Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds

G Model

ARTICLE IN PRESS

ECOENG-4007; No. of Pages 11

Ecological Engineering xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Ecological Engineering journal homepage: www.elsevier.com/locate/ecoleng

Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds Julien Tournebize a , Cedric Chaumont a , Ülo Mander a,b,∗ a b

Hydrosystems and Bioprocesses Research Unit, Irstea, Antony, France Institute of Ecology and Earth Sciences, University of Tartu, Tartu, Estonia

a r t i c l e

i n f o

Article history: Received 20 August 2015 Received in revised form 21 January 2016 Accepted 13 February 2016 Available online xxx Keywords: Artificial wetland Buffer zone Catchment Design recommendations Non-point source pollution Removal efficiency Tile drainage

a b s t r a c t In the context of subsurface drainage, the mitigation of agricultural pollutants means intercepting water flows using green infrastructures such as constructed wetlands (CWs). First, based on a scientific review, this paper analyses how efficiently CWs can remove nitrate and pesticides from the runoff in drained agricultural watersheds. Average efficiency ranges from 20 to 90% and from 40 to 90% for pesticides and nitrate respectively. The main processes involved are based on microbiological activities, for which hydraulic residence time is a key factor. In order to successful implementation of such a wetland system, hydrological diagnosis of water flow and pollutant transfer at different watershed scales should be provided. Typically, the transport and transformation of pollutants shows clear seasonality depending on the application of nitrate (throughout the whole year) and pesticides (only after application periods). We suggest two interception strategies based on field experiments. The “on-stream” strategy means the establishment of free water surface (FWS) CWs directly on streams/ditches and the interception of all drainage flows: a solution suitable for nitrate removal. The “off-stream” strategy requires the establishment of CWs outside of streams/ditches, whereas interception targets only the most polluted water flow, for instance during the period after pesticide application, requiring farmer’s involvement. Suggestions are also made for FWS CW design (a geotechnical survey, topography constraints, etc.) respecting ecological engineering concepts. A following size range is proposed: 76 m3 per drained hectare, equivalent to 1% of the upstream area, given a maximum water depth of 0.8 m. Nevertheless, CWs must be considered as a complementary tool dedicated to transfer reduction, and as part of broader actions aimed at reducing pollutant loading at the plot scale. © 2016 Elsevier B.V. All rights reserved.

1. Introduction To respond to the requirements of the European Water Framework Directive (2000/60/CE) in terms of water pollution by agricultural non-point source pollution, actions can be implemented at different scales (European Union, 2000). Chemical inputs including fertilizers (nitrate; see Tanner and Kadlec, 2013) and pesticides (see Stehle et al., 2011; Vymazal and Bˇrezinová, 2015) are a source of this non-point source pollution. Reducing their use is the first essential stage in limiting the quantities of pesticides or nitrate reaching aquatic environments. For example, The French EcoPhyto Plan (Ministry of Agriculture, 2008) includes a 50% reduction in the quantities of pesticides applied for the next 10 years compared to 2008. However, given that fertilizers and pesticides continue to be

∗ Corresponding author at: University of Tartu, Institute of Ecology and Earth Sciences, Tartu, Estonia. Tel.: +372 7375816; fax: +372 7375825. E-mail address: [email protected] (Ü. Mander).

applied, a portion will always be transferred toward aquatic environments. Complementary actions to the reduction of applications may also be necessary, including setting up buffer zones between the output of agricultural fields and the receptor environments to partially reduce pollutant flows (Correll, 2005; Reichenberger et al., 2007; Gregoire et al., 2009). Among these buffer zones, grass strips have been widely studied (CORPEN GZT, 2007). These impose an untreated zone between the field and the surface waters, thus reducing the quantities of pollutants reaching these waters (runoff, spray drift). Given their high infiltration capacity, grass strips can reduce pollutants from agricultural runoff (Lacas et al., 2005). However, their efficiency drops drastically when the soil gets saturated (limited infiltration) or when the flow to be treated is channelled and no longer diffuse through the grass strip (Souiller et al., 2002). Another case that limits the efficiency of grass strips is the management of waters coming from tile drainage. The hydrological functioning of tile drainage makes it possible to channel flows coming from the entire surface of an agricultural plot or a watershed to a single, easily identified point. Therefore, artificial wetlands (AWs)

http://dx.doi.org/10.1016/j.ecoleng.2016.02.014 0925-8574/© 2016 Elsevier B.V. All rights reserved.

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

2

designed to improve the quality of water drained from agricultural land can be efficient and are easy to implement. Constructed or artificial wetlands are man-made wetlands designed to mimic the biofiltration action of natural wetland systems (Forbes et al., 2004; Vymazal, 2007). Thus in the application of ecological engineering concepts, the goal is to improve the natural function of natural wetlands in order to treat water pollution issues. The term “artificial wetlands” has no legal existence, and nor do the grass strips along water edges. However, it is its function that provides it with its “buffer” or retention role within the watershed. According to the classification proposed by Fonder and Headley (2010), different types of constructed (artificial) wetlands can be distinguished depending on their hydraulic function. They vary from subsurface flow systems (if the water course crosses a porous filter) to free water surface (FWS) systems, which can have shallow and deep sections and range from marshes (intermittent runoff) to lagoons (permanent runoff). For easier understanding and coherence with commonly used terminology, we will hereinafter use the term constructed wetlands (CWs). Since the FWS CWs are dominatingly used for treatment polluted runoff from agricultural watersheds, the further text considers this type of CWs. The main objectives of the study are to: (1) analyse and synthesize results from literature and our earlier studies on the losses of nitrate and pesticides as well as their removal efficiency in FWS CWs purifying runoff from drained agricultural watersheds; (2) suggest polluted flow interception strategies and design parameters for FWS CWs based on field experiments. 2. Material and methods 2.1. Literature review The first part of this paper details the pollution dissipation processes and the efficiency FWS CWs’ treating polluted runoff from drained agricultural watersheds which are (1) reported in the international peer-reviewed literature, and (2) gathered in the previous field studies carried out by the Institut national de recherche en sciences et technologies pour l’environnement et l’agriculture (IRSTEA), France. To find the literature sources via the Thomson-Reuters ISI Web of Science, the combination of keywords “artificial wetland(s)”, constructed wetland(s)”, “nitrate(s)”, “pesticide(s)”, and “tile-drain” or “drainage” have been used. In addition, the analysis is based on studies published by our research group and some other regional reports. The IRSTEA-based field studies have been carried since 2006 on various sizes of plots and watersheds in France (Tournebize et al., 2008, 2015a; Passeport et al., 2011). The specificity of drainage in terms of seasonality and transfer modality is presented in the second part of the paper. Finally, the part three introduces design and implementation aspects and shares experience from literature as well as IRSTEA-based field experiments. 2.2. Experimental sites description Three experimental fields differing in scale were selected: a plot (46 ha, Indre et Loire, described in Passeport et al., 2013), a subbasin (355 ha, Seine et Marne, described in Tournebize et al., 2012), and a watershed (4000 ha, Seine-et-Marne, described in Blanchoud et al., 2013). These sites represent similar drainage conditions: average rainfall about 750 mm, hydromorphic soil (Gleyic Luvisol), crop rotation (mainly winter wheat, rape and barley), high proportion (>80%) of subsurface drainage system (perforated buried PVC pipe every 10 m space at 80 cm in deep due to more clayed layer below). For all three scales, water quality monitoring strategies were similar based on weekly flow weight sampling. All water samples were analyzed by the same laboratories; at IRSTEA for nitrate,

Fig. 1. Average monthly drainage flow (mm) according to the annual rainfall of 693 mm Periods: 1—initiation of drainage; 2—intense drainage season; 3—sporadic spring event; 4—no flow. Dashed arrows show the period of pesticide application for winter and spring cereal crops. Based on GIS ORACLE data from 1998 to 2012 (Tallec et al., 2015).

and at CARSO (subcontractor) for pesticides, screening about 100 molecules, with average quantification limits of 0.01 ␮/L). 3. Dynamics of water and pollutant transfer in tile-drained agricultural watersheds 3.1. Hydrology Knowledge of water pathways and flow at the watershed scale is needed in order to design and establish ecological engineering rules to optimize the purification function of CWs. Choosing to implement a CW requires prior analysis of the quantitative and qualitative dynamics of the waters drained at the output of the watershed. The drainage runoff depends on the rainfall regime, and thus presents inter- and intra-annual variability. The interannual variability is explained by the alternating wet, dry and intermediate years, depending on the runoff precipitated each year. At the intra-annual scale, the behaviour of the drainage discharge at the watershed’s outlet depends on events, presenting a chain of events between the peak discharges followed by a recession phase depending on the precipitation and the state of the soil’s water holding capacity (Tournebize et al., 2008). Over the entire hydrologic year, the drained watersheds in northwestern Europe are typically characterized by three different phases (Fig. 1) regardless of the climatic pattern of year (Tiemeyer et al., 2006; Borin and Tocchetto, 2007; Brown and van Beinum, 2009; Passeport et al., 2010). The first step, called the “initial phase of drainage,” generally appears at the beginning of winter; during this shallow/superficial groundwater recharge phase, rainfall mostly infiltrates, and very little rainfall is returned to the environment via drains. The second phase, called the “intense drainage season,” generally during winter, is characterized by a very high restitution of rainfall. Since the soil is close to hydric saturation, any new water contributed as rainfall is restituted as outlet flow of the drain. Finally, the last phase (from spring to the beginning of fall) corresponds to the recession of superficial groundwater, with the soil becoming decreasingly saturated because of vegetation regrowth and the increase in evapotranspiration demand. For instance, in northern France, the annual mean drained runoff is approximately 180 mm (standard deviation 100 mm) which means very high volumes of water, depending on the size of the basin (Tournebize et al., 2004).

