Hi Dr. Yavitt-
To add on to what Noelle said: One of the key findings of this investigation is that it is nearly impossible to operationalize resilience! We used a combination of qualitative and quantitative data to determine a "resilience score”. While the score is a number in our web diagram, it isn’t replicable. Every group of practitioners would likely come up with similar numbers, but there is no rigorous way of quantifying and replicating our work because:
1. The rankings we used (a scale of 1-5) of the major metrics of resilience are assigned at grossly different temporal and spatial scales and are largely incommensurable. How do you measure modularity and diversity on the same scale? You can’t.
2. The rankings are assigned based on definitions of the 1 and 5 values; most assigned values fall in the middle of those, so it’s hard to distinguish between one person’s “2” ranking and another person’s “3” ranking, and
3. Trying to capture the essence or fundamental nature of a highly complex system with so many moving parts and across such an enormity of scales with incommensurability is an arduous undertaking. I think your question illuminates that very issue: how we derive our values is not transparent to the reader, and that’s a major failing of this approach.
While we acknowledging the major flaws in our methodology, we also emphasize that this is a “first stab” at a very daunting and challenging endeavor! As we mention in our video, understanding thresholds in complex systems is essential for avoiding social-ecological catastrophes. We see our work as a well-developed and a carefully considered starting place for attempting to disentangle key drivers of resilience in these complex systems.
Hope that helps and thanks for giving us the opportunity to explain our work in more detail!

Hi Dr. Bhattacharya, thanks for the question!

The Goulburn-Broken catchment (GBC) in the Murray-Darling Basin in southeastern Australia is an example of a non-resilient social-ecological system. The major industry in the GBC is dairy production, and much of the deep-rooted native vegetation has been removed and replaced with shallow-rooting non-native vegetation as forage. These planted pastures are also irrigated with ground water. The replacement of the native vegetation and the use of irrigation has allowed the water table to rise near to the rooting zone of the pasture plants; this zone also happens to be laden with salt deposits, making any water that reaches the rooting zone deadly to plants. Too little rain in the region would cause drastic declines in dairy output, but too much rain in the region would make pastures salty wastelands. Further reducing resilience, the farms in the region had been so successful through most of the 20th century that they specialized to a degree that any shift in the market could now spell disaster for the dairy industry. This combination of factors has led to a system with low social resilience (little flexibility) and low ecological resilience (very close to a salty threshold). [example summarized from Walker and Salt (2006) “Resilience Thinking”] Fortunately, the residents of the GBC have realized their predicament and are beginning to make small strides toward improving the resilience of the system, but it will likely take a major restructuring of the system to be successful.

There are a few success stories of highly resilient social-ecological systems. One is the island nation of Tikopia. Inhabitants of the island nations have a deep cultural history of careful resource consumption, sustainable and diversified agricultural practices and population control through, for example, infanticide and celibacy. This behavior is likely the result of the islanders’ forward thinking vision -literally! The size and topography of the island makes it easily viewed in its entirety by island inhabitants; so the effects of human activity on the ecosystem were observed and observable by all members of society. Resource consumption and farming practices were adjusted iteratively to maintain visual and experiential indicators of ecological health. That the social-ecological system of Tikopia has continuously existed for more than 3000 years speaks to its deep reservoir of social-ecological resilience. Clearly, we don’t live in a world where infanticide is acceptable or a clear view of a total ecosystem is possible. But that’s not to say that moving towards increased transparency regarding the impact of our resource use and stabilized global populations are out of reach. This example is particularly illuminating because it shows how human and ecological systems are not always at odds. It’s definitely possible to increase human resilience without sapping the resilience of the ecosystem!

Hopefully these examples have expressed the two ends of the resilience spectrum. Most systems lie somewhere between these two, some with greater ecological resilience and others with greater social resilience. For instance, rainforests may be highly resilient ecological systems, but when viewed from a social resilience perspective, we understand that their social resilience is fairly low because they are at great risk for being cut down. Understanding the interplay of social and ecological resilience of systems will help us to make better management decisions and perhaps forecast where limited resources may best be allocated for improving the overall social-ecological resiliency of systems.

Hi Daniel!
There are, in fact, a few definitions floating out there. The earlier ecological resilience thinkers (e.g., Buzz Holling) define it thus: “The ability of a system to absorb disturbance without switching to an alternative stable state”, which is different from the “engineering resilience” definition of “the ability to bounce back to the status quo following disturbance.”

The important difference between these definitions is that the first (ecological resilience) implicitly rejects the idea of linear succession in ecosystems. Rather, it views ecosystems as complex, adaptive, self-organizing, and panarchical. The engineering definition assumes that there is some sort of ecosystem equilibrium or center point that the system should return to. The ecological definition that we subscribe to allows for adaptive transformation through time, in response to disturbance.

For example, a prairie without fire moves towards woody encroachment. At some point that encroachment pushes the system over a threshold and an average prairie fire isn’t enough to remove the woody plants. The system has crossed a threshold. But stochastic fire events before this threshold increase the system’s resilience, which is not its ability to “bounce back” to some arbitrary pre-fire state, but its ability to avoid the woody threshold.

Resistance is another term that has been thrown into the mix of definitions. Resistance is defined as the complement to resilience, and is the “amount of external pressure needed to bring about a given amount of disturbance in the system” (Carpenter et al. 2001, Ecosystems v. 4, p. 766). For example, a system at time A that is disturbed by X will move the system 3 units within its basin of attraction, whereas the same system at time B may require twice the disturbance of X to move 3 units. This is resistance in its most abstract form. Resilience, as compared to this, is measured by the actual size of the basin of attraction that the system is sitting in. With these definitions, you can see that Ecological Resilience does not necessarily include resistance, but that including resistance allows you to define more properties of the system, which is sure to be helpful.

As to which aspect is more important in social-ecological systems, that is a difficult question to answer and probably depends on the properties of the system in question. If you are near the edge of your basin of attraction (i.e., near to a threshold), you may want to attempt to widen the basin somehow (increase resilience), or, you may want to increase resistance of the system. It is difficult to say what would be best.

We hope this helps you think about these different important questions, but realize that we’ve probably added more questions than answers!

-Hannah and Maggi

Great response, Hannah. This topic would clearly be fun and enlightening to discuss with you over beers sometime. . .

Oh, I’d barely claim to be an expert (yet). It’s only my first year, and some of the older cohort are astonishingly well versed on the subject. But I’m still always up for a discussion about resilience, and I love anything from O’Dells or New Belgium!

Hello Dr. Anderson,
Thank you for the question. The Platte River is the research setting of our IGERT. Not all students have research directly related to the Platte, but many do and a lot of our program relates to issues of the Platte River. Therefore, the processes of coming up with the scores was a collaborative, intensive, and iterative one. We broke up into teams, one team focusing on social resilience and one team focusing on ecological resilience. Each person in both teams was in charge of a specific time period or resilience property to research and score. Then each individual presented and defended their initial score decision to the rest of the team in a series of group discussions, which resulted in adjustments to the score. Because there was often disagreement about the “right” score, the final scores used were generated using a rapid prototyping method where each co-author individually and anonymously assigned a score to each variable based on her individual opinion and knowledge of the system. Those were consolidated and averaged to create the group score.

As you can see, there was little actual consensus on some of the scores and there were often tense moments of staunch disagreement, but we needed to be able to move forward in assessing the resilience of the system. This is why we developed the “rapid prototyping” method. With such subjective judgments as assigning a score on a 1 to 5 scale, this method takes into account the opinions of all those involved, and also provides a degree of certainty in the assignment of a value (a higher standard deviation of scores meant lower certainty). This uncertainty may highlight areas that require more research or may point to aspects of a system that are too complex to assign an accurate score.

We feel that this method will be useful for assessing the resilience of other systems in that it can involve multiple stakeholders and forces all involved to consider aspects of the system that they may be unfamiliar with. It provides a way to move forward on difficult issues, and can be an iterative process where you come back and reassess resilience when more information is available or when circumstances or the systems involved change.

Thank you for asking this question! The people involved in the science and the dynamics that happen during meetings are just as important as the outcomes.

Wow- am I glad I asked! As an evaluator on 5 national STEM- related grants, I often use performance and interactive rubrics on a 1 to 5 basis. I am looking forward to reading more about “rapid prototyping”. Like the Platt River system, academic and scientific programs often undergo “sudden change” and must exhibit resilience!

Hi, Dr. Radeloff and thanks for the great question!
Because resilience is so complex and applies to whole systems rather than individual systems parts, it is very difficult to measure. Other attempts to measure resilience consider “specific resilience”, which is the resilience of what (e.g., a forest system) to what (e.g., a forest fire). This is useful because it breaks down the complexity of a system into more palatable bites in order to inform smarter management or policy actions. Unfortunately, this approach is highly specific to an individual system facing a small set of specific, well-documented disturbances.
Our method is a response to this drawback: we wanted to find a way of measuring the overall resilience of a complex social-ecological system, in our case of the lower Platte River. We used a method whereby we investigated changes in Walker and Salt’s nine properties of resilient systems (2006: Resilience Thinking). These included ecological variability, diversity, modularity, acknowledgment of slow variables, tight feedbacks, social capital, innovation, overlap in governance, and ecosystem services. Some of these apply to both social and ecological systems (e.g., tight feedbacks), whereas others apply to only to the social system (e.g., social capital) or the ecosystem (e.g., variability). Walker and Salt’s work builds on decades of research from big resilience thinkers, and their book has illustrative, real-world examples of resilience using a complex systems perspective that we consider more applicable more broadly than the “what to what” approach.
One noteworthy departure of our work from earlier attempts to measure resilience is the inclusion of a quantitative framework. We thought the most useful way to visualize and interpret our scoring system for individual indices in the broader whole system context was to use the spider-web diagrams. This visualization lends itself to an easy interpretation of where resilience is eroded or bolstered, amd where it has been improved, and perhaps how the properties have interacted (e.g., social resilience was only improved at the cost of ecological resilience). It also visually display the emergent nature of resilience. That is, social-ecologial resilience is not a summation of system components but rather a product of all components and their interactions.
Hope that helps address your question!
-Maggi and Hannah

To add to Maggi and Hannah’s comment, some work has been done in the area of disaster resilience in the United States. Cutter, Burton, and Emrich (2010) published a paper in the Journal of Homeland Security and Emergency Management (volume 7: number 1: article 51) entitled “Disaster Resilience Indicators for Benchmarking Baseline Conditions”. The authors developed a quantitative method of generating a disaster resilience score based upon an assessment of various indicators. However, Cutter et al. (2010) note that they were not able to include an ecological component in their approach due to data inconsistency and relevancy issues. Our research takes a different approach so we can look at both the social and the ecological aspects. While this may make resilience harder to assess, we feel that a combined social-ecological perspective is critical.
Thanks for the question!
-Noelle

Hey Erik, thanks for the question.

The cool (and sometimes infuriating) thing about the resilience ball-in-cup model is that it’s applicable to nearly any temporal or spatial scale. For example, environmental changes that take millenia to occur could degrade resilience in complex microbial systems (look up the great oxygenation event!). You could also observe a disturbance over a very short time scale affect a very large area (think volcano). However, in this context (and most others), we define the systems at a scale palatable to human sensibilities(i.e. a watershed over 75 years).

The difficulty is trying to understand resilience as an emergent property of a complex system driven by different processes that occur across a range of temporal and spatial scales (i.e. panarchy).

In terms of “what depth of well are we in?” you first need to define the parameters of your system, which is really difficult. THEN you need to find a way to measure how far you are from that peak/threshold, which is even harder. We attempted to do just that in our poster and, as expected, had a hard time operationalizing any sort of replicable measuring of resilience.

Also, it’s important to note that while this model is really interesting and gives us an elegant way of understanding thresholds, it does have its limitations! I’d almost argue that it belies the complexity that informs it a little too much. Hope that helps!

Check out Resilience Thinking by W.H. Walker or Foundations of Ecological Resilience by Gunderson and Allen (my adviser!) for a more articulate discussion on panarchy.

