Environmental Toxicology and Chemistry, Vol. 22, No. 1, pp. 7±19, 2003 q 2003 SETAC Printed in the USA 0730-7268/03 $12.00 1 .00 DICHLORODIPHENYLTRICHLOROETHANE IN THE AQUATIC ECOSYSTEM OF THE OKAVANGO DELTA, BOTSWANA, SOUTH AFRICA BONTLE MBONGWE,² MELISSA LEGRAND,³ JULES M. BLAIS,² LYNDA E. KIMPE,² J. JEFFREY RIDAL,§ and DAVID R.S. LEAN*² ²Department of Biology, University of Ottawa, 30 Marie Curie Road, Ottawa, Ontario K1N 6N5, Canada ³Dietetics and Human Nutrition, McGill University, St. Anne de Bellevue, QueÂbec, Canada §Saint Lawrence River Institute of Environmental Sciences, Windmill Point, 2 Belmont Avenue, Cornwall, Ontario K6H 4Z1, Canada (Received 21 October 2001; Accepted 15 June 2002) AbstractÐConcentrations of DDT and its metabolites were measured in water, plants, invertebrates, and ®sh from lagoons in the Okavango Delta, Botswana (Africa), where DDT has been used for approximately 50 years. The sampling area was sectioned to distinguish spraying for malaria and for African sleeping sickness. Average concentrations of total DDT (sum of DDT and its metabolites) in the Okavango ranged from 0.009 ng/L in water to 18.76 ng/g wet weight in ®sh. These levels are approximately 1% of those found in piscivorous ®sh from temperate North America. The dichlorodiphenyl ethylene (DDE) metabolite was the most abundant fraction of total DDT. Although total DDT concentrations were higher in areas treated for malaria than areas treated for sleeping sickness, these concentrations were likely driven by factors other than the historic application of the pesticide. Equilibration with air concentrations is the most likely explanation for these levels. Since the mean annual temperature exceeds the temperature of vaporization of DDT, this research points to the need for reliable transport models. Our results showed that total DDT concentration in ®sh was best explained by lipid content of the ®sh and trophic position inferred by d15N, regardless of DDT application history in those areas. The reservoir above Gaborone Dam, an area downstream of the Okavango but where DDT had not been used, was sampled to compare total DDT levels to the treated areas. The two species (a herbivorous threespot talapia and the omnivorous sharptooth cat®sh) from Gaborone had levels higher than those found in the Okavango Delta, but these differences can again be explained using trophic position inferred by d15N rather than by ®sh size or location. KeywordsÐOkavango Botswana DDT Fish Aquatic ecosystem INTRODUCTION Persistent organic pollutants, like DDT, have the capacity One of the dirty dozen chemicals now slated for global to accumulate in organisms and magnify in food chains. Fac- restrictions is DDT [1,2]. In Botswana, use of DDT dates to tors known to contribute to this biomagni®cation include lipid the 1950s for control of the mosquito (Anopheles sp.), which content [4,5], some measure of body size such as length or transmits malaria, and the 1960s for control of the tsetse ¯y weight [4,6], and trophic position [7,8]. (Glossina morsitans centralis), which transmits African sleep- Naturally occurring stable isotopes of nitrogen and carbon ing sickness [3]. Both malaria and sleeping sickness are prev- can be used to determine distinctive food web pathways lead- 15 14 alent in the Okavango Delta (World Health Organization and ing to the top trophic position. The ratio of isotope Nto N 15 15 Ministry of Health, Botswana, Africa, unpublished data). expressed as d N (fraction of N atoms per 1,000 atoms of 14 Although records for the total DDT applied to the distinc- N) are used to characterize an organism's trophic position tive regions do not exist, unpublished reports from the World [7±11]. Historically, food chain structure was established using Health Organization state that DDT (75% wettable powder or stomach content analysis, which Hyslop [12] described as pro- DDT emulsion) was applied to homes in the area of the Oka- viding only a snapshot of the ®sh's present diet. The advantage 15 vango from 1950 to 1970. The Ministry of Agriculture (from of the d N technique, as opposed to the conventional stomach pesticide survey reports by P. Mosupi and B. Koosimile, Plant content analysis, is that the isotope signatures integrate dietary Protection Division, Ministry of Agriculture, Gaborone, Bot- habits over a prolonged period of years [13] and thus is able swana, Africa) estimated DDT applications in the Okavango to accommodate slower-growing species [14]. Peterson and 15 to range between 5,000 and 10,000 kg annually in the early Fry [15] showed that d N is enriched by a factor of 3 to 5½ 1990s. After experimenting unsuccessfully with Fenitrothion, from prey to predator, thereby providing a continuous variable a synthetic pyrethroid, in 1972, DDT applications resumed with which to quantify biomagni®cation of organochlorines in from 1973 until 1997 for malaria control in the form of indoor