Pesticide Residues in Soil and Water from Four Cotton Growing Areas of Mali, West Africa

Safiatou Berthe Dem, Toxicology and Environmental Quality Laboratory, Mali, berthesafiatou@yahoo.com
Jean M. Cobb, Virginia Tech, Virginia, jmcc@vt.edu
Donald E. Mullins, Virginia Tech, Virginia, mullinsd@vt.edu

Abstract

Pesticide residues were determined in soil and water collected from four cotton-producing areas of Mali, West Africa. Pesticides were detected in 77% of the soil samples and included p,p-DDT and its breakdown products, endosulfan I and II, endosulfan sulfate; and profenofos. According to 24 farmers surveyed in the study area, DDT was not used in their fields during the past ten years. Endosulfan II, the most commonly detected pesticide, constituted 65% of the detections with a maximum concentration of 37 ng/g. Residues detected in soil samples were below our quantification limit in the newer cotton-producing region of Kita and intermediate region of San. Residues were detected at reportable levels in Koutiala (older) and Sikasso (intermediate) cotton producing areas. Eight pesticides were detected in water samples: lindane, endosulfan I, endosulfan II, endosulfan sulfate, dieldrin, p,p-DDD, p,p-DDE, and atrazine. All detected pesticides in water had concentrations below our established quantification limit except for atrazine. In spite of low concentration levels, pesticides in these water sources are of great concern because they are used for human and animal consumption. Similarly, plant uptake of pesticides poses health risks to domestic livestock that forage on crop stubble and to consumers of food products from these animals. Further studies in the cotton growing areas of Mali are needed to monitor pesticide residues in soils, water, and living organisms.

Introduction

Intensive agricultural practices often include the use of pesticides to enhance crop yields. However the improvement in yield is sometimes concomitant with the occurrence and persistence of pesticide residues in soil and water (Ware and Whitacre 2004). Pesticides may reach the soil through direct application to the soil surface, incorporation in the top few inches of soil, or during application to crops McEwen and Stephenson 1979). Pesticides can enter ground water resources and surface run-off during rainfall, thereby contributing to the risk of environmental contamination. The fate of pesticides in soil and water environments is influenced by the physico-chemical properties of the pesticide, the properties of the soil and water systems (presence of clay materials, organic matter, pH), climate, biology, and other factors (Singh 2001). The increased use of pesticides has caused pollution of soils and water worldwide. Organochlorine pesticide residues, particularly the oxidized form of heptachlor, remain in soils from South Korea, even though their use has been discontinued since 1980 (Kim and Smith 2001). Residues of toxaphene, DDT, trifluralin, and hexachlorocyclohexane (lindane) have been detected in soils from cotton fields in South Carolina (Kannan et al. 2003). Carvalho et al. 2002 identified chlorinated hydrocarbons and organophosphorus pesticides in coastal lagoons of Nicaragua. The DDTs were the main contaminants detected in soils and water from Banjul and Dakar in West Africa (Manirakiza et al. 2003). A study on surface water quality in Ivory Coast demonstrated the occurrence of organochlorine pesticides at low concentrations (Wandan and Zabik 1996).

In Mali, pesticide use is increasing, but little information is available regarding the environmental impacts resulting from their use (Camara et al. 2000). Few data are available on the specific concentrations and types of pesticides in Malian soil and water. In 1998, newspapers and other media reported on the occurrence of DDT in the tap water in Bamako, the capital city of Mali. In a study conducted from 1992 to 1994 on pesticide residues in water from CMDT (Compagnie Malienne pour le Developpement des Textiles) cotton producing areas, pollution of the Banifing tributary to the Niger River was reported (Coulibaly and Derlon 1994). Pesticides identified were dimethoate, cypermethrin, chlorpyrifos ethyl, profenofos, omethoate, deltamethrin and cyfluthrin. In their study, thirty-three well and surface water samples were collected three times each year during June, September and December; and water pollution by pesticides was frequent in the areas where cotton was intensively grown.

A survey of pesticide application by Malian cotton growers was conducted in four cotton-producing regions at the same time as sample collection for this study. Survey results indicated that farmers with low education levels used highly toxic pesticides. Farmers reported that they did not always use pesticides in an appropriate manner despite the fact that the majority of them received training in pesticide application techniques and pesticide disposal. These improper practices may result in the contamination of the environment by pesticides. A total of 20 pesticides were reportedly applied in the study areas during the past ten years and endosulfan was the most frequently used followed by profenofos, cypermethrin, monocrotophos, and chlorpyrifos.

