|
Thematic Issue on
Emulation techniques for the reduction and sensitivity analysis of complex environmental models
Issue Editors:
Marco Ratto, Andrea Castelletti
Aim
Contributions are invited to a forthcoming thematic issue of Environmental Modelling and Software devoted to survey assess and evaluate available emulation techniques for the analysis of complex environmental models.
This important and expanding area of research represents one of the major advances in the study of complex mathematical models, with applications ranging from sensitivity analysis to model reduction. Computational limitations remain a major barrier to the effective and systematic use of large-scale, process-based simulation models in rational environmental decision-making. Whereas complex models may provide clear advantages when the goal of the modelling exercise is to enhance our understanding of the natural processes, they introduce problems of model identifiability caused by over-parameterization and may not be the best choice for control, management and planning purposes, i.e. when any kind of feedback control, optimization or real-time forecasting is required. Therefore, a combination of techniques for complex model reduction with procedures for data assimilation and learning-based control could help to bridge the gap between science and practical decision-making.
This thematic issue aims at providing a guide and reference for modellers in choosing appropriate approaches and understanding their features. Also, it aims at producing an extensive validation and benchmarking of different emulation approaches, still lacking in the literature. Finally, the illustration of emulation approaches will be completed through the contribution of successful applications in environmental modeling.
The thematic issue will provide a useful benchmark in the academic literature for this important and expanding area of research, and will create an opportunity for dialogue between methodological and user focussed research.
Structure
One position paper will form the heart of the Thematic Issue: the position paper will be the final outcome of the Workshop W8. Complexity reduction strategies for effective use of process based models in environmental decision making, organized by Andrea Castelletti, Rodolfo Soncini-Sessa, Peter C. Young, Hoshin V. Gupta, Marco Ratto at the iEMSs2010 Congress (http://www.iemss.org/iemss2010/page31.html).
Further papers will then be published covering the following topic areas:
Methodology – The focus will be on, but not limited to, model complexity reduction techniques, such as dominant mode analysis, large model emulation and meta-modelling, response surfaces, diagnostic model evaluation, model structure and parameter estimation, model correction and, in the case of partial differential equations, methods such as stochastic collocation on sparse grids. Contributions on specific subtopics, such as design of experiments, factor screening, sparse grids, feature extraction, etc. are also welcome.
Validation and benchmarking – The aim will be to provide an extensive validation and benchmarking of different emulation approaches that can be adopted to attack the same sensitivity/reduction problem at hand. Both the use of analytic test cases and of real applications are welcome.
Application – Applications might include water resources, air quality, ecosystems, land use, energy efficiency, climate modelling and any other relevant environmental topic.
Participation and timeline
The editors would like to extend an open opportunity to the membership of iEMSs, participants to SAMO 2010 -- the conference on sensitivity analysis that will take place at Bocconi University in Milan on July 19-22, 2010, http://samo2010.unibocconi.it/ - and others in the development of the thematic issue, either through specific papers on one of the topics above or through reviewing any of the papers or in helping to write the dedicated position paper.
If you are interested in participating please email your contact details and intended participation (discussion paper, topic, title plus a very brief abstract (max. 1 page)) to
This e-mail address is being protected from spambots. You need JavaScript enabled to view it
not later than 30. May 2010.
Note: After initial selection of responses, we will ask for full papers (approx. 20 pages double spaced lines) up to 31st October 2010.
We will use a three step review process: 1. abstracts of about one page will be reviewed by the Guest Editors, and recommendations will be then made for the scope of the full paper; 2. the full paper will undergo a constructive and comprehensive review process to ensure cohesiveness of the special issue and the best quality output; 3. revised manuscripts will be examined by the Guest Editors and, where necessary, the original reviewers.
|
