Pesticide Behavior in Modified Water-Sediment Systems

J. Pestic. Sci. 41(4), 1–12 (2016) DOI: 10.1584/jpestics.D16-060

Review Article

Pesticide behavior in modified water-sediment systems

Toshiyuki Katagi*

Environmental Health Science Laboratory, Sumitomo Chemical Co., Ltd., 3–1–98 Kasugadenaka, Konohana-Ku, Osaka 554–8558, Japan (Received June 21, 2016; Accepted August 10, 2016)

The standardized laboratory water-sediment study in darkness is utilized as primary information on pesticide behavior to assess its ecotoxicological impacts in the edge-of-field water bodies. The half-lives of pesticide in water and sediment are key param- eters to predict its environmental concentration, and its metabolic profiles help to avoid overlooking unexpected toxicological impacts from metabolites. However, no consideration of environmental factors such as sunlight and aquatic macrophytes is included, and this may lead to a conservative assessment. We review the experimental factors in the existing standardized design and then the effects of illumination and aquatic macrophytes introduced to the water-sediment system. The effects of temperature and the water-sediment ratio should be investigated in more detail and the pesticide behavior is possibly modified by illumination via photodegradation and/or metabolism in phototrophic microorganisms. Aquatic macrophytes play a major role as an additional sorption site and in further pesticide metabolism. © Pesticide Science Society of Japan Keywords: water-sediment, illumination, photolysis, phototroph, aquatic macrophyte.

in these tests gives much information such as mass distribution, Introduction dissipation or degradation rates, profiles of metabolites, and the After the application of pesticide to the field, its entryvia spray extent of bound formation and mineralization, but such a study drift and runoff to the edge-of-field water bodies such as ditch- is very expensive with a long experimental period of up to about es, streams, and ponds could pose a toxicological impact on 3 months. Therefore, the screening method has been proposed aquatic organisms inhabiting them. These water bodies, whether instead, with reference to the ready biodegradability study.9–11) man-made or natural, show the differences in landscape, water Close examination of the available water-sediment stud- flow, physical and chemical properties of bottom sediment, and ies for pesticides and pharmaceuticals, however, has recently aquatic inhabitants.1–3) In order to assess the pesticide impact given rise to some concerns not only on the existing study de- on algae, aquatic invertebrates, and macrophytes, the laborato- sign but also how to mathematically describe their behavior in ry water-sediment study in darkness is primarily conducted as the system. The stagnant irrigation ditch model, represented by a convenient model to understand the pesticide behavior, and OECD 308, does not well simulate the lotic stream/river situ- its half-lives of dissipation or degradation. Data on the forma- ation where advection becomes more important for the dissi- tion percentage of each metabolite are additionally utilized to pation of a chemical.5,12) The low microbial activity in a labo- calculate the predicted environmental concentrations (PEC) in ratory overlying water may cause too conservative evaluation water and sediment.3) The standardized OECD 308 test simu- of biotic degradation,8) while the rapid adsorption of a hydro- lates a shallow or deep surface water respectively associated with phobic chemical to sediment could overestimate its dissipation aerobic or anaerobic sediment,4) and the effect of light may be in water if the water-sediment ratio is much lower than that additionally investigated for a photo-labile pesticide in the Euro- observed in the environment.12) The different dissipation rates pean registration framework.5) An aerobic mineralization study between the laboratory and field studies have been frequently (OECD 309) is utilized alternatively to simulate the behavior of reported for pesticides13) and pharmaceuticals,14) although their pesticide in the surface water of an open water body, and the metabolic profiles are similar. In a sunlit environment, unique effect of diffuse light may be examined if necessary.6,7) The appli- products may be formed by photolysis via rearrangement and cability of these standardized tests has also been discussed from radical reactions,13,15) and the presence of aquatic macrophytes the viewpoint of the PBT (persistent, bioaccumulative and toxic) and phototrophic microorganisms including algae not only gives assessment of pesticide.8) The usage of a radio-labeled pesticide an additional phase for adsorption but also causes further metabolic degradation of the dissolved pesticide.7,13) By the way, the * To whom correspondence should be addressed. transformation of pesticide proceeds in both overlying water E-mail: [email protected] and heterogeneous sediment, and concurrently pesticide and its Published online ●● ●●, ●●●● degradates are transported between two phases and diffuse in © Pesticide Science Society of Japan pore water under adsorption/desorption with sediment.13) The 2 T. Katagi Journal of Pesticide Science

dissipation half-life (DT50) of pesticide in water is easily calculat- diffusion process and basically controlled by physico-chemical ed, while it is troublesome to mathematically estimate its degra- properties of both pesticide and sediment.7,13) Furthermore, any 3,12,16) dation half-life (DegT50) in each phase due to this complex of a water flow, the presence of phototrophic biofilm at the inter- behavior and hence, the several practical mechanistic models face, and the shape of sediment surface may affect the transport have recently been proposed.17) of pesticide between water and sediment. The pond water-sedi- In this review, we briefly discuss the basic factors affecting ment interface covered with microbial mat was used to examine pesticide behavior in the existing water-sediment study design. the effect of a water flow on the diffusion of molecular oxygen.21) Thereafter, the chemical and biological effects of illumination As the water flow over the interface became faster from 0.3 to are respectively discussed from the viewpoints of photolysis as a 7 cm/sec, the diffusive boundary layer thickness decreased by a unique transformation pathway and the action of phototrophic quarter but the oxygen penetration depth in sediment increased algae and microorganisms in water and biofilm. The effect of markedly from 0.05 to 0.15 mm. Although the water-sediment aquatic macrophytes introduced into the illuminated system is interface is generally flat in many studies, the effect of its shape then discussed in relation to the additional adsorption and me- (flat or ripple) on the diffusion of seventeen pharmaceuticals -ap tabolism of pesticides. The behavior of pesticide in the aerobic plied to an overlying water was examined in darkness for the water-sediment system is examined through the survey of litera- Swedish lake water-sediment placed in a recirculating flume.22) ture and regulatory reports, keeping the above factors in mind. The concentration of fluconazole in the pore water at a 3-cm Finally, an overview summary is provided, including issues that sediment depth gradually reached that in the overlying water should be solved. when the interface was flat. A more rapid equilibrium of the concentration in the pore water under the stoss side of the rip- 1. Modification of the study design ple interface than that under the lee side was observed due to a 1.1. Basic experimental factors higher pressure from a water flow, indicating the surface shape The water-sediment ratio and/or the depth of each phase are de- of sediment also as the key factor in diffusion under the lotic fined in the several test guidelines,13) but their effects on pes- conditions. ticide behavior have been scarcely reported. The DegT50 values The experimental temperature is recommended in many test of rimsulfuron in the total water-sediment system showed 40% guidelines to be 20–30°C.7,13) It is frequently necessary to derive relative differences dependent on the depth of each phase, even the DT50 value of pesticide at a reference temperature from the by keeping the guideline water-sediment ratio of 4/1-3/1.17) A experimentally obtained one at a different temperature, not only ten-times increase in water depth reduced the degradation rate as the input parameter to simulate its PEC, but also for the nor- of chlorpyrifos in water to one-ninth, with much lower adsorp- malization of the field data obtained at daily fluctuating temper- tion of the insecticide to the bottom sediment.18) The lower the atures. The Arrhenius relationship was generally taken for this water-sediment ratio and the shallower the water depth, the purpose, and the activation energy (Ea) of 65.4 kJ/mol is recom- faster the observed pesticide degradation. These results indi- mended for an aerobic microbial metabolism in soil through the cate the importance of the pesticide transfer from water to the statistical analysis of the accumulated data.23) The water-sedi- more microbially active sediment phase. The agitation of a wa- ment system consists of two separated compartments connected ter-sediment system may increase the opportunity for microbes to each other by mass transport via diffusion, and either abi- suspended or inhabiting solid surfaces to access pesticide mol- otic or biotic reactions may undergo in each phase. Hydrolysis is ecules. Air introduction by either passing over a water surface or one of the main abiotic reactions, highly dependent on both pH gently bubbling the upper water phase showed insignificant ef- and temperature, with the median Ea value of pesticides around fects on the pesticide fate in the water-sediment system.13) How- 71 kJ/mol (34–135 kJ/mol, n=89, 62 pesticides)24) under alkaline ever, the usage of the aerator shooting jets of water by air, which conditions. The diffusion across the water-sediment interface caused a better mixing of water without sediment turbulence, is also a temperature-dependent process.25) Recently, the tem- enhanced the mineralization of p-nitrophenol by a factor of ca. perature effect on the dissipation of 2,4,6-trinitrotoluene (TNT) 3 against the gentle air bubbling.19) In the U.K. soil slurry of the has been examined in three marine water-sediment suspensions 26) mineral basal salt medium (1/3-1/1, w/v), the gentle shaking in (10/1, v/w) under intact and sterile conditions. The Ea values darkness greatly increased the mineralization of isoproturon and for the diffusion of TNT into the sediment phase were estimated cypermethrin as compared with the static incubation, but with to be 10–30 kJ/mol for two sediments but one did not follow the an insignificant or small effect of the water-sediment ratio.20) Arrhenius relationship. The biotic dissipation of TNT including

