barley rotation

Chemosphere 90 (2013) 2499–2511

Contents lists available at SciVerse ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Mapping field spatial distribution patterns of isoproturon-mineralizing activity over a three-year winter wheat/rape seed/barley rotation S. Hussain 1, M. Devers-Lamrani, A. Spor, N. Rouard, M. Porcherot, J. Beguet, F. Martin-Laurent ⇑ INRA, UMR 1347 Agroecologie, 17 rue Sully, BP 86510, 21065 Dijon Cedex, France

h i g h l i g h t s " Spatio-temporal variability in isoproturon mineralization activity was monitored. " Isoproturon treatment led to high isoproturon mineralization activity. " Isoproturon treatment diminished variability in mineralization activity. " Isoproturon mineralization activity was correlated with soil pH and CEC. " Simple model using pH and CEC can predict Isoproturon mineralization activity.

a r t i c l e

i n f o

Article history: Received 15 June 2012 Received in revised form 1 October 2012 Accepted 5 October 2012 Available online 14 December 2012 Keywords: Isoproturon Mineralization Spatial variability Geostatistical analysis Crop rotation

a b s t r a c t The temporal and spatial variability of the activity of soil microorganisms able to mineralize the herbicide isoproturon (IPU) pesticide was investigated over a three-year long crop rotation between 2008 and 2010. Isoproturon mineralization was higher in 2008, when winter wheat was treated with this herbicide, than in 2009 and 2010, when rape seed and barley were treated with different herbicides. Under laboratory conditions, we showed that isoproturon mineralization was not promoted by sulfonylurea herbicide applied on barley crop in 2010. IPU mineralization was shown to be highly variable at the field scale in years 2009 and 2010. Principal component analyses and analyses of similarities revealed that soil pH and equivalent humidity, and to a lesser extent soil organic matter content and cation exchange capacity (CEC) were the main drivers of isoproturon-mineralizing activity variance. Using a rather simple model that yields the rate of isoproturon mineralization as a function of soil pH and equivalent humidity, we explained up to 85% of the variance observed. Mapping field-scale distribution of isoproturon mineralization over the three-year survey indicated higher variability in 2009 and in 2010 as compared to 2008, suggesting that isoproturon treatment applied to winter wheat promoted isoproturon mineralization activity and reduced its spatial variability. Field-scale distribution of isoproturon mineralization showed important similarity to the distribution of soil pH, equivalent humidity and to a lesser extent to soil organic matter and cation exchange capacity (CEC) thereby confirming our model. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction The use of pesticides in conventional agriculture has attracted much attention due to their potential harmful effects not only on plants, animals and microorganisms but also on humans. The harmful effects of pesticides are linked to their environmental fate, which in turn plays a central role in determining their behavior in natural media, including soils. In recent years, much attention has been paid for predicting the fate of pesticides, and microbial

⇑ Corresponding author. Fax: +33 3 80 69 34 06. E-mail address: [email protected] (F. Martin-Laurent). Present address: Department of Environmental Sciences, Government College University, Faisalabad, Pakistan. 1

0045-6535/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.chemosphere.2012.10.080

degradation has emerged as a main process leading to the dissipation of pesticides after they get into soils. Degradation, which predominantly involves soil microorganism activity, is considered as a key process affecting the dynamics of pesticides in soil environments. It is governed by a number of soil and environmental factors that can induce significant spatial variations in the process. One important lingering environmental question regarding microbial pesticide biodegradation is the extent to which degradation takes place at the field scale. This is an important issue for predicting the fate of pesticides in soils. Isoproturon (IPU), a substituted phenylurea herbicide, is used for controlling pre- and post-emergence annual grasses and broadleaved weeds in wheat, barley and winter rye crops (Ertli et al., 2004; El-Sebai et al., 2007). Being of one of the most extensively

2500

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

used pesticides in Europe, its use has recently been restricted and banned in several European countries, but it still remains extensively used worldwide. (Stangroom et al., 1998; El-Sebai et al., 2007). IPU is relatively recalcitrant in the soil and has a DT50 ranging from 6 to 90 d (El-Sebai et al., 2007). As a result of its widespread repeated use and persistent properties, IPU is frequently detected in water resources at concentrations higher than 0.1 lg L1, the European Union drinking water limit (Spliid and Koppen, 1998; Muller et al., 2002). Ecotoxicological data suggest that IPU and some of its metabolites are harmful not only for microbial communities (Widenfalk et al., 2008), aquatic invertebrates (Mansour et al., 1999), macrophytes (Yin et al., 2008; Knauert et al., 2010; Kumar et al., 2010) and fresh-water algae (Dewez et al., 2008; Vallotton et al., 2009; Dosnon-Olette et al., 2010) but also for humans and animals (Behera and Bhunya, 1990; Hoshiya et al., 1993; Hazarika and Sarkar, 2001; Orton et al., 2009). Lowering IPU contamination in soil and water resources is therefore of particular interest. Microbial biodegradation has been reported as a primary mechanism for the dissipation of isoproturon and of many other phenylurea herbicides from aquatic and telluric environments (Fournier et al., 1975; Gaillardon and Sabar, 1994; Cox et al., 1996; Pieuchot et al., 1996; Bending et al., 2003; El-Sebai et al., 2007; Hussain et al., 2011). Studies have reported that in response to repeated exposure to pesticides, soil microorganisms adapt to the rapid biodegradation of several herbicides including IPU (Cox et al., 1996; Karpouzas et al., 1999; Sorensen and Aamand, 2001; Bending et al., 2003, 2006; El-Sebai et al., 2005; El-Sebai et al., 2007; Hussain et al., 2011). Mineralization of isoproturon has already been reported in agricultural fields repeatedly treated with it in the United Kingdom (Cullington and Walker, 1999), Denmark (Sorensen and Aamand, 2003) and more recently in France (El-Sebai et al., 2005, 2007; Hussain et al., 2011). In addition, several studies have also shown that IPU mineralization activity was not evenly distributed across agricultural fields (Beck et al., 1996; Bending et al., 2001; Walker et al., 2001, 2002; El-Sebai et al., 2007). Beck et al. (1996) determined variability in the degradation time-course of isoproturon on 25 different sampling points within a same field and found that the time required for 50% of the initial IPU concentration to dissipate ranged between 31 and 483 d. More recently, within-field spatial heterogeneity of isoproturon degradation activity was reported in different regions of the world including Denmark, the United Kingdom and France (Bending et al., 2001, 2003, 2006; Walker et al., 2001; Rodriguez-Cruz et al., 2006; El-Sebai et al., 2007). Variations in soil physico-chemical properties such as pH, texture, temperature, moisture or organic matter have been reported to affect the degradation of pesticides (Sorensen et al., 2003; Bending et al., 2003; El-Sebai et al., 2005, 2007, 2011; Vieublé-Gonod et al., 2009). Soil properties affect not only the availability and biodegradability of pesticides (Walker et al., 1992; Welp and Brummer, 1999) but also the diversity, the size and the activity of microbial populations (Smith et al., 1997; Hundt et al., 1998). The spatial variability of IPU degradation within a British agricultural field was correlated (i) with soil pH and microbial biomass (Walker et al., 2001) and (ii) with soil pH and the extent of the proliferation of IPU-degrading organisms (Bending et al., 2003). The same two factors were also found to account for the variability of isoproturon mineralization activity determined for 50 soil samples within an agricultural field in France (El-Sebai et al., 2005, 2007). A significant correlation between pH and IPU degradation has been demonstrated not only in agricultural soils (Walker et al., 2002; El-Sebai et al., 2005, 2007) but also in pure broth cultures (Bending et al., 2003; Hussain et al., 2009, 2011; Sun et al., 2009; Hussain et al., 2011). Similarly, the spatial heterogeneity of IPU-mineralizing activity within an agricultural field was also found to be enhanced by the presence of compost (Vieublé-Gonod et al., 2009).

