard for metribuzine in 2004 (left) and area per km 2 devoted to potatoes in 2004 (right). As can be seen, metribuzine levels exceeded the MTR at numerous monitoring sites in the province of Zeeland (south-west of the Netherlands), but not in other potato-cropping regions. It can also be seen that in certain potato-cropping regions like the provinces of Noord-Holland and Brabant (north-west and south of the Netherlands. respectively) this specifically relevant compound is not monitored at all. Research has been carried out on the relationship between land use and pesticides in surface waters, in terms of both concentrations and standards exceedances (De Graaf & De Snoo, 2003; Tamis et al., 2004). For the purpose of this article we focus on two standards: the ecotoxicological standard (MTR) and the drinking water standard (DWS) and consider two specific issues: - What land use categories have the most (monitored) compounds exhibiting a significant correlation with standards exceedance (MTR and DWS)? - What temporal trends can be identified (MTR)?

4 As a high number of standards-exceeding compounds in a given land use category does not necessarily mean a high number of standards-exceeding monitoring sites or measurements, the following additional question was investigated: - What is the contribution of each land use category to the number of monitoring sites at which standards are exceeded? METHODS To analyse the relationship between land use and the occurrence of pesticides in surface waters the whole of the Netherlands was mapped on a square-kilometre grid. The monitoring data were combined into two-year periods, running from 1997-1998 to 2005-2006. A total of 21 land use categories were distinguished, for which data were aggregated at the km 2 scale level. In 2000 and 2004 a survey of agricultural pesticide use was conducted by Netherlands Statistics (CBS) and this information was used to define the correlations that needed to be calculated between land use and individual pesticides. Correlation between standards exceedances and land use was assessed by means of a one-sided Mann-Whitney U-test, thus to determine, for each period, whether the area devoted to a particular category of land use is higher when there are standards exceedances for a particular pesticide. For each compound a table can then be prepared showing the main land use categories exhibiting a significant relationship with standards exceedances for the compound concerned (Table 1). Table 1: Example of a table of the land use categories in which metribuzine was used in the period 2005-2006, ordered according to the degree of significance of the correlation with MTR exceedance. Very strong: P<0.001, present: 0.01<P<0.05, - = not significant (P>0.05). Land use category Significance Potatoes Very strong Strawberries Present Grassland - Asparagus - For the present article, for each period the following calculations were performed: Conversion of significant correlations between land use categories per standardsexceeding compound and number and percentage of standards-exceeding compounds per land use category. To this end, first the number of compounds with a significant correlation between standards exceedances and land use (category) was counted. Next, this number was compared with the total number of compounds used in the land use category in question, yielding the percentage of standards-exceeding compounds per land use category. Below, the results of this analysis are presented for each period for the various land use categories, in decreasing order of number or percentage of compounds exhibiting a significant correlation with standards exceedance. For the MTR this was done for all periods from 1997-1998 through to 2005-2006, for the DWS for 2005-2006 only.

