PARENT SESSION
46 - Surface Water Pollution
8:00 AM to 6:30 PM, Tuesday, 14 May 2002
Exhibition Area

(46-24) Pesticide contamination in the Lourens River, South Africa: Exposure assessment using GIS.

Dabrowski, James^*,1, Schulz, Ralf^1,2, ¹ University of Stellenbosch, Stellenbosch, South Africa² Technical University Braunschweig, Braunschweig, Germany

ABSTRACT- The characterization of nonpoint-source pesticide input, such as runoff and spraydrift in surface waters and identification of areas of concern in the field are essential for a comprehensive assessment of risk and planning of mitigation measures. To this end, the agriculturally-derived aqueous and particle-associated pesticide contamination of the Lourens River and its tributaries was measured during potential runoff and spraydrift conditions. Runoff sampling was performed in the middle of April during high water conditions following the first rainfall of 9.6 mm d^-1. Rainfall-induced runoff resulted in increased levels of the pesticides endosulfan, deltamethrin, azinphos-methyl, chlorpyrifos, and procymidone up to levels of 0.35, 1.4, 0.6, 0.19 and 9 g L^-1, respectively in water samples. Endosulfan, azinphos-methyl, and chlorpyrifos were detected at maximum concentrations of 273, 152, and 245 g kg^-1 in suspended sediments. Transient pesticide levels exceeded the target water quality limit proposed by the South African Department of Water Affairs and Forestry (DWAF). No increased pesticide levels were detected at a control site upstream of the agricultural area. Partitioning of pesticides was well in accordance with their reported water solubilities, with azinphos-methyl (high water solubility) being more prominent in the water phase and endosulfan and chlorpyrifos (low water solubility) being highly associated with the sediment. The Lourens River received its contamination via tributaries discharging the surrounding farming area. Samples taken at the most downstream site in the Lourens River during days on which spraydrift-borne azinphos-methyl entered tributaries further upstream, showed much lower pesticide levels as opposed to runoff samples (<0.05 g L^-1). In general, contamination depended on landuse and slope characteristics, which was analysed using GIS. Together with detailed knowledge of the spraying programme, GIS analysis enabled predictions of contamination levels in all tributaries and the identification of target sites for risk reduction approaches for both runoff and spraydrift situations.

Key words: runoff, spraydrift, landuse, GIS