In this study, pesticide residues in soil and water were assessed to obtain information on the status of residues in the same four cotton production areas in Mali. One question to be answered was whether pesticide residues levels in the soil and water could be linked to farmer practices and to the time of cotton establishment as a major cultivated crop in these areas. This study was designed to answer these questions by comparing the occurrence of pesticide residues in an older cotton producing area (Koutiala) versus intermediate producing areas (Sikasso, San) and a newer cotton growing area (Kita).

Methods

Study Areas, Sampling and Shipment

Study areas in the Malian cotton production regions were selected based on geographic proximity to villages with higher population densities (assumed to have higher pesticide usage) and the relative time engaged in cotton production. These sites included Koutiala (an older cotton producing area), Sikasso and San (relatively intermediate cotton producing areas) and Kita (a more recent cotton producing area). Three villages within each of these four areas were selected from which 5 samples were collected per village. A total of 60 soil samples were collected (4 areas x 3 villages x 5 samples per village = 60). Eight water samples in total were collected, six from field wells and two from tributaries (surface waters) of the Niger River at the Kouoro Dam and Bani Bridge.

Soil samples were collected in the study areas from June to July 2003. Each soil sample was a composite of 20 subsamples collected at each site using random sampling within a grid. A grid was established by identifying the approximate center of a field and dividing the field into 5 rows, 10 paces apart, with 4 core samples taken per row for a total of 20 cores (Mullins et al. 1971). Soil was collected from the 0-15 cm layer using a 19” Oakfield soil auger. The subsamples were placed into a 16-liter bucket, thoroughly mixed, and sifted through a No.5 (4 mm/0.1575”) brass soil sieve (W.S. Tyler™) at the collection site. After each sample was collected, the soil auger, bucket, sieve and mixing tool were rinsed with tap water and dried before next use. Separate sampling equipment (bucket, soil auger) was used for control samples. One of the 5 samples collected at each village was obtained from an uncultivated field and used as a control. The collected samples were refrigerated at each sampling site and shipped to the Central Veterinary Laboratory (CVL) under ice. Soil samples were then air dried and sieved through a No.20 (850 μm/0.0335”) brass soil sieve (W.S. Tyler™) and refrigerated at about 4°C until shipment to the United States.

Water samples were collected at the CMDT regions of Koutiala and San in June-July 2003. Water sampling bags constructed of rubber were lowered and raised by hand from the wells and surface waters. After collection, the water temperature and pH were measured directly from the sampling bags using a pH meter (Hanna Instruments HI 991301) and Panpeha™ pH paper (Sigma Aldrich) before pouring into a 1/2-gallon glass jar. The water samples were stored at the CMDT refrigerator and transferred under ice to the CVL cold room in Bamako.

In August 2003, soil and water samples were shipped to the United States by airline and delivered to the Virginia Tech Pesticide Residue Laboratory. Soil samples were stored at about -18°C and the freezer was locked in accordance with United States Department of Agriculture (USDA) permit requirements for storage of foreign soil. Water samples were stored at about 4°C at the same facility.

Soil Textural Classification

Soil subsamples were sterilized prior to submission to the Virginia Tech Soil Testing and the Soil Physics laboratories where they were analyzed for pH, percent organic matter, percent clay, and textural class. The reason for sterilization is that neither laboratory was authorized to process unsterilized soil of foreign origin.

Chemicals

High purity pesticide grade solvents (hexane, dichloromethane) and certified ACS reagents were used with the exception of reagent grade acetone and hexane which were distilled and evaluated by GC analysis for purity prior to use. Bulk sand was purchased from a local hardware store and rinsed with multiple solvents and evaluated using GC analysis prior to soil extraction.

Pesticide Standards

Pesticide standards were purchased from Restek and ChemService and stock solutions were prepared using pesticide grade solvents (see Table 1). Spiking solutions for measuring method efficacy (percent recovery) were prepared from stock solutions. Calibration standards in at least three concentrations were also prepared from stock solutions and diluted in hexane. All stock, spiking and calibration standards were transferred to Qorpak™ glass jars after preparation and stored at 4°C. Soil and water samples were screened for a total of 50 pesticides. Calibration curves of working standards were used to evaluate the linearity of the gas chromatograph response each day of analysis and pesticide residues were quantified based on these external standards.

