10Mechanistic models used for prediction should be parsimonious, as models 11 which are over-parameterised may have poor predictive performance. 12 Determining whether a model is parsimonious requires comparisons with 13 alternative model formulations with differing levels of complexity. 14 However, creating alternative formulations for large mechanistic models is 15 often problematic, and usually time-consuming. Consequently, few are 16 ever investigated. In this paper, we present an approach which rapidly 17 generates reduced model formulations by replacing a model's variables 18 with constants. These reduced alternatives can be compared to the 19 original model, using data based model selection criteria, to assist in the 20 identification of potentially unnecessary model complexity, and thereby 21 inform reformulation of the model. To illustrate the approach, we present 22 its application to a published radiocaesium plant-uptake model, which 23 predicts uptake on the basis of soil characteristics (e.g. pH, organic matter 24 content, clay content). A total of 1024 reduced model formulations were 25

The STRATEGY project (Sustainable Restoration and Long-Term Management of Contaminated Rural, Urban and Industrial Ecosystems) aimed to provide a holistic decision framework for the selection of optimal restoration strategies for the long-term sustainable were identified. The importance of stakeholder consultation at a local level and of ensuring that any response is site and scenario specific were emphasised. A value matrix approach was suggested as a method of addressing social and ethical issues within the decision-making process, and was designed to be compatible with both the countermeasure compendia and the decision support system. The applicability and usefulness of STRATEGY outputs for food production systems in the medium to long term is assessed.

No abstract

No abstract

HighlightsWe demonstrate the application of systematic model reduction to a complex crop model.The model was manipulated under software control replacing variables with constants.Redundancy in representation of nitrogen physiology and temperature was found.The level of detail in Sirius is defensible if detailed prediction is required.The approach increases the efficiency and rigour of model evaluation.