JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 1 May 2009
Published in J Environ Qual 38:1031-1041 (2009)
DOI: 10.2134/jeq2008.0355
© 2009 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
Open Access Article
This Article
Free via Open Access
Right arrow OA Abstract
Right arrow Figures Only
Right arrow Full Text (PDF) Free
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
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Charlier, J.-B.
Right arrow Articles by Moussa, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Charlier, J.-B.
Right arrow Articles by Moussa, R.
Agricola
Right arrow Articles by Charlier, J.-B.
Right arrow Articles by Moussa, R.
Related Collections
Right arrow Watershed and Landscape Processes
Right arrow Agricultural Pesticides
Right arrow Tropical Soils
Right arrow Surface Hydrology
Right arrow Water Pollution

TECHNICAL REPORTS

Surface Water Quality

Transport of a Nematicide in Surface and Groundwaters in a Tropical Volcanic Catchment

Jean-Baptiste Charliera,b,*, Philippe Cattana, Marc Voltzc and Roger Moussac

a CIRAD, UPR Systèmes Bananes et Ananas, Capesterre-Belle-Eau, Guadeloupe, F-97130 France
b current address, Université de Franche-Comté-CNRS/UMR 6249 Chrono-environnement, UFR des Sciences et Techniques, 16 route de Gray, F-25030 Besançon cedex, France
c INRA, Laboratoire d'étude des Interactions Sol-Agrosystème-Hydrosystème (LISAH), UMR AgroM-INRA-IRD, Bat. 24, 2 place Viala, 34060 Montpellier cedex 1, France

* Corresponding author (jb.charlier{at}gmail.com).

Received for publication August 7, 2008.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
The aim of this article is to determine how the nematicide cadusafos [S,S-di-sec-butyl O-ethyl phosphorodithioate] contaminates water and soils at two scales, subcatchment and catchment. The study site was a small banana (Musa spp.)-growing catchment on the tropical volcanic island of Guadeloupe in the Caribbean. Two application campaigns were conducted, one in 2003 on 40% of the catchment and one in 2006 on 12%. The study involved monitoring for 100 d the surface water and groundwater flows and the cadusafos concentrations in the soil and in surface and groundwaters in a 2400 m2 subcatchment and a 17.8 ha catchment. The results show that at the subcatchment scale the high retention in the A horizon of the soil limits the transport of cadusafos by runoff, whereas the lower retention of the molecule in the B horizon favors percolation toward the shallow groundwater. Comparing the losses of cadusafos at the subcatchment and at the catchment scales revealed that the nematicide re-infiltrated in the hydrographic network. Two successive phases of stream water contamination were observed, corresponding to two distinct contamination mechanisms: an event-dominated contamination phase (of <30 d) when transport was linked to overland flow during precipitation shortly after application, and a stabilized contamination phase when transport originated mainly from the drainage of the shallow aquifer. Lastly, comparing the losses of the two phases during 2003 and 2006 showed that shallow groundwater, which is promoted in such permeable soils under abundant tropical rainfalls, seems to be the main contributor to stream contamination.


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
POLLUTION from agricultural sources is an important issue in coastal areas and islands in tropical regions like Central America, the Caribbean, and Hawaii. In fact, pollution is often blamed for the degradation of coastal resources, such as fresh and marine water and flora and fauna (Kammerbauer and Moncada, 1998; Rawlins et al., 1998; McDonald et al., 1999; Li et al., 2001; Taylor et al., 2003; Castillo et al., 2006). Particularly in these tropical regions, banana plantations cause sanitary problems and severe diffuse pollution of water resources by pesticides (Henriques et al., 1997; Castillo et al., 2000; Beaugendre and Edmond-Mariette, 2005). Indeed, bananas are grown mostly in zones with high rainfall depth, which leads to washout and leaching of soil-applied pesticides. Moreover, many of the plantations in central America and the Caribbean islands are on Andosols (IUSS Working Group WRB, 2006) or volcanic soils with andic properties. Their large infiltration capacities enhance leaching (Poulenard et al., 2001; Cattan et al., 2007b). Besides, Andosols also present specific properties that influence pesticide retention on the soil matrix: that is, a high organic matter content and large cation and anion exchange capacity that varies according to pH (Wada and Okamura, 1980; Wada, 1989). In this context, it is difficult to predict the transport of soluble elements. Sansoulet et al. (2007), for example, showed that despite fast transport through soil horizons, fertilizer leaching was delayed. This hinders prediction of pesticide fate in such soils and their potential to reach surface waters. Paradoxically, little research has been conducted on soil and water contamination by pesticides in tropical catchments such as those in banana plantations. It then appeared necessary to improve our knowledge of pesticide transport processes in these conditions to help in assessing the associated environmental impact and find efficient means to limit pesticide transfer from the field to the catchment.