Sample Extraction

Soil samples were extracted using a soil-packed bulb column. Each subsample (25 g) was weighed into a glass jar (4 oz), and if appropriate, the soil was fortified at this step, before adding pre-cleaned sand (25 g) and granular sodium sulfate (50 g). The sample mixture was manually shaken for 30 seconds, placed on a roller for 30 seconds, and then allowed to stand for 20 minutes to provide time for the sodium sulfate to adsorb any residual moisture from the soil. The sample mixture was transferred to a 250 mL bulb column and the sample jar was triple rinsed with small amounts (3 to 5 mL) of hexane and transferred to the bulb column. The soil contents were extracted with acetone:hexane (1:1 v/v, 250 mL) and the eluate collected and concentrated to 100 mL using a rotary evaporator (Büchi R 110, Brinkmann™). The soil extract was then subjected to additional cleanup steps (see Cleanup of Soil Extracts).

Water samples were extracted using liquid-liquid extraction (LLE). Each sample (900 mL) was poured through a folded filter paper and measured in a graduated cylinder (1 L) before transfer to a separatory funnel (1 L). Sodium sulfate (10 g) was added to the separatory funnel and dissolved by shaking. The water-sodium sulfate mixture was extracted with dichloromethane (3 x 100 mL) and after each separation; the organic layer was filtered through granular sodium sulfate. The combined dichloromethane layers were evaporated on a rotary evaporator (Büchi R 110, Brinkmann™) to about 5 mL, hexane (15-20 mL) was added, and the evaporation continued to about 2 mL. The concentrated extract was quantitatively transferred to a centrifuge tube and evaporated to 0.5 mL on a nitrogen evaporator (N-Evap® Model No. 111, Organomation) and diluted to 2 mL with hexane. No additional cleanup was needed and the water extracts were ready for GC analysis.

Cleanup of Soil Extracts

The concentrated soil eluate was “washed” by liquid-liquid partitioning with saturated sodium sulfate (25 mL) and distilled water (300 mL) in a separatory funnel (500 mL). After shaking, the aqueous layer was drained into a beaker and the hexane was transferred to a separatory funnel (250 mL). The aqueous layer was returned to the 500 mL separatory funnel and re-extracted with 15% dichloromethane in hexane (40 mL). The organic layers were combined in the 250 mL separatory funnel and gently washed with distilled water (100 mL) for about 30 seconds. After the aqueous layer was discarded, the organic layer was filtered through sodium sulfate, evaporated to near dryness on a rotary evaporator; the sides of the flask rinsed down with hexane (20 mL), and evaporated to about 1 mL. The sample extract was quantitatively transferred to a centrifuge tube, concentrated on a nitrogen evaporator to 0.5 mL, and diluted to 2.0 mL final volume in hexane prior to GC analysis.

Gas Chromatographic Analysis

Soil extracts were analyzed using an Agilent 6890 gas chromatograph equipped with a ⁶³Ni micro-electron capture detector (µECD), capillary column, and Hewlett Packard Chemstation software (GC 2071, Rev.A.06.01). The primary capillary column for separation of pesticides was RTX-5 (Restek, 30 m x 250 µm x 0.25 µm) and confirmation runs were completed using RTX-35 (Restek, 30 m x 250 µm x 0.25 µm). Helium was used as the carrier gas at a constant column flow rate of 1.1 mL/min and the detector makeup gas was nitrogen at a flow rate of 60 mL/min. Samples were injected in the splitless mode with the purge flow to split vent set at 35 mL/min at 1 min and pressure at 15 psi and total flow at 39 mL/min. The injector temperature was either 250°C (RTX-5 injections) or 225°C (RTX-35 injections) and the detector temperature was 350°C. Two different temperature programs were used for the RTX-5 and RTX-35 columns to achieve the separation of pesticides of interest. The temperature program on the RTX-5 capillary column was as follows: 90°C for 0.00 min, 30°C/min to 190°C held for 20 min, 20°C/min to 275°C held for 10 min. For confirmation runs, the temperature program on the RTX-35 column was as follows: 100°C for 2.0 min, 15°C/min to 160°C, 5°C/min to 270°C held for 5 min.

