Observed and Simulated Solute Transport Under Varying Water Regimes: II. 2,6-Difluorobenzoic Acid and Dicamba R. J. Pearson, W. P. Inskeep,*J. M. Wraith, H. M. Gaber, and S. D. Comfort ABSTRACT Significant interestin the fate of agrichemicals in soils hasprompted the development of several transport simulation models.Our primary objective was to evaluatethe simulation model LEACHM for predicting the fate of dicamba (3,6-dichloro-2-methoxybenzoic acid) and a nonre- active tracer, 2,6-difluorobenzoic acid (2,6-DFBA) in fallow and cropped (barley, Hordeum vulgare L.) systems under different water application levels. A field study(1992) was conducted using in situ soil columns on a Borollic Calciorthid (Brocko silt loam, Gallatin Co., MT). Dicamba (~4C-iabeled) and 2,6-DFBA were surface applied rates of 0.26 and 112 kg ha-1, respectively. Solute concentrations were measured at three depths (0.36, 0.66, and 0.96 m) using porous cuplysimetersfor 70 d following chemical application. Time-moment analysis of observed solute breakthrough curves (BTCs) generally showed increasing travel times withincreasing soil depth and decreas- ing waterapplication. Dicamba transport was similar to 2,6-DFBA, with the exception that 40 to 60% of applied dicamba was degraded during transport. The distribution of ~4C remaining in the soil columns showed that the primary degradate of dicamba, 3,6-dichlorosalicylic acid (DCSA), was confinedprimarily to surface samples (0-0.2 while dicamba was found only at lower depths.Thisis consistentwith a much higher sorption coefficient (Ko~) determined for DCSA relative to dicamba. Comparison of observed and predicted (LEACHM) solute BTCs suggested that preferential solute transport occurred especially under high and medium water regimes.Finally, data from three field seasons at the same site suggest that the timing of chemical application relative to initial soil water content and plantstage, the presence of root channels,and temporal changes in soil hydraulic properties in the absence of tillage may significantly affect the degree of preferential flow and subsequent agreement between predicted and observed BTCs. p ~ESTICIDE LEACHING and subsequent contamination of groundwaters from agricultural production is a significant national concern. Recent monitoring pro- grams have detected more than 70 pesticides in ground- waters of 38 states (Ritter, 1990; Parsons and Witt, 1988). In one survey, 17 pesticides were detected at concentrations above health advisory limits. DeLuca et al. (1989) and Clark (1990) conducted well water moni- toring surveys in agricultural regions in Montana and documented that several pesticides (including aldicarb, atrazine, 2,4-D, dicamba, MCPA,picloram, and sima- zine) have migrated into shallow groundwaters, presum- ably through normal agricultural management practices. Dicamba (3,6-dichloro-2-methoxybenzoic acid) commonly used in Montana to control broadleaf weeds in small grain production systems and has been identified as a groundwater contaminant (DeLuca et al., 1989). Dicamba is anionic at most soil pH values (pKa = 1.95) R.J. Pearson, W.P.Inskeep, J.M. Wraith, and H.M.Gaber, Dep. of Plant, Soil and Environmental Science, Montana State Univ., Bozeman, MT 59717-0312; S.D. Comfort, Dep. of Agronomy, Univ. of Nebraska, Lincoln, NE 68583-0915. Contribution of the Montana Agric. Exp.Stn., Journal no. J-5003. Received 10 Apr. 1995. *Corresponding author(usswi@ msu.oscs.montana.edu). Published in J. Environ.Qual. 25:654-661 (1996). and is highly soluble in water (6.5 × 103 mg -l) (Pesticide Manual, 9th ed., 1991). Conse.quently, di- camba has a low sorption coefficient (Burnside and Lavy, 1966; Grover and Smith, 1974) and is highly mobile in most soils (Friesen, 1965; Grover, 1977; Scifres and Allen, 1973). Soil environmental conditions play a sig- nificant role in the fate and mobility ofdicamba. Degrada- tion rates of dicamba in surface soils are generally rapid with half-lives ranging from 13.5 to 45 d (Comfort et al., 1992; Krueger et al., 1991; Smith, 197.4; Smith and Cullimore, 1974). The primary degradate of dicamba is DCSA, which is more persistent and less mobile than the parent compound (Smith, 1973a; Smith, 1974; Com- fort et al., 1992). Consequently, the potential for dicamba to leach out of the rooting zone increases dramatically if soil conditions limit its degradation rate. In a laboratory column and incubation study, Comfort et al. (1992) showed that dicamba transport and subsequent leaching losses could be reduced substantially by delaying water application and providing time for dicamba to degrade to the less mobile DCSA. Because of regional concerns about the potential mobil- ity of broadleaf herbicides, we have been studying the fate and mobility of 2,4-D and dicamba under different soil water regimes, in both laboratory (Comfort et al., 1992; Veeh et al., 1996) and field experiments. The specific objectives of this study were to: (i) monitor the transport and fate of dicamba and a nonsorbing tracer, 2,6-DFBA, under fallow and cropped (barley) conditions and different soil water regimes, and (ii) evaluate the suitability of the computer simulation model LEACHM (Wagenet and Hutson, 1989) for predicting the fate these compounds under field conditions, given site- specific soil and climate input parameters. This is the second of two manuscripts (see Pearson et al., 1996) which compare observed and simulated solute transport over a range of applied water and evapotranspirative demands. MATERIALS AND METHODS Field Design An in situ column experiment was conducted in 1992 on a Brocko silt loam (Borillic Calciorthid) soil to study the transport of t4C-labeled dicamba and a nonsorbing tracer~. 2,6-DFBA. Eighteen of the 24 PVC columns (0.20 m diam., 1.22 m depth) described in the companion paper (Pearson et al., 1996) were used to establish three water regimes (high, medium, and low) on continuous crop and continuous fallow treatments, with three replications per treatment. The crop treatments were seeded to barley (cv. Klages) 1 May 1992, and thinned to eight plants per column (approxi- mately one plant per 38 cm 2) on 12 May to match the plant Abbreviations: DCSA, 3,6-dichlorosalicylic acid; DFBA, 2,6-diflu- orobenzoic acid; PET, potential evapotranspiration; BTCs, breakthrough curves; ET,evapotranspiration. 654 Published July, 1996