Reflections at the Intersection of Science, the Media, and the Public
Department of Earth and Environmental Sciences
Like many seismologists, I spend a great deal of my time at the intersection of science, the media, and public awareness. A major challenge that seismologists (and other scientists) face is how to make our science accessible to the public without misleading them with unwarranted claims of certainty. When scientists make claims of certainty, which are then reported (or misreported) by the media, it may in the short term feed the public’s hunger for certainty in an uncertain world, but in the longer run it undermines the scientific enterprise and does a disservice to the public.
Recent news reports about a supposed discovery of geologically mapped faults that are responsible for earthquakes in the New York City area (and the alarming “news” that one of those faults runs close to the Indian Point nuclear power plant) have given me the opportunity to muse once again on how the reports of seismological studies might be heard and interpreted by the general public.
First, the bottom line: Yes, there is an earthquake hazard in the greater New York City area (albeit less than in California), a hazard that we would be foolish to ignore. But if one claims (without scientific justification) that earthquakes are concentrated on a particular fault and that therefore future large earthquakes will occur along that same fault, then one is also claiming (without scientific justification) that the earthquake hazard is less in other nearby areas. This is a technical result of the way that earthquake hazard probabilities are determined: There is a certain number of earthquakes that are likely to occur in a given region, and if many of them are forecasted to occur in one part of the region, then there will necessarily be fewer forecasted to occur in some other part. The resulting hazard assessment might then be focused on the wrong location, which would not help the public to make informed decisions about issues like what should or should not be done to insure the seismic safety of nuclear power plants.
So what is the “news”? The New York Times, the ScienceDaily website, and other news media recently reported on the results of a Columbia University study of earthquakes in the Greater New York City area. This study by Sykes et al. (2008) presents a case for hypothesized relationships between mapped faults and earthquakes in this highly populated and densely urbanized region.
The New York Times headline: Study Maps Faults for New York Quakes. The ScienceDaily headline: Earthquakes May Endanger New York More Than Thought; Nuclear Power Plant Seen As Particular Risk.
According to the New York Times article, the Columbia University group “mapped out a family of faults responsible for most of the earthquakes.” The Times article also reported that the group “found a previously unidentified boundary, likely a fault, that runs 25 miles to Peekskill, N.Y., from Stamford, Conn., passing within a mile of Indian Point.”
According to the ScienceDaily report the study “found concrete evidence for one significant previously unknown structure: an active seismic zone running at least 25 miles from Stamford, Conn., to the Hudson Valley town of Peekskill, N.Y., where it passes less than a mile north of the Indian Point nuclear power plant.” ScienceDaily also reported that the “Ramapo Seismic Zone” is a “previously known feature” that “runs from eastern Pennsylvania to the mid-Hudson Valley, passing within a mile or two northwest of Indian Point.”
Most readers of these articles would be led to believe that there has been some great new discovery regarding the activity of geologically mapped faults that threaten specific locations in the New York City area. But for me – having spent several decades immersed in the study of seismicity in the Northeast – what is most striking is how little has changed about our knowledge on this issue compared to what was known in the mid 1980s! If there is a “news” story here at all, it is the startling revelation that, after a very thorough analysis of two additional decades of earthquake monitoring, we are left with just as many questions and unresolved issues about faults and earthquakes in the New York City area as we had in the ’80s.
Back in 1978, Aggarwal and Sykes concluded that “seismic activity in the greater New York City area is concentrated along several northeast-trending faults of which the Ramapo fault appears to be the most active,” and they emphasized the importance of how that conclusion needed to be considered in discussions of the seismic safety of the Indian Point nuclear power plant. However, based on our analysis of the data available through 1983, my colleagues and I concluded (Kafka et. al, 1985) that the network seismicity “does not necessarily indicate whether [the Ramapo] fault is either the most active fault in the greater New York City area or if it is the fault along which the largest earthquakes will occur in the future,” and that “the geologic structures associated with most (if not all) earthquakes in this region are still unknown.”
As we demonstrated in 1985, the existence of a “Ramapo Seismic Zone” (RSZ) is very difficult to disentangle from the fact that many of the seismic stations are located in the vicinity of the hypothesized seismic zone, which results in a seismicity map that is biased towards highlighting that zone. If the seismic data for the New York City area is reanalyzed to minimize this bias, the existence of the proposed seismic zone is not so clear, and the distribution of epicenters lends itself to many possible conjectures of hypothetical fault zones, all of which are based on circumstantial evidence. None of these hypotheses can be considered as “concrete evidence” that the site of the Indian Point nuclear power plant is necessarily any more seismically active than many other sites in the study area. (Click here for a more detailed explanation of the complexities of demonstrating the existence of a “seismic zone” when many seismic stations are located in the vicinity of that hypothesized seismic zone.)
Conveying a scientific result as more certain than it really is feeds the appetite of a public hungry for certainty in an uncertain world. By highlighting the discussion of the RSZ (or the newly hypothesized seismic zone between Peekskill and Stamford) as if it was a “discovery” (as opposed to an interesting, but still speculative hypothesis), the impression is given once again that our understanding of the relationship between faults and earthquakes in this region is clear when in fact it isn’t.
Does any of this matter? Is this just an academic discussion? In 2000, I wrote an opinion piece in Seismological Research Letters, entitled Public Misconceptions about Faults and Earthquakes in the Eastern United States: Is it Our Own Fault?, explaining why I don’t think this is just an academic matter.
Based on the availability of new and revised earthquake data resulting from the Columbia University study, my opinion hasn’t changed from what I wrote in 2000. Yes, it does go beyond being just an academic matter when reports are made of discoveries of correlations between faults and earthquakes, when in reality what is found is at best anecdotal evidence.
In this particular case, it does not help the public to make informed decisions about the seismic safety of nuclear power plants when we are not careful to distinguish between our interesting hypotheses supported by anecdotal evidence versus scientifically-tested hypotheses that are well supported by data.