SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 2 December 2005
Published in Soil Sci Soc Am J 70:64-71 (2006)
DOI: 10.2136/sssaj2004.0034
© 2005 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow An erratum has been published
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Vieublé Gonod, L.
Right arrow Articles by Chenu, C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Vieublé Gonod, L.
Right arrow Articles by Chenu, C.
Agricola
Right arrow Articles by Vieublé Gonod, L.
Right arrow Articles by Chenu, C.
Related Collections
Right arrow Spatial Distribution
Right arrow Bioremediation and Biodegradation
Right arrow Soil Biology

Soil Biology & Biochemistry

Spatial Distribution of Microbial 2,4-Dichlorophenoxy Acetic Acid Mineralization from Field to Microhabitat Scales

Laure Vieublé Gonoda, Joël Chadoeufb and Claire Chenua,*

a INRA-Science du Sol, Rte de St Cyr, 78026 Versailles, France
b INRA-Unité de Biométrie, Domaine St Paul-Site Agroparc, 84914 Avignon, France

* Corresponding author (chenu{at}grignon.inra.fr)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Little is known about spatial variability of microbial activity, particularly at microscales. This is especially true for the fate and degradation of pesticides. The objective was to sample soil from micro to field scales and apply geostatistics on the potential mineralization of a widely used herbicide 2,4-dichlorophenoxy acetic acid (2,4-D; C8H6Cl2O3). Soil cores were sampled in the plow layer of a cultivated soil with a systematic sampling procedure. In a first experiment 2,4-D mineralization was measured on 39 crushed cores and we analyzed variability of mineralization at the field scale, from decameter to meter and from meter to decimeter scale. In a second experiment, 432 soil cubes (about 216 mm3) were used to study the variability of mineralization at the "microhabitat" scale from meter to millimeter. The spatial dependence of 2,4-D mineralization was first quantified by computing an empirical variogram function. Spatial independence was then tested by comparing the empirical variogram function to its individual confidence bounds at 95% level obtained under independence assumption by a Monte-Carlo Method.

The potential for 2,4-D mineralization was spatially heterogeneous from field to microhabitat. Mineralization variability increased when the scale decreased from field to microhabitats.

Specifically, the coefficient of variability (CV) of 2,4-D mineralization was of 18.5% at the field scale, 7–22% at the meter scale, 47.9% at the inter-cores scale, and 25–160% at the intra-core scale (microhabitat scale). 2,4-D mineralization was spatially structured only at the microhabitat scale (hot spots). This variability at fine scales should be considered when sampling soil processes involving microbial activities.

Abbreviations: 2,4-D, 2,4 dichlorophenoxy acetic acid • CV, coefficient of variability • MW, molecular weight • {psi}, water potential


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
SOIL DISPLAYS large spatial heterogeneity of its biological and physicochemical characteristics. This prevents the establishment and use of predictive relationships between soil characteristics and microbial processes. Several studies have analyzed the spatial distribution of soil microbial processes and microorganisms at different scales. For example, the spatial variability of pesticide degradation has been investigated in agricultural fields because of the implications for the longevity of these compounds in the soil and for their potential to pollute ground and surface waters (Walker and Brown, 1983; Beck et al., 1996; Bending et al., 2001; Walker et al., 2001). These studies have shown that pesticide degradation displays high spatial variability. For example, Beck et al. (1996) found that the time required for 50% reduction in the concentration of isoproturon ranged from 31 to 483 d in 25 different sampling areas (1 m2) within an agricultural field.

The spatial variability of denitrification has also been extensively studied in soils (Burton and Beauchamp, 1985; Parkin, 1987, 1993; Christensen et al., 1990). Conclusions were similar, in that denitrification in soils has high spatial variability. Parkin (1987) calculated that 25 to 85% of the denitrification in 100-g soil cores was associated with only 0.4 to 0.08% of the soil mass. Likewise, heterogeneity of methanogenesis within single sites has been observed with CV ranging from 40 to 180% (Wachinger et al., 2000) and 340 to 490% (Adrian et al., 1994). Sexstone et al. (1985), Parkin (1987), and Christensen et al. (1990) found that denitrification was mainly localized in hot spots possessing high organic matter contents or in anaerobic zones inside aggregates.