The aim of this paper was to identify, in a small tropical catchment with volcanic soils and mainly cultivated with bananas, the pathways and time course of the surface and groundwater contamination by a nematicide, cadusafos, used in banana plantations. The catchment is on the island of Guadeloupe in the French West Indies. The concentrations of cadusafos were monitored in soil and in surface and groundwaters at the scales of a 17.8 ha catchment and a 2400 m2 subcatchment during two monitoring campaigns of 100 d, each starting after pesticide application. The two campaigns differed by the size of the area where the nematicide was applied: in the first the nematicide was applied over all the banana fields whereas in the second it was applied only in the upper part of the catchment. This enabled us to analyze the effect of varying contributing areas on the contamination dynamics of catchment runoff.


   Materials and Methods
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
 Discussion
 Conclusions
 REFERENCES
 
Study Site
The area studied is the Féfé catchment (Fig. 1 ). It is on the volcanic Caribbean island of Basse-Terre in Guadeloupe (16°03'50'' N, 61°37'12'' W) (Fig. 1a,b), and covers 17.8 ha. It is a mountainous catchment at 318 to 428 m ASL that is divided along its length into two geomorphologically opposed zones (Fig. 1c). The northern half consists of a steep south-facing slope with gradients from 26 to 60%. The southern half consists of a short plateau with an average gradient of 9%, where the stream is perennial. In Guadeloupe, the climate is humid tropical with a maritime influence. There are two seasons: dry in February-March and rainy from July to November. The average annual rainfall in Féfé is 4200 mm. The catchment contains five banana farms. Fifty five percent of the catchment area is used for cultivating bananas, 40% is fallow with patches of flowers and grassland, and 5% consists of the man-made network of roads, platforms, and barns.


Figure 1
View larger version (37K):
[in this window]
[in a new window]

  		Fig. 1. Location of Guadeloupe in the Caribbean (1a) and location of Féfé on the island of Basse-Terre (1b); hydrogeological scheme of the Féfé catchment (1c) showing two overlapping aquifers; hydrological equipment in the Féfé catchment (1d).

 
According to Charlier (2007), the geological context of Féfé is a system of cut-and-fill paleovalleys (Fig. 1c). Lava flows and pyroclastic deposits (nuées ardentes) filled two valleys that run northwest and southeast on a weathered substratum. The catchment is covered by a 4- to 8-m thick formation of lapilli and ash. Thus, the formations of nuées ardentes and lava flows shelter a deep aquifer on which the lapilli deposits shelter a shallow aquifer. Hydrogeological studies have shown that the main stream of the catchment drained the shallow aquifer as well as the deep aquifer downstream from the catchment (Charlier, 2007). As a result, groundwater flow is the main contributor to annual runoff at the catchment scale, and percolation from the shallow aquifer partly recharges the deep aquifer (Charlier et al., 2008). In periods of flooding, runoff on the slopes is mainly Hortonian: the runoff coefficient may reach 35% during a runoff event in a banana field on Andosol (Cattan et al., 2006) and it is enhanced on impermeable surfaces, such as roads, platforms, and barns. Soils over the catchment are classified as Umbric Andosol (IUSS WORKING GROUP WRB, 2006), which Dorel et al. (2000) studied in detail. The saturated hydraulic conductivities in the A (hA) and B horizons (hB) are 30 and 40 mm h–1, respectively (Charlier et al., 2008).

Presentation of the Two Monitoring Campaigns
Two monitoring campaigns of cadusafos concentrations in soil and water were conducted to study the spatial and temporal variations of the contamination mechanisms. The zones of application are shown in Fig. 2 , and the quantities applied as well as the duration of the monitoring periods are given in Table 1 . In the "2003" campaign, cadusafos was applied several times between 3 and 21 Oct. 2003 on a set of banana fields distributed over the whole catchment and covering 40% of the catchment's total surface area; hydrological and chemical monitoring lasted from 3 Oct. 2003 to 11 Jan. 2004. During the "2006" campaign, cadusafos was only applied once on 5 July 2006 in a part of the catchment located upstream and covering 12% of the catchment's total surface area; hydrological and chemical monitoring lasted from 5 July 2006 to 23 Sept. 2006.


Figure 2
View larger version (34K):
[in this window]
[in a new window]

  		Fig. 2. Zones and dates of application of cadusafos for the campaigns in 2003 and 2006.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Dates, surface areas treated, and hydrological monitoring for the campaigns in 2003 and 2006.

 
Hydrological Measurements
Figure 1d shows the location of the devices. Rainfall intensities were measured at four sites using tipping bucket rain gauges (ARG100, Campbell Scientific, Shepshed, Leicestershire, UK), with a sensitivity of 0.2 mm of rain per tip. Measurements from each rain gauge were integrated over a 2-min time step. We did not observe any space dependence of the recorded precipitations upstream and downstream from the small Féfé catchment. Thus, the average catchment rainfall was estimated as the arithmetic mean of the rainfall values at the four gauges.