Such stimulation of the mineralization activity was hypothesised to derive from the microbial characteristics of the compost, the physical properties of the soils and the physico-chemical conditions in the area with high mineralization activity. Since the rate of IPU degradation has been reported to be influenced by a number of soil physico-chemical and microbial properties and also by the repeated use of pesticides, a detailed knowledge of these parameters in relation to spatial variability in the IPU degradation rate is important to better understand IPU dissipation from agricultural soils. In order to tackle this question spatial and temporal variability of IPU-mineralization activity was monitored within an agricultural field by following the mineralization of IPU under a winter wheat/rape seed/barley crop rotation, in connection with the variability of soil physico-chemical and microbiological properties. A special effort was made to search for key parameters driving of IPU-mineralization activity and to develop a simple model predicting IPU-mineralization activity. Mapping field spatial distribution patterns of isoproturon-mineralizing activity was carried out over the three year survey. 2. Materials and methods 2.1. Chemicals Analytical grade IPU (99.0% purity) and 14C-ring-labelled IPU (specific activity 18 mCi mmol1; 99% radiochemical purity) were purchased from Riedel-de-Haen (Germany) and International Isotopes (Munich, Germany), respectively. The commercial product ArchipelÒ was purchased from Syngenta. 2.2. Soil sampling and soil properties The study was carried out on soil collected from an agricultural field located at the experimental farm of the National Institute of Agronomical Research (INRA) of Epoisses (Breteniere, France). Thirty-six separate soil surface samples (0–20 cm) were collected in April 2008, 2009 and 2010 from the plot, following a grid made of 4 columns (with 17, 26 and 24 m in-between each of them) and nine rows (with 20 m in-between each of them). The field had been cultivated under a winter wheat/rape seed/barley rotation and periodically treated with IPU for more than 10 years. Over the three-year time-span (2008–2010), the field was treated with (i) IPU herbicide, applied on Oct. 12th, 2007, several months before the 2008 sampling, at 1.2 kg ha1, (ii) the two herbicides Roundup Typhon, which contains 360 g L1 of glyphosate (Makheshim Agan France) and TreflanÒ, which contains trifluraline (Dow AgroSciences), applied on Aug. 27th, 2008 at 1.5 L ha1 and 2.5 L ha1, respectively; (iii) Colzor TrioÒ (Syngenta) made of a mixture of clomazone (30 g L1), dimethachlor (187.5 g L1) and napropamide (187.5 g L1) applied on Aug. 29th, 2009 at 3.5 L ha1, (iv) the herbicide ArchipelÒ (Syngenta) made of a mixture of mesosulfuronmethyl (0.75 g L1), iodosulfuron-methyl sodium (0.75 g L1) and Mefenpyr-dimethyl, applied on March 24th, 2010 at 0.25 kg ha1, just before the 2010 sampling. The moisture content of each soil sample was estimated before the experiment started. The soil samples were sifted through 5-mm mesh sieves and the samples were sub-divided into two major groups and a minor one. The first major group was air-dried for physico-chemical analysis, whereas the second was kept at 4 °C for biological analysis. The third, smaller group was stored at 20 °C for soil DNA extraction. The physico-chemical properties of the soil samples, i.e. granulometric properties, equivalent humidity, organic matter content, carbon, nitrogen, C/N ratio, pH and cation exchange capacity (Table 1) were determined by the Laboratory of Soil Analysis (INRA, Arras, France) using ISO procedures.

2501

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

Table 1 Descriptive and geostatistical analyses of the physico-chemical parameters [Equivalent Humidity (EH), organic matter content (OM), organic carbon content (OC), nitrogen (N), C/N ratio, cation exchange capacity (CEC), pH and volatile matter (VM)] and biological parameters [microbial C biomass (MCB), number of cultivable bacteria (cfu), maximum percentage of IPU mineralization (A), maximum rate of mineralization (lm), lag phase (k), inflection point abscissa (ti), methanol-extractable residues (ME) and bound residues (BR)] determined from 36 soil samples collected from our experimental field in Epoisses. Parameter

Year

P

Sill

Nugget

EH (g kg1)

2008 2009 2010

Min 22.9 26.8 25.3

Max 28.5 39.6 30.7

25.2 c 30.3 a 28.6 b

Mean

25.0 30.3 28.8

1.6 2.2 1.4

6.1 7.1 4.9

0.7358 0.5128 0.9999

2.4 3.1 1.5

0 1.2 0

OM (g kg1)

2008 2009 2010

26.4 30.3 29.9

38.8 53.7 55.1

32.4 b 37.5 a 37.2 a

32.8 36.5 36.0

3.3 4.8 5.7

10.0 12.8 15.2

0.9999 0.9987 0.7358

10.3 27.0 26.9

0 0 0

31.9 117.7 156

1 1 1

OC (g kg1)

2008 2009 2010

15.3 17.5 17.3

22.4 31.0 31.9

18.8 b 21.7 a 21.5 a

18.9 21.1 20.8

1.8 2.8 3.3

9.8 12.8 15.3

0.9999 0.9987 0.7358

3.2 8.9 9.1

0 0 0

34.0 117.6 155.3

1 1 1

N (g kg1)

2008 2009 2010

1.5 1.7 1.6

1.9 2.2 2.0

1.7 c 2.0 a 1.8 b

1.7 2.0 1.8

0.10 0.13 0.11

6.0 6.5 6.0

0.9999 0.9999 0.9999

0.003 0.013 0.013

0.0075 0.002 0

106.4 48.9 159.1

nd 0.8 1

C/N

2008 2009 2010

8.9 9.2 10.3

13.2 16.9 18.0

10.7 b 11.1 b 12.0 a

10.7 10.7 11.5

0.8 1.5 1.9

7.8 13.7 15.5

0.9999 0.3380 0.5128

0.64 1.95 2.99

0 0 0

36.8 50.8 166

1 1 1

CEC metson (cmol + kg1)

2008 2009 2010

21.0 16.8 15.9

22.0 27.5 27.4

21.3 a 22.7 a 22.5 a

21.0 22.6 22.5

0.5 2.8 2.9

2.4 12.5 12.8

0.3380 0.9999 0.9999

1.5 6.9 8.7

0.74 0 0

40 279.2 321.8

0.5 1 1

pH

2008 2009 2010

7.7 7.3 7.1

7.8 8.0 8.1

7.7 a 7.7 a 7.8 a

7.7 7.7 7.9

0.04 0.19 0.27

0.5 2.5 3.4

0.9987 0.9999 0.9987

0.015 0.029 0.061

0.011 0 0

113 108.5 86.7

nd 1 1

VM (g kg1)

2008 2009 2010

9.0 8.2 7.7

10.0 10.5 9.6

9.5 a 9.3 a 8.4 b

9.5 9.1 8.3

0.7 0.6 0.5

7.4 6.5 5.8

0.9999 0.9987 0.7358

nd nd nd

nd nd nd

nd nd nd

nd nd nd

MCB (mg kg1)

2008 2009 2010

209.7 91.2 199.8

402.7 467.1 413.2

cfu (108 g1)

2008 2009 2010

7.7 7.6 7.7

8.2 8.2 8.1

A (% CO2)

2008 2009 2010

37.2 11.0 4.5

47.9 37.5 53.1

lm (day1)

2008 2009 2010

2.6 0.2 0.1

5.0 1.8 7.3

k (day)

2008 2009 2010

0.3 0.3 2.1

ti (day)

2008 2009 2010

ME (% CO2)

BR (% CO2)

314.8 a 311.3 a 295.7 a

Med

SD

CV

Range 24.9 68.7 89.5

Q 1 0.6 1

318.2 321.5 293.3

41.9 80.2 51.2

13.3 25.8 17.3

0.9999 0.9987 0.9999

1338 6249 2465

498.3 0 0

88 36.2 59.8

0.6 1 1

8.1 7.9 8.0

0.10 0.12 0.08

1.3 1.5 1.1

0.9987 0.9999 0.9999

0.0006 0.0127 0.0069

0.0097 0 0

16.2 30.2 20.6

nd 1 1

42.1 29.5 48.5

2.5 7.1 10.1

5.8 26.2 22.2

0.9987 0.9987 0.0694

2.4 47.3 99.7

3.0 0 0

38.9 36.3 38.8

nd 1 1

4.0 a 0.7 b 3.9 a

4.0 0.7 4.2

0.6 0.4 1.8

15.6 53.7 45.4

0.9987 0.9999 0.9999

0.085 0.119 2.363

0.309 0.013 1.302

115.2 60.6 186.2

nd 0.9 0.5

2.7 7.5 6.3

1.1 c 4.6 a 2.7 b

1.1 4.7 2.2

0.6 1.4 1.6

52.4 29.9 60.2

0.9900 0.9999 0.5128

0.15 1.70 1.10

0.13 0 1.54

97.4 30.9 79.3

nd 1 nd

3.9 9.7 3.9

8.4 29.3 19.2

5.1 c 20.7 a 8.2 b

4.9 20.7 6.5

1.1 5.8 4.0

20.5 27.8 48.2

0.9987 0.9999 0.5128

0.40 29.99 14.32

0.65 2.07 0

78.8 54.8 38.3

nd 0.9 1

2008 2009 2010

2.8 4.1 2.3

9.1 21.1 53.97

5.7 b 9.7 a 7.7 ab

5.7 8.2 3.8

1.5 5.0 10.6

26.7 51.2 137.2

0.9999 0.7358 0.0183

2.29 4.89 11.3

0.07 18.9 0

33.1 nd 37.2

0.9 nd 1

2008 2009 2010

23.8 38.7 23.1

58.6 66.7 61.6

44.1 b 53.0 a 41.9 b

44.1 53.3 41.7

8.1 5.8 11.0

18.3 10.9 26.2

0.9999 0.9999 0.9999

0 30.8 3.3

64.9 0 116.5

0 29.4 3.4

nd 1 nd

8.1 a 7.9 c 8.0 b 42.6 a 27.1 b 45.3 a

Min, Minimum; Max, maximum; Med, median; SD, standard deviation; CV, coefficient of variation; Q, Q-value; nd, not determined. For each variable, significant differences between means over the 3 years are indicated with letters.

2.3. Enumeration of culturable bacteria The culturable bacteria from the collected soil samples were enumerated on ten-fold diluted nutrient agar (NA) medium (DIFCO, France) supplemented with 100 mg L1 of the fungicide cycloheximide. Ten grams (equivalent dry weight) of each soil sample were suspended in 90 mL of distilled sterilised water using a Waring blender (New Hartford, USA) and serially diluted ten-fold down to 106. One hundred microlitres of each of the soil dilutions from 104 to 106 were inoculated on NA medium plates in triplicates and incubated at 20 °C. The number of

bacterial colonies in each plate was counted after 3 weeks’ incubation. 2.4. Microbial C biomass of the soil samples The microbial C biomass of each soil sample was estimated using the chloroform fumigation–extraction method, following the experimental procedure described by Chaussod et al. (1988). Extractable C was measured by an automated UV-persulphate oxidation method in a Dohrman DC80 analyser, and microbial C biomass was estimated using the following formula:

2502

Microbial C biomass ¼

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

ðCfumigated extract  Cunfumigated extract Þ KC

The calculated KC value of 0.38 was used to convert extractable C into microbial C biomass (Wu et al., 1996). 2.5. Molecular analysis of the global structure of the soil microbial community 2.5.1. Soil DNA extraction DNA was extracted directly from each soil sample according to the ISO 11063 method derived from the protocol initially described by Martin-Laurent et al. (2001). Briefly, 250-mg aliquots of each soil sample were suspended in 1 mL of the extraction buffer [100 mM EDTA, 100 mM NaCl, 100 mM Tris (pH 8.0), 2% (w/v) sodium dodecyl sulphate, 1% (w/v) polyvinylpyrrolidone] and then shaken for 30 s at 1600 rpm in a mini-bead beater cell disrupter (Mikro-Dismembrator S; Sartorius AG, Germany). Soil and cell debris were removed by centrifugation at 14 000g and proteins were removed by centrifugation at 14 000g after sodium acetate precipitation. DNA was precipitated with cold isopropanol, washed with 70% ethanol and purified through Sepharose 4B and Polyvinylpyrrolidone (PVPP) spin columns. Soil DNA quality was verified by spectrophotometry by calculating the A260/A280 ratio using a biophotometer (Eppendorf, Germany) and by electrophoretic separation on 1% agarose gels. Soil DNA concentrations were quantified on agarose gels, using DNA standards as previously described (Ranjard et al., 2003). Each DNA sample was diluted to obtain a concentration of 1 ng lL1. 2.5.2. Automated ribosomal intergenic spacer analysis (A-RISA) of the soil samples The bacterial community structure of the soil samples was determined by A-RISA. The targeted intergenic space of the bacterial rRNA operons was amplified with the primers 1522IRD800 (50 TCG GGC TGG ATC ACC TCC TT-30 ) and 132R (50 -CCG GGT TTC CCC ATT CGG-30 ). The PCR mix was prepared in a final 25-lL volume containing 2.5 lL 10X incubation buffer, 0.2 mM of each dNTP, 0.5 lM of each primer, 3.75U of Taq polymerase, 500 ng of T4 gen 32 and 5 ng of the extracted DNA. The PCR reaction was carried out in a thermocycler (PTC 200 gradient Cycler, MJ Research, Waltham, Mass) with the following programme: 1 cycle of 3 min at 94 °C; 30 cycles of 1 min at 94 °C, 30 s at 55 °C, 1 min at 72 °C, and one final cycle of 5 min at 72 °C. The PCR products were verified by migrating 3 lL of each product on 2% agarose gels. For A-RISA fingerprinting, the PCR products were loaded on a 3.7% solidified gel matrix (KB Plus, Li-COR, Bioscience) and run under denaturing conditions for 15 h at 1500 V/80 W on a LiCor DNA sequencer (Science Tec). The gels were analysed using 1D-Scan software (El-Sebai et al., 2007) and the data thus obtained were analysed using Prep RISA software. The resulting data were further analysed using ADE4 software (Thioulouse et al., 1997) allowing a principal component analysis of the A-RISA fingerprint which helps study the global structure of bacterial communities. 2.6. Estimation of IPU mineralization and sorption 2.6.1. IPU-mineralization potential of the soil samples The potential of indigenous soil microorganisms to mineralize IPU in the soil samples was determined by radiorespirometry, using 14C-ring-labelled IPU as previously described (El-Sebai et al., 2007; Hussain et al., 2011). Briefly, 40 g (equivalent dry weight) of each soil sample were treated with 1.5 mg of IPU per kg of soil and 2 kBq of 14C-ring-labelled IPU. The soil samples were moistened up to 80% of their water holding capacity and incubated at 20 °C in the dark for about 70 d in closed respirometer jars

(Soulas, 1993). 14CO2 resulting from the mineralization of 14Cring-labeled IPU was trapped in 5 mL of 0.2 M NaOH solution. The traps were regularly changed over the incubation period and analysed for their radioactivity content by liquid scintillation counting using ACSII scintillation fluid (Amersham). 2.6.2. IPU-mineralization potential of the soil samples treated with isoproturon or with Archipel In order to address the impact of substituted urea (isoproturon) or sulphonyl urea (mesosulfuron methyl), four composite soil samples, namely A, B, C and D, were prepared by pooling the nine soil samples collected from each field row. For each sample, three treatments were applied on 20-g soil samples (equivalent dry weight) pre-treated with isoproturon (1.2 mg kg1) or with Archipel (0.25 mg kg1) or with water (control). The soil samples were kept in the dark at 20 °C for 15 d and then treated with 1.5 mg of IPU per kg of soil and 2 kBq of 14C-ring-labelled IPU as described above in order to measure IPU mineralization. Three replicates were performed (ntotal = 36). 2.6.3. Determination of 14C-IPU residues At the end of the IPU-mineralization assay, residual 14Cring-labelled IPU was extracted and analysed as described by El-Sebai et al. (2005). Briefly, 20 g (equivalent dry weight) of soil were added to 40 mL of methanol and shaken at 180 rpm for 16 h at 20 °C. The samples were then centrifuged for 10 min at 3220g. The supernatant was recovered and 2-mL aliquots of each supernatant were analysed for their radioactivity content by liquid scintillation counting using ACSII (Amersham) scintillation fluid. The soil pellets were also recovered and air-dried at room temperature. Quantitative determination of non-extractable IPU residues, mainly corresponding to bound residues, was carried out by combustion of 0.5 mg of dried methanol-extracted soil samples under O2 flow at 900 °C for 4 min, using a Biological Oxidiser OX-500 (EG&G Instruments, France) as previously described by El-Sebai et al. (2005). 2.7. Data analysis 2.7.1. Estimating mineralization kinetics parameters Kinetics parameters of IPU mineralization were determined by fitting the modified Gompertz model (Zwietering et al., 1990) to the IPU-mineralization curves shown in Fig. 1, using: lM;e

Y ¼ Aee½

A

ðktÞþ1

;

where Y is the mineralization percentage (%), t is time (days), lM is the maximum mineralization rate (% day1), A is the maximum percentage of mineralization (%), k is the lag time (days). This analysis was performed using Sigma PlotÓ 4.0 software. 2.7.2. Statistical analysis A Principal Component Analysis was conducted on the soil physico-chemical properties, microbiological variables and parameters estimated from the mineralization time-course. Data were centred and scaled prior to PCA. The analysis was performed using the dudi.pca function in the ade4 package (Dray and Dufour, 2007) for R (http://www.r-project.org/). An ANalysis Of SIMilarity was performed to test for pairwise differences between years over all the parameters measured. Multiple linear regression models were used to further elaborate the effect of the different physico-chemical and microbiological variables on the main two mineralization time-course parameters lM (the rate of IPU degradation) and A (the maximum amount of IPU degraded). Both forward selection and backward model selection (minimising Akaike Information Criterion) were