Water extracts were analyzed on the same Agilent GC with the RTX-35 column and a temperature program of 50°C for 0.00 min, 10°C/min to 230°C held for 5 minutes, 20°C/min to 280°C held for 12 min. Helium was used as the carrier gas at a constant column flow rate of 1.3 mL/min and the detector makeup gas was nitrogen at a flow rate of 60 mL/min. Samples were injected in the splitless mode with a purge flow to split vent set at 15 mL/min at 1 min and pressure at 15 psi and a total flow of 19 mL/min. The injector temperature was 250°C and the detector temperature was 350°C. The same operating conditions were used with the RTX-5 column to confirm pesticides identified in water extracts.

Method Performance

Control and fortified samples were extracted with each analytical set (about every six samples) using U.S. soil previously extracted using the same method and shown to be free of interfering peaks. Deionized water was used as the pesticide-free matrix for control and fortified water samples. In most cases, the recoveries of detected pesticides ranged between 70 and 99% (5-29% sd) and represent triplicate analyses. Other specific recoveries of pesticides from soil were lower: endosulfan I (52% ± 10% sd), endosulfan II (46 ± 10%), and p,p-DDE (40% ± 9% sd). The recovery for pesticides detected in water ranged from 80% for p,p-DDD to 141% for atrazine.

Quantification Limit

The quantification limit (QL), for pesticides reported in this study, was based upon the lowest concentration that could be consistently and/or reliably recovered (> 70%) in our laboratory from fortified samples (Scholtz and Flory 1999). If this percent recovery could not be achieved, the most consistent pesticide recovery was used to establish the quantification limit. The QL for all pesticides detected in soil in this study was 6 ng/g dry weight with the exception of profenofos which had a QL of 8 ng/g dry weight.

Farmer Survey

Twenty-four farmers in the study area were surveyed to obtain information on their knowledge, attitudes and practices regarding pesticide usage in their fields during the past ten years.  Information from this farmer survey was useful prior to beginning pesticide residue analysis of soil samples in the laboratory. Details about the survey are not reported here.

Results

Soil Textural Classification

Texture distribution of soils in the study area was sandy-loam (58%), loam-sandy (30%), loam (10%) and silt (2%). Soils from the regions of Kita, Sikasso and San were mostly sandy-loam while soils from Koutiala were mostly loam-sandy. Soil pH ranged from 5.0 to 7.2 and percent organic matter ranged from 0.2% to 1.7%.

Pesticide Residues in Soil

Pesticides were identified in seventy-seven percent of the soil samples tested. The pesticides and metabolites that were detected include endosulfan I, endosulfan II, endosulfan sulfate, p,p-DDD, p,p-DDE, p,p-DDT and profenofos. Figure 1 summarizes the occurrence of the pesticides detected in study area soils.

Figure 1.  Percentage occurrence of pesticides found in soil samples from all four cotton growing areas.

The most common pesticide soil residue detected in the cotton growing study area was endosulfan II, constituting 65% of the detections with a maximum of 37 ng/g. The other pesticides (percent occurrence-highest level detected) included p,p-DDE (50%-121 ng/g), endosulfan sulfate (42%-49 ng/g), endosulfan I (33%-10 ng/g), p,p-DDD (28%-below the method quantification limit), profenofos (7%-below our quantification limit) and p,p-DDT (3%-11 ng/g).

Residues detected in soil samples from Kita

Endosulfan (I and II) was the only pesticide detected in soil samples from Kita (see Table 2). The endosulfan I isomer was detected in one out of fifteen samples constituting 7% of the samples. Endosulfan II was detected in 6 samples constituting 40% of those samples examined.

All detected pesticides were below the quantification limit (6 ng/g).
 

Residues detected in soil samples from Sikasso

Six pesticides were detected in samples from Sikasso (see Table 3) including endosulfan (I and II) and the metabolite endosulfan sulfate, p,p-DDT and its breakdown products (p,p-DDE and p,p-DDD). Endosulfan II was detected in fourteen (93%) out of fifteen samples among which 4 samples (27%) were above the quantification limit with a maximum concentration of 26.4 ng/g. Endosulfan sulfate was detected in 11 samples (74%) among which 7 samples (47%) were above the quantification limit with a maximum concentration of 49 ng/g. Endosulfan I was detected below the quantification limit in six (40%) out of fifteen samples. p,p-DDE was detected in 9 samples (60%) among which one sample (7%) was above the quantification limit at a concentration of 20 ng/g. p,p-DDD was detected below the quantification limit in 5 samples (33%). One sample (7%) contained p,p-DDT below the quantification limit of 6 ng/g.
 

