Modification of water movement equations in the PRZM3 for simulating pesticides in soil profile

• Pesticide Root Zone Model (PRZM) was modified concerning water movement equations.
• MC-Cormack method was used to solve Richard's equation.
• 2,4-D concentrations in a silty clay soil were simulated using the modified model.
• The MC-Cormack technique was found to be numerically stable in the modified model.
• Modified model illustrated better results but its running time increased 1.5 times.

In the area of simulating pesticide transportation in soil profile, the majority of models adhere to simplification techniques and due to this, the simulation achieved does not reflect the reality. One such commonly used model in this regard is PRZM3. In the present research, to ensure improved results, in addition to a general software update, modified equations concerning water movement in soil were applied. To achieve this, one of the most exact numerical solutions (MC-Cormack method) was selected for solving water movement equation (Richard's equation). This equation was then rewritten in the form of mobile-immobile (MIM), and the Shuffled Complex Evaluation (SCE) method for calculating mobile-immobile coefficients was also added to the model. Following model modification, this was used to simulate 2,4-D concentration, and the results were then compared with the results of the main model and measured data (Noshadi et al., 2011) in two different treatments (normal irrigation and deficit irrigation). Considering the statistics, in the normal irrigation treatment for PRZM3, the figure for NRMSE (normalized root mean square error), CRM (coefficient of residual mass) and d (index of agreement) accounted for 0.58, 0.78 and −0.47, respectively while the figures reported in the modified model using MC-Cormack method (PRZM3-MC) were 0.79, 0.28 and −0.04, and in the modified model using MIM form (PRZM3-MC-MIM) they were 0.86, 0.23 and −0.06. Regarding deficit irrigation treatment, for PRZM3, the figure for NRMSE, CRM and d accounted for 0.65, 0.52 and 0.08, respectively while the figures reported in the modified model using PRZM3-MC were 0.77, 0.38 and −0.24 and in PRZM3-MC-MIM they were 0.73, 0.36 and −0.24, respectively. Simulation results reveal that compared to PRZM3, results were more accurate after model modification using PRZM3-MC and PRZM3-MC-MIM.