Leading Thoughts: "A Model Approach"

Interim Dean Van Citters talks with Professor Keller outside his lab in the Irving Institute. (Photo by Rob Strong '04)

What brought you to Dartmouth in 2022?

KELLER: The work I'm excited about requires a somewhat uncommon approach. You start with a problem, then let the problem and the decision help you decide on the methods and the team. Thayer has a neat vision of human-centered engineering and a mission-oriented approach that can help to address wicked problems. Dartmouth breaks boundaries, integrating the liberal arts with engineering. That's why I am here.

What research takes place in your lab?

KELLER: I collaborate with many smart researchers at Thayer, Dartmouth, and beyond. We focus on two interrelated tasks: improving our understanding of the Earth system—including the climate system, humans, and built infrastructures—and using this information to inform the design of strategies to manage climate risks. We first identify a climate-related decision that people need to make, such as how much money to spend on decarbonizing a system. We analyze how people or institutions make this decision right now and engage with stakeholder decision-makers to understand their values and mental models to translate those values into objectives. We then approximate the system in a model. Because we do this with a computer, we can test millions of strategies and how they perform in the face of uncertainties. It is better if a strategy fails in a computer model than in real life. We search for ways to improve on current strategy and map out important tradeoffs. In the end, we synthesize everything to come up with guidance for people making decisions.

I imagine a lot of these questions, if solved, would have extraordinary impact on society. How do you whittle it down to one situation to work on now? 

KELLER: The goal is to converge on projects that work for all involved parties. Students and junior researchers want to work on hard, interesting problems relevant to decision-makers and learn cutting-edge methods. In contrast, stakeholders and decision-makers typically care less about cutting-edge methods and want actionable insights, combined with longer-term engagement and decision support. To make this even more complicated, funding is a constant constraint. We do the best we can to find research projects that work for all parties.

What strategies ensure your models support the widest range of communities and can inform good policy?

KELLER: This is an excellent question. Many high-exposure, highly-vulnerable communities struggle to access high-quality information to help them manage climate risks. There are some strategies to mitigate this issue. For one, our lab develops free and open-source tools. In addition, many institutions support organizations that help to link the needs of decision-makers with research and teaching. This provides several benefits. First, decision-
makers can benefit from new insights and tools. Second, students can identify how abstract academic concepts relate to real-world needs and can learn valuable professional skills. Third, researchers can identify new mission-relevant questions. Dartmouth has started on this path, but there is room for growth.

Here is where looking forward is really important. We are training the next generation of leaders. We, as a community, have unprecedented opportunities to improve how we manage climate risks. We face a neat nexus of access to more data, sophisticated tools to better understand the systems and tradeoffs, and a new generation of smart and excited students who want to tackle the challenges.