Table 4 summarizes pesticide residues detected in soil samples from Koutiala. A total of six pesticides were detected. They were: endosulfan (I and II) and the metabolite endosulfan sulfate, p,p-DDT and the breakdown products p,p-DDE and p,p-DDD. Endosulfan II was detected in thirteen (87%) out of fifteen samples among which 3 samples (20%) were above the quantification limit with a maximum concentration of 37 ng/g. Endosulfan I was detected in 10 samples (67%) one of which (10ng/g) was above the quantification limit. Endosulfan sulfate was detected in 8 samples (53%) and 5 samples (33%) were above the quantification limit with a maximum concentration of 49 ng/g. p,p-DDE was detected in 14 samples (93%) and 2 samples (14%) were above the quantification limit with a maximum concentration of 121 ng/g. p,p-DDD was detected below the quantification limit in 12 samples (80%). One sample (7%) contained p,p-DDT (11ng/g) above the quantification limit (6 ng/g).
 

Residues detected in soil samples from San

Five pesticides were detected in soil samples from San (see Table 5). They were: endosulfan (I and II), endosulfan sulfate, profenofos and p,p-DDE. Endosulfan I was detected in 3 samples (20%), profenofos in 4 samples (27%), and p,p-DDE in 7 samples (47%) below the quantification limit. Endosulfan II and endosulfan sulfate were each detected below the quantification limit in 6 samples (40%).
 

Water temperature and pH

Water pH ranged from 4.9 to 7.2 and water temperatures at the time of sampling ranged from 29°C to 33°C.

Pesticide Residues in Water

Occurrence and range of pesticide water residues are summarized in Table 6. Eight pesticides were detected in these water samples: atrazine, lindane, dieldrin, endosulfan I, endosulfan II, endosulfan sulfate, p,p-DDD, and p,p-DDE. All pesticides detected in these samples were below the quantification limit except for atrazine (one sample, 1.4µg/L). For the pesticides detected below the quantification limit, Endosulfan I and p,p-DDE were each detected in 75% of the samples; followed by endosulfan II (62%), endosulfan sulfate (50%), p,p-DDD (37%), atrazine (24%), lindane (12%), dieldrin (12%).
 

Farmer Survey

The farmer survey revealed that a total of 20 pesticides were applied in the cotton growing areas during the past ten years.  Endosulfan was the frequently used pesticide followed by profenofos, cypermethrin, monocrotophos, and chlorpyrifos.  Herbicides were not used extensively.

Results from the residue analyses corroborate the findings of the farmer survey with the detection of endosulfan II in 65% of the soil samples. Forty percent of the fields where these soil samples were collected had a history of endosulfan use. Endosulfan is usually applied as an insecticide in the first two pesticide applications of the cotton-growing season. Technical endosulfan is a mixture of two stereoisomers: 64-67% endosulfan I, and 29-32% endosulfan II (British Crop Protection Council 2003). After release into soil, endosulfan isomers are vulnerable to photo and chemical degradation. The high soil-sorption coefficients of the isomers reduce the potential for volatilization and leaching. Once released into water, endosulfan isomers rapidly hydrolyze under alkaline conditions and more slowly at neutral and acidic pH levels. The half-lives of endosulfan I are 35 and 150 days for pH 7 and 5.5 respectively; the half-lives of endosulfan II are 37 and 187 days for pH 7 and 5.5, respectively (Howard 1991). Given the half-lives for endosulfans and the pH range (5.0 to 7.2) of the soil samples, the residues detected in the soil samples were probably from the current year’s pesticide applications. The fact that endosulfan sulfate was not detected in soil samples from the Kita region supports the idea that the residues came from more recent applications. However, it should be noted that in some cases endosulfans have been reported several months or years after application. In 1964, when 0.38 ppm endosulfan was applied to Colorado soil, 0.04 ppm remained three years later (Mullins et al. 1971). The concentration of endosulfan residues detected in soils in this study is considered to be low. McEwen and Stephenson 1979 reported on the occurrence of endosulfan residue levels of 4600 ng/g and 400 ng/g in orchard and vegetable soils, respectively. In general, the range of the endosulfan concentrations detected in soils from Kita and San were below the quantification limit and below those observed in soils from Sikasso and Koutiala (ND to 49 ng/g dry weight). In all regions, endosulfan II was more prevalent than endosulfan I and endosulfan sulfate (see Figure 2).

