Evaluation of pesticide dynamics of the WAVE-model

Abstract

A validation study of the physical based pesticide leaching model WAVE is presented. The model considers a mechanistic description of 1-D water, solute and heat transport. Linear sorption isotherms and first order degradation sub-models are used to simulate pesticide sorption and transformation. The first order degradation rates are reduced when temperature and moisture stress in the soil profile occur. The model is conceived to describe pesticide fate within rigid mineral soils. Model tests were therefore done using data collected at a sandy (Vredepeel) and a loamy soil (Weiherbach). Both field data and lysimeter data were used to evaluate the performance to describe water, bromide, ethoprophos, bentazone and isoproturon transport in soil. The evaluation procedure presented by Vanclooster et al. (Agric. Water Mgmt., Vol. 44, pp. 1–19) was completely adopted. The measured soil moisture in the sandy soil could only successfully be described after calibrating the hydraulic functions using field observed soil moisture profiles. In addition, the predicted balance terms, such as the soil water drainage, were subject to a lot of uncertainty. Bromide transport in the sandy soil was poorly described with the equilibrium solute transport model. Anomalies were also observed when simulating the transport of the inert tracer in the lysimeter at the loamy site. The fate of the weakly sorbing bentazone component was appropriately described at the Vredepeel field site. However, the retardation of the strongly sorbing ethoprophos and isoproturon components was poorly simulated. Further, the pesticide dissipation varied considerably in time, which could not be accounted for with the first order degradation model. The need for model calibration illustrates the constraints when using mechanistic models such as WAVE to predict field scale pesticide fate and transport. The adoption of a mechanistic model for registration purposes may therefore be subjected to a lot of uncertainty. In addition, processes affecting pesticide fate and transport are still poorly represented within the model. Deficiencies are related to the description of non-linear sorption, time dependent pesticide degradation, and pesticide volatilisation. Future developments with the model should therefore envisage to improve the parametrisation reduce the output uncertainty, and improve process descriptions of essential processes.