Runoff was measured at the subcatchment scale and at the catchment scale. The subcatchment of 2400 m2 was on the northern slope of the catchment and was entirely planted with banana. The gauging station at the subcatchment outlet consisted of a composite weir with a 90° V-notch, 0 to 0.24 m high. Water levels were recorded in 2-min time steps using a manometric probe (Diver, Van Essen Instruments, Delft, the Netherlands). The gauging station at the catchment outlet consisted of a composite weir with (i) a 90° V-notch, 0 to 0.50 m high, (ii) overlaid by a 1.95 m wide, 0.50 to 1.025 m high, rectangular weir, and (iii) for exceptional flood events, the shape of the outlet section above 1.025 m was assumed to be trapezoidal for flow estimations. Water levels were recorded in 2-min time steps using a PDCR1830 depth and level sensor (Campbell Scientific, Shepshed, Leicestershire, UK).

Six shallow piezometers (P1, P4, P7, P8, P10, P12) at depths between 1.5 and 5 m in the lapilli formation and two deep piezometers (FB and FD) between 15 and 30 m in the nuée ardentes and lava formations were monitored manually on a weekly basis. They were located across two transects upstream and downstream from the catchment.

Characteristics of Cadusafos and Application Rate
As stated previously, the pesticide molecule studied was the nematicide cadusafos, which is used in the French West Indies to control banana pests, particularly the nematode Radopholus similis. Since the banana crop is perennial, the nematicide is applied several times a year. It is classified as very toxic and dangerous for the environment according to FOOTPRINT (2006). Studies on the dissipation of cadusafos in Andosol in the Caribbean soils gave values of 9 to 15 ± 1 d for the half-life (DT50) and values between 50 and 620 L kg–1for the distribution organic-C sorption constant (Koc) (Zheng et al., 1994; Lazrak, 2006). Cadusafos was applied in granulated form (Rugby 10G, FMC Corp., Philadelphia, PA) at the base of banana plants; application rate was 6 kg ha–1, which corresponds to standard practice for banana plantations.

Sampling
To study the cadusafos dispersion in soils and water after application, the concentrations were monitored in the A and B soil horizons, in runoff water at the subcatchment and the catchment scales, and in groundwater.

Soil Sampling
The soil was sampled four times during each monitoring campaign, one time before application (24 Sept. 2003 and 27 June 2006) and three times after application (10, 53, and 94 d after the first application in 2003 and 5, 20, and 78 d after the single application in 2006). In 2003, the samples were taken at each sampling time in nine fields distributed over the catchment so as to represent the upstream and downstream areas as well as the northern highland and the southern plateau. In 2006, the samples were taken only in four fields located in the area of pesticide application in the northern highland.

The soil was sampled using a hand auger. For each field, two bulked samples were taken at a depth of 0 to 20 cm in the A horizon (hA) and at a depth of 40 to 60 cm in the B horizon (hB). Each bulked sample was a mix of 20 samples spatially distributed in the field to account for the spatial heterogeneity of cadusafos application, which resulted from application of cadusafos at the base of each banana plant. Accordingly, 12 samples were taken between the rows, 4 between banana plants in the row, 2 at 30 cm upstream from the base of the banana plants, and 2 at 30 cm downstream. For each bulked sample, a fraction of about 500 g was taken, and then frozen at –18°C before analysis.

Water Sampling
Runoff at the Outlet of the Subcatchment (2006).
Water was sampled at the outlet of the subcatchment, upstream from the weir, using an automatic sampler (sampler 900 MAX, American SIGMA, Loveland, CO) comprising 12 glass bottles previously rinsed with distilled water. The flow was intermittent at the subcatchment outlet and occurred only during flood events. Runoff water was sampled three times during each flood event at 2, 7, and 17 min after the water levels in the weir started to rise. This sampling frequency was chosen according to the observed rainfall-runoff time series at the subcatchment outlet to ensure that a sample was taken at each stage of rising, peak, and recession. The bottles were collected the day after each storm event. For each flood event, a bulked sample was then formed that consisted of the sum of the three samples of the event.

Runoff at the Outlet of the Catchment (2003 and 2006).
In 2003, water was sampled at the Féfé catchment outlet with the same equipment as at the subcatchment outlet. However, the sampling strategy differed from that at the subcatchment scale because the catchment outlet flow was continuous. Each bottle of the sampler was filled with six subsamples of 100 mL taken at regular 4-h intervals, which formed an integrated sample for a day. The bottles were collected from the site every 2 or 3 d. From 27 Sept. 2003 to 7 Dec. 2003, each daily sample was selected for analysis. From 8 Dec. 2003 until 11 Jan. 2004, a bulked sample of 3 d was analyzed to restrict analytical costs. Inadequate functioning led to the loss of data between 14 and 30 Nov. 2003, 19 and 22 Dec. 2003, and 5 and 6 Jan. 2004. A sample of runoff water before application was also made up from a set of samples taken during 7 d (19–25 Sept. 2003).

In 2006, water was sampled using an automatic sampler (Simplex Mini, ORI Abwassertechnik, Hille, Germany), comprising one glass bottle. The water sample was collected from the site each day. As in 2003, each daily sample consisted of six subsamples of 100 mL taken every 4 h. From 6 July 2006 to 21 Aug. 2006, each daily sample was analyzed. Then, to restrict the number of analyses, from 22 Aug. to 8 Sept. 2006, we used bulked samples of 2 d and, from 9 to 18 Sept. 2006, bulked samples of 3 d. The sample taken before application was an average of the daily samples taken over 7 d (23–29 June 2006).

Groundwater (2003 and 2006).
Groundwater was sampled using peristaltic pumps: a manual pump for shallow piezometers and an electric pump for deep piezometers. Before sampling, to prevent contamination between each piezometer, the material (plastic pipe) was systematically rinsed with distilled water and then with sampled groundwater. First, wells were purged of at least three well volumes with continuous pumping, and then they were sampled when groundwater regime was stabilized. In 2003, samples were taken every week from 10 Oct. to 20 Nov. 2003. Then, from 20 Nov. 2003 to 9 Jan. 2004, samples were taken twice a month. Sampling before application was conducted on 26 Sept. 2003. In 2006, we were unable to sample groundwater because piezometric levels were constantly deeper than the bottom of the piezometer in the upstream part of the catchment where cadusafos was applied.

Analytical Methods
The soil and water samples were stored frozen at –18°C before being sent for analysis at the Laboratory of Soil Analyses (INRA) in Arras, France. The soil samples were defrosted at 4°C before analysis. The gravimetric moisture content of the soil samples was measured on a subsample of 30 g of moist soil. Moisture content was taken into account to calculate the cadusafos content in the soil. Pure acetone was used for extraction with an Accelerated Solvent Extraction (ASE200 Dionex, Dionex Corp., Sunnyvale, CA) on a subsample of 20 g of moist homogenized soil. The liquid-liquid extraction was conducted using pure hexane. Then, pure hexadecane was added before partial rotative evaporation, followed by total evaporation under a light N flow. The dry residue was collected in 2 mL of hexane. Cadusafos was analyzed by capillary gas chromatograph (GC) using a Varian 3400 (Varian, Inc. Corporate Headquarters, Palo Alto, CA) equipped with a split/splitless injector, a thermo-ionic detector (TSD), and a Restek column RTX 200 (15 m; 0.53 mm; 1 µm). Helium was used as the gas vector at a flow rate of 2 mL min–1. The temperature of the detector was 290°C, that of the injector was 260°C, and that of the stove gradient was 150 to 250°C.

The water samples were defrosted at +4°C before analysis. Cadusafos was extracted from the water by the automaton Autotrace Zymark (Zymark Corp., Hopkinton, MA) for solid phase extraction. An aqueous sample of 200 mL was homogenized: small volumes of organic solvent were added to elute the sample, followed by hexadecane. The sample was placed in a rotative evaporator for partial evaporation and then under a light N flow for total evaporation. The residue was diluted in 2 mL of hexane. Lastly, the analysis of cadusafos was conducted using GC with the same material and under the same chromatographic conditions as for the soil.

The detection limits of cadusafos contents in soil and water were 0.5 µg kg–1 and 0.01 µg L–1, respectively. A value of zero was attributed to amounts below these detection limits. In principle, this can lead to a significant bias with regard to the calculation of quantities when many samples have concentrations below this level. However, this was not the case in our study.

Estimating the Amount of Pesticide Applied and Pesticide Losses
The amount of cadusafos applied in the catchment was calculated on the basis of the recommended Rugby application rate in a field (6 kg ha–1 of active ingredient) and of the areas treated. The amount applied and the fields treated were controlled on the field during application days. Pesticide losses by runoff water were computed on a daily basis from the measured discharges and cadusafos concentrations. When only average concentrations over 2 or 3 d were known, the daily losses were estimated as a pro rata of the total quantity transported over the 2 or 3 d weighted by the observed daily runoff volume.


   Results
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
Discussion
 Conclusions
 REFERENCES
 
Hydrological Characteristics of the Two Monitoring Campaigns
The two monitoring campaigns occurred during two periods with contrasted hydrological conditions. Rainfall depth, discharge, and groundwater levels are plotted in Fig. 3 and 4, and hydrological characteristics are in Table 2

· The 2003 campaign (Fig. 3) was conducted from 3 Oct. 2003 to 11 Jan. 2004, that is, 100 d of monitoring during the cyclonic season. The period was characterized by abundant rainfalls (total of 2185 mm) and generally by high rainfall intensities with an observed hourly maximum of 94.5 mm h–1. Daily rainfall exceeded 25 mm in one-third of the monitored days. This period corresponds to a period of high water levels, with the average daily discharge ranging from 2 to 125 L s–1 and a global runoff coefficient of 42.5% at the catchment scale. Fluctuations of the shallow groundwater (piezometers P1 to P12) were <1.9 m at depths varying initially between 0.6 and 3.4 m depending on the sites. Fluctuations of the deep groundwater at piezometers FD and FB were 2.7 and 13.4 m at depths of 22.6 and 5.8 m, respectively.
· The 2006 campaign (Fig. 4 ) took place from 5 July to 23 Sept. 2006 at the start of the cyclonic season, that is, 80 d. The period was characterized by moderate rainfalls (611 mm). The medium rainfall intensities were lower than during the 2003 campaign: the hourly maximum rainfall intensity was 15.6 mm h–1. Daily rainfalls exceeded 25 mm only on 5 d of the 80. This corresponds to a period of low water levels, with daily discharge ranging from 0.1 to 20 L s–1 and a global runoff coefficient of 22.2% at the catchment scale.