By the way, both the absorption of incident light by chromo- diffusion showed a wide Ea range of 28–103 kJ/mol. In the case phores of dissolved organic matter (DOM) and the scattering by of the stagnant water-sediment system under illumination at 5 suspended matters control the effective irradiance in water, as and 15°C, the Ea values for the mineralization of p-nitrophenol described in Section 1.2.1. Therefore, the water depth, that is a were estimated to be 13–33 kJ/mol.19) Since relevant information light path-length, should also be one of the most important fac- is very limited, the simple application of the Arrhenius relation- tors to determine an illumination effect. ship to the water-sediment system may be inappropriate for the The mass transport through the water-sediment interface is a present. Vol. 41, No. 4, 1–12 (2016) Pesticide behavior in modified water-sediment systems 3

− 1.2. Illumination of water-sediment system rower than that of [TSS], and HCO3 is present at higher concen- − Photodegradation of pesticide (p) can occur via direct and/or trations. [NO2] is much lower by 1–2 orders of magnitude than − indirect photolytic process. The former is reactions from excited [NO3], which may be greatly affected by fertilizer application. states of a pesticide molecule after light absorption, while the Natural water thus contains various types of particulates and latter is those with the photo-produced reactive oxygen species DOM, which reduce the penetration of solar radiation in a con- (ROS), the transfer of energy mainly from the excited triplet centration-dependent manner.42) The light attenuation via ab- state of colored DOM (3CDOM*) to a ground-state pesticide sorption by DOM is much more important than scattering by molecule, or the exchange of an electron or hydrogen between particulate solids in most natural water, and the former effect is them.15,27) By the way, the diurnal pH change of water in the field theoretically described by a light attenuation factor, α(λ), which is known to occur under sunlight as a result of photosynthesis generally increases with decreasing λ: ca. 0.1 (300 nm) and 0.04 by aquatic algae and macrophytes, possibly leading to photo- (600 nm) in an average for ten U.S. river waters.43) Theα (λ) val- synthetically-driven alkaline hydrolysis of some pesticides.7,13) ues were similar for U.S. lake and pond waters,44) and they can There are many kinds of phototrophic bacteria and algae in nat- be conveniently estimated by the concentrations of DOM and ural water, and biofilms not only at the water-sediment inter- suspended matters.28,45,46) S(λ) in Eq. 2 is a function of α(λ), and 28,29) face but also on the surface of aquatic macrophytes are known the larger the α(λ), the smaller the S(λ). The kdirect,p value can to be productive environments for many types of microbes and be conveniently estimated by GCSOLAR based on Eq. 2.43) In algae.7,13) Therefore, these organisms most likely participate in the case of a paddy field, the physical shielding of solar radia- the removal of contaminants, including pesticides from the wa- tion by the canopy of rice leaves becomes more significant with ter-sediment system under illumination. the growth of rice plant,46) and more shielding was measured 1.2.1. Photodegradation of pesticide after heading, with its extent also dependent on a solar zenith 47,48) The photodegradation rate constant of pesticide (kphoto,p), the angle. sum of direct (kdirect,p) and indirect (kindirect,p) photolysis one, can Since the photodegradation of pesticide in distilled water be described below.28) (DW) proceeds only via direct photolysis, the involvement of indirect photolysis in natural water can be conveniently evaluated kkkphoto,p=+ direct,p indirect,p (Eq. 1) when the light attenuation is minimal, by comparing kdirect,p in DW and the observed k (Eq. 1). The average D T values k =2.3Φ × {()Sλ ×× Zλ () ε ()} λ (Eq. 2) photo,p eg 50 direct,p p  p for each of six pesticides in sixteen sterile U.S. surface waters λ were much shorter than that in DW by a factor of 5 at a max- −⋅ kkindirect,p= ⋅ OH,p[]⋅OH ss+ k−⋅ [ CO3 ] ss 49) CO3 ,p imum, showing the involvement of indirect photolysis. The 13 (Eq. 3) ++13 kkO2 ,p[]O2 ss CDOM* ,p[ CDOM* ]ss role of indirect photolysis in the environmental fate has been recently reviewed for pesticides.50) The formation and scaveng-

Φp and εp(λ) are the quantum yield and molar absorption co- ing mechanisms for the reactive species relevant to indirect efficient of pesticide at wavelength λ, respectively. Z(λ) and photolysis are summarized in Fig. 1, based on the accumulated S(λ) are the averaged solar (or artificial light) irradiance and evidence.27–29) The similar mechanisms have been proposed on the light screening factor of natural water, respectively. In in- soil surface,15) but the presence of biofilm with excess water may direct photolys is, kx,p is the reaction rate constant of a species alter such reactions at the water-sediment interface. The most −· X (X=hydroxyl radical ·OH, carbonate radical CO3 , singlet reactive species is ·OH attacking C-H bonds in a relatively non- 1 3 oxygen O2 and CDOM*), and [X]ss means its steady-state con- selective manner at diffusion-controlled rates, and it most likely centration in a water body. The typical concentration ranges of participates in the photodegradation of pesticide. The photo- − − anions, dissolved organic carbon (DOC) and total suspended induced homolytic cleavage of the N-O bond in NO3 and NO2 solids (TSS) in natural water, controlling the photolysis of pes- producing ·OH was confirmed by formation of nitro and ni- ticide, are listed in Table 1. TSS, controlling the light penetra- troso derivatives in the photolysis of chlortoluron51) and mono- 52) −· tion by scattering, shows a wide variation of the concentration, linuron. CO3 is more selective than ·OH toward electron-rich depending on a water body. The range of [DOC] is slightly nar- S- and N-containing chemicals,53) and the negative correlation of

Table 1. Water chemistry of surface freshwater

− − − Type NO3 NO2 HCO3 DOC TSS ref. Lake 0.2–20 0.02 24–270 1.6–7.5 0.8–1.2 29–32 Creek 0.1–13 — 226–273 7.0–15 20–72 30, 31, 33 River 0.1–26 0.002–1.6 10–269 0.5–12 <0.5 19, 30, 31, 34–36 Paddy 0.05 — — 1.8–8.9 — 37, 38 Wetland 0.2–3.6 — — — 44 39–41 Unit: ionic species and TSS (total suspended solid) in ppm; DOC (dissolved organic carbon) in mg C/L. “—”, not available. 4 T. Katagi Journal of Pesticide Science

chloride, isopropanol and benzene derivatives, 10−14–10−19M for 30,56,59) −14 −15 −· 31,59) ·OH; N,N-dimethylaniline, 10 –10 M for CO3 ; furfuryl alcohol, 2,5-dimethylfuran, and 1,4-diazabicyclo[2,2,2] −12 −13 1 32,59,60) octane, 10 –10 M for O2; and 2,4,6-trimethylphenol and trans,trans-hexadienoic acid, 10−15 M for 3CDOM*.32,59)

The kx,p value has been determined for many pesticides by

the competition reaction method against kx,p-known chemicals or probe molecules: 109–1010 /M/sec (·OH),59,61–63) 105–108 −· 54,59,64) 5 1 29,59) 6 9 /M/sec (CO3 ), 10 /M/sec ( O2) and 10 –10 /M/sec 3 59) ( CDOM*). The kindirect,p value of a pesticide can be predicted 28,29) according to Eq. 3 by using [X]ss and kx,p values, and the corresponding estimation program is available.65) Though the inclu- sion of indirect photolysis is very limited to simulate the fate of pesticide in the water-sediment system, the computer model EXAMS (Exposure Analysis Modeling System) considering the reaction with ·OH has been successfully applied.1,66) Since the

