- Volcanic Hazard vs. Volcanic Risk
- Warren D. Huff
- University of Cincinnati
On 10/16/01 I posted the following query to the volcano listserv:
"On the MTU web page at http://www.geo.mtu.edu/volcanoes/hazards/primer/ the following statements are made: "A volcanic hazard refers to the probability that a given area will be affected by a potentially destructive volcanic process (Fournier d'Albe, 1979). A volcanic risk is a loss (this loss can be life, property, jobs, etc.) in the hazardous area (Fournier d'Albe, 1979)." Several of my students question these definitions and suggest that risk refers to the probability of damage. Would anyone care to comment or offer clarification as to the proper use of the terms risk and hazard?"
Quite a few responses were received with what was to me a surprising range of opinions on the subject and I would like to share some of them here. Out of respect for the authors' privacy I have omitted attributions other than those for which I have sought and received permission to cite.
First, let's look at some widely used definitions:
Society For Risk Analysis (SRA)
U.S. Department of Energy (DOE)
U.S. Environmental Protection Agency (USEPA)
Two, very workable definitions that seem reasonable to me are:
- Hazard
- 1. A condition or physical situation with a potential for an undesirable consequence, such as harm to life or limb. (SRA)
- Risk
- 1. A measure of the probability that damage to life, health, property, and/or the environment will occur as a result of a given hazard. (USEPA)
-
- Tom Beer (CSIRO) suggests: Beer, T. and F. Ziolkowski [(1995) Environmental
Risk Assessment: an Australian perspective, Report 102, Supervising
Scientist, Barton, ACT] define risk as the union of a set of likelihoods
and a set
of consequences of the
scenarios under consideration,
over a given time. You might also look at two references out of the many
that address this issue: One is an essay written by Darren Gravley
(see below) and the other is Hill, B.E. et al. , 1998, 1995 eruptions
of Cerro
Negro volcano, Nicragua, and risk assessment for future eruptions:
Geological Society of America Bulletin v. 110, p. 1231-1241.
Warren, I was always taught to use the qualitative 'risk equation':
- risk = hazard * value * vulnerability
-
- risk = 'seriousness' of effects in terms of the value
- hazard -> solely related to the event (ie value independent)
- value = cost (can be economic, or in human lives, no of buildings etc)
- vulnerabilty = probabilty that value will be effected by hazard
The analogy is drawn that, while a large eruption of Olympus Mons has significant hazard, the risk is zero. Vulnerability is a function of likely activity, distance from vent etc. I'm not sure if this the current conventional wisdom, but does clear up some probelms. In terms of the website I guess hazard would mean hazard*vulnerability. I can see why the word risk implies a probability, but I don't think, in this case it does. Maybe 'stake' would be a better descriptor?
There may well be newer and more sophisticated descriptions of risks and hazards elsewhere - there is a book by Russell Blong which covers this topic well - 'Volcanic hazards' and he has a paper in the Encyclopedia of Volcanoes (pp 1215).
Prof. Huff: I suspect you'll get many individual versions of these definitions; in my line of work, I use the following hazard = any process or condition that is potentially dangerous (e.g. an earthquake, a tsunami, groundshaking, ashfall) risk = a combination of vulnerability and hazard (and many types of pseudomathematical relationships are used: multiplicative, additive, it doesn't really matter in my opinion, the point is you must have an intersection of vulnerability and hazard for there to be any risk). Probability can be a funciton of risk, if you can define the likelihood of a hazard existing at any particular point in time - tricky in most geohazard situations.
Hi Warren, I usually use definitions similar to those at the MTU site. Namely, that "hazard" refers to what the volcano has done in the past and will likely do in the future. "Risk", on the other hand, includes human factors. The example I usually use is that in the summit caldera of Mauna Loa, there is a great hazard from innundation by lava flows. But there is essentially no risk because except for a few hikers every few days, nobody is up there, nor are there any roads, cities, pipelines, etc.
I agree with your students. Those are poor. We deal with all kinds of hazards here: . all "Natural Phenomena Hazards" (the key phrase in the DOE world which is dominated by engineers) - earthquake, snow loads, wind, ash fall, floods and lightening . I spend more time with seismic hazards and a little with volcanic hazards. Any seismogenic fault or in NRC terms "Capable Fault" can produce an earthquake or any dormant volcano can produce an eruption. To put those in meaningful terms (that is something an engineer can actually use and translate into design or emergency preparedness), we have to give it a probability of happening and a risk to the area. So anything that can possibly be a hazard is but how big of hazard (or risk)? That's the probability factor plus the reality factor. Probability - For an earthquake this is the annual frequency of peak ground accelerations due to a possible earthquake. These end up being numbers like 10 to the minus 3 which translates into one chance in 1000 or 10 to the minus 6 (1 in 1 million). The engineers can then determine what this means to construction.