Comm. Appl. Biol. Sci, Ghent University, 76/2, 2012 5 Conversion of the number of standards-exceeding compounds per land use category to the number of standards-exceeding monitoring sites per land use category. To this end a calculation was performed to allocate the number of monitoring sites at which the level of a compound exceeded the standards to those land use categories exhibiting a significant correlation, using a simple weighting procedure according to the total area of each land use category. This was done only for the last period (2005-2006). RESULTS Number and percentage of standards-exceeding compounds for MTR and DWS Table 2 shows the number of compounds showing a significant correlation with land use category in 2005-2006, together with the percentage of compounds used in each category that exceeded the standard(s). The results show that it is mainly greenhouse crops, floriculture and bulb-growing where high percentages of the pesticides used exceed water quality standards, with around 10% exceeding the MTR and 20-30% the drinking water standard in the period 2005-2006. With cereals, onions, vegetables, perennial border plants and pulses there are many pesticides exceeding the DWS, but the MTR is exceeded for only one compound or for none. As can be seen from the last column, showing the area (km 2 ) devoted to each land use category, the three categories with the highest number of standards-exceeding pesticides are associated with relatively small areas. Table 2. Number (n) of pesticides exceeding the MTR or DWS per land use category and percentage (%) of total number of compounds used in the category (period 2005-2006) compared with total area devoted to the category (area). The results are presented in decreasing order of number of compounds exceeding the MTR, then ditto for the DWS. Land use category MTR exceedance DWS exceedance Area n % n % km2 Greenhouse crops 12 10.6 21 18.5 77 Floriculture 12 9.5 32 25.4 48 Flow bulbs 10 13.0 26 33.8 200 Arboriculture 5 5.4 12 13.0 98 Strawberries 3 6.3 5 10.4 21 Potatoes 3 5.1 12 20.3 1637 Cabbage crops 2 5.9 7 20.6 84 Fruit-growing 2 3.5 6 10.3 167 Corn 1 7.7 2 15.4 2244 Asparagus 1 3.7 6 22.2 24 Sugarbeet 1 3.1 8 25.0 253 Grass seed 1 3.1 7 21.9 974 Cereals 1 1.7 11 19.0 1826 Onions 11 26.2 255 Vegetable crops 9 19.6 120 Perennial border plants 6 46.2 45 Pulses 6 22.2 119 Leaf vegetables 3 18.8 50 Leeks 2 6.5 30 Paved surfaces 1 14.3 6740 Grassland 1 3.5 9834

6 Trends in number of MTR-exceeding compounds per land use category Figure 3 shows trends in the percentages of compounds exceeding the MTR between 1997 and 2006 for selected land use categories. Throughout this period, greenhouse crops and floriculture consistently rank highest as the categories with the highest number and percentage of standard-exceeding compounds. Indeed, in these crops the percentage of compounds exceeding the MTR is in fact rising, with a surprisingly sharp rise in the case of flower bulbs. in the case of onions, leaf vegetables and leeks, on the other hand, the number of compounds exceeding the MTR declined over the same period. 14.0 Percentage compounds with correlation landuse type 12.0 10.0 8.0 6.0 4.0 2.0 0.0 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Period Greenhouse crops Floriculture Flower bulbs Onions Leaf vegetables Leeks Figure 3: Trends in percentage of compounds exceeding the MTR. Number of standards-exceeding (MTR) monitoring sites per land use category Table 3 shows, for the period 2005-2006, the number of standards-exceeding monitoring sites per land use category together with the total area of each category. As can be seen, the (relatively small-scale) flower-bulb, greenhouse-crop and floriculture segments still feature prominently, accounting for around 50% of the MTR exceeding monitoring sites. Now, though, the land use categories potatoes and cereals, both of which are far larger-scale segments, are also associated with a substantial number of standards-exceeding monitoring sites (31.5% and 4.5%, respectively).

Comm. Appl. Biol. Sci, Ghent University, 76/2, 2012 7 Table 3. Number (n) and percentage (%) of monitoring sites where the MTR standard was exceeded in the period 2005-2006 per land use category and total area per land use category. Land use category Monitoring sites Area n % in km 2 Potatoes 186.7 31.5 1637 Flower bulbs 125.4 21.1 200 Greenhouse crops 108.5 18.3 77 Floriculture 65.8 11.1 48 Cereals 26.5 4.5 1823 Arboriculture 25.0 4.2 98 Fruit-growing 19.8 3.3 167 Corn 16.0 2.7 2244 Cabbage crops 7.3 1.2 84 Sugarbeet 4.0 0.7 974 Grass seed 3.7 0.6 253 Strawberries 2.7 0.5 21 Asparagus 1.7 0.3 24 CONCLUSIONS, DISCUSSION AND RECOMMENDATIONS This study has identified significant correlations between exceedance of pesticide standards and land use. This holds for exceedance of both the Maximum Tolerable Risk level and the drinking water standard. In the case of the MTR, greenhouse crops, floriculture and flower bulbs are associated with the highest number and percentage of pesticides in breach of standards, despite these being relatively small-scale segments. Cereals, onions, vegetable crops, perennial border plants and pulses are associated with numerous pesticides exceeding the drinking water standard. When a correction is made for the acreage devoted to the crops in question, cereals and potatoes also prove to account for a considerable share of monitoring sites where the MTR is exceeded. This is down to the fact that these crops are grown on a substantial scale in the Netherlands. For certain land use categories, including onions, leaf vegetables and leeks, the number and percentage of MTR-exceeding compounds is found to have declined over the years. The land use categories with the highest number and percentage of standards-exceeding compounds (greenhouse crops, floriculture and flower bulbs) show an increase in the percentage of standards-exceeding compounds over the period investigated. Correlation between a particular land use category and occurrence of a particular pesticide in surface waters does not constitute direct proof that the land use category in question is indeed the cause of standards exceedances, but merely provides an indication (De Graaf & De Snoo, 2003; Tamis et al., 2004). A major methodological problem is the existence of quasi-correlations in which the first factor is not causally related to the second factor, but with a third (etc.) factor that is also related (possibly causally) to the second factor. Potatoes and beetroot are almost always grown in the same areas at the same time or in succession, for example, and if a particular pesticide is used on both crops, it cannot be unequivocally concluded which of the crops is responsible for the compound s presence in surface waters. In this case, the third factor is the soil type suitable for both crops. Our analysis of the possible correlation between pesticide standards exceedance and land use categories was carried out at the scale level of square-kilometre grid cells and for amalgamated two-year periods. In the latest version of the Pesticides