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

2503

variables was checked by calculating Pearson’s correlation coefficient. It is important to note that Equivalent Humidity was chosen as a significant explaining variable, but as CEC and total N content values were strongly correlated with Equivalent Humidity values, we could not determine which of these three variables caused the variations of lM and A we observed. We chose to keep Equivalent Humidity in the model because the percentage explained variance by the model was slightly higher with this parameter than with CEC or total N content. However, we should lay the emphasis on the fact that wherever Equivalent Humidity is used as an explaining variable in this manuscript, CEC or total N content could have been used instead. Spatial autocorrelation between residuals was checked by fitting a variogram to Pearson’s residuals. Residuals of the models were found to be normally distributed and independent with regards to spatial distance. 2.7.3. Geostatistical analysis All the parameters considered in this three year-long study were also analysed using geostatistics by calculating a semivariogram for all parameters, thanks to the following formula:

cðhÞ ¼

1 X ½zðxi þ hÞ  zðxi Þ2 2NðhÞ

where c(h) is the experimental semi-variogram value at the distance interval h; N(h) is the number of sample pairs within the distance interval h; z(xi) and z(xi + h) are the sample values at two points separated by the distance interval h (Journel and Huijbergts, 1978; Isaaks and Srivastava, 1989). All pairs of points separated by a distance h (lag h) were used to calculate the experimental semivariogram. The spherical model was fitted to the semi-variogram and the parameters of the model, i.e. sill, nugget and range were determined. In order to estimate spatial dependence, two quality indices were used: Q value, calculated with the formula Q ¼ ðstillnuggestÞ and range. Q values indicate the spatial structure at still the sampling scale and ranges give the limits of spatial dependence. Maps were then computed using ordinary kriging to assess the regional patterns of variation within the field by using the geostatistical package GeoR version 1.6–27 under R.

3. Results 3.1. Characterization of physico-chemical and microbiological parameters in the Epoisses experimental field over a three-year period

Fig. 1. Kinetics of the mineralization of 14C-ring-labelled IPU in 36 sub-site soil samples collected from the Epoisses field along the sampling grid over a three-year crop rotation.

used to select the significant variables accounting for the variations of lM and A. The following model was used:

Z ij ¼ l þ Yeari þ Equiv alent Humidityj þ pHj þ Year  Equiv alent Humidityij þ eij where Zij represents the response variable (lM or A), and Year represents the differences between sampling years (i = 1–3). Note that the Year factor is collinear with IPU treatment, no treatment and phenylurea treatment for 2008, 2009 and 2010, respectively. Equivalent Humidity and pH are continuous explaining variables for each j sample. We also allowed the effect of Equivalent Humidity to be dependent on Year (YearEquivalent Humidity effect). The term eij is the residual error of the model. lM and A were log-transformed prior to the multiple regression. Collinearity between explaining

3.1.1. Physico-chemical properties Descriptive statistics, i.e. minimum, maximum, mean, median, standard deviation (SD) and coefficient of variation (CV) for the physico-chemical parameters (equivalent humidity, organic matter content, organic carbon content, nitrogen content, C/N ratio, CEC and pH) of the 36 sub-site soil samples over 3 years (2008, 2009 and 2010, ntotal = 108) are shown in Table 1. Values varied moderately within the grid pattern of the field over the 3 years surveyed. Three of the soil physico-chemical parameters, i.e. equivalent humidity (EH), nitrogen content (N) and pH were found to vary little, with CV values below 10% over the 3 years. Organic matter content (OM) varied moderately, with CV values ranging between 10% and 15%. C/N ratio and cation exchange capacity (CEC) yielded CV values, within the same range as OM in 2009 and 2010, but below 10% in 2008. Although soil pH varied between 7.1 and 8.1, CV values were low and no significant difference was observed over the 3 years (Table 1). However, the mean ‘‘equivalent humidity’’ and ‘‘nitrogen content’’ values were significantly different between years, with a maximum in 2009 and a minimum in 2008. Similarly, the mean ‘‘organic matter content’’, ‘‘organic

2504

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

carbon content’’ and ‘‘C/N ratio’’ values were significantly different in 2008 as compared to 2009 and 2010 (Table 1). 3.1.2. Microbiological and mineralization kinetics parameters Descriptive statistics for microbiological parameters, i.e. microbial C biomass (MCB), total culturable heterotrophic bacteria (cfu) and mineralization parameters (A, lm, k and ti) of the 36 sub-site soil samples over the 2008–2010 period are presented in Table 1. Microbiological properties displayed higher variability than physico-chemical properties, except the number of culturable heterotrophic bacteria which had a low CV (<5%). Although microbial C biomass displayed moderate variability within the field grid pattern with CV values ranging between 13% and 26%, mean microbial C biomass did not significantly vary over the 3 years of the survey. Most of the 36 sub-site samples showed high IPU-mineralizing ability (Fig. 1). Interestingly, their IPU-mineralizing ability was higher in 2008 and in 2010 as compared to 2009. Indeed, most of the soil samples taken in 2008 and in 2010 mineralized up to 40% of the 14C-ring-labelled IPU within only 3 weeks’ incubation, while less than 20% of the 14C-ring-labelled IPU was mineralized in the 2009 samples. The cumulative percentage of IPU mineralization (A) and the maximum mineralization rate (lM) varied moderately, with CV values of 5.8% and 15.6%, respectively, in 2008 (Table 1). However, these two parameters showed higher variability in 2009 and in 2010 with 25% and 50% CV values, respectively. A and lM values were significantly higher in 2008 (A = 42.6% of 14CIPU evolved to 14CO2, lM = 4.0% of 14C-IPU evolved to 14CO2 per day) and 2010 (A = 45.3% of 14C-IPU evolved to 14CO2, lM = 3.9% of 14C-IPU evolved to 14CO2 per day) as compared to 2009 (A = 27.1% of 14C-IPU evolved to 14CO2, lM = 0.7% of 14C-IPU

60

8

A

50

evolved to 14CO2 per day) (Fig. 2). Lag phase (k) and inflection point abscissa (ti) were the most variable parameters. In addition, the mean values of k and ti were significantly higher in 2009 (k = 4.6 and ti = 20.7) as compared to 2008 (k = 1.1 and ti = 5.1) and 2010 (k = 2.7 and ti = 8.2) (Fig. 2). Methanol-extractable 14C-residues (ME) were relatively low, with values below 10% of the initially added 14C-IPU. Significantly higher amounts of extractable residues were quantified in 2009 as compared to 2008 (p < 0.05). Conversely, non-extractable 14C-residues (BR), corresponding to bound residues, were relatively high: they averaged 45% of the initially added 14C-ring-labelled IPU. As already observed for extractable residues, significantly higher amounts of bound residues were detected in 2009 as compared to 2008 and 2010 (Fig. S2). These observations suggest that in samples with low IPU-degrading ability, IPU or related metabolites formed bound residues. The global structure of the soil microbial communities was assessed by Automated Ribosomal Intergenic Spacer Analysis (ARISA) performed on DNA extracted directly from the soil. Relatively similar A-RISA fingerprints made of 100 bands were obtained from different sub-site samples (data not shown). A Principal Component Analysis (PCA) was applied for the pair-wise comparison of the numbers and the relative intensity of the bands observed in A-RISA profile. The first principal component explained 29%, 32% and 38% of the data variance in 2008, 2009 and 2010, respectively (data not shown), while the second principal component accounted for 21%, 15% and 28% of the variance in 2008, 2009 and 2010, respectively. Most of the sub-site samples were centered on the factorial map. In 2010, several sub-site samples (D6, D4, C7), and A7, A5, A9 and A6 to some extent, were scattered along axis 1 (Fig. S3).

B

6

40 4 30 2

20

0

10 0

2008

0

8

2009

35

C

6

0

2010

2008

2009

2010

D

30 25

4

20 2

15

0

10

-2 -4

5 0

2008

2009

2010

0

0

2008

2009

2010

Fig. 2. Boxes and whisker plots showing the 2008, 2009 and 2010 distribution of isoproturon-mineralization time-course parameters. A–D represent the maximum percentages of mineralization (A), maximum mineralization rates (lm), lag phases (k) and inflection point abscissa (ti) respectively.

