Abstract:
This study reports an investigation of the influence of irrigation frequency on water and
herbicide transport in a Ghanaian field soil. An attempt was also made at simulating the movements
of water and atrazine [2-chloro-4-ethylamino-6-isopropylamino-s-triazine] in the soil under varying
irrigation frequency. The irrigation treatments used in the study were: Treatment 1 (Tl): 150 mm
of water applied in three equal amounts of 50 mm each on days 1, 9 and 17; Treatment 2 (T2): 150
mm of water applied in five equal amounts of 30 mm each on days 1, 5, 9,13 and 17; and Treatment
3 (T3): 150 mm of water applied in nine equal amounts of 16.6 mm each on days 1, 3, 5, 7, 9, 11,
13, 15 and 17 after the commencement of the experiment. Furthermore, the differences in water
storage under the different irrigation treatments were also investigated.
In general, T2 resulted in highest water storage at the end of the experiment, followed by Tl
and then T3. Water loss from the profile was in the reverse order. Treatment 1 caused more
leaching of atrazine and the centres of mass of atrazine at the end of the experiment were in the
order Tl > T2 « T3. Sorption of atrazine decreased with depth. The percentage of atrazine
recovered decreased linearly with time and the amounts of the herbicide recovered at the end of the
experiment were in the order T2 > Tl ~ T3. Thus treatments 1 and 3 resulted in greater loss of
atrazine probably due to deep percolation resulting from the relatively higher amounts of water
applied and partly due to volatilization, respectively.
The simulated water profiles compared favourably with the field-measured ones especially
in Tl and T3 (with R2 values of 0.754 and 0.674, respectively). In all the treatments, the predicted
water contents were higher in the top 0.3 m than the observed values. However, using 1:1 plots it
was noted that the simulated water content values were in good agreement with the measured ones
especially at the initial stages of the experiments. The deviations of the predicted values from the
observed became more apparent with subsequent water applications. The inability of the model to
correctly predict well was ascribed to its simplicity, neglecting the effects of hysteresis during water
redistribution and the type of simplified numerical scheme used to solve the water flow equation.
The patterns of the observed and predicted solute fronts were in good agreement. Using a
Chi-square test, it was found that differences between the observed and predicted leaching fronts
were not significantly different at 1% level.