Other studies (Cambardella et al., 1994; Winter and Beese, 1995; Tessier et al., 1998) have found high spatial variability for total soil microorganisms and demonstrated that the CV of the microbial biomass can be as high as 40% in a cultivated field. Furthermore Morris (1999), Grundmann et al. (2001), Gonod et al. (2003), and Nunan et al. (2003) have identified spatial patterns in the distribution of microorganisms. Morris (1999) for instance, differentiated fungal and bacterial biomass (direct counts by fluorescence microscopy) and observed large variability with fungi and bacteria organized in hot spots of 2-cm diam. Grundmann et al. (2001) showed that the nitrifying microorganisms were clustered in patches of a 250-µm diam. Mobarry et al. (1996) observed that Nitrobacter and Nitrosomonas were also organized in clusters. Grundmann et al. (2001) proposed that the number of cells was the main factor controlling the microbial activity at the microhabitat scale.

Some studies have investigated the mechanisms responsible for spatial heterogeneity of biological properties. Walker et al. (2001) showed that shorter isoproturon half-lives were positively correlated with increasing soil pH and soil microbial biomass. Further studies by Bending et al. (2001) focused on areas with contrasting isoproturon degradation rates. Fast degradation of isoproturon was associated with high 14C-isoproturon mineralization and high 14C microbial biomass compared with slowly degrading soils.

However, few systematic studies, with the exception of those on denitrification or methanogenesis, have been conducted on the spatial variability of soil microbial properties from small to large scales. The variability of pesticide mineralization has mainly been investigated at field scales. In a previous study, we concentrated on 2,4-D mineralization at the microscale (Gonod et al., 2003). We found that the mineralization potential of 2,4-D was extremely heterogeneous among millimeter size soil aggregates. However, we do not know whether this variability is specific to this scale or is present at larger scales.

The mineralization of the herbicide (2,4-D) provides a model process for studying varying scales of spatial variability in soils. 2,4-D is widely used to control dicotyledonous weeds in agriculture and turf grass. While 2,4-D is very mobile in soils with low adsorptive capacities, its leaching potential is reduced due to relatively rapid degradation and mineralization rates in soils (Ka et al., 1994). 2,4-D biodegradation is an aerobic process facilitated primarily by bacteria (Ka and Tiedje, 1994; Kamagata et al., 1997). 2,4-D is degraded by specific microorganisms for which it is a source of C and energy (Gunalan, 1991) and by other microorganisms using this substrate by cometabolism (Fournier, 1980; Robertson and Alexander, 1994). 2,4-D is a small molecule (molecular weight [MW] = 220.8) that is fairly soluble (620 mg L–1 in water at 20°C).

Relatively little information is available about spatial distribution of 2,4-D mineralization. Consequently the objectives of the research were to: (i) analyze the variability of potential aerobic microbial 2,4-D mineralization at spatial scales from field to microhabitat level in a cultivated soil and (ii) determine if this variability was random or structured.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Field Site and Soil Sampling
Soil was sampled in October 2001 from the experimental fields of Institut National de la Recherche Agronomique (INRA) located in the park of the Versailles Palace (France). Samples were taken from the surface horizon of a cultivated eutrochrept (USDA classification) before ploughing. The first 3 cm were discarded on all cores. Fifty-four soil cores were taken (diam. 80 mm, height 100 mm) with minimal disturbance of the soil structure. The soil was a silt loam (33% sand, 50% silt, 17% clay) with 13.5 g C kg–1, 1.27 g N kg–1, and a pH (H2O) of 6.8. The field was under continuous wheat (Triticum aestivum) and had never received 2,4-D. Six plots (noted A to F) comprising cores noted 1 to 9 each were defined in the field according to a systematic sampling procedure (Fig. 1 ). Soil cores were preserved at 4°C for 2 mo and before use, were preincubated at 20°C and in their moisture state (0.19 g water g–1) for 3 wk.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 1. Soil core sampling scheme.

 
Two types of experiments were performed on 2,4-D mineralization and were defined according to spatial patterns: on a field scale at the decameter to the meter scale (inter-plots) and at a meter to decimeter scale (intra-plots) (Exp. 1); and on a "microhabitat" scale at meter to millimeter scales (Exp. 2). For the field scale study, cores were gently crushed, homogenized and passed through a 6-mm sieve. For the microhabitat study, six other soil cores (C1, C5, C9, D1, D5, and D9) were cut into 6-mm slices. Each slice was dissected into 72 soil cubes of 6-mm sides (~400 mg and ~216 mm3 each) with a metal grid and the position of each cube was recorded. The 6-mm dimension was the minimal size we could obtain without disrupting the physical structure of a soil cube.