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

3.2. Linking IPU-degrading activity to physico-chemical and microbiological parameters recorded over a three-year period in the Epoisses field, using multivariate analysis A PCA was performed on the physico-chemical properties and microbiological and mineralization kinetics parameters recorded over the three-year period. The three 1-year periods appeared well separated along axis one, which explained 36.7% of the total variance. An ANOSIM showed that each of the 3 years displayed significantly different overall characteristics (p < 0.001). The analysis of the contribution of each studied parameter to in-between year variance indicated that lM (maximal IPU-mineralization rate) and A (maximal IPU mineralization) were the main two parameters accounting for differences in samples along the first axis of the PCA (Fig. 3, panel A). Interestingly, years 2009 and 2010, while showing a lower IPU-degrading ability than year 2008, were also characterised by a higher overall variance. pH, organic matter content and C/N ratio were the parameters that explained the dispersion of the samples along the second axis of the PCA (Fig. 3, panel A). We calculated Pearson’s correlation coefficients between all the variables over the 3 years (Fig. 3, panel B). We found that the maximal rate of IPU mineralization (lM) and the maximum percentage of IPU mineralization (A) were weakly but significantly positively correlated to the number of culturable bacteria (r = 0.43 and r = 0.34) and inversely correlated to several soil physico-chemical parameters including equivalent humidity (r = 0.61 and r = 0.53), total N (r = 0.63 and r = 0.54), CEC (r = 0.41 and r = 0.40) and volatile matters (r = 0.34 and r = 0.43). Conversely, lag phase (k) and inflection point abscissa (ti) values, which both reflect the initial abundance of the IPU-degrading community in soil samples, were positively correlated to equivalent humidity (r = 0.68 and r = 0.72), total N (r = 0.55 and r = 0.65) and CEC (r = 0.38 and r = 0.42), and inversely correlated to the number of culturable bacteria (r = 0.46 and r = 0.44). Surprisingly, pH, which is a known driver of IPU-degrading activity, was not found to significantly affect the IPU-degrading parameters we measured over the 3 years, while PCA analysis showed that pH strongly contributed to the increase in between-samples variance observed in 2009 and 2010. Consequently, in order to further elucidate the contribution of soil physico-chemical properties to the variations of IPU-degrading parameters over the three-year period, we developed a multiple linear regression model allowing for year-dependent (therefore treatment-dependent too) correlation between soil physico-chemical properties and IPU-degrading parameters. We focused on lM and A (Table 2). Our rather simple model explained 84.5% and 67.1% of lM and A variations, respectively. In both cases, the soil characteristics that significantly contributed to variations in IPUdegrading parameters were pH and Equivalent Humidity. While Equivalent Humidity seemed to positively affect lM values in 2008; it had the opposite effect in 2009 and 2010, as shown by the change in the angle of the regression plane (Fig. 4). The effect of pH on lM variations was on the other hand uniformly weak and positive throughout the 3 years (Fig. 4), as indicated by the absence of significant YearpH interaction in the final model. 3.3. Field-scale patterns of IPU-degrading activity in the Epoisses, field over a three-year period 3.3.1. Physico-chemical properties A spherical variogram model was fitted to each physico-chemical property measured in the experimental field to investigate their field-scale pattern over the 3 years surveyed. The parameters estimated from the variogram, i.e. sill, nugget, range and Q values for all the physico-chemical properties are presented in Table 1. In 2008, physico-chemical properties showed frequent 50-m or more

2505

spatial dependence, except Equivalent Humidity, organic matter, organic carbon and C/N ratio which displayed moderate range values of 24.9, 31.9, 34.0 and 36.8 m, respectively (Table 1). Q values for almost all physico-chemical properties were more than 0.5, except for total N content and pH whose Q values were below 0.5. High Q values indicate that the variance of the physico-chemical property could be explained by the semi-variogram model, further suggesting their highly developed spatial structure. Maps of physicochemical properties were drawn by ordinary kriging based on the data within the range of observed values. The kriged maps of soil pH and Equivalent Humidity are presented in Fig. 5. Organic matter content, organic carbon content, nitrogen and C/N ratio yielded field-scale patterns relatively similar to pH, while Equivalent Humidity had a similar pattern to CEC (Fig. 5). The variation range of all those soil properties was higher in 2009 and in 2010 than in 2008. 3.3.2. Microbiological properties and mineralization kinetics parameters Spatial variation of IPU-degrading parameters was also investigated by applying a spherical variogram model to the mineralization time-course parameters over the 3 years. Q values higher than 0.5 were observed in 2009 and in 2010 for most of the microbiological properties and IPU-degrading parameters, while lower Q values were observed in 2008, which is consistent with the lower variability observed for most of those parameters (A, lM, k and ti) in 2008, as shown in Figs. 1 and 3. Kriged maps were drawn for the mineralization time-course parameters (A, lM, k and ti). Kriged maps for lM are shown in Fig. 5C. Over the 3 years, lM showed similar field-scale patterns, with the highest mineralization rate observed in the top-left corner of the field. The lowest variability in lM was observed in 2008 (i.e. 3.6 < lM < 4.4) while the lowest mineralization rates were observed in 2009 (i.e. 0.4 < lM < 1.2) and the highest variation range in 2010 (i.e. 1 < lM < 5) (Fig. 5). While field-scale maps of A over the 3 years are similar to those of lM, k and ti showed similar but opposite field-scale patterns: the lowest values were recorded in the top-left corner of the field (data not shown). These results are consistent with the strong negative correlations observed between lM and A on the one hand, and ti and k on the other (Fig. 4). 4. Discussion The present study was carried out to assess the field-scale pattern of pesticide-mineralization activity using IPU as a model in relation to biotic and abiotic parameters over a three-year crop rotation, and also to estimate its variability range in relation to herbicide treatment. Our results indicate that over the three-year survey the cumulative percentage of IPU mineralization (A) ranged between 5% and 53% of the 14C-IPU initially added. This observation confirmed that the soil microflora of the field was adapted to IPU mineralization as previously reported (Hussain et al., 2011). We could hypothesise that repeated exposure of the soil microflora to IPU over the last decade in a field cropped with a wheat/rape seed/barley crop rotation (1 IPU treatment treatment every 2 years, 6 treatments over the previous 10 years) led to its adaptation to IPU mineralization. However, we observed that the variability of cumulative percentages of IPU mineralization (A) as well as of maximum mineralization rates (lm) was significantly lower in 2008 (CVA 5.8% and CVlm 15.6%) than in 2009 and 2010 (CVA 26.2% and 22.2%; CVlm 53.7, 45.4, respectively). The relatively high variability of IPU mineralization within our field is in agreement with a number of previous studies reporting spatial variability of the degradation rates of several pesticides including carbofuran (Parkin and Sheloton, 1992), azoxystrobin (Bending

2506

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

A

λ

B

λ

λ

Fig. 3. Panel A. Principal Component Analysis performed on the whole dataset recorded over the three-year period on our experimental field cropped with a winter wheat/ rape seed/barley rotation. A/ Principal component analysis (PC1  PC2, representing 36.7% and 22.8% of the variances, respectively) generated from parameter-clustering by year (2008 in black, 2009 in red and 2010 in green). B/ Representation of the contribution of each measured parameter to the principal component analysis performed on the three-year data set. Parameters are (i) physico-chemical parameters: equivalent humidity (EH), organic matter content (OM), organic carbon content (OC), nitrogen (N), C/N ratio, cation exchange capacity (CEC) and pH; (ii) biological parameters: microbial C biomass (MCB), number of cultivable bacteria (cfu), maximum percentage of IPU mineralization (A), maximum rate of mineralization (lm), lag phase (k), inflection point abscissa (ti) and (iii) pesticide residues: methanol-extractable residues (ME) and bound residues (BR). Panel B. Correlation matrix between the physico-chemical, biological and pesticide-residue parameters measured over the three-year survey carried out on our experimental field cropped with a winter wheat/rape seed/barley rotation. The color scale above the figure indicates correlation intensity and direction. Stars indicate significant correlation between parameters estimated with P values corrected for multiple testing (Bonferroni test). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

et al., 2006), chlorsulfuron (Walker and Brown, 1983), diflufenican (Bending et al., 2006), isoproturon (Walker et al., 2001; El-Sebai et al., 2007) and chlorotoluron (Walker et al., 2002). It is noteworthy

that means of cumulative percentages of IPU mineralization (A) and of maximum IPU mineralization rates (lm) were significantly higher in 2008 and in 2010 than in 2009. On the contrary,

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511 Table 2 Multiple linear regression explaining lM (maximum rate of mineralization) and A (maximum percentage of IPU mineralization) variations.

lM

Df

Mean square

F-value

Year Equivalent Humidity pH YearEquivalent Humidity Residuals % Explained variance

2 1 1 2 97 84.5

6.91 0.70 0.57 0.28 0.03 Model significance

250.75*** 25.491*** 20.821*** 10.003***

A Year Equivalent Humidity pH YearEquivalent Humidity Residuals % Explained variance