To add a bit to what Hannah already explained, I have struggled with the implementation of resilience because it is so difficult to know where thresholds are in your system. There are many examples of thresholds, but in many cases they are known because they’ve been crossed. It’s easy to see them in hindsight, but difficult to predict exactly where they are in systems.

Exactly! I think it’s critical to note here that although we’re better at identifying thresholds in retrospect, that doesn’t mean it’s impossible to identify where they might occur down the road under business as usual status quo. It definitely won’t be easy, but we think it’s important to figure it out!

Thanks for the reply Hannah!

Thanks, Robyn! So glad you enjoyed it.

Hi Dr. Tyre, thanks for the tough question!
Modularity is a tough one to quantify and explain in this context, so I’m glad you asked. I think increased communication among social modules drives higher modularity because processes and behaviors are increasingly interdependent among modules, but the connections between modules are not necessarily stronger. In other words, with increasing modularity, the initial “hierarchy” or structure is reorganized into more flexible connections with more modules, and more overlap of process and activity. I’d argue that communication (also applicable to hydrologic connectivity) informs this modularity, and increases the interdependence without increasing the strength of connections.
PRRIP was scored higher for a variety of reasons. Most importantly, PRRIP fundamentally changed the way many managers, landowners, policy makers and local citizens viewed the Platte. It drove the implementation of habitat reconstruction, adaptive management plans to better accommodate new/changing human water demands and, quite importantly, re-introduced some flow variability to the river.
I hope that answers your question, let me know if you’d like a deeper explanation. The modularity stuff is pretty mind bending!

Hi Ben,
Thanks for the comment. I also found it interesting to see how the social and ecological systems seemed to be inversely related, I imagine that trend would be similar in a historical analysis of most systems. Before people have a large impact, there is great resilience in the ecological system (in most cases) and there is very little point in even assessing social resilience because at that point in time it is irrelevant. Once people become more intertwined with a system, there is often a corresponding decline in various ecosystem attributes, but social resilience (which was essentially 0 before) begins to increase. Then we realize, as a society or community, that we do in fact rely on the ecosystem and thus become concerned about its resilience. My hope is that we can have resilience in both social and ecological systens, maybe we will see this in an assesment of the Platte Rivsr 50 years from now. In this sense, I do think the Platte River is probably representative.

Hi Dr. Waring-
Our framework is adding value to our understanding of how resilience informs an understanding of real social-ecological systems.

Because it’s a novel stab at trying to quantify and track a complex theory in a real social-ecological system, the successes, failures and challenges of our approach should inform better attempts to apply the concept of resilience in real ways.

In my own opinion, ecological theories lose a little of their heft when they lack real social/political perspective implications. This framework helps us move towards making resilience a more powerful tool in real social-ecological systems.

I hope that sheds some light on the value our resilience framework is adding to both the theoretical and real social-ecological system resilience discourse.
Thanks for your question!

Hi Dr. Waring,
I think a lot of us in this program have struggled with this very same question. Why resilience, what is different about resilience? I think for me, at least, resilience is a more organized way to think about systems, and it acknowledges that systems can flip to alternative states, and it also acknowledges (in my opinion more importantly) that systems are not static. A lot of systems require disturbance to maintain their structure and function. I work in prairies, mostly mixed-grass prairie, and I understand that a prairie does not look the same all the time (after it’s been grazed heavily or burned it may look completely broken), but disturbance is necessary for the proper functioning of the system. In other words, the ball NEEDS to roll around in the cup. Resilience thinking is systems thinking, it forces us to consider social aspects of a system, and linkages between different parts of a system that maybe we did not think were important (e.g., does the price of livestock really matter for managing a prairie properly? An ecologist might not think of this, but an ecologist trained in resilience thinking might be more willing and able to acknowledge those linkages).

For the Platte River in particular, the resilience framework forces us to think about the thresholds that the system has crossed (a free river to a managed river) and how we might be able to manage the system for all of its component parts.

I hope this helps! Not to advertise it too much, but Walker and Salt’s “Resilience Thinking” (2006) is an easy read and answers this question much better than I did.

Thanks John! Holding down the fort back at the NREL, still? Hope all is going well for ya!