Date Topics & Readings Notes
Peter Reichert will be presenting Use of Time-Dependent Parameters for Improvement and Uncertainty Estimation of Dynamic Models, some insightful and novel methodological ideas to account for bias (inadequacy of the simulator) in the context of hydrological models. His abstract with references is here. Slide is here.


We joint the Engineering Methodology working group (12:30-14:00, room 203) for a Discussion Session with Tony O’Hagan. His slide is available here.

We will probably discuss issues about the relations between simulators and reality. For that discussion, the paper in the suggested readings list Reified Bayesian Modelling and Inference for Physical Systems (Goldstein and Rougier, 2006) will be very relevant.

We will also discuss Validation. You might like to get familiar with the papers suggested in Tony’s basic reading list and with the issues mentioned in the Gaussian Process Methodology Topics handout that Tony prepared during the Kickoff workshop.

There will also be ample time for general discussion and other questions.

Short organizational meeting (11:00am-12:00pm) to identify concrete areas of interest in which individuals would be willing to get involved and work; we will also plan immediate activities. Here is a preliminary list of "general" areas for people to choose from (non exclusive). Please feel free to add new general areas that you are willing to work in.
11-13-2006 Remote (WebEx, Teleconfererencing and Videoconferencing) presentation of Climate models by Climate Modelers groups in NCAR. Room will be 203, and the hours (Mountain Time) of remote connection appear in read in the definite program (remember to add 2 hours to the times displayed in the program). Margaret will send precise instructions.

There will be a total of 4 presentations of Climate models, in the areas of  Anthropogenic Land Cover, Turbulent Flows, Upper Atmosphere, Planetary Boundary Layer; please plan to attend at least one of them (or more if you are interested) as they might suggest interesting methodological issues to consider in the future by our group.

11-20-2006 No meeting (Thanksgiving)
11-27-2006 Meetings resume. We will have the meeting from 13:30 to 15:30. This is just a punctual change for this one day. We'll devote that session to  "Bayesian Physics and Decision Support in Climate-like Modelling" with Leonard Smith (LSE)" .

Lenny'll derive a few climate models, contrast implications of parametric uncertainty, model inadequacy and empirical (in)adequacy in terms of decision support and the advancement of science. Then he will ask/discuss how Bayesian statisticians might best deploy the skills both in aid of the understanding of the earth system and in response to climate change [and climate-like problems more generally]. 

Lenny's note is available here. Supplementary materails can be found here (.png) and here (.ps).