2 1 1 2 97 67.1

613.03  103 81.14  103 51.71  103 31.03  103 6.58  103 Model significance

93.24*** 12.34*** 7.86** 4.72*

F6,97 = 94.64***

F6,97 = 36.02***

Df: Degree of freedom, % explained variance is defined as the adjusted R2 of the full model. *** p < 0.001. ** p < 0.01. * p < 0.05.

cumulative mean values of lag phase (k) and of inflection point abscissa (ti) were significantly lower in 2008 and in 2010 than in 2009. All these observations reveal that IPU mineralization activity was higher in 2008 and in 2010 than in 2009. This difference in IPU-mineralization activity might be related to the herbicide application schedule for the field over the three-year crop rotation (winter wheat/rape seed/barley. Based on our results showing that lag phase mean values (k) were lower in 2008 than in 2009 and in 2010, we can hypothesise that the IPU treatment applied in 2008 promoted the growth of the microbial community able to mineralize that particular herbicide, thereby decreasing lag phase length, when the size of the pesticide-degrading community increases to reach its optimum. This is in accordance with previous studies showing that repeated atrazine herbicide treatment performed on corn crops led to an increase in the size of the atrazinedegrading community and consequently to a reduction of the lag phase length (Piutti et al., 2002). The result is also in accordance with a study performed under laboratory conditions reporting that repeated treatment of soil microcosms with IPU led to a considerable reduction of lag phase length (El-Sebai et al., 2005). It is noteworthy that the growth of the IPU-degrading community in response to IPU treatment maximised IPU mineralization in almost all sub-site soil samples, thereby lowering the variability of IPU

2507

mineralization activity recorded at the field-scale in 2008. Conversely, in 2009, in the absence of IPU treatment, the IPUdegrading ability significantly decreased while its variability increased. Interestingly, although no IPU was applied in 2010, high IPU mineralization ability was observed. We could hypothesise that the treatment with sulfonylurea herbicides, whose chemical structure displays similarities with substituted phenylureas (sharing the phenyl ring and urea side chain), promoted the activity of the IPU-degrading community. In order to test this hypothesis, IPU mineralization was recorded on soil microcosms treated with IPU, ArchipelÒ or water as a control, and incubated under laboratory conditions (Fig. S4). This experiment further confirmed that IPU treatment significantly reduced lag time (k) and increased the maximal percentage of IPU-mineralization (A) as compared to the control (p < 0.05). However, it also indicated that sulfonylureas did not affect IPU mineralization, thereby infirming our hypothesis suggesting the existence of another explaining parameter. In order to identify the key parameters that influenced IPU-mineralizing activity over the three-year rotation, a multivariate analysis was performed. It further confirmed previous observations by clearly distinguishing year 2008 from year 2009 and placing year 2010 at an intermediary position considering lM (maximal rate of IPU mineralization) and A (maximal IPU mineralization) parameters. Physico-chemical parameters including total N, Equivalent Humidity and CEC were shown to be significantly inversely correlated to the IPU-mineralization activity. Several studies underline the importance of soil pH in the spatial variability of IPU-degrading ability, with rapid IPU biodegradation usually characterised by higher pH values (Walker et al., 2001; El Sebai et al., 2007). Yet, surprisingly, overall the multivariate analysis of our data did not identify pH as a parameter correlated to IPU mineralization. However, for years 2009 and 2010, pH was found to be significantly positively correlated to the IPU-mineralization activity. Indeed this observation further reinforces the hypothesis that the 2008 IPU treatment lowered the effect of intrinsic parameters, including pH, by favouring a homogeneous growth of the IPU-mineralizing community all over the field. However, the impact of these intrinsic parameters on IPU-mineralizing activity became more obvious in the absence of the selection pressure exerted by IPU. In order to get further insight into the contribution of soil physico-chemical properties to isoproturon-mineralizing ability, a multiple-linear regression model was developed. Our rather simple model, which relied on soil pH and Equivalent Humidity, explained 84.5% and 67.1% of the variations in lm and A, respectively. It should be noted that similar modelling of isoproturon-mineralizing activity were

Fig. 4. Contribution of pH and of Equivalent Humidity to the observed variation of lM during the 3 years surveyed. The predicted lM values of the final model (Table 2) for each year are represented by the regression plane. A plain or a dotted line links observed values to the regression plane and hence represents the residuals of the multiple regression model. A plain line is used if the observed value is higher than the predicted value and a dotted line is used in the opposite case.

2508

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

A

7,2 7,4 7,6 7,8 8,0

7,2 7,4 7,6 7,8 8,0

7,2 7,4 7,6 7,8 8,0

18 20 22 24 26

18 20 22 24 26

18 20 22 24 26

3,6 3,8 4,0 4,2 4,4 4,6 4,8

0 1 2 3 4 5

0 1 2 3 4 5

B

C

Fig. 5. Kriged map of soil pH (panel A), equivalent humidity (panel B) and maximum mineralization rate (lM, panel C) of Epoisses field sampled in 2008, 2009 and 2010.

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

obtained with pH and CEC. Interestingly, we observed that Equivalent Humidity affected isoproturon-mineralizing activity positively in 2008, while it had the opposite effect in 2009 and 2010. On the contrary, pH had a uniformly weak, positive effect over the 3 years. These two observations are in agreement with previous reports showing that soil pH and moisture affect pesticide-degrading activity. Our fairly good modelling of IPU-mineralizing activity based on two easily measurable physico-chemical parameters offers interesting research perspectives to predict pesticide biodegradation in different arable soils. Our study showed that IPU-mineralizing activity considerably varied over the field. In order to map field spatial distribution patterns of IPU mineralization, our dataset was analysed using geostatistical tools. As observed in previous studies (LopezGranados et al., 2002; El-Sebai et al., 2007), all the parameters considered were log-normally distributed and could be used for analysis by univariate geostatistical tools. Most of the parameters showed high spatial variability but some of them, including Equivalent Humidity, nitrogen content, pH and number of culturable heterotrophic bacteria were characterised by relatively low coefficients of variation (CV < 10%) as previously observed (Walker et al., 2001; El-Sebai et al., 2007). The physico-chemical and microbiological parameters were analysed using geostatistical tools to explore the spatial variation of the regionalised variables, to perceive the spatial structure of soil parameters and to predict the values of soil attributes at unsampled locations by kriging. Geostatistics has already been used to investigate the spatial variation of IPU mineralization within a French agricultural field (El-Sebai et al., 2007) and to map the field-scale distribution patterns of denitrifiers (Philippot et al., 2009). The in-field spatiallydependent variability of the parameters recorded over our three year-survey was estimated from Q values. In 2009 and in 2010, most of the parameters measured showed significant Q values (Q > 0.8). In 2008, total N, pH, CEC, MCB, cfu, A, lm, k and ti showed low Q values that were either non-significant (Q < 0.5) or moderately significant (0.5 < Q < 0.8). Highly developed spatial structure (Q > 0.8) was observed for all the parameters in 2009 and in 2010, except Equivalent Humidity in 2009 and lm in 2010, which had moderately developed spatial structure (Q  0.5). The existence of varying spatial classes in soil parameters is in accordance with previous studies (Yanai et al., 2001; Lopez-Granados et al., 2002; El-Sebai et al., 2007). Field-scale spatial distribution patterns of maximum percentages of mineralization (A) and maximum mineralization rates (lM) were similar: they were characterised by a patch with a high mineralization rate within the 0- to 30-m strips in the years 2009 and 2010 (Fig. 5). Similar patches were observed in kriged map of lm in 2008, but with a lower extent of variability. These observations thereby reinforce our present results, indicating a lower mineralization activity in 2009 as compared to 2008 and 2010, and a considerably higher variability in mineralization activity in 2010 as compared to 2008. It is noteworthy that the field-scale distribution pattern of lag phase (k) values and of inflection point abscissa (ti) were the opposite of A and lM, with their lowest values in the patches where the highest values of lm and A were recorded (data not shown). Field-scale distribution patterns of isoproturondegrading activity did not share any similarities with the maps representing the genetic structure of the bacterial communities and the maps representing the number of culturable heterotrophic bacteria and microbial C biomass (data not shown). This suggests that the abundance of total heterotrophic bacteria and the overall genetic structure of the bacterial communities were not related to mineralization activity within the field, which is in accordance with previous studies that yielded similar results (El-Sebai et al., 2007). Monitoring the specific abundance of isoproturon degraders by a qPCR assay targeting IPU-degrading genes would have been of