Characteristics of Cores
Three 10-g subsamples from each core were passed through a 200-µm sieve and used for total C and N analysis on an Elemental Analyzer (Carlo Erba Inc, Milan Italy). The initial total microbial biomass of the cores of Plot B was also measured by fumigation extraction on soil aliquots of 5 g according to Vance et al. (1987). Soil solution C and extractable C after fumigation were measured on a Rosemount Dohrman DC 190 analyzer (Dohrman, Santa Clara, CA). A conversion factor of 2.64 was applied to calculate microbial biomass C (Vance et al., 1987). Three replications were performed per core.

Soil Incubation with Carbon-14-Labeled 2,4-D
Combined solutions of 12C 2,4-dichlorophenoxyacetic acid (12C 2,4-D) (Sigma, St. Louis, MO) and 14C-ring labeled 2,4-D (Isotopchim, 5.55.1011 Bq mol–1) were prepared.

Experiment 1
For the incubation with homogenized cores (6-mm sieved soil), 100 g were amended with 14C-2,4-D to provide 7.8 µg of 2,4-D g–1 (1665 Bq per sample) and moisture content adjusted to 0.22 g water g–1 oven dry soil ({psi} = p–1.6 kPa). Amended aggregates were loosely packed in Mason jars. These were incubated at 20°C for 15 d; water loss was minimized by placing wet paper in the jars. Evolved 14CO2 was trapped in 0.2 M NaOH according to the reaction 2NaOH + 14CO2 -> Na214CO3 + H2O. The Na214CO3 was measured using liquid scintillation. Sodium hydroxide solutions were changed after 1, 2, 3, 4, 8, and 15 d of incubation.

Experiment 2
The second incubation was done on 72 soil cubes per core, each cube being placed in a well of a microtiter plate. Each well was amended with a 14C 2,4-D solution to provide 7.8 µg 2,4-D g–1 (185 Bq per well) and a moisture content of 0.41 g water g–1 (2.24 times the field capacity, {psi} > –0.001 MPa). This moisture content was chosen to enable an adequate amount of water for biological activity that could be added in a reproducible way (73 µL per well). The microtiter plates were incubated at 20°C and water loss was minimized with wet paper. The evolved 14CO2 was trapped in a filter paper impregnated with a solution saturated with barium hydroxide (56 g L–1) placed on top of the microtiter plates. The filters were changed after 2.7, 6.7, 11, and 23.7 d. The radioactive spots were analyzed using a Phosphorimager (Phosphorimager 445 IF, Molecular Dynamics, Sunnyvale, CA) with reference to standards described by Gonod et al. (2003).

Geostatistical Analysis
Analysis of spatial dependence of 2,4-D mineralization at different scales was performed using the "spatial" package (Venables and Ripley, 1999) developed under the R environment (R Development Core Team, 2003).

Spatial dependence was measured using the semi-variogram method (Cressie, 1991). This measures the mean value of half the squared difference of 2,4-D mineralization measured on paired points separated by a given distance:

[1]
where n(h) is the number of pairs of samples, and (xi, xi+h) is the paired positions separated by the distance lying at an interval [h, h+{delta}h], z(xi) and z(xi+h) parameters were obtained from the paired points, (xi, xi+h). Intersamples distances (h) ranged from 6 mm (distance between two adjacent soil cubes) to 10 m (diagonal of the field).

The initial step was to use a Monte-Carlo method to test evidence of a spatial correlation. A short description of variogram testing can be found in Ribeiro et al. (2003) while a more complete description, developed for point processes can be found in Diggle (1983). Its use for analyses of grids of continuous data by the Greig-Smith approach is described by Cressie (1991). Under the assumption that observed data are independant and identically distributed, any permutation of the observed values among the observation points has the same probability as the observed data. The test has then been performed in the following way:

We computed N independent permutations of the data among their locations.
For each permutation, we computed a variogram and obtained N variograms of {gamma}j(h).
For each chosen distance h, we computed the confidence bounds of {gamma}(h) under the independence assumption by first ranking the N values {gamma}j(h). This allowed us to denote the ordered values of {gamma}[j](h). The largest (resp smallest) integer m (resp M) was computed so that m < {alpha}/N and [resp M > (1 – {alpha})/N]. The confidence interval [{gamma}[m](h), {gamma}M(h)] was used at a level of at least 1 – 2{alpha}. This gives the exact levels where m = {alpha}/N and M = (1 – {alpha})/N.
Individual confidence bands were used because at each value h you get a confidence interval of known level, and by linking these limits this can be compared with the observed variogram.