8 Atlas monitoring data are now available at the level of individual monitoring sites and per individual monitoring year, which means future analyses can be performed with higher resolution in both space and time. The monitoring site level also makes it possible to use alternative types of geographical mapping that tie in better with local-level water management, in particular mapping based on pre-defined catchment areas (afwateringseenheden, GAF90). Since a pesticide originating from an agricultural plot will move downstream within the catchment area, an analyses could be made to identify potential correlations between standards-exceeding monitoring data and land use categories further upstream. Another possible methodological improvement is to factor in partial correlations (quasi-correlations) by means of multiple regression. ACKNOWLEDGEMENTS We are grateful to Nigel Harle of Gronsveld for his translation of the Dutch manuscript. REFERENCES CBS (2010). Netherlands Statistics, Statline, January 2012. Ctgb. College voor de Toelating van Gewasbeschermingsmiddelen en Biociden (2011). Jaarverslag 2010. Wageningen. Compendium voor de Leefomgeving (2011). www.compendiumvoordeleefomgeving.nl Graaf, H.J. de & Snoo G.R. de (2003). Technische beschrijving van de Atlas "bestrijdingsmiddelen in het Nederlandse oppervlaktewater" en een verkenning koppeling van meetgegevens aan landgebruik. CML report 36, Leiden: Institute of Environmental Sciences (CML). OECD (2009). OECD Environmental Data Compendium. OECD, Paris. Snoo G.R de & Vijver M.G. (2012). Bestrijdingsmiddelen en waterkwaliteit. Leiden: Institute of Environmental Sciences (CML). Tamis, W.L.M., Zelfde M. van 't & Hoefsloot P. (2004). Technische rapportage van het project Bestrijdingsmiddelen in het oppervlaktewater en koppeling van meetgegevens aan grondgebruik MEBOL. CML report 38, Leiden: Institute of Environmental Sciences (CML). Zelfde M. van 't, Tamis W.L.M. & Vijver M.G. (2010). Technische rapportage van de update van de bestrijdingsmiddelenatlas met gegevens van het jaar 2007. Leiden: Institute of Environmental Sciences (CML). Zelfde M. van 't, Musters C.J.M., Tamis W.L.M. & Vijver M.G. (2010). Technische rapportage van project: Bestrijdingsmiddelenatlas Kader Richtlijn Water (KRW) proof. Leiden: Institute of Environmental Sciences (CML). Zelfde M., van 't & Vijver M.G. (2008). Technische rapportage van de update van de bestrijdingsmiddelenatlas met gegevens van de jaren 2005/2006. Leiden: Institute of Environmental Sciences (CML). Zelfde, M. van t, Tamis, W.L.M., Vijver, M.G. & Snoo, G.R. de (2011) The contribution of neighbouring countries to pesticides levels in Dutch surface waters. Proceedings of the 63rd International Symposium on Crop Protections Vol. III. Communications in Agricultural and Applied Biological Sciences (pp. 867-878). Gent.