[X]ss values highly depend not only on the natural water to be Fig. 1. Formation (solid lines) and dissipation (dotted lines) profiles assessed but also on irradiance, the relative contribution of di- of reactive species participating in indirect photolysis of pesticide in the rect and indirect photolysis was more conveniently examined aquatic environment. ISC, inter-system crossing; CDOM, colored dis- for sixteen pesticides in the sunlit U.S. surface waters by using solved organic matters with the superscripts 1* and 3* meaning the excit- 59) n+ the probe molecules as quenchers. Direct photolysis predomi- ed singlet and triplet, respectively; M , metal cation; MxOy, metal oxides. nated for dinitroaniline herbicides, but more than half of kphoto,p 3 was attributed to kindirect,p for other pesticides with CDOM* and + 1 −· −· log kCO3 with the Hammett σ was reported for anilines, phen- O2 as the main contributors. Both ·OH and CO3 played some 54) oxides and phenylureas. The DegT50 values of six pesticides role in the photodegradation of several of the pesticides, and in the U.S. surface waters negatively correlated in the Pearson’s the quenching of these species by CDOM at high concentrations − − analysis with either [NO3] or [HCO3], but there was no correla- (26–38 mg C/L) may account for their insignificance. 49) tion with [CDOM]. In contrast, the kphoto,p values of atrazine 1.2.2. Photo-induced pH change and 2,4-D in four U.S. basin waters were proportional to either In the laboratory system consisting of the German river water- − 55) [NO3] or [CDOM]. These results show that the two anions act sediment (10/1, v/v), the insignificant pH change of water was as a radical source and that CDOM has the dual character of a reported under exposure to a fluorescent light limiting the sig- radical scavenger and a photo-sensitizer (Fig. 1). nificant algal growth.19) In contrast, the light-dependent pH − [·OH]ss was found to be proportional to [NO3] and decreased change between pH 8 and 10 was observed in the U.K. lake with increasing [CDOM] in natural water at low alkalinity, and water-sediment system (3/1, v/w) after incubation at >400 nm − 2− it became lower due to the scavenging effects of HCO3/CO3 at (16 hr light per day) for 4–5 weeks, and the addition of Elodea higher alkalinity.29,30,56) Through the photolysis of benzene de- canadensis enhanced its change.67) The other stream water- rivatives as probe molecules, it was found in the ·OH produc- sediment system (10/1, v/w) in the presence of the water milfoil − − tion that the relative contribution of NO2 and NO3 highly de- (Myriophyllum spicatum) showed an insignificant pH change of pended on the river water quality35) and that CDOM dominantly water at >290 nm, but the absence of sediment caused an al- 35,57) 68) participated via H2O2-independent mechanism. Some ionic kaline shift of pH to ca. 9. Therefore, not only the growth of species of metal hydroxides, such as [Fe(OH)]2+, participated photosynthetically active algae and microorganisms but also in the photo-induced Fenton reaction producing ·OH.13) Fer- the presence of sediment likely controls the photo-induced rous/ferric ions in montmorillonite could produce ·OH under pH change of water. The pH of paddy water changed diurnally, illumination, showing the importance of clay minerals for in- with its maximum (9–10) at noon, and the CO2 concentration direct photolysis,58) but its contribution may be lessened by the showed the opposite profile.69–71) These changes are pronounced −· light scattering effect. [CO3 ]ss is known to be a product of not in the early growing season due to less shading by the rice can- − 2− − 31) 72) only [HCO3+CO3 ] but also [NO3] and [CDOM]. The pho- opy. A similar pH change was observed in the outdoor tank to-induced formation rates of 3CDOM* in five Italian lake wa- microcosm consisting of the U.K. lake water-sediment system 1 ters were found to be higher than those of O2 by a factor of with about 30% coverage of its water surface by aquatic macro- 2–5, and the suspended as well as dissolved organic molecules phytes.34) with a size smaller than 0.1 µm only contributed to their for- 1.2.3 Contribution of phototrophic microorganisms 32) mation. [X]ss in natural water can be estimated by the reac- Various kinds of phototrophic algae and microorganisms exist tions with probe molecules whose rate constants for each reac- in biofilm at the water-sediment interface. The DNA fragments tive species have been determined: probe(s), [X]ss for X; n-butyl were isolated from the Norwegian meromictic lake water sedi- Vol. 41, No. 4, 1–12 (2016) Pesticide behavior in modified water-sediment systems 5 ment and amplified by the polymerase chain reaction for de- ing the possible release of some photo-sensitizers from the algae. naturing gradient gel electrophoresis (DGGE) analysis.73) The By the way, the algal degradation of aniline under irradiation presence of several cyanobacteria was confirmed in the biofilm at >350 nm was faster in the presence of dead cells than live (top 3-mm sediment layer) by the homology of 16S ribosomal ones for four species.81) Aniline was most slowly degraded in RNA fragments. In relation to the metabolism of fludioxonil the presence of the live diatom Nitzschica hantzschiana, having by phototrophic communities, several green algae, cyanobacte- a hard cell wall, and its degradation with the live cyanobacteria and diatom species were identified in a U.K. lake sediment rium Anabaena cylindrica was greatly suppressed under an N2 surface.74) The phototrophic biofilm, collected from the Chi- atmosphere. Under irradiation, aniline was insignificantly taken nese water treatment plant and lake, showed site-specific algal up by any algal species, while the carboxylic acids and pigments, inhabitation with diatoms predominant (38–66%), followed by such as chlorophylls, characterized by absorption and IR spec- green algae (17–44%) and cyanobacteria (3–14%).75) About half tra, were released from these algae. The authors confirmed the 1 of a U.K. river sediment surface, reconstituted in the illuminated photo-induced formation of O2 and ·OH, indicating that these fluvarium channel, was covered by several diatom species for chemicals released from algae act as photo-sensitizers to pro- 4 weeks, with the remaining 20% by filamentous algae.76) The duce ROS. DGGE analysis confirmed various Proteobacteria in the biofilm, Since many herbicides are generally applied to paddy fields colonized on unglazed ceramic discs dipped into the Taiwanese and may contaminate the tail waters, their behavior in this man- river, throughout all seasons, in addition to forty diatom species made outdoor water-sediment system should be examined to in fourteen genera.77) Although about four times algal biomass assess the impact on the surrounding aquatic environment. was measured in summer as compared with spring, and the About forty varieties of aquatic organisms including algae and double for bacteria in autumn than spring, the degradation rates duckweed were collected from the waters of Japanese paddy of three carbamate insecticides by the biofilm did not correlate field plots, and the appearance frequency of Lemna, Closterium, with these biomass changes. Furthermore, the degradation rate and Euglenida spp. was greatly affected by agronomical prac- of alachlor was dependent on the microbial consortia collected tices such as fertilization.82) Chlamydomonas and Nitzschia spp. from the river and sewage treatment plant and not proportion- appeared dominantly in Japanese paddy water after flooding, al to the autotrophic index defined as the Chla content divided but their population gradually decreased with rice growth due by the ash-free dry mass of the biofilm.78) These results show to less irradiance interfered by the rice canopy.83) In the cor- that the biofilm is abundant in various phototrophic algae and responding paddy soil, green algae were the major species fol- microorganisms but that specific species among them are most lowed by cyanobacteria, but with diatoms as a minor species. likely to be involved in the degradation of pesticide. From ten sites of Korean fields, fourteen genera of cyanobacteria Light with wavelengths of >400 nm, where most pesticides were isolated, among which Nostoc and Anabaena spp. were pre- have no absorption, is generally used in the degradation study dominant.84) The DGGE analysis for 275 bacterial isolates from by phototrophic microorganisms to exclude the effect of photo- Japanese paddy water showed that α-Proteobacteria constantly degradation.67) In a German river water-sediment system under appeared with the bacterial inhabitants varying during cultiva- illumination (12 hr per day) and air bubbling, the time for 10% tion.85) Therefore, the paddy field is abundant in various aquatic mineralization of p-nitrophenol was 4 days at 20°C and 2000 lx microorganisms including algae, and the contribution of these but longer (17–23 days) at 15°C and 300 lx.19) The higher irradi- phototrophs on pesticide degradation is most likely in sunlit ance likely enhanced the mineralization if the temperature ef- shallow water. fect was calibrated by the Arrhenius relationship.23) In contrast, The partition of pesticide to biofilm and algae from water, fol- the four-times difference in fluorescent light intensity showed an lowed by metabolism, can be conveniently estimated by its log insignificant effect on the biodegradation of isopyrazam in U.K. Kow. The presence of hydrophilic extracellular polymeric sub- lake water inoculated with sediment microorganisms.79) The stances in biofilm caused a weaker response in the adsorption 7) growth of specific degraders under illumination may account coefficient of pesticide against log Kow. The low correlation be- for the different effects of light. The relative contribution of -pho tween log BCF (bioconcentration factor) in algae and log Kow todegradation and algal metabolism was examined for seven was reported for pesticides (r2=0.28).86) The existing 69 BCF polycyclic aromatic hydrocarbons (PAHs) by using two green values of 43 non-polar pesticides in 24 algal species were re- algae, Selenastrum capricornutum and Chlorella sp., under expo- analyzed separately for cyanobacteria, green algae, and diatoms, sure to white (fluorescent lamp at 310–750 nm) or gold (incan- but the correlation coefficients were only slightly improved to descent lamp at 400–780 nm) light, both of which gave similar 0.32–0.46 (data not shown). Biodegradation of pesticide in bio- cell growth.80) The comparable transformation of photo-stable film and various algae via phase-I and -II reactions with species PAHs by the live algae irrespective of the light source indicated differences has been reported7,86) and at least in part accounts for the importance of the algal metabolism. In contrast, the dead these lower correlations. cells markedly enhanced the photodegradation of benzo[a] 1.3. Introduction of aquatic macrophytes anthracene and benzo[a]pyrene even after 1-day exposure under Aquatic macrophytes are classified based on their ecology into the white light with formation of oxidized photoproducts, show- emergent (rooted in sediment with foliage extending into air), 6 T. Katagi Journal of Pesticide Science floating-leaved (rooted in sediment with leaves floating on the 2. Effect of photolysis water surface), submerged (roots attached to or associated with The illumination effect on the DT50 values of pesticide in the substrates and all parts under water) and free-floating (floating water-sediment system is summarized in Table 2. Either a fluo- on or under the water surface) types.87) In the case of floating- rescent tube with a maximal emitting wavelength of 300–400 nm leaved and free-floating macrophytes if densely developed, or a Xenon arc lamp equipped with cut-off filters (>290 nm) is the dissipation of pesticide susceptible to photodegradation in generally utilized as a light source. The illumination effect on water may be significantly retarded due to the light shielding DT50 should be first considered not only from the photolytic by their leaves, as demonstrated for several pharmaceuticals in susceptibility of pesticide but also from its hydrolytic instability 88) w hydroponic plant reactors. Rooted macrophytes may change and adsorption to sediment. Longer water-phase DT50 (DT50 ) the redox profiles of bottom sediment around roots via radial values under illumination than in darkness were observed for oxygen loss, whose extent is dependent on the root length89) and photo-stable benalaxyl104) and metalaxyl-M,68) which may indi- hence, they may modify the redox reactions and microbial me- cate reduced microbial activity by light but without any contri- tabolism of pesticide in the sediment. bution of photolysis. The comparative studies of clomazone98) More rapid dissipation of prosulfocarb resistant to photoly- and thiamethoxam68) in the water-sediment and DW, under il- sis was observed in an outdoor ditch study than in a laboratory lumination or in incubation in darkness, showed that indirect water-sediment system, and it was successfully simulated by photolysis was more predominant than direct photolysis, but TOXSWA inverse modeling when its sorption to macrophytes still with 2–3 times higher rate constants of microbial degrada- was considered.90) The partition profiles of nine non-polar pes- tion than photodegradation. Photo-induced geometrical isom- ticides were studied for a day by using Chara globularis, Elo- erization but without formation of the intramolecular cycliza- dea nuttallii, and Lemna gibba, whose organic matter contents tion product was reported for fluoxastrobin in the illuminated (o.m.c.) were 36, 86, and 87%, respectively.91) Atrazine and linu- water-sediment and hence, its rapid partition to the sediment ron showed non-linear adsorption isotherms, while the other was considered to result in less contribution of photolysis.105) w pesticides followed a linear one. The o.m.c.-normalized linear The hydrophobic esfenvalerate rapidly dissipated with a DT50 of adsorption coefficient was negatively correlated with the water <1 day, but no clear effect was observed on the DT50 value in 2 sys 101) solubility of each pesticide (r =0.86). The eight-day concentra- the water-sediment system (DT50 ) under illumination. Since tion factors (CFs) of five pesticides were reported for Elodea the photo-induced decarboxylation product was not detected, crispa in outdoor tank microcosms.92) Although no uptake of its rapid adsorption to the sediment caused much less effect of mecoprop was observed likely due to its complete ionization in photo-degradation than microbial degradation. The moderate 93) the tested water (pKa=3.86 ), the log CF values of the other adsorption of pyraclostrobin to the sediment resulted in almost 13) 2 undissociated pesticides correlated well with log Kow (r =0.99, the same contribution of photo- and microbial degradation with 14 106) n=4). The relative uptake rate constants (kup) of six C-labeled three photoproducts detected in the illuminated system. phenol derivatives by the axenic L. gibba were estimated by the Metaflumizone (log Kow=4.2–4.9) showed a rapid partition to sys compartmental model kinetic analysis considering the phase- sediment, and the DT50 under illumination was more than fifty II conjugations and bound formation, and the low correlation times shorter via rapid geometrical isomerization followed by