2509

interest, but unfortunately the IPU-degrading genes have remained unknown up to know. Interestingly, field-scale spatial distribution patterns of several physico-chemical parameters such as organic matter content, organic carbon content, C/N ratio, CEC and Equivalent Humidity were similar to those of mineralization activity (lM and A). In 2009 and in 2010, patches with the highest values were found in the same strip (0–30 m) as already observed for lM. In addition, these observations were reinforced by Pearson’s coefficient of correlation (r), which showed that 2009 and 2010 lm values were considerably correlated with organic matter content (r = 0.611 and 0.440), organic carbon content (r = 0.612 and 0.440) and C/N ratio (r = 0.625 and 0.566). Thus, our results point out the importance of soil organic matter content in determining the spatial variability of pesticide mineralization activity. They are in accordance with a number of previous studies reporting organic matter as an important driver of pesticide biodegradation in soils by favouring the proliferation of microorganisms and increasing their activity (Spark and Swift, 2002; Saison et al., 2006; Briceno et al., 2007). Moreover, in 2009 and in 2010, soil pH was correlated with lM (r = 0.493 and 0.453). Field-scale spatial distribution patterns of pH were also patchy, with highest values in the strip where the highest lM values were recorded. Soil pH is admittedly a main driver of microbial pesticide biodegradation in soils (Bending et al., 2001; El-Sebai et al., 2005, 2007) as well as in pure cultures (Hussain et al., 2009; Sun et al., 2009; Hussain et al., 2011). Interestingly, field-scale spatial distribution patterns of cation exchange capacity (CEC) and Equivalent Humidity in 2009 and in 2010 were also patchy, with lowest values where lM values were highest. This is in agreement with multivariate analyses, which reveal a negative correlation of these two parameters with lm in 2009 and in 2010 (CEC r = 0.578 and 0.709; Equivalent Humidity r = 0.498 and 0.606). Altogether our results indicate that although physico-chemical parameters are moderately correlated with mineralization, they increase isoproturon degradation potential and are involved in biodegradation rate heterogeneity (Vinther et al., 2001; El-Sebai et al., 2007). The multivariate analysis further reinforced these observations by identifying soil pH, Equivalent Humidity and CEC as the main drivers of the isoproturon-mineralization activity. The modelling of lM as a function of soil pH and equivalent humidity represents up to 85% of observed variance. It therefore offers interesting perspectives for predicting pesticide biodegradation at the field-scale using simple parameters. 5. Conclusions Pesticide-mineralization activity varied spatially and temporally in our experimental field over a three-year crop rotation. The pesticide treatment schedule was identified as a main driver of IPU mineralization activity. At the field scale, IPU-mineralizing activity heterogeneity was found to be driven by soil pH, Equivalent Humidity and CEC. We can therefore conclude that soil physicochemical parameters in combination with pesticide treatment schedules should be considered for predicting the variance in the level of pesticides biodegradation in arable soils. Acknowledgements The PhD work of Sabir Hussain was funded by Higher Education Commission (Pakistan) and the Doctoral School of the University of Burgundy (France). Nicolas Mugnier-Jolain (UMR Agroéecology, INRA) is thanked for giving access to the experimental field of Epoisses part of the long-term assay « PIC adventices ». Marie-Hélène Bernicot and the team of experimental farm of

2510

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511

Bretenières is also thanked for their helpful assistance in the field. We would like to thank Annie Buchwalter for editing the manuscript. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chemosphere. 2012.10.080. References Beck, A.J., Harris, G.L., Howse, K.R., Johnston, A.E., Jones, K.C., 1996. Spatial and temporal variation of isoproturon residues and associated sorption/desorption parameters at the field scale. Chemosphere 33, 1283–1295. Behera, B.C., Bhunya, S.P., 1990. Genotoxic effect of isoproturon (herbicide) as revealed by three mammalian in vivo mutagenic bioassays. Indian J. Exp. Biol. 28, 862–867. Bending, G.D., Shaw, E., Walker, A., 2001. Spatial heterogeneity in the metabolism and dynamics of isoproturon degrading microbial communities in soil. Biol. Fert. Soils 33, 484–489. Bending, G.D., Lincoln, S.D., Sorensen, S.R., Morgan, J.A.W., Aamand, J., Walker, A., 2003. In-field spatial variability in the degradation of the phenyl-urea herbicide isoproturon is the result of interactions between degradative Sphingomonas spp. and soil pH. Appl. Environ. Microb. 69, 827–834. Bending, G.D., Lincoln, S.D., Edmondson, R.N., 2006. Spatial variation in the degradation rate of the pesticides isoproturon, azoxystrobin and diflufenican in soil and its relationship with chemical and microbial properties. Environ. Pollut. 139, 279–287. Briceno, G., Palma, G., Duran, N., 2007. Influence of organic amendment on the biodegradation and movement of pesticides. Crit. Rev. Environ. Sci. Technol. 37, 233–271. Chaussod, R., Houot, S., Guiraud, G., Hetier, J.M., 1988. Size and turnover of the microbial biomass in agricultural soils: laboratory and field measurements. In: Jenkinson, D.S., Smith, K.A. (Eds.), Nitrogen Efficiency in Agricultural Soils. Elsevier, London, pp. 312–326. Cox, L., Walker, A., Welch, S.J., 1996. Evidence for the accelerated degradation of isoproturon in soils. Pestic. Sci. 48, 253–260. Cullington, J.E., Walker, A., 1999. Rapid biodegradation of diuron and other phenylurea herbicides by a soil bacterium. Soil Biol. Biochem. 31, 677–686. Dewez, D., Didur, O., Vincent-Heroux, J., Popovic, R., 2008. Validation of photosynthetic-fluorescence parameters as biomarkers for isoproturon toxic effect on alga Scenedesmus obliquus. Environ. Pollut. 151, 93–100. Dosnon-Olette, R., Trotel-Aziz, P., Couderchet, M., Eullaffroy, P., 2010. Fungicides and herbicide removal in Scenedesmus cell suspensions. Chemosphere 79, 117– 123. Dray, S., Dufour, A.B., 2007. The ade4 package: implementing the duality diagram for ecologists. J. Stat. Soft. 22 (4). El-Sebai, T., Lagacherie, B., Cooper, J.F., Soulas, G., Martin-Laurent, F., 2005. Enhanced isoproturon mineralization in a clay silt loam agricultural soil. Agron. Sustain. Dev. 25, 271–277. El-Sebai, T., Lagacherie, B., Soulas, G., Martin-Laurent, F., 2007. Spatial variability of isoproturon mineralizing activity within an agricultural field: geostatistical analysis of simple physicochemical and microbiological soil parameters. Environ. Pollut. 145, 680–690. El-Sebai, T., Devers-Lamrani, M., Lagacherie, B., Rouard, N., Soulas, G., MartinLaurent, F., 2011. Isoproturon mineralization in an agricultural soil: impact of temperature and moisture content. Biol. Fert. Soils 47, 427–435. Ertli, T., Marton, A., Foldenyi, R., 2004. Effect of pH and the role of organic matter in the adsorption of isoproturon on soils. Chemosphere 57, 771–779. Fournier, J.C., Soulas, G., Catroux, G., 1975. Degradation microbienne de l’isoproturon dans des modeles de laboratoire. Chemosphere 4, 207–214. Gaillardon, P., Sabar, M., 1994. Changes in the concentrations of isoproturon and its degradation products in soil and soil solution during incubation at two temperatures. Weed Res. 34, 243–250. Hazarika, A., Sarkar, S.N., 2001. Effect of isoproturon pretreatment on the biochemical toxicodynamics of anilofos in male rats. Toxicology 165, 87–95. Hoshiya, T., Hasegawa, R., Hakoi, K., Cui, L., Ogiso, T., Cabral, R., Ito, N., 1993. Enhancement by non-mutagenic pesticides of GST-P positive hepatic foci development initiated with diethylnitrosamine in the rat. Cancer Lett. 72, 59– 64. Hundt, K., Wagner, M., Becher, D., Hammer, E., Schauer, F., 1998. Effect of selected environmental factors on degradation and mineralization of biaryl compounds by the bacterium Ralstonia picketii in soil and compost. Chemosphere 36, 2321–2335. Hussain, S., Sorensen, S.R., Devers-Lamrani, M., El-Sebai, T., Martin-Laurent, F., 2009. Characterization of an isoproturon mineralizing bacterial culture enriched from a French agricultural soil. Chemosphere 77, 1052–1059. Hussain, S., Devers-Lamrani, M., El Azhari, N., Martin-Laurent, F., 2011. Isolation and characterization of an isoproturon mineralizing Sphingomonas sp. strain SH from a French agricultural soil. Biodegradation 22, 637–650. Isaaks, E.H., Srivastava, R.M., 1989. An Introduction to Applied Geostatistics. Oxford University Press, Oxford. Journel, A.G., Huijbregts, C.J., 1978. Mining Geostatistics. Academic Press, London.

