I prefer to make up my own quotes and attribute them to very smart people, so that I can use them to win arguments — Albert Einstein
Summary: Attribution is a science that seeks to explain an observable phenomenon by something that may be responsible for it. It attempts to do so by dividing the world in two — a world of observable that is being influenced and a world that is the influencer. Attribution gets progressively challenging as the role of probability in shaping the observable increases.
Attribution refers to the process of identifying the factors responsible for an observed phenomenon. Take the example of epidemiology.
In epidemiology, attribution is the process of determining the cause(s) of a health outcome or disease. This involves identifying and quantifying the factors that contribute to the occurrence of a disease or health condition within a population. For example, attributing lung cancer (the observable phenomenon) to smoking (the cause).
In general, a single factor may not be able to explain the totality of the observable. In such cases, the notion of attributable fraction is used to measure the part of the observable phenomenon that can be attributed to various causes.
In another example, the concept of attribution is of significant interest in the context of weather and climate, where it is used to discern the causes of observed climatic variations that include societally and economically devastating extreme events. The approach of attribution in climate science relies on dissecting the totality of the Earth System into an internal component that is being influenced by the external component (the influencer).
For instance, scientists use attribution analysis to assess how much human-induced changes in the atmospheric constituents (e.g., carbon dioxide, ozone etc., treated as an external forcing, the influencer) contribute to the changes in intensity and frequency of extreme weather events like hurricanes, heatwaves, and floods (the phenomenon).
The approach for attribution involves comparing observed data with model simulations that are done with and without the inclusion of human induced changes, thereby isolating the impact of human influence.
The separation of the Earth System between internal and external is arbitrary. One can put variations in Earth’s atmosphere and ocean in the box tagged internal and changes in the atmospheric constituents in the box tagged as external. One can then treat what is in the external box as a known forcing (i.e., not worry about causality of its origin) and attempt to quantify to what extent variations in the internal box are linked to the former.
One can also put Earth’s atmosphere only in the box that is tagged internal and put variations in the ocean and atmospheric constituent in the box tagged external. In this paradigm, the exercise of attribution is an attempt to explain variations in the atmosphere when variations in the ocean and atmospheric constituents both are prescribed.
The boundary between what components in the Earth System get tagged as internal and what get tagged as external is an artificial construct and depends on the question one is posing. The question — What changes in the frequency of climate extremes is due to anthropogenic causes — slices the Earth System into influencer and influencee (which is not a word in in the English lexicon but so what) in one way. Asking the question — what changes in the frequency of climate extremes is due to anthropogenic causes during an El Niño — requires a different strategy for slicing the Earth System.
If you are not a climate scientist, you will think this discussion about climate attribution is not terribly interesting. There are, however, aspects of attribution that transcend the specifics of a single area of interest.
The necessity for the science of attribution stems from the fundamental role of randomness in how the world operates. The future of stocks is uncertain, knowing if it is going to rain two weeks from now starts to resemble flip of a coin. Life is known for throwing curve balls and hence we are advised to “Hope for the best, prepared for the worst, and unsurprised by anything in between.” Without the role of randomness, there would not be a need for the science of attribution.
The very existence of attribution science underscores the significant role of randomness. The greater the effort required to identify causes, the more pronounced the influence of randomness in the system. One could even envision the effort needed to find causality and the role of randomness as complementary variables — similar to Heisenberg’s uncertainty principle — in that their magnitudes are inversely proportional.
On that note would the world not be a boring place if randomness does not have a role in shaping our future? In the absence of randomness and with the ability to understand the causality of it all, we would have missed out constructing superstructures of religious and philosophical traditions. The world would have been a boring monochromatic place.
Humans have an innate desire for certainty and understanding, which stems from our need to predict and control our environment to enhance our success for survival and reproduction. Uncertainty can lead to anxiety, discomfort, and stress, because it challenges our ability to make informed decisions and anticipate future events. This aversion to uncertainty makes the science of attribution particularly attractive and important.
The science of attribution is also synonymous with prediction. If we can attribute an observable phenomenon to an external cause, and if that cause can be anticipated ahead of time, we can predict its consequences. For example, in epidemiology, if we can attribute an outbreak of a disease to a specific pathogen and understand the conditions that lead to its development and spread, and subsequently monitor it, we can predict future outbreaks. This allows public health officials to implement preventive measures, such as vaccinations or quarantine protocols, to mitigate the impact of the disease.
Attribution science also has significant ethical and policy implications. Attributing extreme weather events to climate change can influence public policy, disaster preparedness, societal practices, and international agreements on climate action. It can also affect legal decisions, such as liability for damages caused by climate-related events.
In the grand tapestry of our desire for understanding, attribution serves as the thread that weaves together the intricate patterns of cause and effect, the influencer and the influencee. It attempts to transform the chaos of randomness into a coherent narrative, allowing us to predict, adapt, and thrive. Whether deciphering the mysteries of climate change, unraveling the complexities of disease outbreaks, the science of attribution empowers us to see beyond the surface. It is a testament to our innate curiosity and relentless pursuit of knowledge, bridging the gap between uncertainty and insight. As we continue to refine our tools and expand our understanding, attribution will remain a cornerstone of our quest to make sense of the world and lead us to a better future.
Ciao, and thanks for reading.