Water flow patterns and pesticide fluxes in an upland soil in northernThailand H. C IGLASCH a , W. A MELUNG a , S. T OTRAKOOL b & M. K AUPENJOHANN a a Berlin University of Technology, Department of Soil Science, Salzufer 11–12, 10587 Berlin, Germany, and b Chiang Mai University, Institute of Soil Science and Conservation, Faculty of Agriculture, Chiang Mai 50200, Thailand Summary Rapid percolation of water through soil facilitates both the recharge and the contamination of ground- water reservoirs. We have studied the variation of water flux and pesticide leaching through a soil in northern Thailand. At a depth of 55 cm, two pits were equipped with tensiometer-controlled glass suction lysimeters that were connected to a novel on-line solid-phase extraction device. Nine insecticides varying in water solubility from 10 2 to 10 þ6 mg l 1 were applied on the soil surface, and leaching was monitored for 8 weeks. Measured water fluxes were compared with simulated values. Total recovery ranged from traces (malathion, triazophos) to 1.3% (dimethoate) of the applied amount, showing a decreasing retardation with increasing polarity of the substances. All pesticides were detectable in the soil solution during the first rain after application. Due to fingering, 83% of the leachate was transported through 38% of the area at leaching rates of < 2 mm per day. A new adaptation of the Simpson Index revealed that the diversity of the flow pattern increased exponentially with decreasing rates of seepage water flux (R 2 ¼ 0.80). No such correlation was found when leaching was faster, indicating that the flow pattern switched from a fingering- to a matric-dominated flux. No long-term leaching of insecticides was observed. The two profiles studied behaved similarly in terms of both water and pesticide transport. Therefore we suggest that the flow pattern is a stable property of the soil that can be accurately described by our combination of novel experimental setup and statistical analysis of the flow field. Introduction In the last four decades, the population in the uplands of northern Thailand has increased, and farming there has shifted from subsistence to cash crops. One result of this shift has been the loss of approximately half of Thailand’s forest. Deforesta- tion of the slopes has led to severe soil erosion, and water supply there has been affected by an increased demand for irrigation and by pollution with agrochemicals. To solve the problem of soil erosion and to establish a cropping system that is more sustainable than vegetable farming, fruit orchards such as lychee plantations were introduced. Nevertheless, lychee is a cash crop produced for international markets, and its produc- tion requires application of large quantities of agrochemicals. Local lychee farmers were reported to be poisoned by organo- chlorine pesticides, indicating careless handling of these sub- stances (Stuetz et al., 2001). Although pesticides have also been found in Thai food and ground and surface waters (Baun et al., 1998; Thapinta & Hudak, 2000), we are not aware of any studies of the flow pathways of water and contaminants in Thai soils. The soil of the study area swells and shrinks little, and it is dominated by microaggregates, giving it the structure of sand. Nevertheless flow patterns of soil water and its dissolved agro- chemicals may still be highly heterogeneous (Flury, 1996). Most of the preferential flow in the tropics under the typical intense rainstorms passes through animal burrows such as termite galleries (McGarry et al., 2000) and other regions of the soil that are exceptionally conductive (fingering infiltra- tion, Hillel & Baker, 1988; Reichenberger et al., 2002). Prefer- ential flow may contribute to groundwater recharge even when evaporation exceeds precipitation. However, pesticide concen- trations also peak in preferential flow, inevitably linking recharge and potential pollution of groundwater reservoirs and thereby bringing about the urgent need for a better under- standing of the temporal and spatial variation of water and pesticide fluxes in such soil. All common approaches to assess water and contaminant fluxes in the field have limitations when preferential flow occurs. In macroporous soils, where water bypasses the soil Correspondence: H. Ciglasch. E-mail: holger@ciglasch.de Received 1 March 2004; revised version accepted 17 December 2004 European Journal of Soil Science, December 2005, 56, 765–777 doi: 10.1111/j.1365-2389.2005.00712.x # 2005 British Society of Soil Science 765