Karpouzas, D.G., Walker, A., Froud-Williams, R.J., Drennan, D.S.H., 1999. Evidence for the enhanced biodegradation of ethoprophos and carbofuran in soils from Greece and the UK. Pestic. Sci. 55, 301–311. Knauert, S., Singer, H., Hollender, J., Knauer, K., 2010. Phytotoxicity of atrazine, isoproturon, and diuron to submersed macrophytes in outdoor mesocosms. Environ. Pollut. 158, 167–174. Kumar, S., 2010. Effect of 2,4-D and isoproturon on chromosomal disturbances during mitotic division in root tip cells of Triticum aestivum L. Cytol. Genet. 44, 79–87. Lopez-Granados, F., Jurado-Exposito, M., Atenciano, S., Garcia-Ferrer, A., de la Orden, M.S., Garcia-Torres, L., 2002. Spatial variability of agricultural soil parameters in southern Spain. Plant Soil 246, 97–105. Mansour, M., Feicht, E.A., Behechti, A., Schramm, K.W., Kettrup, A., 1999. Determination photostability of selected agrochemicals in water and soil. Chemosphere 39, 575–585. Martin-Laurent, F., Philippot, L., Hallet, S., Chaussod, R., Germon, J.C., Soulas, G., Catroux, G., 2001. DNA extraction from soils: old bias for new microbial diversity analysis methods. Appl. Environ. Microb. 67, 2354–2359. Muller, K., Bach, M., Hartmann, H., Spiteller, M., Frede, H.G., 2002. Point- and nonpoint-source pesticide contamination in the Zwester Ohm catchment, Germany. J. Environ. Qual. 31, 309–318. Orton, F., Lutz, I., Kloas, W., Routledge, E.J., 2009. Endocrine disrupting effects of herbicides and pentachlorophenol: in vitro and in vivo evidence. Environ. Sci. Technol. 43, 2144–2150. Parkin, T.B., Shelton, D.R., 1992. Spatial and temporal variability of carbofuran degradation in soil. J. Environ. Qual. 21, 672–678. ˇ uhel, J., Saby, N.P.A., Chèneby, D., Chronˇáková, A., Bru, D., Arrouays, D., Philippot, L., C Martin-Laurent, F., Šimek, M., 2009. Mapping field-scale spatial patterns of size and activity of the denitrifier community. Environ. Microb. 11, 1518–1526. Pieuchot, M., Perrin-Ganier, C., Portal, J.M., Schiavon, M., 1996. Study on the mineralization and degradation of isoproturon in three soils. Chemosphere 33, 467–478. Piutti, S., Hallet, S., Rousseaux, S., Philippot, L., Soulas, G., Martin-Laurent, F., 2002. Accelerated mineralisation of atrazine in maize rhizosphere soil. Biol. Fert. Soils 36, 434–441. Ranjard, L., Lejon, D.P.H., Mougel, C., Schehrer, L., Merdinoglu, D., Chaussod, R., 2003. Sampling strategy in molecular microbial ecology: influence of soil sample size on DNA fingerprinting analysis of fungal and bacterial communities. Environ. Microb. 5, 1111–1120. Rodriguez-Cruz, M.S., Jones, J.E., Bending, G.D., 2006. Field-scale study of the variability in pesticide biodegradation with soil depth and its relationship with soil characteristics. Soil Biol. Biochem. 38, 2910–2918. Saison, C., Degrange, V., Oliver, R., Millard, P., Commeaux, C., Montange, D., Le Roux, X., 2006. Alteration and resilience of the soil microbial community following compost amendment: effects of compost level and compost-borne microbial community. Environ. Microb. 8, 247–257. Smith, E.A., Prues, S.L., Oehme, F.W., 1997. Environmental degradation of polyacrylamides. 2. Effects of environmental (outdoor) exposure. Ecotoxicol. Environ. Saf. 37, 76–91. Sorensen, S.R., Aamand, J., 2001. Biodegradation of the phenylurea herbicide isoproturon and its metabolites in agricultural soils. Biodegradation 12, 69–77. Sorensen, S.R., Aamand, J., 2003. Rapid mineralisation of the herbicide isoproturon in soil from a previously treated Danish agricultural field. Pest Manage. Sci. 59, 1118–1124. Sorensen, S.R., Bending, G.D., Jacobsen, C.S., Walker, A., Aamand, J., 2003. Microbial degradation of isoproturon and related phenylurea herbicides in and below agricultural fields. FEMS Microb. Ecol. 45, 1–11. Soulas, G., 1993. Evidence for the existence of different physiological groups in the microbial community responsible for 2,4-D mineralization in soil. Soil Biol. Biochem. 25, 443–449. Spark, K.M., Swift, R.S., 2002. Effect of soil composition and dissolved organic matter on pesticide sorption. Sci. Total Environ. 298, 147–161. Spliid, N.H., Koppen, B., 1998. Occurrence of pesticides in Danish shallow ground water. Chemosphere 37, 1307–1316. Stangroom, S.J., Collins, C.D., Lester, J.N., 1998. Sources of organic micropollutants to lowland rivers. Environ. Technol. 19, 643–666. Sun, J.Q., Huang, X., Chen, Q.L., Liang, B., Qiu, J.G., Ali, S.W., Li, S.P., 2009. Isolation and characterization of three Sphingobium sp strains capable of degrading isoproturon and cloning of the catechol 1,2-dioxygenase gene from these strains. World J. Microbiol. Biotechnol. 25, 259–268. Thioulouse, J., Chessel, D., Dole´dec, S., Olivier, J.-M., 1997. ADE-4: a multivariate analysis and graphical display software. Stat. Comput. 7, 75–83. Vallotton, N., Eggen, R.I.L., Chevre, N., 2009. Effect of sequential isoproturon pulse exposure on Scenedesmus vacuolatus. Arch. Environ. Contam. Toxicol. 56, 442– 449. Vieuble-Gonod, L., Benoit, P., Cohen, N., Houot, S., 2009. Spatial and temporal heterogeneity of soil microorganisms and isoproturon degrading activity in a tilled soil amended with urban waste composts. Soil Biol. Biochem. 41, 2558– 2567. Vinther, F.P., Elsgaard, L., Jacobsen, O.S., 2001. Heterogeneity of bacterial populations and pesticide degradation potentials in the unsaturated zone of loamy and sandy soils. Biol. Fert. Soils 33, 514–520. Walker, A., Brown, P.A., 1983. Measurement and prediction of chlorsulfuron persistence in soil. Bull. Environ. Contam. Toxicol. 30, 365–372. Walker, A., Moon, Y.H., Welch, S.J., 1992. Influence of temperature, soil-moisture and soil characteristics on the persistence of alachlor. Pestic. Sci. 35, 109–116.

S. Hussain et al. / Chemosphere 90 (2013) 2499–2511 Walker, A., Jurado-Exposito, M., Bending, G.D., Smith, V.J.R., 2001. Spatial variability in the degradation rate of isoproturon in soil. Environ. Pollut. 111, 407–415. Walker, A., Bromilow, R.H., Nicholls, P.H., Evans, A.A., Smith, V.J.R., 2002. Spatial variability in the degradation rates of isoproturon and chlorotoluron in a clay soil. Weed Res. 42, 39–44. Welp, G., Brummer, G.W., 1999. Adsorption and solubility of ten metals in soil samples of different composition. J. Plant Nutr. Soil Sci.-Z. Pflanzenernahr. Bodenkd. 162, 155–161. Widenfalk, A., Bertilsson, S., Sundh, I., Goedkoop, W., 2008. Effects of pesticides on community composition and activity of sediment microbes – responses at various levels of microbial community organization. Environ. Pollut. 152, 576– 584.

2511

Wu, J., Brookes, P.C., Jenkinson, D.S., 1996. Evidence for the use of a control in the fumigation-incubation method for measuring microbial biomass carbon in soil. Soil Biol. Biochem. 28, 511–518. Yanai, J., Lee, C.K., Kaho, T., Iida, M., Matsui, T., Umeda, M., Kosaki, T., 2001. Geostatistical analysis of soil chemical properties and rice yield in a paddy field and application to the analysis of yield-determining factors. Soil Sci. Plant Nutr. 47, 291–301. Yin, X.L., Jiang, L., Song, N.H., Yang, H., 2008. Toxic reactivity of wheat (Triticum aestivum) plants to herbicide isoproturon. J. Agric. Food Chem. 56, 4825– 4831. Zwietering, M.H., Jongenburger, I., Rombouts, F.M., Vantriet, K., 1990. Modeling of the bacterial-growth curve. Appl. Environ. Microb. 56, 1875–1881.