Many conceptually different models have been developed to simulate water flow and contaminant transport in the vadose zone. Parameters in these models are often estimated from readily available data using pedotransfer functions (PTFs). While many PTFs have been developed, it is generally not known which PTF is the most suitable for a specific site. Selection of the best model is always uncertain at best. In this presentation we advocate the use of several different models, rather than searching only for the best model. 

One method of using conceptually different models is ‘model abstraction’ - a methodology for reducing the complexity of a simulation model while maintaining the validity of the simulation results with respect to the questions that the simulation is to address. Simpler models can be used along with more complex models to reduce uncertainty in the simulations, to make the modeling effort and its results more explicable and transparent to end users, and to enable more efficient use of available resources in data collection and computations. An example of systematic model abstraction was developed for the case of infiltration in a variably saturated soil when the complex original model gave inexplicable simulation results. Model abstraction produced a transparent process description with the simpler model.

Another way to take advantage of model multiplicity is to combine results obtained with different models using weights. This approach, called multimodeling, is appropriate when different PTFs are available to simulate unsaturated flow. The similarity in results with the different PTFs may require dimensionality reductions.  Application of the multimodeling approach to infiltration showed that monitoring soil water regime in combination with multimodel simulations can be a viable approach for simulating unsaturated flow.

Overall, systematic use of multiple models can substantially improve the worth of monitoring data and the use of monitoring resources.