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

3

Fig. 3. The nitrogen cycle in wetlands. 1—ammonification, 2—volatilization, 3—nitrification, 4—dissimilative nitrate reduction to ammonia (DNRA), 5—anaerobic ammonia oxidation (ANAMMOX), 6—denitrification, 7—assimilation, 8—N2 fixation.

Fig. 2. The scale effect of nitrate and pesticide concentrations in the outflow of three IRSTEA experimental with equivalent land use pattern (>80% in agricultural use; Seine-et-Marne department, France). (a) For nitrate, (b) for 26 different pesticides (Tournebize et al., 2008; Tournebize et al., 2015b). Average (asterix), median (bold line), 25 and 75% quantile (box) and min/max values (whiskers) are shown.

3.2. Nitrogen and pesticide runoff In Europe, about ninety percent of the nitrogen losses drained from agricultural plots are generated by nitrate (Billy et al., 2013; Tournebize et al., 2015a,b). The similar high percentages have been found for other agricultural regions (Randall et al., 1997; Nolan et al., 2012; Tanner and Kadlec, 2013; Groh et al., 2015). Nitrate ions are highly soluble and are therefore found preferentially in all agricultural waters. Nitrate seasonality is, however, marked by the position of increased nitrate stock in the soil profile at the moment of the rainfall event. Considering the outlet concentrations of the nitrate from the agricultural drainage network, Billy et al. (2010, 2013) and Tournebize et al. (2008) show three nitrate transport modes. In autumn, for low-level drainage runoffs, the NO3 − –N concentrations measured are very high (>100 mg/L). This is a result of the leaching of nitrate from the soil surface to the drain in early winter. During the intense drainage season, despite dilution during high-water episodes, the mean concentration oscillates around 13 mg NO3 − –N/L. This period strongly contributes to annual exports (>60%), due to the substantial drainage runoff. Finally, occasional peaks in concentration can be observed during the spring, consecutive to agricultural practices (fertilization). These peaks also reflect the typical leaching process of nitrate from the soil surface to the drain. The interannual nitrate fluxes, whether or not from the drained agricultural plots, are correlated with two factors: hydrology (the cumulation of winter water runoff) and the remnant nitrogen storage in soil at the beginning of winter (related to agricultural practices). Based on three experimental fields, the same concentration ranges (3–20 mg NO3 -N/L; Fig. 2) are found at all spatial scales from the outlet of agricultural plots to the sub-catchment and the watershed. Given their nitrogen application levels, all of the agricultural plots of a watershed contribute to nitrate leaching.

Pesticide exports via agricultural drainage are often less than 0.5% of the dose applied and rarely exceed 3% (Kladivko et al., 2001; Boithias et al., 2014). The quantities exported are on the order of several grams of pesticides (all pesticides combined) per hectare and per year, i.e. three orders of magnitude less than the quantities of nitrate exported. Generally, the first high-flow events after application are the most concentrated in pesticides (Kladivko et al., 2001; Branger et al., 2009), and therefore they present the highest risk of pesticide transfer. Occasionally, when flows resume after periods of low or no discharge, high pesticide concentrations can be measured, even after long periods since the last applications. Then the general behaviour of pesticide transfer is restricted to the period after applications, meaning that certain flows present no risk of transfer (except for remnant persistent pesticides such as Atrazine). Exports depend on the position of the pesticides in the drained soil. The pesticides located above the drain may be washed out more rapidly than those located at the inter-drain soil (Branger et al., 2009). In contrast to nitrate, pesticide concentrations depend clearly on the spatial scale. In the upstream part of the watershed the concentrations are significantly higher (>1 ␮g/L) than in the lower course area (Fig. 2). The diversity of agricultural practices by different farmers is the reason for this effect of large-scale dilution. Dissipating in the watershed, the concentrations decrease, even if the flows remain conservative. The hydrological diagnosis and prior knowledge of transfer modes and seasonality (herein we have used the example of agricultural drainage) are essential prerequisites to any analysis of the preservation of water resources. Therefore, we show that in tiledrainage context it means that: (1) concerning non-point source nitrate pollution, all flows export nitrate, at similar concentration ranges regardless of the spatial scale of the intervention; (2) concerning non-point source pesticide pollution, for the majority of chemicals, only the flows after application present high risks of transfer, with the highest concentrations at the direct outlet of agricultural plots, upstream in the watershed. 4. Removal of agricultural pollutants using constructed wetlands 4.1. Nitrate 4.1.1. Dissipation processes The nitrogen cycle of CWs (Fig. 3) is well known (Mitsch and Gosselink, 2007), and a large number of studies have investigated

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11 4

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

the efficiency of CWs in treating wastewater, industrial residual waters, and agricultural waters (Tanner et al., 2005; Vymazal, 2005). The engineering of CWs is increasingly advanced while attempting to fulfil the conditions underlying ecological engineering principles. Input flows can be forced through a variety of filtering media, whereas the outlets can be located at an elevation that favours the installation of different oxygenation conditions (Kadlec and Wallace, 2008). One of nitrate’s main elimination pathways in CWs is denitrification (Reddy and Patrick, 1984; Tanner et al., 2005). This microbial anaerobic process depends on the ratio of available carbon (C/N ratio). It transforms nitrate into two successive gases: N2O (a powerful greenhouse gas) and N2. If oxygen is available during the last stage of denitrification, N2O will be the only gas produced, replacing N2. In CWs, hydraulic management (filling, emptying etc.) can be the source of oxygen introduction. The production of N2O depends on the hydrologic conditions and management of CWs (McPhillips and Walter, 2015). Vymazal et al. (2006) and Mander et al. (2014a,b) compiled the denitrification rates observed in various studies. The N2 O emission values varied from 0.003 to 1.02 g N m2 /year, being less than 2% of the total nitrogen entering CWs. Taking the maximum value measured, a CW emits on the order of 10 kg N2O-N/ha/year, which seems relatively low compared to the nitrogen emitted from agricultural fields. However, considering the very high radiative power of this greenhouse gas (298 times more than an equivalent amount of CO2; IPCC, 2007), these values are high and might hinder the success of the CW’s implementation. Further investigations to obtain a better knowledge of N2O emissions processes and the N2/N2O ratio are necessary to limit this negative factor of CWs. CWs are often represented as systems in which the superficial layer of sediments (approximately 0–5 mm) is anoxic or even anaerobic, whereas the conditions are aerobic in the upper part of the water column. Over-oxygenation is observed in the water column the closer one comes to the surface. Moreover, around plant roots (the rhizosphere), a thin aerobic layer is created following transfer of oxygen by plants during photosynthesis by aerenchyma processes (Mitsch and Gosselink, 2007). Denitrification therefore preferentially takes place in the sediment–water contact layer. The second frequently mentioned pathway for nitrate ions is plant uptake. This natural phenomenon occurs during the vegetation period, preferably in spring and the beginning of summer. Even if the biomass produced is sometimes high, depending on the type of macrophyte, the proportion of nitrate taken up remains low compared to overall annual nitrogen fluxes. Pulou (2011) showed that the removal of nitrogen (7%) by vegetation is secondary compared to denitrification (93%), which occurs throughout the year. 4.1.2. Nitrate removal efficiency of CWs A wide range of efficiency has been reported by different articles in the literature concerning the evaluation of different CWs treating agricultural drainage runoff (Kovacic et al., 2000; Miller et al., 2002; Koskiaho et al., 2003; Crumpton and Helmers, 2004; Tanner et al., 2005; Appelboom and Fouss, 2006; Borin and Tocchetto, 2007; Beutel et al., 2009; Díaz et al., 2012, Tanner and Kadlec, 2013; Groh et al., 2015; Tournebize et al., 2015a,b) (Fig. 4). This type of black-box evaluation is carried out by calculating the reductions in nitrate concentrations or flows by comparing inputs and outputs of CWs. For example, for 11 natural wetlands and CWs treating urban or industrial waste waters, Kadlec (1994) found nitrate reduction varying from 38 to 96%, whereas input concentrations varied from 0.1 to 18 mg NO3− –N/L. Hammer and Knight (1994) calculated a mean efficiency of 44% for 17 CWs. A long term monitoring of a CW