(r=0.55) of log kup with log Kow was slightly improved (r=0.76) the formation of photoproducts than that in darkness, although when the fraction of an undissociated species was taken into ac- different water-sediment systems were used.105) The direct pho- 94) count for each phenol (pKa=7.15–9.98). Approximately ten- tolysis of fludioxonil in water is moderate with a half-life of 10 times uptake of phenol and benzoic acid derivatives through days, but DT50 values in both water and sediment were great- the shoot and leaves of Myriophyllum elatinoides was observed ly reduced by illumination at >290 nm.105) Greater formation by the water exposure, as compared with their root uptake by of bound residues and higher mineralization were observed 95,96) the sediment exposure. In the water exposure of this water under illumination for these pesticides. The DT50 values of uni- 14 milfoil to C-phenol derivatives, the correlation of log kup, es- conazole-P were greatly reduced by photo-induced E/Z isom- 100) timated similarly to the duckweed study above, with log Kow erization followed by intramolecular cyclization. However, was slightly higher (r=0.66) but was not improved by consid- the kinetic analysis using the rate constants of the photo- and eration of log D instead.96) Re-analysis of the log BCF values86) microbial degradation underestimated both its dissipation and for twelve non-polar pesticides in nine species of aquatic mac- the formation of photoproducts in the experiment. These results 2 rophytes showed the low correlation with log Kow (r =0.43), and may imply additional photodegradation proceeding at the illu- a significant variation of BCF up to two orders of magnitude minated water-sediment interface. was observed for each pesticide among these species. These re- In addition to the microbial metabolism of chlorothalonil sults indicate that hydrophobicity of pesticide is a primary fac- in darkness, its significant UV absorption with maxima at 314 tor, but the dissociation and species differences, for example in and 325 nm resulted in the accelerated dissipation from water the o.m.c. and metabolic activity, additionally control the BCF in via direct photolysis.97) Since the microbial degradation is aquatic macrophytes. slow, direct photolysis plays a major role in the dissipation of imidacloprid105) and oxolinic acid108) in illuminated systems. The Vol. 41, No. 4, 1–12 (2016) Pesticide behavior in modified water-sediment systems 7

Table 2. Laboratory water-sediment study of pesticide in darkness and under illumination

2) 3) 6) Sediment Water DT50(L/D)

1) 4) 5) Pesticide log Kow type oc% pH DOC TSS Cond Illumination system water ref.

a Chlorothalonil 2.9 C(5/1) — 7.6 — — 30/20 F, 340(max)/C — <1/3 97 a Po(5/1) — 7.6 — — 30/20 F, 340(max)/C — <1/2.1 97 Isopyrazam 4.1–4.4 L(5/1)c 4.2 — 18.3 15 60/20 EC, >400/16 10.8/228 2.5/5.3 44 b Clomazone 2.5 Pa(4/1) 0.41 — — — 35/30 U, 300(max)/C 15.7/23.6 — 98 a Ethofenprox 6.9 Pa(2/1) 0.31 — — — 29/28 U, 300(max) /C 3/— — 99 a Uniconazole-P 3.8 Po(3/1) 6.8 8.0 19.8 8.0 30/20 X, >290/8 1.4/109.3 0.6/6.9 100 a Esfenvalerate 6.2 Po(3/1) 5.3 7.1 — 170 15/20 X, >290/8 11.9/25.5 — 101 a Po(3/1) 2.7 8.4 — 140 7/20 X, >290/8 13.9/7.4 — 101 Pyriproxyfen 5.4 L(2/1)a 3.4 6.9 7.8 10 30/20 X, >290/12 3/9 — 102 a Flumioxazin 2.6 Po(3/1) 1.9 6.9 — 43 30/20 X, >290/8 2.1/3 1.2/1.6 103 L(3/1)a 4.9 7.9 — 16 30/20 X, >290/8 0.2/0.2 0.2/0.2 103 c Benalaxyl 3.7 Re(4/1) — — — — 90/25 F, —/C — 40/20 104 Chlortoluron 2.5 L(3/1)b 5.5 7.7 12.1 4.8 28/20 F, >400/16 28/106 — 67 Lufenuron 5.1 L(3/1)b 5.5 7.7 12.1 4.8 28/20 F, >400/16 — — 67 Pinoxaden* −1.1 L(3/1)b 5.5 7.7 12.1 4.8 28/20 F, >400/16 86/129 — 67 Prometryn 3.3 L(3/1)b 5.5 7.7 12.1 4.8 28/20 F, >400/16 51/225 — 67 Propiconazole 3.7 L(3/1)b 5.5 7.7 12.1 4.8 28/20 F, >400/16 10/71 — 67 Fludioxonil 4.1 L(3/1)b 5.5 7.7 12.1 4.8 28/20 F, >400/16 3/64 — 67 a Po(3/1) 4.8 — — — 100/20 X, —/12 19/>1000 1.7/6.7 105 a Ri(3/1) 1.2 — — — 100/20 X, —/12 25/>1000 1.8/6.4 105 Benzovindiflupyr 4.3 L(—) 0.9 6.7 — 3.1 100/20 F, —/16 83/676 7/39106 L(—) 4.5 7.9 — <0.1 102/20 F, —/16 283/463 4/19 106 Saflufenacil 2.6 — — — — — — 3.6/71 2.8/56 107 Imidacloprid 0.57 —(9/1)b 2.1 — — — 30/22 X, >290/C — <5/>30 105 b Fluxaproxad 3.1 Ri(3/1) 4.3 7.3 3.2 — 57/20 X, >290/13 145/417 10/7 106 S(3/1)b 2.7 7.2 3.1 — 57/20 X, >290/13 116/694 13/12 106 b Oxolinic acid 0.95 Po(10/1) 2.6 8.1 — — 72/rt F, —/C 9.5/98.7 — 108 Thiamethoxam −0.13 S(2/1)#a 5.0 8.3 — — 56/20 EC, —/16 — 12/16 68 Metalaxyl-M 1.75 S(2/1)#a 5.0 8.3 — — 56/20 EC, —/16 — 33/28 68 1) Application to water. *, de-esterified diketone metabolite. 2) Type: C, creek; L, lake; Pa, paddy; Po, pond; Re, reservoir; Ri, river; S, stream.# , stream sediment plus M4 medium solution. oc%, organic carbon %. The value in the parentheses is the water-sediment ratio (a, v/v; b, v/w; c, w/w). 3) DOC and TSS in ppm. 4) Experimental period in day/temperature (°C). rt, room temperature. 5) Light source of illumination (EC, environmental chamber; F, fluorescent bulb; U, UV lamp; X, xenon lamp), emitted wavelength in nm/irradiation period (C, continuous; x, x hr/day). 6) DT50, light/dark in day. “—”, not available.

