- 4.4 basic considerations about use of numerical models

Mathematical models of ecological, hydrogeological and geochemical systems may be used to simulate the movement of water, and the fate and transport of pollutants within water bodies. Models take a variety of forms and the question(s) that need to be answered (e.g. what is the likely chemical status of a groundwater body?), the data availability and the time and funds available are all relevant considerations in deciding what complexity of model is used. In general the more complex the model, the greater the data requirements and the greater the time and costs needed to complete it. As a consequence, the accuracy of a robust numerical model may be greater than that which can be achieved using a simpler model. However, in the context of water body characterisation under the WFD there are many questions that may be answered adequately with a simple model.

An iterative approach is recommended, where assessors begin with simple conceptual understandings or analytical models and shift to mathematical models only where water bodies appear to be at risk, or where a detailed programme of measures is being developed. In many cases simple analytical models will be adequate to allow an assessment of contaminant behavior, however in certain situation more complex numerical models will be required.

Assessors may use numeric models to make predictions about combined point and diffuse source pollution effects on the wider groundwater body and on dependent surface waters and ecosystems, and to predict the effects of abstractions and artificial recharges on water resources. In addition, development of a numeric model helps assessors to:

• identify data and knowledge limitations;
• predict the impacts from a number of pollution pressures on remote receptors;
• predict the impacts from a number of abstraction or artificial recharge pressures on water resources, including any impacts on surface water bodies and dependent aquatic ecosystems;
• make predictions on the fate and transport of pollutants;
• include spatial and temporal variability in model predictions (which is often not possible with simpler analytical models).