Fig. 4. Removal efficiency of pesticides and nitrate from subsurface drainage flow, based on comparison of input/output mass removal. Explanation of Koc categories is presented in text. N—number of analyses. Box-whiskers plot symbols see in Fig. 2.

system in France showed the average removal efficiency was 50%, and the variation (40–70%) was depending on hydrological years (Tournebize et al., 2015a,b). For the drainage context (N = 34), the average removal efficiency reaches 42%. Natural wetlands generally present better efficiency (> 65%; Hammer and Knight, 1994). Several reasons can explain this difference in efficiency. Very often, natural wetlands are more stable than CWs and therefore stabilized in terms of the microorganisms’ functioning. In FWS CWs, the reducing conditions that are favourable to denitrification actions need time to become established, which could also explain the low efficiency of CWs compared to natural wetlands (see Craft, 1997; Fisher and Acreman, 2004; Batson et al., 2012). A key parameter to be optimized in order to mimic natural processes, demonstrated in a meta-analysis reported by Fisher and Acreman (2004), is the residence time related to the hydraulics of the wetland. Generally, hydraulic management is greatly intensified for CWs (the input flow is very frequent and sizeable) compared to the limited connection with the hydrographic network of natural wetlands (the input flow is reduced over time and characterized by peaks). Diverse vegetation in wetlands, encourages nitrate removal contributing organic carbon needed by denitrifying microorganisms (Reddy and Patrick, 1984; Craft, 1997; Yepsen et al., 2014). Denitrification needs time and too high hydraulic loading does not enable to remove nitrate and complete the process, thus, hydraulic retention time has direct impact on the nitrate reduction rate (Kadlec and Wallace, 2008; Bastviken et al., 2009). For instance, Spieles and Mitsch (2000) measured lower efficiency rates following heavy rainfall events, resulting in shorter hydraulic retention times in CWs. Several other factors are known for their effect on the fate of nitrate in CWs. The elimination of nitrate increases with temperature (Kadlec, 2005) because the denitrifying activity of microorganisms is greatly reduced by low temperatures. Therefore, seasonal variability is generally noted with higher efficiency values during the warmest and wettest seasons (Spieles and Mitsch, 2000). The nature of available carbon from vegetation is also an important factor controlling nitrate reduction rate. For instance, as shown by Pulou et al., 2012, very labile carbon from watercress (Nasturtium officinale) supports denitrification better than reed (Phragmites australis). The latest investigations also show that the ANAMMOX process (Fig. 3) can be significant in CWs (Ligi et al., 2014). These need more detailed investigations, however, and will open new opportunities for nitrogen removal in CWs.

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

Fig. 5. Processes involved in pesticide retention in a shallow surface flow constructed wetland. The red numbers indicate results from the IRSTEA experiments on S-Metolachlor (Hoyos-Hernandez, 2010) and Expoxiconazole (Passeport et al., 2011).

4.2. Pesticides 4.2.1. Dissipation processes The assessment of the purification potential of CWs to dissipate pesticide pollution is quite recent. However, the pesticide problem is complicated by the great diversity of uses and pesticides’ properties in terms of their transfer and dissipation. Pesticides can be eliminated from the water column through transfer and transformation processes (Gregoire et al., 2009). The transfer of pesticides from the water column to solid surfaces corresponds to adsorption phenomena. The suitable solid surfaces for this process in CWs mostly include sediments (Runes et al., 2003). However, dead and living vegetation or filter materials in subsurface flow CWs can also be considered to be a potential structure for adsorption. Adsorption rates reaching 55% for Chlorpyrifos (Moore et al., 2002), 50% for S-Metolachlor (Hoyos-Hernandez, 2010), 48% for Lambda-cyhalothrin and Imidacloprid (Mahabali and Spanoghe, 2014) and 28% for Fipronil (Peret et al., 2010) were observed on CW sediments. The adsorption process can be reversible, however, which is notably valid for molecules presenting a high level of solubility or a low adsorption coefficient. In this case, the opposite phenomenon, desorption, causes the return of adsorbed molecules to the water column. Adsorption–desorption should therefore only be considered as a temporary phenomenon that delays the transfer of peaks of pesticide concentrations through the CWs and attenuates peaks in these concentrations. For vulnerable aquatic environments, this reduces the acute toxicity of the pollution. Certain molecules can also be removed by plants and thus transferred to the interior of plant tissues (phytoaccumulation) in lower quantities (e.g. 0.5% for Lambda-cyhalothrin and Imidacloprid) (Feurtet-Mazel et al., 1996; Miglioranza et al., 2004; Mahabali and Spanoghe, 2014). Pesticides can thus be adsorbed, removed or released into the water column during the decomposition of vegetation. Transformation processes are those that produce new molecules (metabolites or degradation by-products) from the so-called parent pesticide molecule in which certain bonds have been cut. The molecules thus formed, even though generally less toxic than the parent molecules can, however, themselves present strong toxic properties. As in the soil (Zhang et al., 2000), the degradation of pesticides has multiple sources in CWs: due to the effect of light (photodegradation), water molecules (hydrolysis), and particularly microorganisms (biodegradation; Fig. 5). Photodegradation is favoured in shallow waters that do not excessively attenuate the solar radiation. The microbial processes of pesticide degradation appear to dominate in CWs (Hijosa-Valsero

5

et al., 2010). The biodegradation of pesticides is generally greater in aerobic than in anaerobic environments. Wetland sediments are more subject to reductive conditions. However, aerobic zones subsist, notably in the rhizosphere (Mitsch and Gosselink, 2007). Alternating between the presence and absence of water, which can be part of the management mode selected, allows one to implement successive reductive and oxidative conditions resulting in different oxidoreduction reactions. Moreover, the degradation rate has sometimes been shown to be higher in reductive than in oxidative conditions (Accinelli et al., 2005). Under aerobic conditions, high rates of atrazine mineralization (70–80%) have been observed in CW sediments (Anderson et al., 2002). Interaction between pesticides and nitrate removal processes is one of the challenges for further studies. There are already evidences that pesticides like Difenoconazole (fungicide), Deltamethrin (insecticide) and Ethofumesate (herbicide) at very high concentrations (500 mg/kg) in soil can significantly inhibit soil ˜ microbial processes including denitrification (Munoz-Leoz et al., 2013). 4.2.2. Pesticide retention efficiency of wetlands The assessment of CWs acting as the “black box” to remove pesticide pollution has shown very promising results, with a mean efficiency generally higher than 60% (Gregoire et al., 2009; Stehle et al., 2011; Vymazal et Bfezinova 2015). However, the removal rate of different pesticides varies (Blankenberg et al., 2007; Maillard et al., 2011; Vymazal and Bˇrezinová, 2015), and therefore “average removal” is a term that can be used as a general characteristics of CWs’ efficiency. As for nitrate, the input–output efficiency of CWs varies between negative values and 100% (Stehle et al., 2011). Studies presenting negative efficiencies indicate the observation of higher output than input concentrations in CWs. It should be noted that occasionally the concentration values are close to or less than the limit of quantification when measurement uncertainty is high. The detachment of biofilms that have accumulated pesticides or the desorption of certain molecules can thus contribute to the appearance of higher output than input concentrations between CWs. Although real, this phenomenon cannot be generalized to all pesticides. Long-term studies would be necessary in order to quantify this. The fate of degradation products such as pesticide’s metabolites is challenging in terms of sustainability. The biodegradation of molecules is a slow process that is favoured by long retention times. The hydraulic functioning of a CW, most particularly the residence time, is a key factor in optimising biological processes. Similarly, vegetation has direct and indirect effects on the dissipation of pesticides. By airing sediments, vegetation increases microbial activity. By creating roughness, it slows the flows and thus increases the hydraulic retention time of pesticides (Brix, 1997) and favours particle sedimentation. Poorly distributed, it can also generate hydraulic short-circuits that have the opposite effect (Jenkins and Greenway, 2005). Decomposing vegetation provides organic carbon to microorganisms (Moore et al., 2007). It also serves as an adsorbing surface for pesticides and can at times remove some. It can also assist in developing biofilms in which pesticide biodegradation can take place, and it helps stabilize sediments (Brix, 1997). Literature analysis showed high variability of pesticide concentrations in the inflow to CWs. Therefore, removal efficiencies are presented in mass removal values calculated as difference between the inflow and outflow parameters (Fig. 4). Field experiments found in literature on CWs treating drainage water can be divided based on two types of drainage—controlled and uncontrolled. In the first type, drainage flow is hydraulically controlled as inflow into a experimental CW insuring about 8 days retention times (see Alvord and Kadlec, 1996; Braskerud and Haarstad, 2003; Blankenberg et al., 2007; Hunt et al., 2008). The uncontrolled type represents a system