− hydrolytically unstable flumioxazin at an environmentally rel- its half-life inversely proportional to [NO3], most likely via radi- w sys 34) evant pH resulted in insignificant differences in 50DT and DT50 cal reaction with ·OH. In the case of ethofenprox, the ester between light and dark conditions, but the major formation of derivative was formed via photo-induced oxidation at the benzyl the azetidine-2-one derivatives in the light showed a greater carbon, showing the involvement of radical oxidation.99) contribution of photolysis than biotic degradation.103) Cyan- In an outdoor tank microcosm with the water-sediment ratio sys traniliprole is unstable at alkaline pH or under illumination in of 6/1(v/v), the DT50 values of isoproturon, chlorpyrifos, and natural water, but the usage of different water-sediment systems permethrin in the formulations were determined to be 48–65, makes it difficult to determine which route predominates under 20 and ca. 4 days, respectively.92) The effect of sunlight was illumination.105) By the way, indirect photolysis was most likely minimal for isoproturon and chlorpyrifos from the comparable sys for the following pesticides in the illuminated water-sediment DT50 in the laboratory water-sediment studies, but significant sys sys system. The DT50 value of pyriproxyfen was shortened by a fac- for permethrin showing much longer laboratory DT50 of ca. 12 tor of 3 with formation of not only the hydroxylated derivative weeks.13,93) Both the susceptibility of permethrin to photodeg- at 4-position of the terminal phenyl ring but also the photo- radation and its residence near the water surface after the spray cleaved product followed by oxidation.102) The photodegradation application to water13) are likely to result in these differences. of isopyrazam at >290 nm was more rapid in natural water with In a greenhouse marsh water-sediment system (4/1, v/v) with 8 T. Katagi Journal of Pesticide Science the flooded/ebb cycles, fipronil underwent photo-induced -de Both fludioxonil and propiconazole dissipated more rapidly sulfinylation with more oxidation at the sulfinyl moiety than in in the illuminated system than in darkness with their residues darkness, indicating the involvement of both direct and indirect lower both in water (0–12%) and sediment (40–65%), showing photolysis.109) The photodegradation rate constant of chlortol- the importance of both sorption and metabolism. The effect of − − uron increased in DW with [NO3], while the effect of NO3 was illumination was minimal for lufenuron, which is resistant to reduced by half in the outdoor tank microcosms at the corre- both photolysis and biotic degradation. The effect of algae was sponding concentrations.51) The formation of its nitrated prod- most likely on benzovindiflupyr, which is resistant to hydroly- − ucts in the field was only observed at the higher [NO3], so that sis and photolysis, since more rapid dissipation was observed in DOM in natural water likely scavenged the produced ·OH and the two illuminated water-sediment systems with formation of 106) NO2· radicals. The formation of oxidized products was reported hydroxylated derivatives. By the way, the contribution of pho- for rice pesticides such as thiobencarb in the field studies, most totrophic communities isolated from a U.K. lake water-sediment probably via indirect photolysis.110) was examined for degradation of fludioxonil.74) The microbial The literature survey clearly shows that not only the photo- communities showed a 75% degradation of fludioxonil in 7 days chemical reactivity of pesticide but also its adsorption to sedi- under illumination at >400 nm (18 hr per day) but only 7% in ment determines its DT50 values under illuminated conditions, darkness, and among them heterotrophic bacteria and fungi while the relevant mechanism controlling the pesticide behavior showed no degradative activity. The algal fraction showed a 50% is difficult to evaluate due to insufficient product identification degradation in 7 days with higher activities in Chlorophyta and and the lack of information on the quality of an overlying water. Cyanophyta, from each of which eight and four species degrad- These data are necessary to determine whether direct or indirect ing 48–80% of this fungicide were isolated as pure cultures. photolysis operates in the photodegradation of pesticide and to Therefore, these results show that phototrophic algae and micro- identify the relevant reactive species in any indirect photolysis. organisms are the key contributors to pesticide degradation in Furthermore, the experimental design such as an application the illuminated water-sediment system. method, the characteristics of water-sediment system, and cli- mate should be carefully taken into account in the interpretation 4. Effect of aquatic macrophytes of outdoor microcosm studies. Laboratory illuminated water-sediment studies examining the effect of aquatic macrophytes are more limited, as summarized 3. Effects of algae and phototrophic microorganisms in Table 3. The presence of E. canadensis insignificantly affected w sys It should be noted that the contribution of phototrophic micro- the DT50 value of isopyrazam at >400 nm, and the DT50 in- organisms to pesticide behavior cannot be appropriately evalu- creased from 11 to 37 days due to a lower metabolism in the ated unless the water-sediment system is pre-incubated under il- sediment.34) An outdoor tank microcosm study with 30% of the lumination necessary for their growth. No illumination effect on water surface covered by several macrophyte species showed w sys DT50 but a shorter DT50 than that in darkness was reported for DT50 values similar to those in the illuminated non-macrophyte photo-stable fluxaproxad in the pre-incubated system under il- laboratory system. By the laboratory introduction of E. canaden- 106) w lumination, which likely indicates the adsorptive and/or met- sis the insignificant effect on the 50DT of benzovindiflupyr was abolic contribution of phototrophic microorganisms. The pho- similarly observed, but its dissipation in the total system was tosynthesis-driven pH elevation by phototrophs such as algae accelerated by a factor of 2 to 4 with greater formation of the w sys 106) may account for the much shorter DT50 and DT50 under illu- hydroxylated derivatives and sediment bound residues. In the mination for saflufenacil, which is photo-stable but unstable at outdoor microcosm with aquatic macrophytes, the fate profiles alkaline pH.107) In the illuminated fluvarium channel, more hy- of this fungicide were similar to those in the corresponding lab- drolytically unstable trans-permethrin was less distributed in the oratory system but with slightly slower dissipation in water. The biofilm than thecis -isomer,76) being accounted for by the above different effects for two structurally analogous fungicides may mechanism. Under the non-UV light exposure at >400 nm, the be at least in part accounted for by the sorption and metabolism 14 DT50 values of C-isopyrazam markedly decreased and the fun- by macrophytes and/or the population of fungicide-degrading gicide amounted to only 30% of 14C in the sediment, indicating microorganisms specific to each microcosm. In the compara- its sorption and metabolism by algae at the sediment surface.34) tive laboratory study using chlortoluron, prometryn, propicon- 67) sys Comparative studies using the fluorescent light at >400 nm have azole and fludioxonil, the effect of E. canadensis on the DT50 recently been conducted for several 14C-pesticides.67) Photosyn- was found to be minimal and each residue in the extractable thesis-driven pH elevation caused the very rapid hydrolysis of 14C from the macrophyte after 28 days greatly varied with pesti- pinoxaden with a slight effect of algae on the dissipation of the cide (19–83%). The adsorption and/or uptake of thiamethoxam sys hydrophilic diketone metabolite. The DT50 values were reduced and metalaxyl-M to M. spicatum were evident from the shorter w 68) to about one-fourth by the illumination for chlortoluron and DT50 values than those in the absence of the macrophyte. prometryn, each of which was the main component (72–87%) in In the laboratory water-sediment systems with each 14C-pesti- the sediment, indicating the increased sorption to sediment but cide applied to water,34,67) the radioactivity in the macrophytes with less contribution of the algal and microbial metabolism. reached a maximum (5–68%) mostly after 7 days, indepen- Vol. 41, No. 4, 1–12 (2016) Pesticide behavior in modified water-sediment systems 9

