Maximum number of students: 18. If the number of enrollments will be higher than 18, students of the ECGS -programme are prioritized
B.Sc. in aquatic studies, environmental studies, environmental economics or other relevant field. Basic understanding of statistical notions. Courses ECGS-014 and ECGS-501 are recommendable.
After the course the students:
- Understand the principles of probabilistic decision analysis
- Can structure different types of environmental management problems in causal decision-analytic format
- Are able to build simple decision analytic models using a Bayesian network software
Environmental management problems are typically laden with different types of uncertainties arising from various sources. Probabilistic decision analysis can add our understanding concerning the nature of the management problems and help us to identify the formally optimal decisions, given the prevailing level of knowledge and uncertainties. By requiring precise definition of the management objectives, decision analysis forces us to be clear and transparent on what we are actually deciding about and choosing between. This way the method has a great potential for supporting evidence-based decision-making and the related discussion in the society.
A graphical Bayesian Network (BN) is a visual probabilistic model for causal inference and decision analysis. BNs can integrate qualitative and quantitative data in the same analysis, still providing numerical estimates and indicators about the system in focus. The approach can help with problem framing and structuring, and aid in diagnostic analysis of your system of interest. It can also be utilized for predicting and divergent optimization tasks, for example for finding optimal management strategies.
This course is an introductory course to the environmental decision analysis and Bayesian networks. We will start from the basics, thus all you need to be able to follow, is the understanding of basic concepts of statistics, which you probably have learned already in the high school (such as the notions of average, frequency and percentage calculation). Students from a wide variety of scientific disciplines are encouraged to enroll. The more multi-disciplinary group we have, the better!
The case examples of the course are mostly related to the social-ecological systems of the Baltic Sea. It is recommendable to take the courses ECGS-014 (Diagnosis of environmental problems in aquatic ecosystems) and ECGS-501 (Management of environmental problems in aquatic ecosystems) before this course.
Estimated work hours: 135 h
Lectures, exercises, project model and report (small groups), final seminar
Materials provided in Moodle
Grade 0-5 based on:
- Home exercises: accomplishment of all the exercises is required to pass the course. Their quality makes 20% of the course grade. Not returning them in time lowers the course grade.
- Course project (group work): accomplishment is required to pass the course. The grade of the work makes 50% of the course grade (given that the work load among the group members have been equal).
- Self-evaluation report of the project group: mandatory, quality makes 10% of the grade.
- Participation to activities and discussions in class: 20% of the grade.
- Attendance at the lectures: attendance of 80% is required to pass the course. Non-attendance needs to be compensated with additional work as agreed with the responsible teacher. Attendance at the final seminar (the last course day) is mandatory.
- Principles and elements of decision analysis
- Structuring and framing divergent environmental decision-making problems
- Probabilistic causal inference and knowledge integration with Bayesian networks
- Learning to use a Bayesian network software (incl. modelling exercises in class room and at home).
- Course project (in small groups): developing a Bayesian network model to answer to a chosen environmental management -related analytical question. Writing a report, including a self-evaluation of the group.
- Final seminar
The course is an optional part of the “Baltic Sea Studies” module.