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

6

Fig. 6. The concept of drained flow interception through constructed wetlands; (a) on-stream interception; (b) off-stream interception.

which is depending on natural conditions such as rainfall intensity (Miller et al., 2002; Passeport et al., 2013; Tournebize et al., 2013; Vallée et al., 2015). Both approaches demonstrated similar average retention efficiency—32% and 39% for controlled and uncontrolled conditions respectively. Adsorption coefficient Koc is one of the key parameters describing removal potential of pesticides in CWs (Vymazal and Bˇrezinová, 2015). In Fig. 4, pesticides are grouped based on Koc values to: low Koc (<400 mL/g); moderate Koc (400 - 1000 mL/g) and strong Koc (>1000 mL/g). The first group of pesticides with low Koc value and average removal efficiency of 25% is represented by MCPA, Bentazone, Matalaxyl, Isoproturon, Chlortoluron, Metamitrone, SMetolachlor, Ethofumesate, Atrazine, and Metazachlor) with low Koc values. Pesticides with moderate and strong Koc values show higher removal potential of 49% and 51%, respectively. This group contains pesticides such as Boscalid, Chlorothalonil, Napropamide, Tebuconazole, Azoxystrobine, Propyzamide, Propiconazole, Fenpropimorph, Epoxiconazole, Chlorpyrifos, Prosulfocarbe, Difflufenilcanile, Aclonifen, and Pendimethaline. CWs have real potential to reduce the concentrations and flows of agricultural pollutants in concentrated flows. However, the efficiency results show that the objective of 100% abatement is hard to achieve. The purification performance of CWs depends on hydrological conditions and seasonality. CWs should therefore not be considered as a permit to pollute but rather a complementary tool to actions implemented at the agricultural plot scale to reduce the pressure of pollution (reduction of inputs). Microbial processes are the main actors in nitrate or pesticides dissipation. Sorption allows only pesticide trapping from the water column and the rest of processes (plant uptake, hydrolysis, photolysis) have secondary importance (Passeport et al., 2013). Living vegetation and litter as additional source of organic matter favours the elimination of pollutants. Purification efficiency is highly correlated with hydraulic residence time within the CW. 5. Feedback on design assistance The conceptual development of CWs is part of a wider ecological engineering process, which can be defined as the application of ecological principles to managing the environment. Here, ecological engineering serves water quality. Setting up CWs requires one to sacrifice some of the land farmed by the farmer, which is why the overall process should involve all the actors in a

co-construction process (Tournebize et al., 2012). Available land is often limited and does not necessarily make it possible to treat all water volumes satisfactorily. If the wetland to catchment ratio (WCR) is too small (<1%), the water and pollutant retention times in the CW risk being too short for good water purification. Koskiaho and Puustinen (2005) demonstrate, that in boreal conditions, for substantial (>20%) N and P load reductions the WCR should be more than 2%. This may result in making a choice between which flows to treat. Within the objective of eliminating the maximum quantity of pollutants with the minimum volume of water, it may well be advantageous to give priority to intercepting and treating the most concentrated flows. We therefore propose two strategies depending on the water quality parameters targeted. 5.1. Intercepting flows: On-stream or off-stream interception strategy Depending on the surface area available for implanting an CW, it must be implemented with either on-stream or off-stream interception in relation to the agricultural drainage collector, at the output of a watershed (Fig. 6; see also Chen et al., 2007). One or another of these strategies is more or less better adapted depending on the transfer mode of the targeted pollutant (nitrate or pesticide) (Passy et al., 2012; Tournebize et al., 2013). 5.1.1. On-stream interception of flows: Favouring nitrate removal When nitrate concentration is the most important management objective, its transfer mode provides information on two aspects: all flows export nitrate and the concentrations are similar at all spatial scales. Therefore, if the area available is sufficiently broad, the CW can be planted in an on-stream formation, i.e., at the outlet of a collector or through an agricultural ditch by widening it. All drainage flows transit through the CW. This strategy has been tested in Seine-et-Marne (Tournebize et al., 2013) for a situation in which the CW’s primary objective was to stock drainage waters for irrigation purposes. Some on-stream strategies use additional woodchips-filled bioreactors in the bottom of streams to treat nitrate (Robertson and Merkley, 2009; Moorman et al., 2015). 5.1.2. Off-stream interception of flows: Favouring pesticide removal The transfer dynamics of pesticides show that the first flows following pesticide applications are the most highly concentrated.

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

7

Fig. 7. Suggested localization of constructed wetlands according to nitrate (one big one, whatever the scale) and pesticide removal (small and several ones upstream), based on the example at Rampillon, Seine-et-Marne, France (Tournebize et al., 2012).

These are generally autumn flows (after the application of herbicides) or spring flows (applications of herbicides and fungicides) outside of the intense drainage season (Fig. 2). These flows can also be associated with high nitrate concentrations. In terms of land area, it can therefore be advantageous to target the waters intercepted at only these high-risk periods. For smaller available surface areas, a CW could be located in off-stream interception in relation to the agricultural drainage ditch and be associated with the hydrologic control (a gate) of water flow in the agricultural ditch. This was tested in two agricultural catchments, in the French Indre-etLoire department (Passeport et al., 2013) and in the Seine-et-Marne department (Tournebize et al., 2012). The gate can be operated by the farmer himself immediately after pesticide treatments. This has the double advantage of involving the farmer in the processes of his impact on the land and of hydraulically targeting the first high waters after application. The first flows are thus redirected through the CW. Closing the water one month after application will extend the pesticides’ residence time in the CW and thus promote implementation of the expected natural water treatment processes. 5.2. Location in the catchment area The location of the CW depends on the parameter targeted (nitrate or pesticide) and the catchment’s hydrological functioning. At this stage, it is important to know the water pathways in order to determine the initial locations. It can therefore be useful to obtain the drainage plans with farmers or local authorities that have intervened in any new drainage network over 20 ha. However, it must be noted that this is often difficult in the field. Later one must encourage non-agricultural spaces that could ideally be placed within the drainage network. Finally, if this remains insufficient, the possibility of requiring agricultural land directly from farmers should be entertained. This is why their involvement in coconstruction processes should be established from the beginning of project planning. Depending on the parameter being targeted (nitrate or pesticide), one could recommend positioning CWs as far upstream of catchments as possible, near the pollution source, at the outlet of the sub-catchment (<100 ha) (Fig. 7a and b) (van der Valk and Jolly, 1992). However, in particular situations, more or less difficult contexts, in cooperation with various actors, a CW at the outlet of the catchment (>100 ha), managed collectively, could be a good compromise (Tournebize et al., 2012).