In a second step, if independence was rejected, we modeled the variogram by adjusting it to an exponential model: {gamma} (h) = c [1 – exp(–h/{lambda})].

Parameter c describes the curve asymptote and estimates the variance of the process, whereas {lambda} describes the spatial dependence between measures. Webster and Oliver (1992) noted that for such a model, two observations are always dependent, even if it decreases quickly with their separating distance and a practical range has to be defined as 3{lambda}.


   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Experiment 1: Spatial Variability of Homogenized Cores at Field Scale
To study the spatial variability of mineralization at the decameter to meter scale (inter-plots comparison) and at the meter to decimeter scale (intra-plot comparison), all cores were considered individually. Kinetics of 2,4-D mineralization in crushed cores showed typical degradation patterns; a lag phase, an exponential phase and a plateau. The example of Plot B is presented in Fig. 2 . Mineralization in the 39 cores varied between 5 and 30% of the added 14C after 2 d of incubation (CV 48.5%) between 22 and 52% after 3 d of incubation (CV 18.5%) and between 40 and 70% at the plateau (CV < 10%). Results for the spatial independence test are presented in Fig. 3 . Bins of (0 m, 0.6 m), (0.6 m, 1 m), (3.4 m, 3.8 m), (3.8 m, 4.2 m), (4.2 m, 4.6 m), and (4.6 m, 5 m) were used to estimate the experimental variogram. The number of corresponding pairs, printed on the observed curve (thick line), varied between 34 and 97; the curve showed a gradual increase, except between the two last lags. Independent simulations (N = 2000) were used to build the individual confidence band of the variogram under spatial independence; these are presented as dotted lines. The observed curve is slightly below the observed confidence band for distances <2 m, this corresponds to the mean value of half the squared difference between mineralization values of adjacent cores that were lower than expected under independence. For distances larger than 2 m no spatial structure of 2,4-D mineralization was detected. The CVs of C and N contents, calculated from the 39 analyzed cores, were <20%.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 2. Kinetics of 2,4-D mineralization of each of the nine crushed cores in the Plot B (B1 to B9).

 


View larger version (17K):
[in this window]
[in a new window]
  		Fig. 3. Semi-variogram of 2,4-D mineralization after 3 d of incubation from the 39 soil cores that were used to study the field scale from decameter to m levels (Exp. 1).

 
Comparison of the mineralization CVs within each plot showed that variability of 2,4-D mineralization was between 31 and 64% after 2 d of incubation and between 7 and 22% after 3 d of incubation (Table 1). Based on a comparison between the observed variogram and its confidence band, we concluded that 2,4-D mineralization was not spatially structured due to the observed curve lying inside the confidence band. Variability of other tested variables (C and N contents, C/N, and total microbial biomass at sampling) was low (CV < 10%) per plot (Table 2). We found no correlation between these variables and the extent of 2,4-D mineralization at the beginning of the experiment.


View this table:
[in this window]
[in a new window]
  		Table 1. Inter- and intra-plots variability of 2,4-D mineralization after 2 and 3 d of incubation for the soil cores from plots A to F (Exp. 1).

 

View this table:
[in this window]
[in a new window]
  		Table 2. Intra-plot variability of C and N contents, C/N, and microbial biomass for the soil cores from plots A to F (Exp. 1).

 
Experiment 2: Spatial Variability at Microhabitat Scale
The mean kinetics of 2,4-D mineralization obtained from the 72 soil cubes of each core attained a plateau after 7 d of incubation. 2,4-D mineralization in the 72 cubes of one core after an incubation period of 3 d (exponential phase) was very heterogeneous (Table 3). Within a single core, CV of mineralization ranged from 25 to 160%.