Of the pesticides detected, p,p-DDE, a metabolite of p,p-DDT representing 50% of the detections, was the second most frequent compound found in soils. Only 5% of the sixty soil samples had DDE residues above the quantification limit. Technical DDT is primarily a mixture of p,p-DDT (70%) and o,p-DDT (<30%), (British Crop Protection Council 2003). DDT may undergo dehydrochlorination under alkaline condition to non-insecticidal DDE. In soil, DDT is biologically degraded to the stable and toxic metabolites DDE and DDD (Harner 1999). The half-life of DDT is estimated to be 15 years in soil, 350 days in surface waters and 31 years in ground water (Howard 1991). Figure 3 compares the distribution of DDT and its metabolites in Malian soils.

There was no previous history of DDT use reported in the four Malian agricultural regions according the 24 farmers who were surveyed. It is interesting to note the occurrence of DDT or its degradation products in the Sikasso, Koutiala and San regions, although it has reportedly not been used in those areas. Conversely, neither DDT nor its degradation products were detected in the Kita soil samples and this suggests that it has probably never been used in this region. In Sikasso and Koutiala, p,p-DDT, p,p-DDE and p,p-DDD were all detected although the presence of a small proportion of p,p-DDT suggests that this pesticide has probably not been used for many years in these areas. DDT residues detected in soils in this study are similar to other reports. Cotton field soils in the U. S. (South Carolina, Georgia) contained DDT residues ranging from 0.11 ng/g to 45 ng/g dry weight (Kannan et al. 2003). Conversely, other studies have reported DDT residues as high as 3000 ng/g and 5000 ng/g in alfalfa and desert soils (Ware et al. 1971). In our study, p,p-DDE accounts for an average of 96% of the total DDT concentrations and p,p-DDT accounts for 7% of the total DDT concentrations. The ratio of DDT/DDE may be used as an indicator of the approximate time of DDT application if the half-life of DDT for the specific environment is known. No such half-life information is available for Malian soils. The time elapsed since DDT application to date of measurement can be calculated from the expression: t = T_1/2 x ln(C_f/C_i)/ln1/2 where T_1/2 is the half life of the substance in that environment (time for 50% of degradation from initial concentration), C_i is the initial concentration and C_f is the final concentration of a substance which has not degraded. Calculation of the DDT/DDE ratio in the Koutiala region was 0.08 suggesting that this lower DDT/DDE ratio is indicative of an older DDT application (Tavares et al. 1999). Assuming that half-life values for DDT in soils under tropical conditions is 672 days as reported in a study in Brazil (Racke 2003). DDT may have been applied in the study area of Koutiala in 1997.

Soil samples taken from five fallow fields in Koutiala and Sikasso tested positive for one or more pesticides. This finding supports the high land-use pressure reported for these two regions. Due to the increases in human and animal populations, fallow periods have been reduced both in length and area or even abandoned in many farming systems (Diarra 2003; Kaya et al. 2000).

Profenofos was detected only in samples from the San region where this pesticide is used in the first two annual treatments of cotton plants. However, profenofos was not detected in the other regions probably because it is not applied in these areas. It should be noted that if profenofos had been applied in previous years, it would probably have degraded by the time of sampling because its half-life in soil is only one week (British Crop Protection Council 2003).

Other pesticides reportedly used by farmers in the study area during the past ten years were not detected in the soil samples. Their non-detection status is probably related to their low soil persistence. The half-life (days/pesticide half-life) for these pesticides is: cypermethrin (13 weeks) chlorpyrifos (7-15 days), monocrotophos (1-5 days), dimethoate (7-16 days), atrazine and lindane (240 days), (British Crop Protection Council 2003; Howard 1991).

Eight pesticides were detected in water samples (lindane, endosulfan I, endosulfan II, endosulfan sulfate, dieldrin, p,p-DDD, p,p-DDE and atrazine). Among those detected, dieldrin and DDT were not reported in the farmer survey as being applied in the study areas. Therefore, they may have originated from earlier applications given their reported high persistence in soil and water. All pesticides detected in the water samples had concentrations below the quantification limit except for atrazine (1.4 µg/L) that was detected in one surface water sample. Atrazine may be released into the environment at application points where it is used as an herbicide. The half-life of atrazine in natural water is 10-105 days (British Crop Protection Council 2003). Therefore, atrazine residues detected in surface water in this study may originate from recent use and from water runoff.

Endosulfan detected in surface and ground waters probably originated from runoff water or from the rinsing of application equipment near water sources. Endosulfan has a very low leaching tendency (British Crop Protection Council 2003). Atmospheric fallout or drift during pesticide application and practices such as washing spraying equipment in surface waters could be the source of the contamination of soil and water. The farmer survey indicated that none of the wells were sealed in the survey areas and they could be easily contaminated with runoff water. No Malian drinking water norms exist for pesticides. Therefore, we compared the results of this research with the United States Environmental Protection Agency (US EPA) norms. The Maximum Contaminant Level (MCL) is the highest level of a contaminant that is allowed in US drinking water. Of the pesticides detected in Malian water samples, US EPA MCLs are established for atrazine and lindane at 3.0 µg/L and 0.2 µg/L, respectively. Pesticide residues detected in this study do not exceed the drinking water MCLs of the US EPA for atrazine (1.4 µg/L) and lindane (below the quantification limit).

Conclusion

Numerous studies have demonstrated the persistence of organochlorine pesticides in soils from past agricultural usage in the United States, Canada (Bidelman and Leone 2004) and worldwide (Gong et al. 2004). The results from this study demonstrate the occurrence of several persistent and moderately persistent pesticides at low concentrations in soil and water from Malian cotton growing areas. Pesticide residues were detected in 46 of the 60 soil samples from Malian cotton growing areas. Seven pesticides or their metabolites were detected in soil samples and included: endosulfan I, endosulfan II, endosulfan sulfate, p,p-DDD, p,p-DDE, p,p-DDT and profenofos. The high persistence of a pesticide like DDT is evidenced by residues detected in soil samples with no recorded history of DDT application during the past ten years. The concentrations of pesticide residues detected in this study are relatively low compared to other studies. The soil residues detected in the Kita and San regions were below the quantification limit and below residues levels detected in Koutiala and Sikasso. Koutiala, the oldest cotton growing area, appeared to have the highest frequency of residue detection in the study areas; the highest residue concentrations were also observed in this region. Of the eight pesticides detected in water samples, all had concentrations below the quantification limit except for atrazine (1.4 µg/L) detected in one surface water sample. The results of our study suggest that contamination by pesticides in the four cotton-producing areas is not as severe as might be anticipated. However, this study was limited to samples collected from a small number of farmers in only one phase of the growing season. At the same time, the significance of the presence of endosulfans and DDTs in soils and water in close proximity to human and animal populations should not be discounted. Further residue studies in other agricultural areas of Mali are needed in order to assess the levels of pesticide residues in Malian soils, water and living organisms. 

Acknowledgements

Funding for this project was provided by the Integrated Pest Management/Collaborative Research Support Program (IPM/CRSP), USAID Grant # LAG-4196-G-00-3053-00. We would like to thank USAID-Mali Mission and the OIRED-VA Tech for their financial support and assistance with this project. We would like to thank the Compagnie Malienne pour le Développement des Textiles (CMDT) for their assistance in this research. A number of individuals at institutions in Mali provided invaluable assistance in our research: Dr. Saïdou Tembely, Ms. Traoré Halimatou Koné, Mr. Fousséni Diallo and Ms. Maïga Habiba Ahamadou from the Central Veterinary Laboratory; Ms. Gamby Kadiatou Touré from the Institute d’ Economie Rurale (IER); Mr. Issa Sidibé from the Opération Haute Vallée du Niger (OHVN); and Mr. Soumaré Boubacar, Mr. Oumarou Aya and Mr. Jean Pierre Derlon from the CMDT. We would also like to thank Ms. Pat Hipkins, Drs. Keith Moore and Carlyle Brewster for their advice on conducting the farmer survey and Ms. Sandra Gabbert for her assistance in some of the laboratory work.

References

Bidelman, T.F., Leone, A.D. 2004. Soil-air exchange of organochlorine pesticides in the Southern United States. Environ. Poll. 128: 49-57.

British Crop Protection Council. 2003. The Pesticide Manual, Tomlin,C. (ed.), 13^th ed., United Kingdom, pp. 1344.

