Dear Colleagues,

We wish to invite your participation in a new seminar series on Quantitative Environmental Decision Analysis. The goal of this seminar series is to foster study and discussion among the community of scholars at Penn State working on quantitative approaches to environmental decision analysis. This domain includes especially environmental applications of statistical decision theory and Bayesian analysis. The forum aspires to serve as a workshop setting for presentation and discussion of research in progress. It will also feature discussion of selected readings, and visits by the occasional guest speaker. The goal is to create a setting that both challenges and inspires participants, including especially graduate students, to produce high-quality innovative scholarship in the area of environmental decision analysis. This series is sponsored jointly by the Earth System Science Center, a unit within the Earth and Environmental Systems Institute, College of Earth and Mineral Sciences; and by the Department of Statistics in the Eberly College of Science. The series is organized and led jointly by Arthur Small (Associate Professor, Meteorology), Murali Haran (Assistant Professor, Statistics), and Klaus Keller (Assistant Professor, Geosciences).

We will kick off this series this Friday, November 16, 3-5 p.m. in 529 Walker. This opening session will feature three short talks on the theme, "What Makes a `Better' Prediction System? Integrating Economic and Statistical Measures of Prediction Quality." The speakers will introduce three themes that tie into a recently-launched project:

Arthur Small, Meteorology: "Measuring `Quality' in Prediction Systems"

Klaus Keller, Geosciences: "The Value of Improved Predictions: An Application Concerning the Effect of Climate Change on the MOC"

Andrew Kleit, Meteorology and Energy & Mineral Engineering: "The Value of Improved Predictions: Using Temperature Forecasts in Electricity Trading"

The end of the first session will be set aside for a brief discussion of organizational issues. Three issues of note are the schedule of sessions for the spring semester; the value and logistics of organizing this study group as an official course for graduate credit; and the scheduling of presenters and discussants for upcoming sessions. A second session will be held on Friday December 7, 3-5 p.m., also in 529 Walker. This session will feature a presentation by John Yorks, M.S. student in Meteorology. Details will be forthcoming. During Spring 2008 the group will meet approximately once per month, generally 3-5 p.m. on a Friday afternoon, on dates to be determined.

Topics. This series will welcome a range of topics related to environmental decision analysis, including:
* Decision analysis under uncertainty. Environmental applications of statistical decision theory and Bayesian analysis.
* Statistical estimation problems characteristic of environmental decision contexts.
* Quantification of uncertainty in parameter estimates and model outputs.
* Estimation of model parameters from non-stationary processes. Examples include the incidence of precipitation, floods, or tropical cyclones in systems subject to decadal-scale climate variation or long-run climate change.
* Extreme value statistics. Estimation of probabilities of low-frequency, high-consequence events, particularly from limited data.
* Probabilistic forecasting and prediction.
* Design of decision systems that integrate data from multiple sources at diverse spatial and temporal scales.
* Appropriate techniques for incorporating domain-specific knowledge (e.g., from meteorology or hydrology) into decision models, particularly in settings with limited data.
* Computation of the value of information for prediction and decision systems. Integration of economic and statistical measures of prediction quality. Incorporation of value of information measures into the data collection and system design process.
* System architecture; data management. Issues of interoperability.
* Risk communication. Information design applied to environmental decision systems. Incorporation of Bayesian and value-of-information concepts into design of user interfaces.
* Behavioral responses of individuals, households and organizations to information about environmental risks. Perception, judgment and decision-making under environmental uncertainty.
* Economics and industrial organization of environmental decision systems.
* Other topics of interest to participants.

Again: The central goal is to provide a workshop environment oriented towards fostering and encouraging research in progress, particularly on projects involving graduate students and post-docs.

The ANGEL group: We have set up an ANGEL group associated with this seminar. This ANGEL group serves as a mailing list for seminar-related announcements, a bulletin board for the calendar of meetings and related events, and a repository for seminar-related articles and readings. The seminar series, and this ANGEL group, are open by invitation to Penn State faculty, researchers, students, and others interested in quantitative environmental decision analysis. Interested participants can add themselves to the ANGEL group:
1. Log in to ANGEL at
2. Look for the section "My Groups". Click on the link "Find a Group".
3. In the keyword search box, enter the name of the group (or some shorter version like "environmental decision").
4. Enroll in the group with the case-sensitive PIN: Bayes
For more details on this procedure, see:

We are also organizing for the distribution of slides and audio recordings via the ESSC web site. Check soon at the ESSC Seminar page.

We hope you will join us.


Murali Haran, Statistics
Klaus Keller, Geosciences
Arthur Small, Meteorology