Figure 3
View larger version (43K):
[in this window]
[in a new window]

  		Fig. 3. Campaign in 2003 (application dates [showed by arrow] between 3 and 21 Oct. 2003): time series of rainfall, discharge at the catchment outlet, groundwater depths, and cadusafos concentrations.

 

Figure 4
View larger version (26K):
[in this window]
[in a new window]

  		Fig. 4. Campaign in 2006 (date of application [showed by arrow] 5 July 2006): time series of rainfall, discharges, and cadusafos concentrations at the subcatchment outlet and at the catchment outlet.
 

View this table:
[in this window]
[in a new window]

  		Table 2. Hydrological characteristics of the time series during the application campaigns in 2003 and 2006.

 
These observations are in accordance with those of Charlier et al. (2008) who studied the main hydrological processes in the Féfé catchment in 2003 and 2004. Hence, using the same method of hydrograph separation between surface runoff and baseflow, the comparison of the 2003 and 2006 discharge series at the catchment scale confirmed that (i) the wetter the initial conditions, the larger the surface runoff coefficient and (ii) groundwater flow is the main contributor to catchment runoff whatever the year since it amounted to two-thirds of runoff although runoff volume was sevenfold higher in 2003 than in 2006.

Persistence of Cadusafos in the Andosol
Table 3 shows the evolution in cadusafos concentrations for the upper soil horizon hA and the deeper horizon hB of the treated fields during the 2003 and 2006 campaigns. The initial cadusafos concentrations in the soil before application were small and similar for all fields sampled in 2003 and 2006, ranging between 8.4 and 11.3 µg kg–1 for hA, and between 1.1 and 4.4 µg kg–1 for hB. These concentrations before application indicate that cadusafos is persistent in the soil since the previous applications dated more than 3 mo. The new applications at the start of the monitoring periods caused a significant increase in soil concentrations: 5 to 10 d after application, the observed maxima of average field concentrations were 80 ( ± 124) and 38 ( ± 23) µg kg–1 for hA, and 28 ( ± 30) and 13 ( ± 14) µg kg–1 for hB in 2003 and 2006, respectively. The observed peaks were simultaneous for the two horizons, which suggests that the molecule migrated rapidly in depth.


View this table:
[in this window]
[in a new window]

  		Table 3. Average cadusafos concentrations in soils for the A horizon (hA) and the B horizon (hB) of the treated fields.

 
Table 3 also shows that there was a large spatial variation of cadusafos concentrations in each horizon among the set of sampled fields and that statistically significant differences in soil concentration between hA and hB were detected for three sampling dates among four in 2003 and for two sampling dates among four in 2006. As expected, concentrations in the upper horizon hA on which cadusafos was applied were larger than in hB, which was contaminated by compounds leaching from hA. Besides, soil concentrations did not differ statistically between the abundant rainfall period in 2003 and the moderate rainfall period in 2006. This could be explained by the fact that in both years no drought occurred and consequently soil moisture content always remained close to saturation, which ensured similar degradation rates.

Transport Processes
Surface Runoff
Cadusafos transport by surface runoff could be analyzed using water concentration data from the subcatchment outlet in 2006 (Fig. 4), since at this scale there was no contribution of groundwater to runoff. First, note that the total losses of cadusafos by runoff at the subcatchment outlet during 20 d represented 6.4 g ha–1, that is, 0.1% of the total mount applied in 2006. The maximum concentration in runoff water was 1100 µg L–1 and was observed during the first flood event after application. Then concentrations fell rapidly during the next flood events to 3.7 µg L–1 (Fig. 4). Figure 5 represents the cumulative losses of cadusafos and the cumulative volume of surface runoff at the subcatchment scale as a function of time. We observed that 90% of cadusafos losses during the monitoring period occurred the first 6 d after application, which cumulated 3.6 m3 of surface runoff (i.e., 1.5 mm depth) under 63 mm of rainfall. Between 6 and 9 d after application, the last 10% of cumulative cadusafos losses occurred despite an increase in cumulative surface runoff volume from 3.6 to 13.4 m3 d–1 under 86.2 mm of rainfall. This pattern indicated that the amount of active ingredient available to runoff water decreased rapidly after application.


Figure 5
View larger version (20K):
[in this window]
[in a new window]

  		Fig. 5. Cumulative cadusafos losses and cumulative surface runoff volume after application at the subcatchment scale (Campaign in 2006).

 
Groundwater
Local contamination of the shallow groundwater was monitored in 2003 by piezometers P1 to P12 (Fig. 3), situated in fields where cadusafos was applied. The contamination pattern was similar for all piezometers. First, before application, cadusafos concentrations in all piezometers were below the detection limit (0.01 µg L–1), except for P7, where a low concentration of 0.02 µg L–1 was found. This suggests that the persistence of the molecule in shallow groundwater is limited and does not exceed the delay between two applications. Second, after application, groundwater concentrations increased significantly for the first sampling dates, 6 to 7 d after application, with a mean concentration peak of 15.34 ( ± 10.02) µg L–1. Then, 2 wk after this peak, the concentration dropped to reach a concentration plateau of an average of 0.29 µg L–1 and varying between 0.01 and 1.88 µg L–1.

Regarding deep groundwater, in Fig. 3, initial concentrations before application in the deep piezometers FB and FD were below the detection limit, as for the shallow groundwater. After application, two concentration peaks very close to the detection limit were observed (maximum value of 0.05 µg L–1). In both piezometers, however, the molecule became undetectable 6 wk after application.

Catchment Runoff
Identification of Two Phases of Water Contamination after Application. During each of the two monitoring periods, 2003 and 2006, two successive phases of water contamination by cadusafos could be distinguished at the catchment outlet (Fig. 3 and 4). The first phase started after the first application and exhibited an erratic variation of water contamination, with maximum daily mean concentrations corresponding to the storm events. This phase lasted about 28 d in 2003 and 16 d in 2006 and exhibited maximum daily mean contamination peaks of 1.5 µg L–1 in 2003 and 0.26 µg L–1 in 2006. We named it the "event-dominated contamination phase." During the event-dominated phase in 2006, the concentration peaks at the catchment outlet mostly corresponded to surface runoff events at the subcatchment outlet (Fig. 4). This can be explained by the large concentrations of cadusafos in overland flow that occurred during 10 d after application in 2006 (Fig. 4). Moreover, as can be seen on Fig. 3 and 4, during 2 wk after application, catchment runoff exhibited episodic nil concentrations of cadusafos. Between 10 and 16 d after application in 2006 (Fig. 4), when no precipitation occurred, these nil concentrations corresponded to periods during which catchment runoff came from groundwater only.

The first phase was followed by a "stabilized contamination phase" exhibiting smaller fluctuations around an average contamination level of about 0.07 µg L–1 in 2003 and 0.06 µg L–1 in 2006. This second phase started 29 and 17 d after the first application in the 2003 and 2006 campaigns, respectively. The cadusafos concentrations in runoff water at the catchment outlet varied little although the relative contribution of groundwater and surface runoff to stream runoff varied largely with flood occurrence. In 2006, the two phases can be clearly distinguished: the stabilized phase started the 17th day after application, just after a few days without any flood event, and thus without overland flow at the subcatchment outlet.

The cadusafos losses at the catchment outlet 78 d after application corresponded to 0.03% of the total amount applied in 2003 and to 0.01% in 2006. Moreover, it must be stressed that the losses arose mainly during the stabilized phase, which delivered 65% of the total 2003 losses and 69% of the 2006 losses. This was expected because, even if pesticide concentration peaks remained higher in the event-dominated phase, the duration of the stabilized phase was larger. Nevertheless, these results should be nuanced by the fact that the sampling strategy at the catchment outlet (daily mean concentration constituted of six samples every 4 h) probably underestimated concentration peaks because the small flood events may last only a few hours.

Buffering Effect of the Contamination Level between Subcatchment and Catchment Scales.
The maximum concentrations at the subcatchment outlet were almost 4000-fold higher than at the catchment outlet, which indicates the existence of dilution and/or buffering effects between the subcatchment and catchment scales. Even if this buffering effect should be nuanced by the probably slightly underestimated concentration peaks at the catchment scale (due to the sampling strategy detailed above), we have considered that this effect was partly responsible for the concentration decrease from upstream to downstream. To test this buffering hypothesis, we compared the cumulative loads of cadusafos supposed to be transported in 2006 via the overland flow downstream from the treated fields and the loads observed in runoff at the catchment outlet. The loads of cadusafos transported from the treated fields were estimated by assuming that the loads, expressed per surface area from the subcatchment of 2400 m2, were similar for all the treated fields of the catchment (19,400 m2). Results show that 8 d after application, the cumulative load of cadusafos from the treated fields was 60-fold larger than that at the catchment outlet (cumulative loads of 12.4 and 0.2 g, respectively). This pattern suggests that cadusafos infiltrated in the ditch and hydrographic networks at a very high rate before reaching the catchment outlet. Along with the dilution effect by surface runoff on nontreated fields (fallow and grassland) and by noncontaminated groundwater, these results point out that the buffering effect due to cadusafos seepage in the ditches is likely to be an important process explaining the changes in contamination level between the field and catchment scales.