View this table:
[in this window]
[in a new window]
  		Table 3. Inter- and intra-cores variability of 2,4-D mineralization after 3 d of incubation for the soil cubes (Exp. 2).

 
Knowing the initial position of the cubes allowed us to establish maps of the mineralization potential of 2,4-D in cross-sections of the six soil cores. Two examples are given in Fig. 4 for cubes incubated for 3 d. Cubes with different mineralization potentials were not distributed randomly within the soil slices, but rather grouped into centimetric hot spots.



View larger version (44K):
[in this window]
[in a new window]
  		Fig. 4. Examples of the maps of spatial distribution of 2,4-D mineralization sites. Maps are slices of the cores C1 (Exp. 1) and D5 (Exp. 2) after a 3-d incubation.

 
Results of the spatial analysis are presented in Fig. 5 . Bin widths of 3.3 mm regularly spaced were used to estimate the experimental variogram. The semi-variance increased from 0.93 for a distance of 6 mm to a semi-variance of 1.94 at 30 mm. Independent simulations (N = 2000) were used to compute the individual confidence band under the assumption of spatial independence of mineralization values at the 95% level, which is presented as dashed lines. Experimental points for distances <20 mm were below the confidence limits, so that the assumption of spatial independence was rejected with mineralization values between samples at short distance having significantly closer values than under independence. An exponential variogram, represented as a thick line in Fig. 5, was fitted to the experimental points. Its parameter values were equal to a c of 1.98 for the sill, whereas {lambda} was 8.96, leading to a practical range of 26.88 mm. This was in good agreement with the observed hot spots of mineralization on maps of 2,4-D mineralization potential.



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 5. Semi-variogram of 2,4-D mineralization in slices of soil cores after 3 d of incubation (Exp. 2).

 

   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The objective of this work was to study the potential mineralization of 2,4-D in a silty soil from field to microbial habitat scale. An experimental constraint of the incubations was differences in water contents for crushed cores (0.22 g of H2O g–1 soil) and soil cubes (0.41 g of H2O g–1 soil). However, the patterns of mineralization kinetics were the same in both cases.

Statistical Analysis
Statistical Method
We chose the Monte-Carlo variogram method to study 2,4-D mineralization potential at different scales. The experimental field was chosen for its homogeneity, which had no known trend of anisotropy. This allowed us to use the isotropic variogram as a statistical test. Where anisotropy was suspected, this could be explored using directional variograms.

Sampling Procedure
We chose a sampling procedure where we focused on several identified scales, the millimeter scale where microbial community differences were expected, the decimeter scale where agricultural practices were suspected of homogenizing the soil and the decameter scale where larger pedologic effects could be expected. As a counterpart, this led us to a non-systematic exploration of spatial dependence at all distances. This effect appeared clearly on the variogram curve (Fig. 3) with large segments where no data appeared. In the same way, this choice led us to explore only six plots in the field. Even where the number of cores used to explore the variogram was more important, the value of the variogram at large distances has to be interpreted with caution. As an example, the Monte-Carlo test of spatial independence is an exact conditional test of measured values in the cores but spatial independence was not rejected. However, since this was determined on only six locations with 39 cores and 42 pairs at the largest distance, the power of the test may not be very high.

Choice of a Monovariable Study
2,4-D mineralization in the field was analyzed in two steps. A temporal component was studied first by looking at mineralization curves for different soil cubes to detect the period for which variability was at its maximum among the samples. As can be seen on mineralization curves, mineralization was low at the beginning and no spatial variability of the samples could be detected at this scale. The spatial component was then studied by analyzing the spatial variability at the date where variability among samples was at its maximum, that is, at the middle of the exponential phase of the mineralization curve.

Toward a Multivariate Study
Understanding the evolution of the 2,4-D mineralization in the field is a spatiotemporal problem as can be seen in Fig. 3 and 5, the shape of the curve depending on the core. An initial step for studying this could be through exploratory data analysis, using for example principal component analysis. The advantage of using such a method is in resolving data without making many assumptions about the underlying process. On the other hand, hypothesis testing with such methods are not straightforward, but would be interesting to apply to suggest statistical models. From a practical point of view, methods like principal component analysis decompose the process in orthogonal linear components whereas kinetics described in Fig. 3 or 5 suggest highly nonlinear variations between curves. Initial mineralization rates were low for every core. The slopes during the exponential phase were also core dependent but were not linked to the date at which mineralization began. In this respect, interpreting results of a principal component analysis would be difficult.