Table 3. Effect of macrophytes in laboratory water-sediment study of pesticide under illumination

14 2) 3) C distribution DT50 Ratio Pesticide mac1) A: sed B: sed, mac system water ref. Isopyrazam E (4∼6) 80(60) 40(26), ∼10(12) 0.29 1.1 34 Chlortoluron E (∼4) 72(28) 45(28), 31(7) 1.2 — 67 Lufenuron E (∼4) 84(28) 36(28), 66(7) — — 67 Prometryn E (∼4) 76(28) 48(28), 28(7) 1.1 — 67 Propiconazole E (∼4) 78(28) 75(28), 5(7) 1.1 — 67 Fludioxonil E (∼4) 67(28) 50(28), 24(7) 1.0 — 67 Benzovindiflupyr# E (—) 84(61) 77(61), 5(7) 2.0 1.2 106 E (—) 73(60) 55(102), 41(46) 4.1 — 106 Thiamethoxam M (0.4) — — — 12 68 Metalaxyl-M M (0.4) — — — 2.6 68 The experimental conditions are same as described in Table 2.# , two water-sediment systems. “—”, not available. 1) mac, macrophyte. E, Elodea Ca- nadensis; M, Myriophyllum spicatum. The value in the parentheses is the ratio (%) of macrophyte (g, wet weight)/sediment (g, dry weight). 2) Maximum distribution percent of the applied 14C in sediment (sed) and macrophyte. A, water-sediment system ; B, A+macrophyte. The value in the parentheses is the period of incubation in day. 3) DT50 (system A)/DT50 (system B).

dent of their log Kow (2.5–5.1), followed by the gradual de- concrete channel consisting of a German water-sediment (ca. crease with half-lives of >1-2 months. The distribution profiles 2/1, v/v) vegetated with E. nuttallii, and each concentration in of benzovindiflupyr in the macrophyte were dependent on the water, sediment, and macrophytes in the up- and downstream origin of the water-sediment,106) and the macrophyte residue sites was monitored under a flow rate of ca. 10 cm/sec for 2 hr.111) rapidly decreased for hydrophilic thiamethoxam.68) The early The percentage of mass removed by the vegetated streams cor- 2 uptake of pesticide to macrophytes followed by its gradual related well with the log Kow of each pollutant (r =0.84), and desorption has also been reported in the tank microcosm for their major sink was macrophytes (7–34% of the applied mass) benzovindiflupyr106) and in the stream mesocosms for three fun- but with about 1% retention in the sediment. In the same sys- gicides.111) Less adsorption of pesticide to sediment in the pres- tem applied with the mixture of four pesticides, each aqueous ence of macrophytes, as listed in Table 3, and the accumulated concentration at the inlet was reduced by more than 90% at the evidence as above show that the adsorption/desorption of pes- outlet, and the multi-regression analysis showed that the re- ticide proceeds for the macrophytes, concomitant with its me- duction rates were positively and negatively correlated with the tabolism. aquatic plant density and the pesticide water solubility, respec- The role of the associated periphyton on E. canadensis was tively.112) The longitudinal dispersion coefficients of penflufen, examined for 14C-fludioxonil, by using an axenic macrophyte pencycuron and triflumuron in this system, estimated by the successively washed with sterile Hoagland’s medium and 0.1% one-dimensional advection-dispersion solute transport model, NaOCl and periphyton separately collected on polypropylene increased by a factor of 2–4 in the presence of E. nuttallii, likely plates in the water-sediment system for 28 days.74) More than via wake turbulence behind stems and leaves along with a water 70% of the fungicide in water was degraded in 7 days by the flow.113) On the reduced peak concentration of each pesticide, macrophyte, irrespective of sterilization, and periphyton. Since the contribution of dispersion decreased from 90% to 60–70% much less 14C was distributed after 6 hr in the macrophyte by the elongation of the macrophyte height from about half to (1–2%) than in water (62%) and periphyton (36%), the periphy- all of the water depth, and the adsorption effect became more ton may be more involved in the biotic degradation of pesticide important for more hydrophobic pesticides. In a 50-m-long in an aquatic environment. The dissipation half-lives of the an- vegetated ditch applied with atrazine or λ-cyhalothrin assum- ti-inflammatory drug ibuprofen (pKa=4.3) under illumination ing the worst-case storm runoff (5%) event, about one-month of were 16–44 days in the full-strength growth solution containing monitoring showed that the aquatic plants were their main sink, each of four free-floating and submerged macrophytes.88) From followed by sediment, especially for the more hydrophobic pyre- not only the moderate photo-stability of ibuprofen but also the throid.114) Bifenthrin and λ-cyhalothrin were introduced to the low uptake of its ionized species to the plants at neutral pH, the 650-m ditch system mainly vegetated with Ludwigia sp., suppos- possible participation of phototrophic periphyton was supposed ing a storm runoff (1%) event, and their aqueous concentrations for the oxidative degradation of this drug. decreased in proportion to the distance from the inlet with their The importance of aquatic macrophytes controlling pesticide masses dominantly partitioned to the aquatic plant.115) These in- behavior has been reported in natural water-sediment systems vestigations indicate that aquatic macrophytes are potentially a such as streams, ditches and constructed wetlands. A mixture more important sink for pesticides than the bottom sediment in of three fungicides and two biocides was applied to a 45 m-long vegetated lotic streams and ditches. 10 T. Katagi Journal of Pesticide Science

By the way, the constructed wetlands have been utilized as had better be conducted in parallel. Concerning the contribu- a management practice to reduce pesticide contamination via tion of indirect photolysis, the possible involvement of the re- 1 3 agricultural runoff. The maximal aqueous concentrations of actions with not only O2 and CDOM * but also ·OH and/or −· azinphos-methyl, chlorpyrifos and endosulfan were reduced by CO3 should be investigated if necessary by taking the measured more than 90% at the outlet of the flow-through wetland (ca. 0.4 water quality into account, based on the basic information on ha) covered with several emergent macrophytes, as compared the photo-reactivity of the pesticide. The distribution of micro- with those at the inlet.40) About a 10% residue of azinphos-meth- organisms such as algae, bacteria, and fungi in these illuminated yl remained in these macrophytes on a mass basis.41) The tail systems, conveniently examined by using antibiotics, would help water contaminated with fluometuron, aldicarb and endosulfan to understand the pesticide degradation profiles more in detail. from a cotton field was introduced to the non-vegetated open The role of aquatic macrophytes as the pesticide sink in the wa- (Po) and vegetated (Pv, with several emergent macrophytes and ter-sediment system has been confirmed in the laboratory, out- water milfoil) ponds, and the concentration of each pesticide door microcosm and field studies. However, the adsorption/de- in water was monitored.39) The dissipation of any pesticide was sorption profiles of pesticide to aquatic macrophytes have never faster in Pv than in Po and more accelerated by the algal bloom been systematically investigated and furthermore, their meta- in Po. These observations clearly show the importance of algae bolic information is not sufficient. Along with these examina- and aquatic macrophytes as a pesticide sink. tions, the appropriate design of the macrophyte-water-sediment system had better be refined as a higher-tier study for pesticide. Conclusion References The fate of any pesticide in a natural water body associated with bottom sediment is controlled physically, chemically, and bio- 1) K. L. Armbrust: Environ. Toxicol. Chem. 19, 2175–2180 (2000). logically in a complex manner. Although the experimental tem- 2) D. G. Karpouzas, A. Ferrero, F. Vidotto and E. Capri: Toxicol. perature is one of the most important physical factors, its effect Chem. 24, 1007–1017 (2005). 3) European Food Safety Authority Panel on Plant Protection Prod- on the water-sediment DT50 has scarcely been examined, in con- ucts and their Residues: EFSA J. 13: 4176, 145 pp. (2015). trast to the aerobic soil metabolism of pesticide whose tempera- 4) OECD Guideline for the testing of chemicals 308 (2002). ture dependence is well described by the Arrhenius relationship. http://www.oecd-ilibrary.org/environment/test-no-308-aero- The corresponding information is indispensable to convert the bic-and-anaerobic-transformation-in-aquatic-sediment- DT50 values of pesticide in the water-sediment study to those at systems_9789264070523-en (Accessed 7 Sept., 2016). the desired temperature, used as key input values in the sophis- 5) Off. J. Eur. Union L93, 85–152 (2013). http://eur-lex.europa.eu/ ticated simulation of PECs. The temperature controls hydrolysis, legal-content/EN/TXT/?uri=uriserv:OJ.L_.2013.093.01.0085.01. diffusion, and microbial metabolism differently. The investiga- ENG&toc=OJ:L:2013:093:TOC (Accessed 7 Sept., 2016). tion of the temperature effect on each process is necessary, but 6) OECD Guideline for the testing of chemicals 309 (2004). http:// it may be a wiser and convenient way instead to primarily know www.oecd-ilibrary.org/environment/test-no-309-aerobic- mineralisation-in-surface-water-simulation-biodegradation- the overall effect for typical classes of pesticides. The existing test test_9789264070547-en (Accessed 7 Sept., 2016). guidelines for the water-sediment system suppose a lentic water 7) T. Katagi: J. Pestic. Sci. 38, 10–26 (2013). body. However, a water flow may affect many factors controlling 8) K. Solomon, M. Matthies and M. Vighi: Environ. Sci. Eur. 25: 10, pesticide dissipation, such as the mass transport between water 17 pp. (2013). and sediment phases, the growth of phototrophic microorgan- 9) T. Junker, C. Paatzsch and T. Knacker: Sci. Total Environ. 408, isms with the bioavailability of pesticide, and the contribution of 3803–3810 (2010). aquatic macrophytes as a pesticide sink. Therefore, these issues 10) E. Baginska, A. Haiß and K. Kümmerer: Chemosphere 119, 1240– should be further examined by considering the agitating method 1246 (2015). of water and the design of a container, together with the water- 11) L. Gutowski, E. Baginska, O. Olsson, C. Leder and K. Kümmerer: Chemosphere 138, 847–855 (2015). sediment ratio. 12) J. F. Ericson, R. M. Smith, G. Roberts, B. Hannah, B. Hoeger and J. Illumination of the water-sediment system may add to the Ryan: Integr. Environ. Assess. Manag. 10, 114–124 (2014). pesticide degradation photochemical pathways, accelerated al- 13) T. Katagi: Rev. Environ. Contam. Toxicol. 187, 133–251 (2006). kaline hydrolysis and metabolism by phototrophs at the inter- 14) M. Radke and M. P. Maier: Water Res. 55, 63–73 (2014). faces of water-sediment and/or water-macrophytes. The photo- 15) T. Katagi: Rev. Environ. Contam. Toxicol. 182, 1–189 (2004). lytic effect is mentioned as an optional study in the E.U. registra- 16) S. Jeffery, J. A. Harris, R. J. Rickson and K. Ritz: Soil Biol. Biochem. tion and other regulatory guidelines for photo-labile pesticides. 39, 1226–1229 (2007). However, pesticides susceptible to alkaline hydrolysis should 17) M. Honti and K. Fenner: Environ. Sci. Technol. 49, 5879–5886 also be taken into account, since the photosynthesis-driven pH (2015). 18) J. W. H. van der Kolk and S. J. H. Crum: Sci. Total Environ. elevation by algae and aquatic macrophytes is known to proceed 134(Suppl. 2), 1429–1437 (1993). under illumination. In order to distinguish the contribution of 19) W. Kalsch, T. Knacker, G. Danneberg, G. Studinger and C. Franke: photolysis and algal metabolism, illuminated studies using light Int. Biodeterior. Biodegradation 44, 65–74 (1999). sources with different emission spectra (>290 and >400 nm) Vol. 41, No. 4, 1–12 (2016) Pesticide behavior in modified water-sediment systems 11

20) K. A. Fenlon, K. C. Jones and K. T. Semple: Chemosphere 82, 163– gadic, J. Bohatier and J. Einhorn: Environ. Sci. Technol. 43, 3148– 168 (2011). 3154 (2009). 21) B. B. Jørgensen and D. J. Des Marais: Limnol. Oceanogr. 35, 1343– 52) S. Nélieu, L. Kerhoas, M. Sarakha and J. Einhorn: Environ. Chem. 1355 (1990). Lett. 2, 83–87 (2004). 22) Z. Li, A. Sobek and M. Radke: Environ. Sci. Technol. 49, 6009–6017 53) J. Huang and S. A. Mabury: Chemosphere 41, 1775–1782 (2000). (2015). 54) S. Canonica, T. Kohn, M. Mac, F. J. Real, J. Wirz and U. Von Gunt- 23) European Food Safety Authority: EFSA J. 622, 1–32 (2007). en: Environ. Sci. Technol. 39, 9182–9188 (2005). 24) T. Katagi: Rev. Environ. Contam. Toxicol. 175, 79–261 (2002). 55) O. S. Furman, M. Yu, A. L. Teel and R. J. Watts: Chemosphere 93, 25) H.-H. Liu, L.-J. Bao and E. Y. Zeng: Trends Anal. Chem. 54, 56–64 1734–1741 (2013). (2014). 56) R. G. Zepp, J. Hoigné and H. Bader: Environ. Sci. Technol. 21, 443– 26) M. A. Chappell, B. E. Porter, C. L. Price, B. A. Pettway and R. D. 450 (1987). George: Mar. Pollut. Bull. 62, 1736–1743 (2011). 57) S. E. Page, W. A. Arnold and K. McNeill: Environ. Sci. Technol. 45, 27) D. Vione, M. Minella, V. Maurino and C. Minero: Chemistry 20, 2818–2825 (2011). 10590–10606 (2014). 58) F. Wu, J. Li, Z. Peng and N. Deng: Chemosphere 72, 407–413 28) J. T. Jasper and D. L. Sedlak: Environ. Sci. Technol. 47, 10781– (2008). 10790 (2013). 59) T. Zeng and W. A. Arnold: Environ. Sci. Technol. 47, 6735–6745 29) M. W. Lam, K. Tantuco and S. A. Mabury: Environ. Sci. Technol. (2013). 37, 899–907 (2003). 60) R. G. Zepp, N. L. Wolfe, G. L. Baughman and R. C. Hollis: Nature 30) P. L. Brezonik and J. Fulkerson-Brekken: Environ. Sci. Technol. 32, 267, 421–423 (1977). 3004–3010 (1998). 61) S. A. Mabury and D. G. Crosby: J. Agric. Food Chem. 44, 1920– 31) J. Huang and S. A. Mabury: Environ. Toxicol. Chem. 19, 2181–2188 1924 (1996). (2000). 62) H. Shemer, C. M. Sharpless, M. S. Elovitz and K. G. Linden: Envi- 32) M. Minella, F. Romeo, D. Vione, V. Maurino and C. Minero: Che- ron. Sci. Technol. 40, 4460–4466 (2006). mosphere 83, 1480–1485 (2011). 63) L. Wojnárovits and E. Takács: Radiat. Phys. Chem. 96, 120–134 33) M. L. Hladik and K. M. Kuivila: J. Agric. Food Chem. 57, 9079– (2014). 9085 (2009). 64) J. Huang and S. A. Mabury: Environ. Toxicol. Chem. 19, 1501–1507 34) L. H. Hand and R. G. Oliver: Environ. Toxicol. Chem. 29, 2702– (2000). 2712 (2010). 65) M. Bodrato and D. Vione: Environ. Sci.: Processes Impacts 16, 732– 35) K. Takeda, H. Takedoi, S. Yamaji, K. Ohta and H. Sakugawa: Anal. 740 (2014). Sci. 20, 153–158 (2004). 66) K. L. Armbrust, Y. Okamoto, J. Grochulska and A. C. Barefoot: J. 36) T. Hirata, H. Ii, M. Hasebe, N. Egusa, Y. Sakamoto, T. Kumekawa, Pestic. Sci. 24, 357–363 (1999). K. Nishiyama, N. Sakai and H. Iwasaki: J. Jpn. Soc. Civ. Eng. 614/II, 67) K. A. Thomas and L. H. Hand: Environ. Toxicol. Chem. 30, 622– 97–107 (1999) (in Japanese). 631 (2011). 37) T. Hitomi, I. Yoshinaga, A. Miura, K. Hamada, E. Shiratani and K. 68) K. Traisup: PhD Thesis, Univ. York, Environ. Dep., 225 pp. 2012. Takaki: Agric. Hortic. 82, 505–508 (2007) (in Japanese). 69) Y. Usui and T. Kasubuchi: Jpn. J. Limnol. 74, 15–20 (2013) (in Japa- 38) S. Shim, B. Kim, Y. Hosoi and T. Masuda: Water Sci. Technol. 52, nese). 233–241 (2005). 70) A. Ito: Soil Sci. Plant Nutr. 33, 449–459 (1987). 39) M. T. Rose, F. Sanchez-Bayo, A. N. Crossan and I. R. Kennedy: 71) J. N. Seiber, M. P. Catahan and C. R. Barril: J. Environ. Sci. Health B Chemosphere 63, 1849–1858 (2006). 13, 131–148 (1978). 40) R. Schulz and S. K. C. Peall: Environ. Sci. Technol. 35, 422–426 72) W. G. Johnson and T. L. Lavy: J. Environ. Qual. 24, 487–493 (2001). (1995). 41) R. Schulz, C. Hahn, E. R. Bennett, J. M. Dabrowski, G. Thiere and 73) S. P. C. Tankéré, D. G. Bourne, F. L. L. Muller and V. Torsvik: Envi- S. K. C. Peall: Environ. Sci. Technol. 37, 2139–2144 (2003). ron. Microbiol. 4, 97–105 (2002). 42) D.-P. Häder, H. D. Kumar, R. C. Smith and R. C. Worrest: J. Photo- 74) K. A. Thomas and L. H. Hand: Environ. Toxicol. Chem. 31, 2138– chem. Photobiol. B 46, 53–68 (1998). 2146 (2012). 43) R. G. Zepp and D. M. Cline: Environ. Sci. Technol. 11, 359–366 75) F. Fang, W.-T. Lu, Q. Shan and J.-S. Cao: Carbohydr. Polym. 106, (1977). 1–6 (2014). 44) J. S. Winterle, D. Tse and W. R. Mabey: Environ. Toxicol. Chem. 6, 76) I. J. Allan, W. A. House, A. Parker and J. E. Carter: Environ. Sci. 663–672 (1987). Technol. 39, 523–530 (2005). 45) W.-C. Liu, R.-S. Wu, E. M.-Y. Wu, Y.-P. Chang and W.-B. Chen: 77) C.-J. Tien, M.-C. Lin, W.-H. Chiu and C. S. Chen: Environ. Pollut. Paddy Water Environ. 8, 267–275 (2010). 179, 95–104 (2013). 46) K. L. Armbrust: J. Pestic. Sci. 24, 69–73 (1999). 78) A. Paule, A. Biaz, J. Leflaive, J. R. Lawrence and J. L. Rols: Water 47) A. Ito: Jpn. J. Crop. Sci. 38, 355–363 (1969) (in Japanese). Air Soil Pollut. 226, 1–14 (2015). 48) A. Isoda, T. Yoshimura, T. Ishikawa, Y. Nakamura, H. Nojima and 79) L. H. Hand and H. J. Moreland: Environ. Toxicol. Chem. 33, 516– Y. Takasaki: Tech. Bull. Fac. Hortic. Chiba Univ. 43, 39–43 (1990) 524 (2014). (in Japanese). 80) L. Luo, P. Wang, L. Lin, T. Luan, L. Ke and N. F. Y. Tam: Process 49) D. F. Wallace, L. H. Hand and R. G. Oliver: Environ. Toxicol. Chem. Biochem. 49, 1723–1732 (2014). 29, 575–581 (2010). 81) L. Wang, C. Zhang, F. Wu and N. Deng: J. Photochem. Photobiol. B 50) C. K. Remucal: Environ. Sci.: Processes Impacts 16, 628–653 (2014). 87, 49–57 (2007). 51) S. Nélieu, F. Perreau, F. Bonnemoy, M. Ollitrault, D. Azam, L. La- 82) M. Yamazaki, Y. Hamada, T. Ibuka, T. Momii and M. Kimura: Soil 12 T. Katagi Journal of Pesticide Science