   Discussion
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
Conclusions
 REFERENCES
 
This study attempted to determine the mechanisms of contamination of water and soil compartments by a nematicide, cadusafos, in a cultivated catchment of a volcanic tropical region. Our observations at the subcatchment scale show that surface runoff must be considered as a main pathway for water contamination by cadusafos. This phenomenon has been widely observed at this scale for different types of pesticides in different climates (Leonard, 1990; Lennartz et al., 1997; Donald et al., 1998; Louchart et al., 2001; Leu et al., 2004). As already reported, most pesticide losses via this process occurred during the first rainfall events following the application. However, in our case, the total losses of pesticide by surface runoff were small, representing <0.1% of the total amount applied on the field, which is much lower than the losses usually observed at the same scale (Leonard, 1990; Gentry et al., 2000; David et al., 2003). These results are consistent with those obtained by Saison et al. (2008) on an experimental field in the same pedoclimatic context. Those authors explained the low losses by a rapid depletion of cadusafos in the very first centimeters of the soil, due to both a rapid degradation of cadusafos after application and its migration downward once the applied cadusafos granules had dissolved, which took a week. This situation differs from other study site where most soil-applied pesticides remain in the surface soil layer for several months; this is notably the case where the methods of pesticide application (spraying for instance) and soil conditions are different (e.g., Leonard, 1990; Louchart et al., 2001).

In parallel, our results also show that the contamination by cadusafos affected mainly the shallow aquifer, which is a consequence of the high infiltrability of the Andosols. The concentration peaks of the shallow groundwater (reaching 15 µg L–1) and the shallow piezometric levels suggest the existence of relatively homogeneous and rapid percolation processes within the soil, which cause the contamination of the shallow aquifer in less than a week. Later, these percolations favor the dilution of the contamination by the transfer of less polluted water (at a concentration level of 0.29 µg L–1). These results are coherent with the high percolation rates in the deep soil horizons (40 mm h–1) and the absence of vertical hydraulic discontinuity in the shallow aquifer (Charlier et al., 2008). In contrast, the very low concentration found in the deep piezometers reveals the separation that exists between a shallow aquifer exposed to high and rapid contamination and a deep aquifer that is barely affected or not at all by pesticide percolation (piezometers FD and FB in lava flows and nuées ardentes). This separation is generally linked to the geological structure of the site (e.g., Fenelon and Moore, 1998). In the Féfé catchment, a weathered ash layer over 1 m thick with low porosity at the base of the lapilli formation likely filtered the percolating pesticides.

The analysis of cadusafos concentration variations at the catchment outlet showed large and constant contamination of runoff water during 100 d after application. The observed daily mean concentrations of cadusafos, ranging from maximum values of 1.5 µg L–1 in 2003 and 0.26 µg L–1 in 2006 during the first rainfall events after application to 0.09 and 0.06 µg L–1 after 80 d, were well within the range of those measured by Castillo et al. (2006) in the water of a ditch draining 12 ha of a banana plantation in Costa Rica, namely from 0.17 to 0.48 µg L–1 for a week after application and above 0.02 µg L–1 a month afterward. On a larger scale, in the Suerte basin in Costa Rica (38 200 ha with 15% planted with banana), the concentrations of cadusafos in surface water were also similar, that is, between 0.10 and 1.00 µg L–1 depending on the sampling site (Castillo et al., 2000). It therefore seems that the phenomenon of water contamination by cadusafos applied in banana plantations is not site-specific.

This study also provides insights into events that are likely to happen in catchment runoff contamination. In particular, the observation of two phases in the contamination dynamics of catchment runoff water highlights the different contributions of surface runoff and groundwater flow to catchment pollution. The contamination of surface runoff just after application was fast with large concentration peaks but decreased strongly (1000-fold) in <10 d whereas the contamination of the groundwater increased more slowly from hillslopes to the stream (between 16 and 28 d). That change from the first to the second phase may be related to the time transfer of the contaminated groundwater below the treated field to the outlet of the catchment. Consequently, in a first phase lasting 16 to 28 d after first application, erratic variation of water contamination of catchment runoff (concentration peaks reaching 1.5 µg L–1) closely followed the flood events when surface runoff occurred. Indeed, during a week after application in 2006, large concentrations in overland flow ranging from 1100 to 3.7 µg L–1 were observed downstream from the treated fields. After that phase, catchment runoff concentrations stabilized near 0.1 µg L–1 due to the convergence of pesticide levels in surface and groundwater flows. A likely explanation of this result is that the concentration levels of the groundwater and the surface runoff were similar during this second phase, which can be confirmed in part by our observations. In fact, 8 d after the first application, the contamination of surface runoff at the subcatchment outlet had already largely decreased to a few micrograms per liter. If we refer to the concentrations measured in surface runoff in similar conditions (Saison et al., 2008), they declined to 0.2 µg L–1 20 d after application. If we also account for possible dilution effects due to surface runoff from the nontreated fields and noncontaminated groundwater, it is therefore likely that surface runoff water contributing to catchment runoff was contaminated at nearly the same level as groundwater, namely 0.3 µg L–1.