However, any potential location should be validated by studies on the topography of the site, guaranteeing the possibility of temporary water storage and the characterization of soil layers (geotechnical study), ensuring natural hydraulic conductivity during the construction of the CW. 5.3. Sizing and development Several publications consider sizing and location of CWs in agricultural landscapes (Lesta et al., 2007; Moreno-Mateos et al., 2010; Stringfellow et al., 2013; Tomer et al., 2013). In general, sizing of the CWs should be based on ecological engineering criteria, although preference could be given to some particular function. In our case, the improvement of water quality, without preventing other functions (hydraulic buffers, increased biodiversity) has been highlighted (Mitsch and Cronk, 1992). One of the key principles of ecological engineering is to leave time for CWs to self-organize while avoiding over-engineering of their construction (Mitsch et al., 2012). Residence time can be lengthened by increasing the size or the volume of the CW. The water heights generally recommended are on the order of 0.5–0.9 m. When land availability precludes this, it is possible to set up dikes so as to increase the water’s pathway through the CWs (Persson, 2000). It is important to avoid positioning the flow input and output structures so that they face one another, which would increase the hydraulic dead zone and limit the use of some of the CW zones. The shape of the CW naturally depends on the available site. For reasons of convenience, the rectangular shape is generally more frequently used, since it is the simplest to construct and best observes the rectilinear plot limits. However, freer and more meandering shapes can be constructed. Flow input and output should be located at the edges of the hydraulic pathway. If the configuration allows, the creation of small dikes is recommended, because this increases the flow path (Figs. 8 and 9). This optimization limits the dead zones, which allows circulation within a greater volume. This slower circulation is conducive to higher contact times in the system (water/vegetation/sediment) for better efficiency. One must keep in mind that increasing hydraulic retention time (and therefore pesticide retention time) is a key consideration in sizing. Scholz et al. (2007) developed the concept of FWS integrated constructed wetlands (ICW), which explicitly combines the objectives of cleansing and managing water flow from farmyards and fertilized fields

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11 8

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

Fig. 8. Example of a rectangular constructed wetland, integrating an input protected by riprap rock fill, reed-grown beds, an outlet and embankments which lengthen the flow path. The example of the Rampillon basin No. 5 experiment is provided. Designed by IRSTEA-Antony and CIAE Nemours (Design office).

with that of integrating the wetland infrastructure into the landscape and enhancing its biological diversity. The effectiveness of the reduction in concentrations and flows is closely correlated with residence times in CWs. However, the mean 50% objective observed in the literature review seems realistic. Tanner et al. (2010) found that for New Zealand CWs designed to treat tile drainage, 5% of WCR value is necessary for 53% nitrate removal. In a more recent study, Tanner and Kadlec (2013) propose a WCR of about 2% to reach 50% of NO3− –N removal (Fig. 10).

Garnier et al. (2014) showed that 8% of WCR value contributes to removing 50% of total annual nitrogen fluxes. However, with optimal conditions for denitrification and in some deeper parts of CWs also for ANAMMOX (see Ligi et al., 2014), the target of 50% nitrate removal may be possible with a 1% wetland-to-catchment ratio. Our experiments also show that this objective can be met with a minimum residence time of 2 days for nitrate and five times longer for pesticides. With these hypotheses, Tournebize et al. (2015b) proposed a number of design rules dedicated to the French drainage

Fig. 9. Successional steps of basin No. 5 at the Rampillon experimental catchment. (a) Before construction (cultivated plot); (b) during construction; (c) overview of the reedbed; (d) the reedbed seen from above, with small dikes. (Source: IRSTEA).

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

9

plant species with smaller stems reduce erosion on slopes, reduce weeds, and create habitat diversity. Conclusively, the inclusion of design and management is one of the most important challenges to integrating other ecosystem service benefits in particular wetland systems. 6. Conclusions

Fig. 10. The constructed wetland volume to catchment ratio (%) vs. percentage nitrate-N mass removal in the subsurface drainage context. Modeled based on a simple first-order dynamic model that accounts for hydrology (the tanks-in-series approach), nitrate concentrations, internal hydraulic efficiency and water temperature. Adapted from Tanner and Kadlec (2013).

context, leading to an average volume of 76 m3 of CW per drained hectare upstream. This ratio needs to be refined for each local context, and a pedological study should be included in the hydrological diagnosis. Taking a mean depth of 0.8 m, this volume corresponds to approximately 1% of the area of the upstream contributing watershed. 5.4. Multifunctionality of CWs Most of CWs are multifunctional systems that potentially provide several ecosystem services: from provisional, regulation and habitat functions to recreational and socioeconomic services (Millennium Ecosystem Assessment, 2005; Yang et al., 2008). In addition to water quality, runoff and greenhouse gas regulation, other well-known services of CWs include supporting habitat and biodiversity functions (Ghermandi et al., 2010; Fennessy and Craft, 2011) and biomass production (Wild et al., 2001; Maddison et al., 2009; Meerburg et al., 2010). Sometimes these functions have a conflicting character. Hansson et al. (2005) showed that among the features used to design new wetlands, area, depth and shoreline complexity have fundamental and sometimes conflicting effects on nutrient retention and biodiversity. However, it is possible to direct the ecosystem function of a specific wetland in desired directions. In urban areas and for recreational/educational purposes, the shape and layout of a CW should vary to suit the landscape and to satisfy aesthetic requirements as a water feature (Persson et al., 1999). Nevertheless, there are only a few examples of the successful design of CWs that actually provide multiple services and also have a high recreational and aesthetic value (Bays et al., 2000; Brix et al., 2007; Mitsch et al., 2012). The maintenance of these complex systems is more costly, and poor management can lead to clogging of the system and the failure of several functions (Brix et al., 2007). Self-designing capacity of ecosystems and minimum intervention to the ecosystem succession are the main principles in ecological engineering (Odum, 1989; Mitsch and Gosselink, 2007). Therefore, smart choice of plant species plays an important role: in many cases it depends on the water depth the plants will inhabit. Most commonly, macrophytes such as common reed (P. australis), bulrushes (Scirpus spp), cattails (Typha spp) and water lilies (Nymphea spp) are selected for their resistance to the substantial variations in water depth. In periods of senescence, these plants are a source of carbon, which is favourable to denitrification. They also resist weed invasion and tend to dominate other species in terms of nutrients (Kadlec and Wallace, 2008). In semi-submerged areas or on baffles,

CWs have true potential to reduce the flow of agricultural contaminants such as nitrate and pesticides. Developments using green infrastructures undeniably contribute to reducing contaminant transfer within agricultural landscapes. A realistic objective of 50% can be expected by devoting 1% of the upstream contributing surface area to a CW. Since area scale does not play important role in nitrate losses from drained catchments, we suggest for nitrate removal one large on-stream FWS CW at the lower course of watershed. In opposite, pesticides runoff has been significantly lower in small upstream catchments, therefore, several small off-stream FWS CWs in upstream parts would be the most optimal solution. The hydraulic residence time is the priority factor determining water treatment efficiency in CWs. For better performance, that must be regulated by users. The other factors play only a stimulating role, supporting microbiological pathways of natural purification. Their effectiveness will vary greatly depending on the season, the hydrological flows and the pollutants’ properties. It is important to integrate the understanding of CWs as a complementary action to those responsible for reducing agricultural pollution and not use them as permission to pollute. Due to the high environmental hazard of N2 O emissions and the limited knowledge of its fluxes from FWS CWs, more detailed study is needed. In particular, the N2 /N2 O ratio must be highlighted. Moreover, possible accumulation of pesticides’ metabolites or bound residues in CWs should be deeper studied in order to avoid the time bomb effect. Acknowledgements The study was promoted by ONEMA (The French National Agency for Water and Aquatic Environments), the French technical group for buffer zones (www.zonestampons.onema.fr). The field experiments were financially supported by AESN, Life ARTWet (ENV/F/000133), PIREN Seine. The authors thank also AQUI’Brie association for their support. This study was also supported by the Estonian Research Council (grant no. IUT2-16) and by the EU through the European Regional Development Fund (Centre of Excellence ENVIRON). References Accinelli, C., Screpanti, C., Vicari, A., 2005. Influence of flooding on the degradation of linuron, isoproturon and metolachlor in soil. Agron. Sustainable Dev. 25 (3), 401–406. Alvord, H.H., Kadlec, R.H., 1996. Atrazine fate and transport in the Des Plaines Wetlands. Ecol. Model. 90, 97–107. Anderson, K.L., Wheeler, K.A., Robinson, J.B., Tuovinen, O.H., 2002. Atrazine mineralization potential in two wetlands. Water Res. 36 (19), 4785–4794. Appelboom, T.W., Fouss, J.L., 2006. Methods for removing nitrate nitrogen from agricultural drainage waters: a review and assessment. In: Proc. of ASABE Annual Meeting, Paper # 062328. Bastviken, S.K., Weisner, S.E.B., Thiere, G., Svensson, J.M., Ehde, P.M., Tonderski, K.S., 2009. Effects of vegetation and hydraulic load on seasonal nitrate removal in treatment wetlands. Ecol. Eng. 35 (5), 946–952. Batson, J.A., Mander, Ü., Mitsch, W.J., 2012. Denitrification and a nitrogen budget of created riparian wetlands. J. Environ. Qual. 41, 2024–2032. Bays, J., Dernian, G., Hadjimiry, H., Vaith, K., Keller, C., 2000. Treatment wetlands for multiple functions: Wakodahatchee Wetlands, Palm Beach County, Florida. In: Proc. Water Environ. Feder. WEFTEC 2000, session 51-60, pp. 15–37 (23). Beutel, M.W., Newton, C.D., Brouillard, E.S., Watts, R.J., 2009. Nitrate removal in surface-flow constructed wetlands treating dilute agricultural runoff in the lower Yakima Basin, Washington. Ecol. Eng. 35, 1538–1546.