Another development could be through developing a spatiotemporal kinetics model where the parameters and the residual term are random fields. Using a known kinetics model with random field representation of the model characteristics will allow spatial predictions, and for optimization of sampling procedures. Identification of the model could be done by multivariate geostatistics on vector kinetic parameters estimated core by core. Two main problems that need to be considered are: (1) the potential nonstationarity of the parameters as their statistical properties will depend on the kinetics model adjustment; and (2) the choice of admissible models. Coregionalization models (Goulard and Voltz, 1992) can be useful modeling tools and the measured spatial dependence at different scales as done in this paper can be a useful initial tool for choosing a class of model.

Experiment 1: Spatial Variability at Field Scale
Mineralization potentials differed widely from one core to another, particularly during the exponential phase. Consequently, 2,4-D mineralization appeared to be heterogeneous at the field scale. Studies dealing with denitrification have obtained similar results and have highlighted significant variability at the field scales (Burton and Beauchamp, 1985; Parkin, 1987; Robertson et al., 1988). Walker and Brown (1983) showed spatial variability for the degradation between plots of simazine (C7H12ClN5) and metribuzin (C8H14N4OS) to be 6.9 and 21%, respectively. They found spatial CV to be 23% for simazine and 24.8% for metribuzin within the same plot. However, Parkin and Shelton (1992) observed that the kinetics of carbofuran (C12H15NO3) mineralization showed a similar low level of variability when plowed and unplowed fields were compared. They showed that ploughing lowers spatial heterogeneity of carbofuran mineralization. This follows other studies that have shown that ploughing homogenizes soil at the decameter to meter scale (Röver and Kaiser, 1999; Nunan et al., 2002). Although we observed a variability of 2,4-D mineralization, no spatial structure of mineralization was found at a scale larger than a few centimeters.

At a millimeter scale, Gonod et al. (2003) suggested that the 2,4-D mineralization potential was due to the initial amount of microbial biomass, which was probably related to the level of degrading microorganisms. Soulas (1993) found that the proportion of 2,4-D degraders (by metabolism or cometabolism) in the total microbial biomass is rather constant and would therefore increase or decrease with population size. In the latter study, 2,4-D mineralization potential also correlated with easily decomposable C, likely because of cometabolism. But at decameter to decimeter scales, we found no correlation between the mineralization potential and either C content or microbial biomass. It could mean that other variables are involved in the control of mineralization or that the tested variables have a synergistic effect.

Experiment 2: Spatial Variability at Microhabitat Scale
In a previous study, Gonod et al. (2003) found that the mineralization potential of 2,4-D was very heterogeneous among individual aggregates of different size classes, 2 to 3.15, 3.15 to 5, and 5 to 7 mm of diameter. After 5 d of incubation, CVs of 2,4-D mineralization were 67.1% for 2- to 3-mm aggregates, 37.8% for 3- to 5-mm aggregates and 41.3% for 5- to 7-mm aggregates (Gonod et al., 2003). Here, we considered slightly larger units and we also found a very large spatial variability between 6 by 6 by 6 mm cubes.

The maps of spatial distribution of sites of 2,4-D mineralization showed that the soil cubes with high mineralization potential were not distributed randomly in the soil, but rather as systematic hot spots organized at centimeter scales, confirming preliminary work by Gonod et al. (2003). A few authors have obtained distribution maps of microbial counts or activities at this scale. Parkin (1987) and Parkin et al. (1987) demonstrated the occurrence of hot spots of denitrification with centimeter diameters. Morris (1999) dissected soil samples in centimeter-size cubes and showed that there was an important spatial variability of the fungal and bacterial biomasses. Morris (1999) identified hot spots as well as "cold spots" with a size of approximately 2 cm. In our study, these results were confirmed by the geostatistical analyses, which showed that mineralization was spatially structured at a scale of a few centimeters. The sum of previous studies using geostatistical analyses at such fine scales (Grundmann and Debouzie, 2000; Nunan et al., 2002) also concluded that heterogeneity occurred over short distances.

