INTRODUCTION

The existence of a large number of catchment hydrology models, evident from even a cursory glance at the literature, is likely to cause trepidation or confusion for even expert modellers, let alone practitioners of hydrology who merely require something ‘off the shelf’ which they can use with confidence. Typically, available catchment models often come with exaggerated claims for the breadth of their applicability, and little or no in-depth discussion of their inherent assumptions and consequent limitations. Two questions will therefore be addressed in this chapter:

1. How can an appropriate model for my catchment be chosen, given an intended application? and

2. How can an appropriate model be constructed (or an existing model be modified) if none exists at present?

There would appear to be several reasons for the present wide range of models, including:

• a diverse range of catchments and purposes (for example, forecasting or regulatory support) which in turn implies interest in many different kinds of processes;

• availability of different levels of information or data quantity and quality; and

• the fact that catchments are complex systems, having a huge number of potentially significant processes, and consequently ‘emergent behaviour’ (defined later on) which is not evidently a simple sum of the component parts.

Taken together, these three factors imply that to represent catchment behaviour efficiently, much of what is deemed to be of secondary importance must inevitably be either ignored or greatly simplified by using specific assumptions.