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11 10

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

Billy, C., Billen, G., Sebilo, M., Birgand, F., Tournebize, J., 2010. Nitrogen isotopic composition of leached nitrate and soil organic matter as an indicator of denitrification in a sloping drained agricultural plot and adjacent uncultivated riparian buffer strips. Soil Biol. Biochem. 42 (1), 108–117. Billy, C., Birgand, F., Ansart, P., Peschard, J., Sebilo, M., Tournebize, J., 2013. Factors controlling nitrate concentrations in surface waters of an artificially drained agricultural watershed. Landscape Ecol. 28 (4), 665–684. Blanchoud, H., Barriuso, E., Nicola, L., Schott, C., Roose-Amsaleg, C., Tournebize, J., 2013. La contamination de l’Orgeval par les pesticides, une préoccupation de longue date, L’observation long terme en environnement. In: Loumagne, C., Tallec, G. (Eds.), Exemple du bassin versant de l’Orgeval. QUAE, Irstea, Antony, pp. 159–174. Blankenberg, A.-G.B., Haarstad, K., Braskerud, B.C., 2007. Pesticide retention in an experimental wetland treating non-point source pollution from agriculture runoff. Water Sci. Technol. 55 (3), 37–44. Boithias, L., Sauvage, S., Srinivasan, R., Leccia, O., Sánchez-Pérez, J.M., 2014. Application date as a controlling factor of pesticide transfers to surface water during runoff events. Catena 119, 97–103. Borin, M., Tocchetto, D., 2007. Five years water and nitrogen balance for a constructed surface flow wetland treating agricultural drainage waters. Sci. Total Environ. 380, 38–47. Branger, F., Tournebize, J., Carluer, N., Kao, C., Braud, I., Vauclin, M., 2009. A simplified modeling approach for pesticide transport in a tile-drained field: the PESTDRAIN model. Agric. Water Manage. 96 (3), 415–428. Braskerud, B.C., Haarstad, K., 2003. Screening the retention of thirteen pesticides in a small constructed wetland. Water Sci. Technol. 48, 267–274. Brix, H., 1997. Do macrophytes play a role in constructed treatment wetlands? Water Sci. Technol. 35 (5), 11–17. Brix, H., Koottatep, T., Laugesen, C.H., 2007. Wastewater treatment in tsunami affected areas of Thailand by constructed wetlands. Water Sci. Technol. 56 (3), 69–74. Brown, C.D., van Beinum, W., 2009. Pesticide transport via sub-surface drains in Europe. Environ. Pollut. 157 (12), 3314–3324. Chen, Q., Shan, B., Yin, C., Hu, C., 2007. An off-line filtering ditch–pond system for diffuse pollution control at Wuhan City Zoo. Ecol. Eng. 30, 373–380. CORPEN GZT, 2007. Les fonctions environnementales des zones tampons, Les bases scientifiques et techniques des fonctions de protection des eaux. d. l. é. Ministère de l’écologie, du développement durable et de l’aménagement du territoire et Ministère de l’agriculture et de la pêche. CORPEN GZT, Paris, pp. 176. Correll, D., 2005. Principles of planning and establishment of buffer zones. Ecol. Eng. 24 (5), 433–439. Craft, C.B., 1997. Dynamics of nitrogen and phosphorus retention during wetland ecosystem succession. Wetlands Ecol. Manage. 4, 177–187. Crumpton, W.R., Helmers, M., 2004. Integrated drainage-wetland systems for reducing nitrate loads from tile drained landscapes. In: Self-Sustaining Solutions for Streams, Wetlands, and Watersheds, Proceedings of the 12–15 September 2004 Conference (St. Paul, Minnesota USA). Díaz, F.J., O´ıGreen, A.T., Dahlgren, R.A., 2012. Agricultural pollutant removal by constructed wetlands: implications for water management and design. Agric. Water Manage. 104, 171–183. European Union, 22 December, 2000. Water framework directive (2000/60/CE). EU Off. J. (OJ L 327). Fennessy, S., Craft, C., 2011. Agricultural conservation practices increase wetland ecosystem services in the Glaciated Interior Plains. Ecol. Appl. 21 (3), S49–S64. Feurtet-Mazel, A., Grollier, T., Grouselle, M., Ribeyre, F., Boudou, A., 1996. Experimental study of bioaccumulation and effects of the herbicide isoproturon on freshwater rooted macrophytes—(Elodea densa and Ludwigia natans). Chemosphere 32 (8), 1499–1512. Fisher, J., Acreman, M.C., 2004. Wetland nutrient functioning: a review of the evidence. Hydrol. Earth System Sci. 8 (4), 673–685. Fonder, N., Headley, T., 2010. Systematic classification, nomenclature and reporting for constructed treatment wetlands. In: Vymazal, J. (Ed.), Water and Nutrient management in Natural and Constructed Wetlands. Springer Science + Business Media B.V., Dordrecht, pp. 191–220. Forbes, E.G.A., Woods, V.B., Easson, D.L., 2004. Constructed Wetlands and their Use to Provide Bioremediation of Farm Effluents in Northern Ireland. A Review of Current Literature. Agri-Food and Bioscience Institute, Belfast, UK, pp. 41, Literature report. Garnier, J., Billen, G., Vilain, G., Benoit, M., Passy, P., Tallec, G., Tournebize, J., Anglade, J., Billy, C., Mercier, B., Ansart, P., Azougui, A., Sebilo, M., Kao, C., 2014. Curative vs. preventive management of nitrogen transfers in rural areas: lessons from the case of the Orgeval watershed (Seine River basin, France). J. Environ. Manage. 144, 125–134. Ghermandi, A., van den Bergh, J.C.J.M., Brander, L.M., de Groot, H.L.F., Nunes, P.A.L.D., 2010. Values of natural and human-made wetlands: a meta-analysis. Water Resour. Res. 46, W12516. Gregoire, C., Elsaesser, D., Huguenot, D., Lange, J., Lebeau, T., Merli, A., Mose, R., Passeport, E., Payraudeau, S., Schütz, T., Schulz, R., Tapia-Padilla, G., Tournebize, J., Trevisan, M., Wanko, A., 2009. Review: mitigation of agricultural nonpointsource pesticide pollution in artificial wetland ecosystems. Environ. Chem. Lett. 7, 205–231. Groh, T.A., Gentry, L.E., David, M.B., 2015. Nitrogen removal and greenhouse gas emissions from constructed wetlands receiving tile drainage water. J. Environ. Qual. 44 (3), 1001–1010. Hammer, D.A., Knight, R.L., 1994. Designing constructed wetlands for nitrogen removal. Water Sci. Technol. 29 (4), 15–27.