aquatic food webs. 13 residual spraying at a rate of 2 g/m2 once a year. Furthermore, Isotopes of carbon (d C) re¯ect an organism's diet and have regions within the Okavango sprayed for control of anopheles been used to trace the source of carbon from prey to predator had also been historically sprayed to control the tsetse ¯y. In in complex food webs [16]. These authors found that the frac- contrast, the tsetse regions had only been sprayed for control tionation rate of carbon isotopes is relatively low (0±3½) com- of the tsetse. Following pressure from international organi- pared to that of nitrogen (3±5½) isotopes [13]. If food prey zations, DDT in Botswana was banned in 1997. options have a distinctive d13C signature, this isotope has the advantage of providing a time-integrated measure of assimi- * To whom correspondence may be addressed lated diet independent from stomach contents [17]. ([email protected]). The objectives of this study are to compare levels of total 7 8 Environ. Toxicol. Chem. 22, 2003 B. Mbongwe et al. Fig. 1. Map of the Okavango Delta showing the malaria (M) and tsetse or Xakanaka (T) areas where sampling took place. DDT in the water and pelagic biota (plankton to ®sh) in areas ®lled with sediments of various types, notably, wind-blown with different histories of DDT application; to establish a re- brown and white desert sands [19±24]. Maximum and mini- lationship between inferred trophic position (d15N), carbon mum average temperatures in ambient air during the summer source (d13C), lipid content, length and weight, versus total range from 30.5 to 33.78C and 14.8 to 19.28C [22]. Water DDT tissue concentration for selected ®sh species; and to com- temperatures vary between 178C in winter (July) to 358Cin pare biomagni®cation factors, de®ned as the slope of log DDT the summer rainy season (January±February) [23,24]. concentration versus d15N [8] with those of published studies To study the fate and persistence of DDT in the Okavango in temperate regions. Delta, we chose two regions with distinct historical patterns of DDT spraying [25]. The location of the Okavango and the MATERIALS AND METHODS sampling sites are shown on Figure 1. Xakanaka lagoon was Study site chosen as a site where DDT was used exclusively for tsetse The Okavango Delta is an alluvial fan located between 198 ¯y control from the 1960s to the late 1970s. Three other la- and 208S and 228 and 248E in the northern part of Botswana. goons (Ikoga, Samuchima, and Guma) were chosen as sites The fan has an approximate surface area of 25,000 km2 [18]. where DDT was used for Anopheles mosquito control since The area of the wetland is approximately 12,000 km2 [19]. It the 1950s until 1997 [19]. A review of literature on DDT and lies on the edge of the Kalahari Desert and is formed as a other pesticides used in the delta suggests that these three result of the discharge of the Okavango River at the southern lagoons were also treated historically for tsetse control. It extremity of the East African Rift system [20]. The entire should therefore be noted that distributions of DDT in Xak- Okavango River system lies within the Kalahari Basin partly anaka are the only ones that can be solely attributed to tsetse DDT in Botswana, Africa Environ. Toxicol. Chem. 22, 2003 9 ¯y control activities, while the other three locations may be ®sh was intact in the stomach (not yet digested), individual affected by DDT used for both tsetse and Anopheles control analysis was run. For plant material or when samples were activities. digested, we opted for bulk analysis. The reservoir above Gaborone Dam, located in the capital city of Botswana, where DDT use has not been recorded, was Stable isotope analysis also chosen as control site. Although the area has moderate Zooplankton, ®sh muscle (excluding skin), and plant tissue industrial activity, it is a nonagricultural area and therefore were freeze-dried in the laboratory for stable isotope analysis excludes the possibility for the reservoir to be contaminated but not for chemical analysis. Both wet and dry (freeze-dry) directly by DDT use for the control of agricultural pests but weights were determined. Samples were ground with a mortar does not exclude upstream or atmospheric sources. Throughout and pestle, and a 200-mg subsample was transferred into tin this paper, we refer to Gaborone as the control area, Xakanaka capsules. All samples were analyzed by the G.G. Hatch Stable as the tsetse area, and the combined lagoons that received DDT Isotope Laboratories, Department of Earth Sciences, Univer- for both Anopheles and tsetse control as the malaria area. sity of Ottawa (Canada) for carbon and nitrogen, with an au- Sampling tomated CE Instrument (Model EAÐ1110; elemental C and N analyzer) coupled to a Finnigan Mat deltaPLUS IRMS with Most of the sampling was conducted in July and August a Con¯ow II interface (Finnigan Mat, Bremen, Germany).