Building a modelling toolkit for prediction of catchment behaviour

Since 1999 the CRC for Catchment Hydrology has been developing a modelling toolkit for the prediction of catchment behaviour. The modelling toolkit is underlain by a development, management and deployment environment that supports module-based modelling services such as data loading, computation and analysis, and display. Modules are constructed through cloning of existing modules, or from scratch. Models can be built by selecting and linking relevant modules on a visual 'canvas'. This process is supported by a control and documentation system that draws upon the module code base for parameter and interface information. For deployment of models, options include command line operation of individual modules, stand-alone executable applications, and module libraries that allow testing of alternative modules with a given data set. Development of the modelling toolkit has included consideration of a broad range of social, technical and communication aspects that range beyond the provision of modelling tools. Key issues include conceptual development, user modelling practices and information management and dissemination. The modelling toolkit has a range of users, such as toolkit developers, model developers and users of applications, and a range of ways that it can be used. Each user has particular information needs, varying from parameter and data information requirements of individual modules, through user manuals for models, to meta-level information on selection of appropriate modules from those available in libraries. For deployment of some or all of the toolkit via the WWW, a system of information supply and user management must be implemented to provide relevant information and functionality, maintain security and provide user-group specific support. This paper will report on the toolkit, its range of users and uses, and development of a toolkit information system.