Feb. 03, 2022 -- Resilience engineering - what can we learn from social-ecological resilience?
As part of rolling out this website, I'm going to revisit some old projects. From 2019 to 2020, I got the chance to work with a great group of folks (led by Dr. David Yu at Purdue) across institutions and countries to ask: How do we assess resilience in designed systems like infrastructure? What happens when those systems contain social, ecological, and technological components? We published out finding in Risk Analysis - "Toward General Principles for Resilience Engineering" https://doi.org/10.1111/risa.13494
Let's talk about what we found. With the help of two awesome reviewers at Risk Analysis, we identified ways the rich social-ecological resilience lit can inform engineering and design. We know what it means for a system to be resilient - it is able to withstand shocks and maintain its function or state. We sort of know how to measure it (although this is HARD). Social-ecological resilience has proposed design principles for systems to make them more resilient. So what can we learn from these principles and apply to infrastructures? Our experience of the world is mediated by coupled natural-human infrastructures composed of built env's, natural systems, and human institutions. Under climate change, these infrastructures are being tested more than ever. The challenge of design is preparing for these tests.
By comparing resilience-enhancing principles in resilience engineering and soc-ecol resilience, we identify eight principles for enhancing design of infrastructure-dependent systems. For those familiar with Ostrom's Rules, you'll see her influence here.
Principle 1: Recognize system context matters - no infrastructure by itself is "resilient" - it is embedded in social-ecological-technological context. So design must expand boundaries. And this means no panaceas - just because it works one place, doesn't mean it'll work another!
P2: Foster social capital - human connections matter! Trust, broad participation, collaboration - this all makes a difference when systems are trying to recover after a disaster, and they make it more likely that systems will be just and equitable when they are functioning.
P3: Maintain diversity - redundancy & functional div. Redundancy means if one goes down, the rest still works because there's a backup. Functional div. means components vary - we know human systems work better w/ diversity - also true of systems w/ humans & physical components!
P4: Manage connectivity - connectivity improves communication, resource sharing, and collaboration! BUT if a system is too connected, failure in one place can produce cascading failures - a case in point is the ongoing COVID-19 pandemic.
P5: Encourage learning-by-doing - under a changing climate, every year is likely to be unprecedented. How do we prepare to make decisions under uncertainty? By learning through practice and sharing this knowledge collectively through "social learning" (back to P4)
P7 (don't worry, P6 is next): Address the problem of fit - how well does the structure of the managing system align with the structure of the system being governed? governance works best when the scale & structure of the manager matches ("fits") the system.
P6: Embrace polycentric control - polycentric control means semi-autonomous mangers work in concert at different scales - this encourages self-correction & diversity & is tied to "fit", since polycentric managers can be positioned to match the structure of the system being governed.
P8: Manage for complexity - this one is tricky, and we proposed a few sub-principles to operationalize it, including consideration of multiple scales and their links, understanding robustness-vulnerability tradeoffs, and attention to interdependencies in infrastructures.
We conclude by pointing out that there is no magic formula for resilience - instead, by paying attention to these design principles, we can better understand why some infrastructures are more resilient than others - and we can ensure all are more resilient moving forward.