Hansson, L.-A., Brönmark, C., Nilsson, P.A., Åbjörnsson, K., 2005. Conflicting demands on wetland ecosystem services: nutrient retention, biodiversity or both? Freshwater Biol. 50 (4), 705–714. Hijosa-Valsero, M., Matamoros, V., Sidrach-Cardona, R., Martin-Villacorta, J., Becares, E., Bayona, J.M., 2010. Comprehensive assessment of the design configuration of constructed wetlands for the removal of pharmaceuticals and personal care products from urban wastewaters. Water Res. 44 (12), 3669–3678. Hoyos-Hernandez, C., 2010. Degradation du S-metolachlor dans une zone tampon humide artificielle en fonction de l’activité microbienne, des conditions d’oxydoréduction et de différentes sources de carbone, Vol. Master II Ingénierie biologique de l’environnement. Université Paris Est Créteil, France, pp. 62, Master Thesis, in French. Hunt, J., Anderson, B., Philips, B., Tjeerdema, R., Largay, B., Beretti, M., Bern, A., 2008. Use of toxicity identification evaluations to determine the pesticide mitigation effectiveness of on-farm vegetated treatment systems. Environ. Pollut. 156, 348–358. IPCC, 2007. Climate Change, 2007. The Physical Science Basis. Cambridge University Press, Cambridge. Jenkins, G.A., Greenway, M., 2005. The hydraulic efficiency of fringing versus banded vegetation in constructed wetlands. Ecol. Eng. 25 (1), 61–72. Kadlec, R.H., 1994. In: Mitsch, W.J. (Ed.), Wetlands for Water Polishing: Free Water Surface Wetlands. Global Wetlands: Old World and New. Elsevier, Amsterdam, pp. 335–349. Kadlec, R.H., 2005. Nitrogen farming for pollution control. J. Environ. Sci. Health A 40 (6–7), 1307–1330. Kadlec, R.H., Wallace, S.D., 2008. Treatment Wetlands, second ed. CRC Press, Boca Raton, FL. Kladivko, E.J., Brown, L.C., Baker, J.L., 2001. Pesticide transport to subsurface tile drains in humid regions of North America. Crit. Rev. Env. Sci. Tech. 31, 1–62. Koskiaho, J., Ekholm, P., Räty, M., Riihimäki, J., Puustinen, M., 2003. Retaining agricultural nutrients in constructed wetlands—experience under boreal conditions. Ecol. Eng. 20, 89–103. Koskiaho, J., Puustinen, M., 2005. Function and potential of constructed wetlands for the control of N and P transport from agriculture and peat production in boreal climate. J. Environ. Sci. Health A 40 (6–7), 1265–1279. Kovacic, D.A., David, M.B., Gentry, L.E., Starks, K.M., Cooke, R.A., 2000. Effectiveness of constructed wetlands in reducing nitrogen and phosphorus export from agricultural tile drainage. J. Environ. Qual. 29, 1262–1274. Lacas, J.G., Voltz, M., Gouy, V., Carluer, N., Gril, J.J., 2005. Using grassed strips to limit pesticide transfer to surface water: a review. Agron. Sustainable Dev. 25 (2), 253–266. Lesta, M., Mauring, T., Mander, Ü., 2007. Estimation of landscape potential for construction of surface-flow wetlands for wastewater treatment in Estonia. Environ. Manage. 40 (2), 303–313. Ligi, T., Truu, M., Oopkaup, K., Nõlvak, H., Mander, Ü., Mitsch, W.J., Truu, J., 2014. Genetic potential of N2 emission via denitrification and ANAMMOX from the soils and sediments of a created riverine treatment wetland complex. Ecol. Eng. 80, 181–190. Maddison, M., Mauring, T., Remm, K., Lesta, M., Mander, Ü., 2009. Dynamics of Typha latifolia L. populations in treatment wetlands in Estonia. Ecol. Eng. 35 (2), 258–264. Mahabali, S., Spanoghe, P., 2014. Mitigation of two insecticides by wetland plants: feasibility study for the treatment of agricultural runoff in Suriname (South America). Water Air Soil Pollut. 225, 1771. Maillard, E., Payraudeau, S., Faivre, E., Grégoire, C., Gangloff, S., Imfeld, G., 2011. Removal of pesticide mixtures in a stormwater wetland collecting runoff from a vineyard catchment. Sci. Total Environ. 409 (11), 2317–2324. Mander, Ü., Dotro, G., Ebie, Y., Towprayoon, S., Chiemchaisri, C., Nogueira, S.F., Jamsranjav, B., Kasak, K., Truu, J., Tournebize, J., Mitsch, W.J., 2014a. Greenhouse gas emission in constructed wetlands for wastewater treatment: a review. Ecol. Eng. 66, 19–35. Mander, Ü., Tournebize, J., Kasak, K., Mitsch, W.J., 2014b. Climate regulation by free water surface constructed wetlands for wastewater treatment and created riverine wetlands. Ecol. Eng. 72, 103–115. McPhillips, L., Walter, M.T., 2015. Hydrologic conditions drive denitrification and greenhouse gas emissions in stormwater detention basins. Ecol. Eng. 85, 67–75. Meerburg, B.G., Vereijken, P.H., de Visser, W., Korevaar, H., Querner, E.P., Blaeij, A.T., van der Werf, A.K., 2010. Surface water sanitation and biomass production in a large constructed wetland in the Netherlands. Wetlands Ecol. Manage. 18 (4), 463–470. Miglioranza, K.S.B., de Moreno, J.E.A., Moreno, V.J., 2004. Organochlorine pesticides sequestered in the aquatic macrophyte Schoenoplectus californicus (C.A, Meyer) Sojak from a shallow lake in Argentina. Water Res. 38, 1765–1772. Millennium Ecosystem Assessment, 2005. Ecosystems & Human Well-being: Synthesis. Island Press, Washington, DC. Miller, P.S., Mitchell, J.K., Cooke, R.A., Engel, B.A., 2002. A wetland to improve agricultural subsurface drainage water quality. Transactions ASAE 45 5, 1305–1317. Ministry of Agriculture, 2008. The ECOPHYTO 2018 Plan for the reduction of pesticide use over the period 2008–2018., pp. 21, http://agriculture.gouv.fr/IMG/ pdf/PLAN ECOPHYTO 2018 eng.pdf (verified September 2014). Mitsch, W.J., Cronk, J.K., 1992. Creation and restoration of wetlands: some design considerations for ecological engineering. Adv. Soil Sci. 17, 217–259. Mitsch, W.J., Gosselink, J.G., 2007. Wetlands. John Wiley and Sons, Hoboken, NJ, USA, pp. 582.

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014

G Model ECOENG-4007; No. of Pages 11

ARTICLE IN PRESS J. Tournebize et al. / Ecological Engineering xxx (2016) xxx–xxx

Mitsch, W.J., Zhang, L., Stefanik, K.C., Nahlik, A.M., Anderson, C.J., Bernal, B., Hernandez, M.E., Song, K., 2012. Creating wetlands: primary succession, water quality changes, and self-design over 15 years. BioScience 62 (3), 237–250. Moore, M.T., Cooper, C.M., Smith, S., Cullum, R.F., Knight, S.S., Locke, M.A., Bennett, E.R., 2007. Diazinon mitigation in constructed Wetlands: influence of vegetation. Water Air Soil Pollut. 184 (1–4), 313–321. Moore, M.T., Schulz, R., Cooper, C.M., Smith, S., Rodgers, J.H., 2002. Mitigation of chlorpyrifos runoff using constructed wetlands. Chemosphere 46 (6), 827–835. Moorman, T.B., Tomer, M.D., Smith, D.R., Jaynes, D.B., 2015. Evaluating the potential role of denitrifying bioreactors in reducing watershed-scale nitrate loads: a case study comparing three Midwestern (USA) watersheds. Ecol. Eng. 75, 441–448. Moreno-Mateos, D., Mander, Ü., Pedrocchi, C., 2010. Optimal location of created and restored wetlands in Mediterranean agricultural catchments. Water Resour. Manage. 24 (11), 2485–2499. ˜ ¨ Munoz-Leoz, B., Garbisu, C., Charcosset, J.-Y., Sanchez-Pérez, J.-M., Antiguedad, I., Ruiz-Romera, E., 2013. Non-target effects of three formulated pesticides on microbially-mediated processes in a clay-loam soil. Sci. Total Environ. 449, 345–354. Nolan, B.T., Malone, R.W., Gronberg, J.A., Thorp, K.R., Ma, L.W., 2012. Verifiable metamodels for nitrate losses to drains and groundwater in the Corn Belt, USA. Environ. Sci. Technol. 46 (2), 901–908. Odum, H.T., 1989. Ecological engineering and self-organization. In: Mitsch, W.J., Jørgensen, S.E. (Eds.), Ecological Engineering: An Introduction to Ecotechnology. John Wiley & Sons, New York, NY, pp. 79–101. Passeport, E., Benoit, P., Bergheaud, V., Coquet, Y., Tournebize, J., 2011. Epoxiconazole degradation from artificial wetland and forest buffer substrates under flooded conditions. Chem. Eng. J. 173 (3), 760–765. Passeport, E., Tournebize, J., Chaumont, C., Guenne, A., Coquet, Y., 2013. Pesticide contamination interception strategy and removal efficiency in forest buffer and artificial wetland in a tile-drained agricultural watershed. Chemosphere 91 (9), 1289–1296. Passeport, E., Tournebize, J., Jankowfsky, S., Promse, B., Chaumont, C., Coquet, Y., Lange, J., 2010. Artificial wetland and forest buffer zone: hydraulic and tracer characterization. Vadose Zone J. 9 (1), 73–84. Passy, P., Garnier, J., Billen, G., Fesneau, C., Tournebize, J., 2012. Restoration of ponds in rural landscapes: modelling the effect on nitrate contamination of surface water (the Seine River Basin, France). Sci. Total Environ. 430, 280–290. Peret, A.M., Oliveira, L.F., Bianchini Jr., I., Regali Seleghim, M.H., Peret, A.C., Mozeto, A.A., 2010. Dynamics of fipronil in Oleo Lagoon in Jatai Ecological Station, Sao Paulo-Brazil. Chemosphere 78 (10), 1225–1229. Persson, J., 2000. The hydraulic performance of ponds of various layouts. Urban Water 2 (3), 243–250. Persson, J., Somes, N.L.G., Wong, T.H.F., 1999. Hydraulics efficiency of constructed wetlands and ponds. Water Sci. Technol. 40 (3), 291–300. Pulou, J., 2011. Les anciennes cressonnières de l’Essonne: Effets de la recolonisation des zones humides artificielles sur la dynamique de l’azote. AgroParisTech, pp. 212, Ph.D. Thesis. Pulou, J., Tournebize, J., Chaumont, C., Haury, J., Laverman, A.M., 2012. Carbon availability limits potential denitrification in watercress farm sediment. Ecol. Eng. 49, 212–220. Randall, G.W., Huggins, D.R., Russelle, M.P., Fuchs, D.J., Nelson, W.W., Anderson, J.L., 1997. Nitrate losses through subsurface tile drainage in Conservation Reserve Program, alfalfa, and row crop systems. J. Environ. Qual. 26 (5), 1240–1247. Reddy, K.R., Patrick, W.H., 1984. Nitrogen transformations and loss in flooded soils and sediments. Crit. Rev. Env. Control 13 (4), 273–309. Reichenberger, S., Bach, M., Skitschak, A., Frede, H.-G., 2007. Mitigation strategies to reduce pesticide inputs into ground- and surface water and their effectiveness: a review. Sci. Total Environ. 384 (1–3), 1–35. Robertson, W.D., Merkley, L.C., 2009. In-stream bioreactor for agricultural nitrate treatment. J. Environ. Qual. 38 (1), 230–237. Runes, H.B., Jenkins, J.J., Moore, J.A., Bottomley, P.J., Wilson, B.D., 2003. Treatment of atrazine in nursery irrigation runoff by a constructed wetland. Water Res. 37 (3), 539–550. Scholz, M., Harrington, R., Carroll, P., Mustafa, A., 2007. The Integrated Constructed Wetlands (ICW) concept. Wetlands 27 (2), 337–354. Souiller, C., Coquet, Y., Pot, V., Benoit, P., Réal, B., Margoum, C., Laillet, B., Labat, C., Vachier, P., Dutertre, A., 2002. Capacités de stockage et d’épuration des sols de dispositifs enherbés vis-à-vis des produits phytosanitaires Première partie: Dissipation des produits phytosanitaires à travers un dispositif enherbé; mise en évidence des processus mis en jeu par simulation de ruissellement et infiltrométrie. Etude et Gestion des Sols 9 (4), 269–285. Spieles, D.J., Mitsch, W.J., 2000. The effects of season and hydrologic and chemical loading on nitrate retention in constructed wetlands: a comparison of low- and high-nutrient riverine systems. Ecol. Eng. 14 (1–2), 77–91.

