Postdoctoral Research Fellowship in Biology: Management for an uncertain world: robust decision theory in face of regime shifts
This research will explore methods for more robust management of ecosystems when the underlying dynamics are uncertain. This work seeks to bridge mathematical and biological methods that have hitherto been developed largely in isolation, such as optimal control (from decision-theoretic work) and early warning signals (from resilience work) as well as machine learning approaches (from statistics and computer science). To anchor this work in the biology of a real world problem, examples and applications will come from the ecosystem dynamics and economic concerns of marine fisheries. The three main objectives of this project are: (1) Develop an approach to integrate early warning signals into a decision-theory framework (2) explore how nonparametric Bayesian inference can account for structural uncertainty in ecological dynamics, (3) explore how active learning approaches in each of these areas could be developed to design adaptive management policies that actively decrease uncertainty.
Effective decision-making in face of deep uncertainty is a fundamental challenge across many disciplines. Methods that can improve the ability to select robust actions under such circumstances may have far reaching value. Such real-world applications will only emerge as they are percolated through and developed by a larger community of researchers and managers. With this in mind, this project also seeks to promote a more immediate impact to the scientific community, the science education community, and the public at large through the use and development of an online open laboratory notebook. The online notebook will chronicle the daily advances in this research while also exploring mechanisms and technology that can advance scientific infrastructure, collaboration and dissemination of results. See http://carlboettiger.info/lab-notebook.
