Position paper on declarative modelling in ecological and environmental research

Simulation models play a central role in ecosystem research and management. They integrate understanding, and enable us to predict future behaviour and to explore alternative policy scenarios. However, there are major problems with current practice, including the cost of developing models, rapid obsolescence, duplication of effort, and opaqueness of model structure. The fundamental reason for this is that models are implemented in a conventional programming language, such as Fortran or C++. Declarative modelling is emerging as a fundamentally different way of addressing the modelling process. The core idea is to separate the representation of model design (the specification of the model) from the tools used to conduct simulations. Once this has been achieved, the model design can be edited, displayed, interrogated, analysed and transformed in ways impossible using existing approaches. This publication explores the basic concepts behind declarative modelling, illustrates its feasibility by reference to an existing visual modelling environment, explores its potential utility in large-scale ecosystem research projects, and compares it with the current trend to develop 'integrated modelling frameworks'.