Although the contamination dynamics changed between the two phases, the average concentration level in stream runoff remained similar between the two phases for each monitoring period: 0.14 and 0.05 µg L–1 in the first phase and 0.07 and 0.06 µg L–1 in the second phase for 2003 and 2006, respectively. In fact, the highest surface runoff concentrations in the first phase were compensated by the low base flow concentrations whereas in the second phase the decrease in the surface runoff water contamination was compensated by an increased and more continuous contribution of the groundwater contamination. This finding differs markedly from the contamination dynamics observed in many other environments (e.g., Thurman et al., 1991; Ng et al., 1995; Louchart et al., 2001) where, at the catchment scale, large changes in concentration levels with time were observed even several weeks after the pesticide application. Finally, despite the similar range of cadusafos concentrations in stream runoff between the two monitoring periods, the total loads in 2003 were higher than in 2006: losses of 0.03 and 0.01% of total amount applied, and maximum concentration peaks at 1.5 and 0.26 µg L–1, respectively. This change of loads was consistent with the higher hydrological flows in 2003 than in 2006 and the different amount applied, 40.2 kg in 2003 and 11.6 kg in 2006. However, even if pesticide loads were lower during the event-dominated contamination phase, pesticide concentrations remained higher in this phase, which may have a different biological impact on the stream.

To verify the main hypothesis about the pesticide transport processes highlighted above, a modeling approach should be developed to quantify the spatial and temporal variability of pesticide losses in soil, and surface and groundwaters. Implementing such a runoff pesticide model in a cultivated context of permeable and saturated soils under abundant rainfalls will require taking into account some specific processes. Hence, to simulate water flows and pesticide transport, a runoff pesticide modeling strategy should include the following elements: (i) the rainfall redistribution by banana plant (Cattan et al., 2007a) that influenced runoff at the plot scale (Charlier, 2007) and the fate of fertilizers (Sansoulet et al., 2007) and nematicide (Saison et al., 2008) and (ii) the surface/groundwater exchanges to simulate the re-infiltration of overland flow in ditches on hillslopes as well as the aquifer drainage by the stream. The development of such a runoff pesticide model calls for further studies to better understand the degradation and the mobilization processes of the nematicide in soil horizons.


   Conclusions
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
REFERENCES
 
The aim of this study was to identify the mechanisms of contamination of soil and water compartments at the both subcatchment and catchment scales by the nematicide cadusafos. The study highlights the environmental impact of pesticide application in cultivated areas in pedoclimatic conditions where rainfall is abundant, soils permeable, and organic matter content high.

The first important result is the distinction of two contamination phases, corresponding to the different contributions of surface and groundwater flows to runoff contamination at the catchment outlet. This pattern seems to be promoted in a context where high infiltration rates in fields and ditches lead to rapid pesticide lixiviation in deep horizons, and then in shallow groundwater by percolation. To manage water resources, this information should be considered to limit pesticide dispersion in groundwater as well as in streamflow.

A second result is that the contamination levels of surface water, as well as shallow and deep groundwaters, reflect the geological structure of the Féfé catchment: that is, a shallow aquifer in the most recent deposits that is rapidly exposed to pollution and a deeper aquifer that is relatively protected from the pollution coming from the treated fields. These results confirm the importance of knowing the geological structure and the major hydrological processes to better study the water contamination dynamics of basins on permeable substrates, for which the water balance exhibits the large contribution of groundwater flows to the stream contamination.

In some situations where water pollution is largely due to surface runoff on hillslopes, one of the solutions to limit pollution in streams is to encourage infiltration uphill, downstream from the fields, for example by implementing grassy strips (see for a review Lacas et al., 2005) or by favoring infiltration in the hydrographic network. Indeed in the case of Féfé catchment, it seems that re-infiltration of runoff in the hydrographic network buffered contamination peaks. But given the rapid propagation of percolation flow to groundwater and the large contribution of groundwater to catchment runoff, increasing the infiltration of contaminated surface water would principally aggravate the contamination of the shallow aquifer. It would somewhat lower the runoff concentration peaks at the catchment outlet without strongly limiting the total loads of pesticides in runoff water. Finally, the only conceivable way of reducing pollution is to reduce the treated areas and/or the rates of application in the fields.


   ACKNOWLEDGMENTS
 
We are grateful to Germain Onapin, Colbert Béhary, and Claire Eme for field and laboratory work. This work was carried out with the financial support of the Guadeloupe Region (FWI), the Ministère de l'Ecologie et du Développment Durable (France), the European Community under the project "Assessment of water-pollution risks associated with agriculture in the French West Indies: management at the catchment scale", and the ANR- Agence Nationale de la Recherche - The French National Research Agency under the Programme Agriculture et Développement Durable, project ANR-05-PADD-010, GeDuQuE. We thank the Laboratory of Soil Analyses in Arras for all chemical analyses of soil and water samples.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Conclusions
 REFERENCES
 





This Article
Free via Open Access
Right arrow OA Abstract
Right arrow Figures Only
Right arrow Full Text (PDF) Free
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
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Charlier, J.-B.
Right arrow Articles by Moussa, R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Charlier, J.-B.
Right arrow Articles by Moussa, R.
Agricola
Right arrow Articles by Charlier, J.-B.
Right arrow Articles by Moussa, R.
Related Collections
Right arrow Watershed and Landscape Processes
Right arrow Agricultural Pesticides
Right arrow Tropical Soils
Right arrow Surface Hydrology
Right arrow Water Pollution

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
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome