Data-Based Mechanistic Modelling: Natural Philosophy Revisited?
Professor Peter Young, Lancaster Environment Centre, Lancaster University
Thursday 01 December 2011, 1200-1300
LEC Training Rooms 1 And 2
Centre for Sustainable Water Management Seminar
This seminar will outline the main aspects of Data-Based Mechanistic (DBM) modelling, a largely inductive approach that harks back to the scientific methods used in the era of Natural Philosophy that preceded the Twentieth Century. DBM modelling recognises that 'reductionist' approaches to modelling many natural systems, particularly at the holistic or macro-level (global climate, river catchment, macro-economy), often results in very large simulation models that suffer from ambiguity or 'equifinality' and so are not fully identifiable from the available data. In considering such systems, the DBM modelling approach is not exclusively inductive: it covers the whole range of model-based scientific inference, from hypothetico-deductive simulation modelling when data are scarce, to inductive modelling when suitable data become available, where possible using 'large model emulation' to help bridge the gap between the results obtained using these two modelling procedures. The approach is illustrated by an example concerned with the modelling of solute transport and dispersion in a wetland area, based on data from a tracer experiment.