11

Stehle, S., Elsaesser, D., Gregoire, C., Imfeld, G., Niehaus, E., Passeport, E., Payraudeau, S., Schafer, R.B., Tournebize, J., Schulz, R., 2011. Pesticide risk mitigation by vegetated treatment systems: a meta-analysis. J. Environ. Qual. 40 (4), 1068–1080. Stringfellow, W.T., Karpuzcu, M.E., Spier, C., Hanlon, J.S., Graham, J., 2013. Sizing mitigation wetlands in agricultural watersheds. Water Sci. Technol. 67 (1), 40–46. Tallec, G., Ansart, P., Guérin, A., Delaigue, O., Blanchouin, A., 2015. Observatoire Oracle. Irstea, http://dx.doi.org/10.17180/OBS.ORACLE. Tanner, C.C., Kadlec, R.H., 2013. Influence of hydrological regime on wetland attenuation of diffuse agricultural nitrate losses. Ecol. Eng. 56, 79–88. Tanner, C.C., Nguyen, M.L., Sukias, J.P.S., 2005. Nutrient removal by a constructed wetland treating subsurface drainage from grazed dairy pasture. Agric. Ecosyst. Environ. 10 (1–2), 145–162. Tanner, C.C., Sukias, J.P.S., Yates, C.R., 2010. New Zealand Guidelines: Constructed Wetlands Treatment of Tile Drainage. NIWA Information Series No 75, National Institute of Water and Atmospheric Research Ltd., Hamilton, N.Z., pp. 48. Tiemeyer, B., Kahle, P., Lennartz, B., 2006. Nutrient losses from artificially drained catchments in North-Eastern Germany at different scales. Agric. Water Manage. 85 (1/2), 47–57. Tomer, M.D., Crumpton, W.G., Bingner, R.L., Kostel, J.A., James, D.E., 2013. Estimating nitrate load reductions from placing constructed wetlands in a HUC-12 watershed using LiDAR data. Ecol. Eng. 56, 69–78. Tournebize, J., Kao, C., Nikolic, N., Zimmer, D., 2004. Adaptation of the STICS model to subsurface drained soils. Agronomie 24 (6–7), 305–313. Tournebize, J., Arlot, M.-P., Billy, C., Birgand, F., Gillet, J.-P., Dutertre, A., 2008. Quantification et maîtrise des flux de nitrate: de la parcelle drainée au bassin versant. Ingénieries, pp. 5–25. Tournebize, J., Gramaglia, C., Birmant, F., Bouarfa, S., Chaumont, C., Vincent, B., 2012. Co-design of constructed wetlands to mitigate pesticide pollution in a drained catch-basin: a solution to improve groundwater quality. Irrig. Drain. 61, 75–86. Tournebize, J., Passeport, E., Chaumont, C., Fesneau, C., Guenne, A., Vincent, B., 2013. Pesticide decontamination of surface waters as a wetland ecosystem service in agricultural landscapes. Ecol. Eng. 56, 51–59. Tournebize, J., Chaumont, C., Fesneau, C., Guenne, A., Vincent, B., Garnier, J., Mander, Ü., 2015a. Long-term nitrate removal in a buffering pond-reservoir system receiving water from an agricultural drained catchment. Ecol. Eng. 80, 32–45. Tournebize, J., Chaumont, C., Marcon, A., Molina, S., Berthault, D., 2015b. Guide technique à l’implantation des zones tampons humides artificielles (ZTHA) pour réduire les transferts de nitrate et de pesticides dans les eaux de drainage. Irstea-ONEMA, 44 p. Vallée, R., Dousset, S., Schott, F.-X., Pallez, C., Ortar, A., Cherrier, R., Munoz, J.-F., Benoit, M., 2015. Do constructed wetlands in grass strips reduce water contamination from drained fields? Environ. Pollut. 207, 365–373. van der Valk, A.G., Jolly, R.W., 1992. Recommendations for research to develop guidelines for the use of wetlands to control rural nonpoint source pollution. Ecol. Eng. 1 (1–2), 115–134. Vymazal, J., 2005. Horizontal sub-surface flow and hybrid constructed wetlands systems for wastewater treatment. Ecol. Eng. 25 (5), 478–490. Vymazal, J., 2007. Removal of nutrients in various kinds of constructed wetlands. Sci. Total Environ. 380, 48–65. Vymazal, J., Bˇrezinová, T., 2015. The use of constructed wetlands for removal of pesticides from agricultural runoff and drainage: a review. Environ. Int. 75, 11–20. Vymazal, J., Greenway, M., Tonderski, K., Brix, H., Mander, Ü., 2006. Constructed wetlands for wastewater treatment. In: Verhoeven, J.T.A., Beltman, B., Bobbink, R., Whigham, D.F. (Eds.), Wetlands and Natural Resource Management. Ecological Studies, vol. 190. Springer, Berlin, Heidelberg, pp. 69–96. Wild, U., Kamp, T., Lenz, A., Heinz, S., Pfadenhauer, J., 2001. Cultivation of Typha spp. In constructed wetlands for peatland restoration. Ecol. Eng. 17 (1), 49–54. Yang, W., Chang, J., Xu, B., Peng, C.H., Ge, Y., 2008. Ecosystem service value assessment for constructed wetlands: a case study in Hangzhou, China. Ecol. Econ. 68 (1–2), 116–125. Yepsen, M., Baldwin, A.H., Whigham, D.F., McFarland, E., LaForgia, M., Lang, M., 2014. Agricultural wetland restorations on the USA Atlantic Coastal Plain achieve diverse native wetland plant communities but differ from natural wetlands. Agric. Ecosyst. Environ. 197, 11–20. Zhang, R., Krzyszowska-Waitkus, A.J., Vance, G.F., Qi, J., 2000. Pesticide transport in field soils. Adv. Environ. Res. 4 (1), 57–65.

Further reading Girard, M.-C., Walter, C., Rémy, J.-C., Berthelin, J., Morel, J.-L., 2005. Sols et environnement, Chapter 16 Les zones humides et leurs sols. In: Sols et environnement. P. Dunod, France.

Please cite this article in press as: Tournebize, J., et al., Implications for constructed wetlands to mitigate nitrate and pesticide pollution in agricultural drained watersheds. Ecol. Eng. (2016), http://dx.doi.org/10.1016/j.ecoleng.2016.02.014