At DOE, they translate these into hazard classes. We have to figure the probability of an earthquake happening that produces a "certain ground motion." At XXXXX, high ground motions are less probable than low ground motions (that is, big earthquakes occur less frequently than small earthquakes as is everywhere in the world). We use our earthquake record to calculate these return periods for all sizes of earthquakes (B value). These then help determine the construction classes or "Performance Classes (PC)." For example, if the probability of the event is (10 to the minus 5) 10-5, it is PC 4 [that is a big earthquake so big design FOR a facility that can do some real environmental damage like a reactor here is the actual risk factor]; 1 x 10-4 is PC 3; 5 x 10-4 is PC 2 and 10-3 is PC 1. A reactor is built for PC-4, general use is PC-1. Here you can see the risk. You don't build an office building for a 10-5 event but you also don't build a reactor for a 10-3 earthquake. So to us hazard becomes the probability factor and the risk is more tied to reality. At XXXXX we have a volcanic hazard but only in ash. It's important because of the HEPA filters on plants and design loads on the roofs. If they get clogged during an ash fall (like St. Helen's). We don't want the roof to collapse either. But in an office building the risk to safety and health is small compared to a waste storage facility. So to me every volcano that is active or dormant poses a hazard. But then you must determine what the hazard is and how often it will occur. Mt Rainier (and other Cascade volcanoes) are high hazards because the probability of eruption is extremely high. It is more of a hazard to Seattle because of mudflows and less to XXXXX because of the ash. but it is still a hazard to us and we must design and do emergency preparedness in proportion to the real hazard.
Hi Warren: I am a Ph.D. student in volcanology at the University of Canterbury in New Zealand. I have attached an essay I wrote on the terminology of hazad, risk, and disaster. I found there is no proper definition and it is dependent on the perception of the individual or a group of individuals. Nonetheless, there are several referenced definitions in this essay and my own definitions. Enjoy! Cheers, Darren Gravley Gravley.pdf
Regarding the following statement: "A volcanic hazard refers to the probability that a given area will be affected by a potentially destructive volcanic process (Fournier d'Albe, 1979). A volcanic risk is a loss (this loss can be life, property, jobs, etc.) in the hazardous area (Fournier d'Albe, 1979)." I would venture to say that the definitions are switched. The hazard is the agent or type of loss/damage (e.g., pyroclastic flow, building destruction), and the risk is a mathematical evaluation of the occurrence of such hazard.
Dear Prof. Huff - This is a confusing topic, as there's many definitions in the literature to choose from and quite a few are not at all consistent with each other. Basically, "hazards" can be probability weighted if you want, whereas risk must include the probability of the event. Risk usually has a resulting consequence/impact, with the metric for risk (lives lost, building damage, crop loss) is usually defined at the outset of the analysis. I have used these terms to assess a general risk from tephra falls in Nicaragua. What's unusual here is that we've applied a specific forecasting model to define annual probabilities of an eruption, and stochastic-process models to calculate a range of potential hazards. Thus, we used risk as P[eruption] * [Hazard] = mm tephra/yr, and show how different confidence levels could be used for decision making. Many traditional hazard analyses use a generalized probability based solely on recurrence interval of 1-2 past events, with no process linkage to recurrence. In addition, hazards are qualitatively assessed (high, medium, low likelihood zones, etc.). But more recent analyses are using process models to derive a range of hazards, and assigning probabilities/statistical measures to these realizations. Thus, a 50% chance of lava flow inundation may be nothing more than the mean of 100 realizations for that model. But this is a conditional probability for the hazard, ultimately dependent on the annual probability of an eruption. The definitions on the MTU page are accurate, but potentially confusing. IMHO, the emphasis on hazard should be the volcanic process itself, whereas risk convolves some consequence of the hazard along with its likelihood of occurrence. My bias is that hazard does not involve the probability of the event whereas risk does, as this definition is more consistent with the usage of risk outside geology (risk = probability * hazard * consequence).
Greetings. I have had similar discussions - - at least there seems to be more agreement on the usage of "risk" and "hazard" than for prediction vs. forecast! A hazard may exist with no risk involved - - for example, a slope with a high potential for failure but presenting no risk to human activities. A risky hazard must threaten someone's person or property. Your students are right, in the sense that the hazard must have the potential to cause harm - - thus it presents a risk. Sharks are a hazard but they present no risk to non-swimmers.
