Ashton Drew

Drew_profileThe purpose of my research is to improve decision-making; particularly urgent decisions that must be made despite data-limitations, high uncertainty, and conflicting values. My applied research has contributed to practical decision models that have helped individuals and agencies effectively and efficiently leverage available resources towards the best possible decisions for their particular and unique contexts.   Just as critical, these decision models allow clients to avoid counterproductive decisions, secure in the confidence that these alternatives have been fairly assessed rather than ignored. Working primarily in the fields of conservation planning and natural resource management, my research draws upon ideas and methods from the fields of natural science, social science, and data science. I am involved in applied, collaborative projects dealing with a wide range of data and design questions that cross multi-disciplinary boundaries, but primarily focus on applied statistics, decision theory, and systems modeling.

Knowledge Elicitation, Statistical Encoding, and Integration in Decision Models

Expert knowledge is often the only immediate source of knowledge from which to propose conservation management actions. A considerable portion of my methodological and theoretical research has addressed the rigorous collection and application of expert knowledge within adaptive management settings, especially as integrated into Bayesian model frameworks. Structured elicitation techniques are necessary to avoid common pitfalls of expert bias, inaccuracy, and uncertainty. I have tested alternative approaches to integrating expert knowledge into conservation planning to set population and habitat objectives, to design protected areas, to support Recovery Planning, and to quantify the biodiversity impacts of commercial agriculture. The breadth of these experiences and collaborations also led to the publication of two textbooks and recommendations to improve the rigor of expert knowledge applications in landscape ecology. My present efforts in this area include collaborations: (1) to further test and expand on simple elicitation techniques developed by Rebecca O’Leary, and (2) to develop guidelines for use of Sama Low Choy’s scenario-based elicitation techniques when used to support Bayesian mixed model regressions.

Complicated versus Complex Representations of Ecological Systems

Often conservation projects exhibit strong planning and design, but weak delivery with failure to fully implement the decision-outcome-learning feedback loops required to achieve adaptive management. Although this failure is often blamed on agency resource limitations, I wish to investigate another possible explanation: the failure of adaptive management models to address complexity. Ecology and conservation papers recommend adaptive management for situations where uncertainty is high and risk is low – but this contradicts advice from organizational development theory (e.g., work by Glenda Eoyang, David Snowdon). Organizational development theory distinguishes between complicated (ordered) and complex (unordered) systems and recommends adaptive management only for complicated systems. Yet, many adaptive management plans address challenges of constantly evolving natural-human systems, without considering the complex, unordered nature of these systems. I observe that when the temporal and spatial scale of decision models exceeds that of the managed ecological-environmental systems (and the social systems setting the values and priorities for management), then uncertainty remains high as new patterns emerge faster than adaptation can occur. Adaptive management will always fail under these conditions. To that end I am exploring (1) the degree to which complexity can be reduced by adjusting spatial and/or temporal scales of decision models, and (2) how common decision frameworks, such as Structured Decision Making, can be updated to better incorporate complexity. I recently initiated collaborations to explore theoretical aspects of adaptive management within complex, coupled systems, as well as statistical methods to gather and analyze the relevant data. Meanwhile, I am exploring applied aspects of the same research questions through my involvement in a National Socio-Environmental Synthesis Center think-tank developing algorithms and tools for endangered species recovery prioritization.

To learn more, visit: