Pesticide leaching experiments using widely spaced sampling sites may
not adequately characterize chemical leaching behavior such as nonunif
orm flow between sampling points. We conducted this study to determine
the three-dimensional variability of atrazine and chloride movement w
ithin a small volume of soil (2,700 cm(3)) under field conditions, A 1
-m(2) area of Williamson silt loam (coarse-silty, mixed, mesic, Typic
Fragiochrept) was sprayed uniformly with atrazine (1.1 kg ai/ha) and c
hloride (80 kg/ha). We used the Leaching Estimation and Chemistry Mode
l (LEACHM) to simulate chemical movement, After 6.5 cm of rainfall dur
ing a 29-d period, we sampled 36 squares (5 by 5 cm) in the central 30
- by 30-cm portion of the treated area at six depth increments (0 to 2
, 2 to 5, to 5, 5 to 10, 10 to 15, 15 to 21, and 21 to 30 cm) and dete
rmined atrazine and Cl- concentrations. We recovered 26% of the applie
d atrazine and 138% of the applied chloride. Low atrazine recovery may
have been due to leaching beyond 30 cm and/or degradation while exces
s chloride recovery is attributed to high background concentrations, C
oefficients of variation (CVs) for atrazine significantly increased wi
th depth and ranged from 26 to 353%, while CVs for Cl- were independen
t of depth and ranged from 32 to 66%. Derived atrazine concentration i
sograms illustrated highly nonuniform herbicide transport. Although LE
ACHM overestimated atrazine movement in the upper 15 cm, it was fairly
accurate in the lower 15 cm. The overall trend in Cl- flow was adequa
tely predicted, even though the predicted Cl- concentrations were unde
restimated. LEACHM could not accurately predict nonuniform flow or the
variability in solute concentrations between points, However, its pre
diction of the atrazine center oi. mass (about 4.7 cm) agreed well wit
h the derived isograms. These findings demonstrate that localized noni
deal solute transport may be missed in larger sampling schemes and in
simulation models.