Camara, M., Haidara, F., Traoré, A. 2000. Etude socio-économique de l'utilisation des pesticides au Mali. INSAH: Les Monographies Saheliennes 12, Comité Permanent Inter-Etats de Lutte Contre la Sécheresse dans le Sahel (CILSS) ; Bamako, Mali, pp. 108.

Carvalho, F.P., Villeneuve, J.P., Cattini, C., Tolosa, I., Montenegro-Guillén, S., Lacayo, M., Cruz, A. 2002. Ecological risk assessment of pesticide residues in coastal lagoons of Nicaragua. J. Environ. Monit. 4: 778-787.

Coulibaly, A., Derlon, J.P. 1994. Pollution des Eaux du Fleuve Banifing par lesIintrants Agricoles Chimiques, Compagnie Malienne pour le Developpement des Textiles (CMDT). Bamako, Mali.

Diarra, Z. 2003. Communication personnelle sur l'usage des terres, les pesticides utilisés et les cas d'intoxication dues aux pesticides en zone CMDT, M'pessoba.

Gong, Z.M., Tao, S., Xu, F.L., Dawson, R., Liu, W.X., Cui, Y.H., Cao, J., Wang, X.J., Shen, W.R., Zhang, W.J., Qing, B.P., Sun, R. 2004. Level and distribution of DDT in surface soils from Tianjin, China. Chemosphere 54: 1247-1253.

Harner, T., Wideman, J.L, Jantunen, L.M.M., Bidelman, T.F., Parkhurst, W.J. 1999. Residues of organochlorine pesticides in Alabama soils. Environ. Poll. 106: 323-332.

Howard P.H. 1991. In: Handbook of Environmental Fate and Exposure Data for Organic Chemicals Vol. III : Pesticides. Lewis Publishers, Chelsea, Michigan. pp. 684

Kannan, K., Battula, S., Loganathan, B.G., Hong, C.S., L, W.H., Villeneuve, D.L., Sajwan, K., Giesy, J.P., Aldous, K.M. 2003. Trace organics contaminants, including toxaphene and trifluralin, in cotton field soils from Georgia and South Carolina, USA. Arch. Environ. Contam. Toxicol. 45: 30-36.

Kaya, B., Hildebrand, P.E., Nair, P.K.R. 2000. Modeling changes in farming systems with the adoption of improved fallows in southern Mali. Agric. Systems 66: 51-68.

Kim, J.-H., Smith, A. 2001. Distribution of organochlorine pesticides in soils from South Korea. Chemosphere 43: 137-140.

Manirakiza, P., Akinbamijo, O., Covaci, A., Pitonzo, R., Schepens. P. 2003. Assessment of organochlorine pesticide residues in West African city farms: Banjul and Dakar case study. Arch. Environ. Contam. Toxicol. 44: 171-179.

McEwen, F.L., Stephenson, G.R. 1979. The Use and Significance of Pesticides in the

Environment. John Wiley & Sons, New York. pp. 538

Mullins, D.E., Johnsen, R.E., Starr, R.I. 1971. Persistence of organochlorine insecticide residues in agricultural soils of Colorado. Pest. Monit. J. 5: 268-275.

Racke, K.D. 2003. Pesticide Science: Environmental Fate and Effects of Pesticides. Coats, J. R., Yamamoto, H. (eds.) American Chemical Society, Washington D. C.

Scholtz, T. G., Flory, D. A. 1999. Clearing up the confusion. Environ. Protect. 10: 37-41.

Singh, R.P. 2001. Comparison of organochlorine pesticide levels in soil and groundwater of Agra, India. Bull. Environ. Contam. Toxicol. 67: 126-132.

Tavares, T.M., Beretta, M., Costa, M.C. 1999. Ratio of DDT/DDE in the All Saints Bay, Brazil and its use in environmental management. Chemosphere 38: 1445-1452.

Wandan, E.N., Zabik., M.J. 1996. Assessment of surface water quality in Cote d'Ivoire. Bull. Environ. Contam. Toxicol. 56: 73-79.

Ware, G.W., Estesen, B.J., Cahill, W.P. 1971. DDT moratorium in Arizona-agricultural residues after 4 years. Pest. Monitor. J. 5: 276-280.

Ware, G.W., Whitacre, D.M. 2004. The Pesticide Book, MeisterPro Information Resources, Willoughby, Ohio. pp. 496.