In this study, incubation was done under optimal conditions, ensuring adequate water and oxygen. Furthermore 2,4-D is soluble and has a low MW. Consequently we propose there are two factors responsible for the spatial variability among soil cubes or aggregates: (i) the distribution of the degrading microorganisms and (ii) the distribution of organic compounds used for cometabolism of 2,4-D. In this study as well as in a previous work (Gonod et al., 2003), we could not measure C content on cubes or aggregates before incubation because the method is destructive. However, using other aggregates of the same soil, we found that C was heterogeneously distributed among aggregates at millimeter scales (Gonod et al., 2003). Other authors have shown that microorganisms can have an uneven spatial distribution at millimeter scales (Robertson et al., 1997; Grundmann and Normand, 2000; Chenu et al., 2001; Nunan et al., 2002, 2003).

Implications of 2,4-D Mineralization Distribution at the Microscale
To estimate the impacts of the observed microscale variability of 2,4-D mineralization potential for soil sampling, we considered the soil cubes of all cores as a total population (72 cubes per slice x 6 cores = 432 cubes). Using the percentage of mineralization after 3 d of incubation, we determined the minimum number of soil cubes necessary to obtain a mineralization average equivalent to that obtained with the total sample of the 432 soil cubes. Practically, we sorted out subsets of "n" soil cubes from the population of 432 soil cubes and we compared their average mineralization potential to that of the total population mineralization (Student or Mann–Whitney test, P = 0.05). We increased "n" until these means were not significantly different. This procedure was performed 30 times. The results showed that the number of soil cubes representative of the total sample was relatively small, that is, equal to 4, and corresponded to a mass of soil of about 1.6 g (4 x 400 mg). Provided soil is carefully homogenized at the millimeter scale, a few grams of soil would thus be sufficient for a study of 2,4 D mineralization. If we wanted to preserve the soil structure, we would have to sample a larger quantity of soil because, as our results showed, mineralization occurs in soil in hot spots at centimeter scales.

In this study, we measured optimal 2,4-D mineralization, with all soil units being amended with the same amount of substrate. However, in situ, this is not the case and all soil units do not receive the same amount of pesticide. When a pesticide solution is applied to the field, it will be concentrated at the soil surface and leach within large soil pores, and between aggregates or preferential flow paths. We hypothesize that because of the spatial variability of mineralization potential at millimeter scales, this would result in large differences in 2,4-D mineralization.

To simulate this, we assumed that soil cubes having a high potential of mineralization (hot spots of mineralization) were either not in direct contact with the pesticide (e.g., within soil clods, not exposed to preferential flow paths) or in direct contact with it (e.g., on the periphery of clods). In the first case, the mineralization was equal to 1.8 ± 1.2% of added C. To carry out these estimates, we calculated the average 2,4-D mineralization, after 3 d of incubation of the 43 soil cubes (1/10th of the total population), which had the lowest mineralization potential among the initial population of 432 soil cubes. In the other case, where 2,4-D was in contact with the soil cubes with a high potential of mineralization, mineralization was greater and attained 74.9 ± 8% of added C (average of the 43 soil cubes for which mineralization was the highest). Thus the spatial distribution of hot spots may have a strong impact on actual mineralization rates of pesticides in the field.

This study was performed on a plowed soil, which likely to have a homogenizing effect on soil properties and microbial activities from the meter to field scale. Studies are needed from pastures and unmanaged ecosystems. It seems likely that the spatial variability of 2,4-D degradation would be much greater at field scale for these systems. Furthermore, alternative results might be expected for an arable soil that has a history of 2,4-D application because 2,4-D degrading populations should be more abundant and so, it seems likely that spatial variability of 2,4-D degradation would be reduced.


   ACKNOWLEDGMENTS
 
We are very grateful to D.L. Jones and P. Roberts for the corrections of the manuscript.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
L. Vieublé Gonod and C. Chenu, present address: INAPG-INRA, Bâtiment EGER, 78850 Thiverval Grignon, France.

Received for publication January 27, 2004.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow An erratum has been published
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Vieublé Gonod, L.
Right arrow Articles by Chenu, C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Vieublé Gonod, L.
Right arrow Articles by Chenu, C.
Agricola
Right arrow Articles by Vieublé Gonod, L.
Right arrow Articles by Chenu, C.
Related Collections
Right arrow Spatial Distribution
Right arrow Bioremediation and Biodegradation
Right arrow Soil Biology

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome