Probabilistic assessments of asteroid impact risks: problematic



At the Kluge Center of the Library of Congress’s recent “ScholarFest” event, a person in the audience asked astrobiologist David Grinspoon if Earth was due for an extinction-scale asteroid impact. Grinspoon replied that impacts are non-probabalistic events. Just because an extinction-scale impact event hasn’t happened lately doesn’t mean that one will happen soon.

Grinspoon is right on. Yet the community of scientists who work on finding, tracking, and characterizing near-Earth asteroids and identifying those that might pose a risk of impacting with Earth remain deeply wedded to their mathematical methods of calculating impact risks, which – not necessarily intentionally – can convey an impression of regularity where none is proven.

This commitment to probabilistic risk assessment was evident at the so-called “first international workshop on potentially hazardous asteroid assessment,” held July 7-9 at NASA’s Ames Research Center (ARC). The workshop was organized by a new ARC research group, supported by NASA’s Near Earth Object Observations Program (one of two NASA science programs that funds my work), that is studying the physics of asteroid impacts with Earth. A video recording of the workshop is archived here. This was a workshop for expert-to-expert communication. My commentary here is not intended as a critique of participants’ abilities to communicate across the expert-non-expert boundary. I’m simply raising some points that you all can “talk among yourselves” about.

The purpose of the workshop was to advance understanding of potentially hazardous asteroids (PHAs) through modeling of their atmospheric entry/breakup, risk assessments of surface impact (land and tsunami), and characterization of pre-entry properties, toward “developing reliable predictive and assessment tools enabling decision makers to take appropriate mitigation action in the event of pending PHA strike.”

Workshop participants were asked to address these questions: What are the likely physical characteristics for different classes of potentially hazardous asteroids (PHAs) and how can they be measured? What is the dependence of surface damage on PHA size and impact trajectory? What is the range of damage effects and potential for casualties from impacts (both land and tsunami)?

ARC’s David Morrison said characterizing asteroid impact risks in terms of human deaths per year is “a good way to communicate with decision makers.” (I disagree.) He also said “it is prudent to assess the nature of the threat” by means of statistical analyses. (It may be prudent to do so among experts, but such assessments are not necessarily meaningful to non-experts – such as decision makers.)

Elizabeth Pate-Cornell, Jason Reinhardt, and Matthew Daniel of Stanford University’s Engineering Risk Research Group presented interesting material at the workshop that, in my humble opinion, is strictly for discussion among fellow experts. When Steinhardt started talking about such things as “backcasting” and “recency bias,” he lost me.

Steinhardt said risk=probability + consequences. Risk communication expert Peter Sandman says risk=hazard + outrage. Steinhardt’s conception of risk is mathematical. Sandman’s is psycho-social. Steinhardt said, “We can quantify asteroid risk.” (I’m not convinced. Well, I can accept that perhaps statisticians can quantify this risk to their own satisfaction. But I don’t believe such quantifications will be widely meaningful to people who don’t live by the numbers.) Matthew Daniel commented, “uncertainty quantification can be hard to interpret.” (Amen.)

I’m not criticizing the work of this group, which is top-notch.* I am commenting on the value of probabilistic risk assessment in communicating with decision makers about “actionable” risks.

After hearing presentations on an array of approaches to modeling asteroid entry and breakup, Peter Brown of Western University (Toronto) said experts now need to compare and validate these various models. Mark Boslough of Sandia National Laboratory said he’s skeptical that any amount of data on fragmentation generated by computer models will enable accurate predictions of how actual impacting objects will break up (presumably because every impact is unique?).

My colleague Leviticus Lewis of the Federal Emergency Management Agency sent a message to workshop participants when he said, “There will be no formulas in my presentation.” (Most people got the message, I think, based on questions and comments afterward.) He explained how the U.S. national incident management system works, and he explained what disaster planners and emergency responders will need to know about predicted impacts. (Al Harris later commented that Lewis had “told scientists what they need to know.”)

When the experts, and the media, talk about asteroid impact risks, they tend to speculate about “what if” an impact occurs over New York, or Los Angeles, or maybe Paris (in the U.S., at least, they don’t tend to speculate about the world’s most populous cities, such as Delhi, Mexico City, or Sao Paulo….). Somebody at the workshop wondered, “What if something explodes over the Vatican?” I often wonder why – if asteroid impacts with Earth are randomly distributed (see the bolide-impact map released last year) – the experts tend to focus on big cities when most of the world’s surface is covered by water or unpopulated/sparsely populated land. Maybe it’s due to concerns about injuries and fatalities? Then again, I’m not a mathematician, so maybe I’m missing something.

I’m not convinced of the existence of widespread public concern over when the next catastrophic asteroid impact with Earth will occur. I am convinced of the media’s attention to this subject. Impending doom is a popular topic, in fiction and nonfiction media. (If you want to worry about impending doom, read Kathryn Schultz’s story, “The really big one,” in the July 20 issue of The New Yorker. It’s not about asteroid impacts.)

As I told a reporter recently, I don’t worry about the next asteroid impact with Earth. I worry about my neighbor’s tilting hemlock tree falling on my house during the next derecho.

* See Jason C. Steinhardt, Matthew Daniel, and M. Elizabeth Pate-Cornell, “Probabilistic Analysis of Asteroid Impact Risk Mitigation Programs,” Probabilistic Safety Assessment and Management (PSAM) 12, June 2014, Honolulu, Hawaii.


One Response to “Probabilistic assessments of asteroid impact risks: problematic”

  1. J.L. Galache (@JLGalache) Says:

    Always a pleasure to read your pieces, Linda. Wish I had been able to attend this meeting!

    You and I have discussed these very issues on several occasions, and I’m sad to see we still find ourselves at an impasse. Maths has never been the language to use when communicating with the lay person, but there are no common metaphors that we can use, because no other natural disaster compares to asteroid strikes.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: