Atrazine movement in a dark red latosol of the tropics

Atrazine displacement was studied in a soil profile as a function of water movement and time after herbicide application, taking into account possible influence of preferential flow on leaching. The experiment consisted of two 7 x 7m plots of a dark red latosol (Kanduidalfic Eutrudox), located at Piracicaba, SP, Brazil (22o 43'S and 47o 25'W), 250km inside continent, at an altitude of 580m. One plot was previously treated with 1,000 kg/ha of lime, in order to increase base saturation to 88%, and 500kg/ha of gypsum. Each plot was instrumented with tensiometers, neutron probe access tubes and soil solution extractors, in order to monitor water and atrazine flows. Atrazine was applied at the high rate of 6 kg/ha of active principle. Results showed intensive leaching of atrazine in the whole soil profile, up to the instrumented depth of 150cm, already ate the first sampling, seven days after herbicide application. The limed plot showed much higher atrazine leaching losses than the other plot. The atrazine adsorption capacity of the soil is very low, its maximum value being of the order of 10%, for the 0-15cm surface layer.


INTRODUCTION
In modern agriculture, mainly under intensive land use conditions, manual and animal labors were almost totally replaced by machines and the use of agrochemicals to control pests, diseases and weeds. The use of these chemicals is, in most cases, performed in inadequate and abusive ways. A technology introduced to help the fanner increasing productivity, can also contribute causing severe damage to the environment. Among the several agrochemicals used, atrazine was chosen for this study, since according to BAILEY et al. (1968), it presents base characteristics and can, therefore, be adsorbed by positively charged clay particles. According to these authors, a basic compound is 10% associated when soil pH is one unit above the value of the product pKa, 90% associated when the pH is Sci. agric., Piracicaba, 54(Numero Especial), p.116-120, junho 1997 one unit below pKa, and 100% associated or completely dissociated when the pH is two units above or below pKa, respectively. Although the pKa value for atrazine is 1.68, the increase of soil pH due to liming develops conditions for low atrazine adsorption because the increase in the soil cation exchange capacity is counteracted by a much stronger decrease of atrazine positive charges, and the final result is a lower adsorption. Under these conditions there is an increase in the risk of ground-water contamination through the leaching process.
According to Green & Obien (1969), in the case of soils of low adsorption capacity, changes in soil water content alter significantly the herbicide concentration in soil solution. Bacci et al. (1989) also observed for a given soil profile, that 130 days after the application of 2 kg/ha of atrazine, the chemical was found in all soil layers, down to the 3.2 m depth. Smith et a/. (1992) also observed a rapid downward atrazine movement, to the 0.8 m depth, in 1.15 m long sandy soil columns, submitted to 35 mm irrigation. Guth et al. (1977) verified that the main atrazine transport mechanism is the aqueous phase mass transport. Zins et al. (1991) studied the effect of alfalfa roots on atrazine and alachlor movements, in silty soil columns, concluding that the presence of roots facilitated pesticide movement as a consequence of the development of macropores, due to root growth. This paper has the aim of studying atrazine displacement in a soil profile, as a function of soil water movement, after herbicide application, and the possible influence of preferential flow on herbicide leaching losses.

MATERIALS AND METHODS
The experiment was carried out on a dark red latosol (Kanduidalfic Eutrudox), known as "terra roxa estruturada", at the county of Piracicaba, SP, Brazil (22°43'S and 47°25W), 250km inside continent, at an altitude of 580m. Two plots of 7 x 7m, separated by a distance of 8m, were instrumented with i) a neutron probe access tubes of 2m length, to measure soil water contents at the depths of 0.20, 0.50, 1.00 and 1.50 nr, 9 tensiometer sets, each composed of six tensiometers with cups at 0.35, 0.65, 0.85, 1.15, 1.35 and 1.65 m below soil surface in order to observe the direction of water flow in the different portions of the profile, and iii) 9 soil solution extractor sets, each composed of four porous cups at 0.20, 0.50, 1.00 and 1.50 m below soil surface ( Figure 1). Plot N° 1 received 1,000 kg/ha of lime and 500 kg/ha of gypsum, which were manually incorporated into the 0.10m surface layer, to increase base saturation to 88%. Plot N° 2 was left in its natural base saturation condition . Table 1 characterizes the plots from the chemical and physical points of view.

RESULTS AND DISCUSSION
Total soil water potential head Y (cm 1120) distributions can be seen in Figure 3. They indicate downward water flow during the whole experimental period. This was a desired condition in order to maximize pesticide leaching and have an extreme situation. Soil water content q (cm3.crtf3) profiles show a slight influence of the textural B horizon (0.40-0.60m) of the profile (Figure 4), due to its higher clay content. The extreme variability of q measurements did not allow the calculation of soil water flux densities, as already discussed for this soil by Reichardt et al. (1993).   Comparing data of the three lines, for the 1st sampling (7d), leaching of atrazine was fast and variable, reaching the depth of 1.50 m at concentrations of 5, 2 and 200 ppb for line 1, 2 and 3, respectively, with maximum of 55 ppb at depth 0.50 m for line 1, 130 ppb at depth 100 for line 2, and with no maximum for line 3. Atrazine concentration increased along time at the 1.50 m depth, for line 1 and 2, and decreased for line 3, where the highest concentration occurred at the first sampling date. These results indicate that leaching losses were greater at the plot side corresponding to line 3, suggesting conditions of preferential flow at this side. For plot N° 2 ( Figure 6) atrazine profiles show a very different behavior as compared to plot N° 1. Here, concentrations were much lower for all depths at all times. The highest concentrations were found at the depth of 0.20 m, and they decreased rapidly along time. Variation between lines was significantly lower, indicating a greater homogeneity in the leaching process.
Plot N° 1, which received lime and gypsum, presented greater leaching since, because according to BAILEY et al. (1968), the increase in pH reduces atrazine adsorption capacity, promoting leaching. This fact associated to the high soil water content levels, affects significantly the herbicide concentration in soil solution and contributes to fasten atrazine movement in the soil profile (GREEN & OBIEN, 1969). Saturated hydraulic conductivity measurements, performed before herbicide application, show that the free drainage of plot No 2 is 3.5 times greater than of plot N° 1. On the 3r d and 7th day after atrazine application, plots received rainfalls of 34.4 and 34.9 mm, totalizing a water 10) excess of 55 mm, and it is possible that a great part of the atrazine of plot No 2 was leached to depths greater than 150 cm, by mass flow. GUTH at al. (1977) state that the main transport mechanism of atrazine is mass flow.
Results obtained for adsorption (Figure 7 With respect to desorption , it was observed that 65% of the adsorbed amount in soil of the 0-0.30 m layer turned free by desorption. This indicates that besides having a very low adsorption capacity, the small amounts that are adsorbed can easily be released to soil solution.
Although it was not possible to quantify atrazine flux densities, it is concluded that for the kind of soil tested, under high soil water content and pH conditions, atrazine leaching losses may be significant, reaching soil layers below root zone, with consequent risk to groundwater contamination. The presence of preferential water flow paths increase this leaching potential.