Sci. Plant Nutr. 47, 587–599 (2001). 310–316 (2006). 83) Y. Fujita and H. Nakahara: Jpn. J. Limnol. 60, 67–76 (1999) (in 101) R. Kodaka, T. Sugano and T. Katagi: J. Pestic. Sci. 34, 27–36 (2009). Japanese). 102) R. Kodaka, S. E. Swales, C. Lewis and T. Katagi: J. Pestic. Sci. 36, 84) J.-D. Kim and C.-G. Lee: J. Microbiol. Biotechnol. 16, 240–246 33–40 (2011). (2006). 103) A. Shibata, R. Kodaka, T. Fujisawa and T. Katagi: J. Agric. Food 85) N. Nakayama, A. Okabe, K. Toyota, M. Kimura and S. Asakawa: Chem. 59, 11186–11195 (2011). Soil Sci. Plant Nutr. 52, 305–312 (2006). 104) M. Liu, D. Liu, Y. Xu, X. Jing, Z. Zhou and P. Wang: J. Agric. Food 86) T. Katagi: Rev. Environ. Contam. Toxicol. 204, 1–132 (2010). Chem. 63, 5205–5211 (2015). 87) P. A. Chambers, P. Lacoul, K. J. Murphy and S. M. Thomaz: Hydro- 105) European Food Safety Authority: Conclusion regarding the peer biologia 595, 9–26 (2008). review of the pesticide risk assessment of the active substance. 88) V. Matamoros, L. X. Nguyen, C. A. Arias, V. Salvadó and H. Brix: “Publications, conclusion on pesticides” in http://www.efsa.europa. Chemosphere 88, 1257–1264 (2012). eu/en/search/site (Accessed 7 Sept., 2016). 89) F. Mermillod-Blondin, D. Lemoine, J.-C. Boisson, E. Malet and B. 106) European Food Safety Authority: Draft Assessment Report. http:// Montuelle: Freshw. Biol. 53, 1969–1982 (2008). dar.efsa.europa.eu/dar-web/ provision (Accessed 7 Sept., 2016). 90) P. I. Adriaanse, J. J. T. I. Boesten and S. J. H. Crum: Pest Manag. Sci. 107) Pest Management Regulatory Agency: Health Canada. Proposed 69, 755–767 (2013). Registration Decision. PRD2009-18. http://www.hc-sc.gc.ca/cps- 91) S. J. H. Crum, A. M. M. van Kammen-Polman and M. Leistra: spc/pest/part/consultations/index-eng.php (Accessed 7 Sept., Arch. Environ. Contam. Toxicol. 37, 310–316 (1999). 2016). 92) R. H. Bromilow, R. F. de Carvalho, A. A. Evans and P. H. Nicholls: 108) H.-T. Lai and J.-J. Lin: Chemosphere 75, 462–468 (2009). J. Environ. Sci. Health B 41, 1–16 (2006). 109) S. S. Walse, P. L. Pennington, G. I. Scott and J. L. Ferry: J. Environ. 93) Pesticide Properties DataBase: http://sitem.herts.ac.uk/aeru/ppdb/ Monit. 6, 58–64 (2004). en/index.htm (Accessed 7 Sept., 2016). 110) S. A. Mabury, J. S. Cox and D. G. Crosby: Rev. Environ. Contam. 94) T. Fujisawa, K. Ichise-Shibuya and T. Katagi: J. Pestic. Sci. 35, 456– Toxicol. 147, 71–117 (1996). 463 (2010). 111) C. Stang, D. Elsaesser, M. Bundschuh, T. A. Ternes and R. Schulz: 95) D. Ando, T. Fujisawa and T. Katagi: J. Pestic. Sci. 37, 342–346 J. Environ. Qual. 42, 1889–1895 (2013). (2012). 112) D. Elsaesser, C. Stang, N. Bakanov and R. Schulz: Bull. Environ. 96) D. Ando, T. Fujisawa and T. Katagi: J. Agric. Food Chem. 63, 5189– Contam. Toxicol. 90, 640–645 (2013). 5195 (2015). 113) C. Stang, M. V. Wieczorek, C. Noss, A. Lorke, F. Scherr, G. Goerlitz 97) J.-W. Kwon and K. L. Armbrust: J. Agric. Food Chem. 54, 3651– and R. Schulz: Chemosphere 107, 13–22 (2014). 3657 (2006). 114) M. T. Moore, E. R. Bennett, C. M. Cooper, S. Smith Jr., F. D. 98) P. L. Tomco and R. S. Tjeerdema: Pest Manag. Sci. 68, 1141–1147 Shields Jr., C. D. Milam and J. L. Farris: Agric. Ecosyst. Environ. 87, (2012). 309–314 (2001). 99) M. Vasquez, T. Cahill and R. Tjeerdema: J. Agric. Food Chem. 59, 115) E. R. Bennett, M. T. Moore, C. M. Cooper, S. Smith Jr., F. D. 7874–7881 (2011). Shields Jr., K. G. Drouillard and R. Schulz: Environ. Toxicol. Chem. 100) R. Kodaka, T. Sugano and T. Katagi: Environ. Toxicol. Chem. 25, 24, 2121–2127 (2005).