By Hans Propsma
The largest forest fire recorded at Yellowstone National Park in the United States occurred in 1988. It began as multiple smaller fires before converging into several large fires. The fires burned a total of 800,000 acres (~3,200 km²) from June until November (U.S. National Park Service, 2025). Once the situation was under control, an investigation was launched to understand how a fire of this magnitude could have started and why the existing firefighting methods were unable to stop it.
At the time, the standard policy of the U.S. Forest Service was the “10 a.m.” fire policy, which stated that any fire had to be extinguished no later than 10 a.m. on the day following its detection. Although this policy appeared logical, it had a critical flaw: dense forests cause fast-growing fires, which are virtually impossible to stop by the following day at 10 a.m. due to their exponential growth. Following the 1988 fires, the U.S. Forest Service adopted the “let-burn” fire policy. Under this approach, smaller fires were contained but allowed to burn freely within designated boundaries to prevent the system from reaching a critical point at which massive fires could occur (Forest History Society, 2025).
This example of a forest fire shows a system (i.e., the forest) that naturally tunes itself to a critical point (having a high saturation of trees in a forest), which eventually causes a rare high-impact event (the massive forest fires in 1988). This is called self-organized criticality (SOC), and numerous systems can be found that self-organize to a point of criticality throughout nature and human behavior. This article will first highlight the basic concepts of SOC, and then show two examples of the application of SOC to safety.
Core SOC Concepts
The term self-organized criticality originates from the field of physics, and more specifically, statistical mechanics (AKA statistical physics), which uses mathematical frameworks, statistical methods, and probability theory to analyze large assemblies of microscopic entities. SOC was coined by scientist Per Bak and his colleagues in 1987 and introduced a metaphor commonly referred to as the sandpile model. In this model, individual grains of sand are slowly added to a pile, resulting in occasional landslides that redistribute the sand.
Notably, the sizes of these landslides follow a power-law distribution, meaning that both small and large landslides arise from the same underlying conditions. A power law is a statistical term describing a functional, nonlinear relationship between two quantities in which a relative change in X produces a proportional relative change in Y, regardless of their initial size. This relationship is evident in Bak’s example between the frequency and size of landslides.
Furthermore, when these quantities are plotted on a log–log graph, the power-law relationship becomes visible as a straight line (see Figure 1). Another important characteristic of this metaphor is the absence of a theoretical maximum landslide size; as more grains are added to the pile, the average landslide size increases. This occurs because rare but extremely large events skew the statistics upward.
Self-organized criticality helps explain the occurrence of incidents once a system reaches a certain threshold. In the sandpile example, each additional grain moves the system closer to a point of criticality. When that point is reached, a landslide occurs. The system is also interdependent, meaning that small changes in one area can influence distant parts of the system. This interdependence can produce cascading behavior, where disturbances trigger chain reactions of varying magnitude (Bak et al., 1988).
Figure 1 Log-log plot of the frequency of avalanches of a certain size: D(E) and the size of the avalanches: E
Application 1: Critical Infrastructure Cascading Failures
The first safety-related application of self-organized criticality examined in this article concerns the management of electrical grids, particularly the potential for cascading failures. The reliability of the electrical grid is essential to modern civilization and should therefore be considered a key safety issue. For example, during the New York blackout of August 14–15, 2003, accidental deaths increased by 122%, while nonaccidental deaths (i.e., disease-related deaths) rose by 25% (Anderson & Bell, 2012). Furthermore, the economic damage was estimated to be between seven and ten billion USD (ICF Consulting, 2003). These statistics demonstrate that large-scale blackouts are not merely inconvenient but have significant societal consequences, including increased mortality and substantial economic losses.
It is important to first understand how power transmission systems operate before examining the application of SOC. Power transmission systems consist of numerous components that interact in complex ways. When a component reaches its operating limits, built-in protection mechanisms are triggered to disconnect it from the grid. Components may also fail due to aging, fire, weather, poor maintenance, design flaws, or faulty operating settings. When one component fails, the load it was carrying is transferred to other components.
Blackouts typically occur when an initial component failure leads to a load increase that triggers a series of cascading failures throughout the system. Such cascades may remain localized, affecting only a limited area; however, in worst-case scenarios, they can propagate across the entire grid and cause a total blackout. These grid interactions form a tightly coupled system in which local component failures can rapidly cascade throughout the broader network, a key characteristic of systems operating near a point of criticality.
Modern societies place significant demands on electrical grids, creating continuous pressure to increase grid loading. These demands stem from economic, societal, and engineering factors. For instance, the rapid development of artificial intelligence requires substantial computational power, while technologies such as air conditioning and washing machines have become commonplace. Collectively, these factors may gradually drive the electrical grid toward a point of criticality through self-organizing processes.
Crucially, the increase in grid loading is not the result of a single centralized decision. Rather, it emerges from ongoing responses to rising demand and economic pressures. Protection mechanisms disconnect failing components, while operational practices aim to restore service and maximize utilization. Together, these processes can slowly reshape the grid’s operating state, pushing it toward conditions in which small disturbances may trigger failures of varying magnitudes.
Analyses of blackout sizes across different power grids indicate that they do not follow linear or exponential distributions but instead adhere to a power-law distribution. Smaller blackouts occur more frequently, whereas larger blackouts are less common; however, they occur far more often than would be predicted by linear or exponential statistical models. This pattern suggests that the same mechanisms responsible for smaller outages can, under certain conditions, generate large-scale cascading failures.
A critical factor in power transmission systems is grid loading. When the grid operates under low load, the probability of failure is reduced, and any resulting damage to other components is typically minimal. However, as previously discussed, power grids are increasingly likely to operate under higher loads, which elevates the risk of failure. Components already functioning near their limits are more vulnerable when required to absorb additional load from a failing component, thereby increasing the likelihood of cascading overloads.
Overall, power transmission systems operate within a complex and dynamic balance between societal and economic demands and the operational limits of the grid. Upgrades and maintenance are therefore essential for continued reliability. In this respect, electrical grids share similarities with the earlier SOC example of forest fires: increased fuel saturation (trees) parallels rising grid loads, while the acceptance of smaller fires resembles load shedding or localized outages. Likewise, large-scale fires are analogous to widespread cascading blackouts.
However, it should be noted that power-law behavior in electrical grids does not automatically imply the presence of SOC; similar patterns may also emerge from optimization-based design principles. Nevertheless, the combination of slow, demand-driven loading, operational thresholds, and cascading failures makes SOC a useful conceptual framework for understanding the intrinsic risks associated with operating power grids (Dobson et al., 2007).
Application 2: Workplace Accidents
The second safety related application of self-organized criticality that will be examined in this article will be about workplace accidents. Research done by a group of American scientists shows a power-law distribution occurring between lost time and the number of cases, this shows that self-organized criticality is applicable to workplace accidents. The research considered the data from 28 different labor categories varying from mining and construction to educational and health services, and for all of the 28 categories that were analyzed a power-law distribution was present (Mauro et al., 2018).
All this data strongly suggests that workplace accidents are governed by self-organized criticality, and that there is a common underlying cause of accidents irrespective of their impact. This research closely aligns with other studies in the field of safety sciences, even though those studies do not explicitly mention self-organized criticality. Take for example the research of Heinrich into workplace accidents, the results of the study that analyzed the 28 different labor categories is consistent with the accident triangle Heinrich originally proposed. The application is not limited to the theories of Heinrich but is also applicable to Perrow’s normal accident theory, Dekker’s drift into failure and Vaughan’s normalization of deviance. In the case of workplace accidents there are similarities to other previously mentioned applications of SOC. Think back to the sandpile model, each unsafe act of work is similar to the grain of sand being added to the pile, and the landslides in the sandpile model are similar to accidents occurring (Geraghty, 2026).
Conclusion
The origin of self-organized criticality might be in the field of physics, but as demonstrated in this article, SOC can also be applied to analyze safety related issues. Applications vary from natural disasters, electrical grids and workplace accidents to name a few. The key being the presence of a power-law distribution. Within the field of safety sciences Self-organized criticality can be utilized as a framework to understand systems and activities, and provide insights into nature of safety failures.
Sources:
Anderson, G. B., & Bell, M. L. (2012). Lights out. Epidemiology, 23(2), 189–193. https://doi.org/10.1097/ede.0b013e318245c61c
Bak, P., Tang, C., & Wiesenfeld, K. (1988). Self-organized criticality. Physical Review. A, General Physics, 38(1), 364–374. https://doi.org/10.1103/physreva.38.364
Dobson, I., Carreras, B. A., Lynch, V. E., & Newman, D. E. (2007). Complex systems analysis of series of blackouts: Cascading failure, critical points, and self-organization. Chaos An Interdisciplinary Journal Of Nonlinear Science, 17(2), 026103. https://doi.org/10.1063/1.2737822
Forest History Society. (2025, 2 december). U.S. Forest Service Fire Suppression. https://foresthistory.org/research-explore/us-forest-service-history/policy-and-law/fire-u-s-forest-service/u-s-forest-service-fire-suppression/#:~:text=In%201935%2C%20the%20Forest%20Service,eliminate%20fire%20from%20the%20landscape
Geraghty, T. (2026, 5 januari). Safety-Organised criticality. Psych Safety. https://psychsafety.com/safety-organised-criticality-socy/
ICF Consulting. (2003). A coordinated attack on a power grid: economic damages and vulnerability to terrorist attacks. In Issue Paper On The Northeastern Blackout. https://www.solarstorms.org/ICFBlackout2003.pdf
Mauro, J. C., Diehl, B., Marcellin, R. F., & Vaughn, D. J. (2018). Workplace accidents and self-organized criticality. Physica A Statistical Mechanics And Its Applications, 506, 284–289. https://doi.org/10.1016/j.physa.2018.04.064
U.S. National Park Service. (2025, 22 April). 1988 Fires – Yellowstone National Park. https://www.nps.gov/yell/learn/nature/1988-fires.htm
Home | Centuria Study Association 
Leave a comment