Warren: Here is my opinion on the "hazard vs risk" question. I have been involved in these issues for some time, but my definition may not be accepted by all (many?) "A volcanic hazard refers to a potentially destructive volcanic process that may affect a given area² "A volcanic risk is the probability of loss (this loss can be life, property, jobs, etc.)within the hazardous area.²
Russell Blong addresses this quite well in an article he wrote in the Encyclopedia of Volcanoes. Here's the reference: Blong, R.J. (1999) Volcanic hazards and risk management. In: Encyclopedia of Volcanoes, Ed. Harald Sigurdsson, Academic Press, 1999, 1215 – 1227.
Hi Warren, As an English major, statistician and amateur volcanologist here's my take on this terminology question. The issues your students raise are worthy of consideration. The sentence from the next paragraph in the primer, "Tilling and Lipman (1993) estimate that 500 million people will be at risk from volcanic hazards by the year 2000." uses the terms "risk" and "hazard" appropriately for this context. So we could approach the terminology dispute with this usage as the gold standard. "Risk" can be an objective measurement or a subjective determination of the likelihood of an adverse outcome resulting from certain events or circumstances. "Hazard" more usually refers to a specific danger; the cause-effect relationship is that the existence of a hazard produces a risk. The degree of risk depends upon factors such as the potential imminence or destructiveness of the hazard. Thus I would describe "volcanic hazards" in terms of the specific potentially destructive processes themselves (lahar, CO2, tsunami...) and "[volcanic] hazard zones" as areas potentially endangered by such processes. "Volcanic risk" would represent an assessment of the degree of such danger, whether expressed numerically or qualitatively (color code, etc). Actually, when already discussing "volcanic hazards", simply speaking of "risk" is sufficient. The term "volcanic risk" would be useful as an umbrella term to summarize the overall degree of danger from the entire constellation of volcanic hazards in a given circumstance. Your students, of course, are keying on the buzzword "probability" which I've tried to avoid. But all things considered, I'd have to come down on their side anyway.
Hi Warren, Saw your enquiry in the IAVCEI Newsletter and thought you might be interested in a New Zealand perspective. Here, we tend to use NATURAL HAZARDS as natural events having a probability of occurrence in a given period of time. RISK is only present when something of human value is at stake. Hence RISK = HAZARD x VALUE x VULNERABILITY where vulnerability indicates the varying susceptibility of different things to destruction by the same hazard, and is strongly dependent on prediction, evacuation etc. If you want to follow this up, pages 370-371 in an article in the Journal in Geodynamics in 1985 covers this. Its entitled Volcanic Hazards of North Island, New Zealand - Overview and appears in vol. 3, pages 369 - 396.
Dear Colleague, Following your message on Volcano list about the terminology for hazard and risk, please read my opinion. According to the UNESCO practice, RISK = HAZARD X VULNERABILITY, with HAZARD referring to the physical event produced by a volcanic eruption and VULNERABILITY, including a consideration of the consequences for people, buildings, infrastructure and economic activity. Vulnerability to an eruption is a status resulting from human action or from an inherent situation such as poverty. It describes the degree to which a society is threatened by the impact of Natural Hazards.
Dear Warren, I guess I have most sympathies with your students. I have always regarded the hazard as the potential for a destructive event to occur, with no weighting given to that event - i.e. whatever it is, it is simply listed as an equal. Conversely, risk to me is an attempt to attach a probability or likelihood to that event occurring, but loaded towards a human perspective. For example, on an active (dormant) volcanic island, there may be hazards including tuff ring or maar eruptions, effusive eruptions, caldera collapse, landslide, etc. Whether or not the island is uninhabited and remote, or thinly rurally populated or close to a large city, the hazards are the same. However, the overall risks are low in the uninhabited situation, higher for the thinly rurally populated case, and high in the last case. Within each of the three situations cited, the risks should be individually weighted according a combination of those most likely to occur plus those most likely to cause serious damage. Not well phrased but I hope you get my drift.
Dear Warren, My understanding of the term volcanic hazard was confirmed by a quick visit to the Cascades Volcano Observatory website. A volcanic hazard is a natural hazard, a process or event, which may endanger life or property. The list of hazards at the CVO website include such things as ash clouds, pyroclastic flows and lahars. Risk is indeed the term that relates to probability of property loss.
Dear Warren. I just read your e-mail on VOLCANO. From a purely "health and safety" point of view the terms "hazard" and "risk" can be defined as follows. HAZARD - the danger and damage liable to objects or persons from encountering a particular substance or event - for example a volcanic eruption. RISK - the chances of this happening and affecting objects or persons. Should the likelihood of the event happening be once a week or every hundred thousand years the HAZARD will still maintain the same - time independent. However the differences in RISK between a volcano erupting once a year on a deserted island and once every century close to a major city can be appreciable - even though the type and severity of eruption (the HAZARD) are identical.
Hi, Warren. I haven't read d'Albe, but I would say the word "projected" was left out of the sentence so it should read "... risk is a projected loss in the hazardous area...."
Dear Mr Huff, For me, the term hazard has to be used when considering only the phenomenon. It is the probability of occurrence of a damaging event over a given area for a given time period. The term risk is used to express the potentiality of an area to be affected by a phenomenon, taking into account the elements at risk (values, historical buildings, population, lifelines, etc) and their vulnerability/fragility regarding the event. For our research group, risk is the result of the interaction between Hazard, Element at Risk, Vulnerability, Resilience. And space and time of occurrence are parameters that modify the dimension of Risk as all elements are evolving depending on space and time consideration. Finally, the term hazard can be used regarding the phenomenon itself and the term risk is used when one looks at the potential impact of an event on the population and its environment (social, economic, physical, etc.).
Hi Warren, I edit technical manuals for the quality in business industry. Hazard is defined as "A possible source of danger, i.e. a fire hazard; to expose to danger or harm." In other words, we already know there is a danger of loss or injury at the site. Risk is defined as "the possibility of suffering harm or loss; danger. A danger or probability of loss to an insurer. A factor, element or course involving uncertain danger; hazard." Or, you take a risk when you go into a volcanic area with known hazards. A better/best word might be casualty: "an unfortunate accident, especially one involving lost of life." "A volcanic (risk) refers to the probability that a given area will be affected by a potentially destructive volcanic process (Fournier d'Albe, 1979)". A volcanic (casualty) is a loss (this loss can be life, property, jobs, etc.) in the hazardous area (Fournier d'Albe, 1979)." So, if you just switch the words, your sentences should escape the scrutiny of your students. I hope this helps!
Hi Warren: Those MTU definitions are pretty poor -- but these terms are mis-used by most people. "Hazards" refer to hazardous things volcanoes do (eruptions, mass movements, etc) and has no relationship to people and their structures. Hazards occur all the time at remote volcanoes with no risk..... "Risk" refers to the impact of volcanic hazards on the built environment or on peoples lives (or birds or bees or anything else deemed to be "at risk"). You can attach probabilities to either: "What are the probabilities that XXX Volcano will erupt next year?" (i.e. probabilities of a "hazard" occurring) "What are the probabilities that an eruption of XXX Volcano will destroy the village of XXXXX?" (probabilities of risk for a certain area). Not good enough definitions for the textbook I'm writing - but they give an idea of correct usage?
Dear Warren, Yes I think you and your students are correct, as far as I am concerned "risk" refers to a situation where there is a chance/probability that damage or loss may occur, it does not mean that damage or loss has occurred, or will occur in the future.
Dear Colleague, As I understand the risk come from affect the hazard (as you described) with the consequences (vulnerability times cost): R=C*V*H Then through risk assessment you could evaluate the losses (human lives or material value) expected within a pre-defined return period of the hazard (you know its probability). In this way the Risk should be the most useful tool for the decision makers.
Warren, It's great that your students have the intellectual curiosity to question the definitions of "risk" and "hazard" on the MTU website. What is or is not "proper use" of such terms is in clearly in the eye of beholder. The differences in terminology between "hazard" and "risk" have been long debated, not only among geoscientists but also among social scientists, risk-analysis experts, and other specialists. There is a huge and inconsistent literature in the natural hazards, social science, and risk-analysis communities. The renowned British volcanologist George L.P. Walker, in a 1982 review paper on volcanic hazards, stated (surely tongue in cheek?): "In principle, volcanic risk could be eliminated by the total abandonment of all volcanic areas. However, this is not realistic?" Walker was simply noting the obvious that volcanic phenomena only are "hazardous" if they adversely affect people, property, and (or) the environment. Most (all?) volcanologists would agree that there is a clear distinction between "hazard" and "risk," but that is where general agreement ends. You and your students would do well to review the other definitions, contradictions, and vagaries in terminology, rather than only questioning the definitions on the MTU website. This could be a fun class project. A good place to start might be to compare/contrast the definitions in various chapters in the recently published "Encyclopedia of Volcanoes" (Sigurdsson et al., 2000) that treat volcano hazards, volcanic crises, emergency management, etc. A lot of smart people, including your students, have given thought to this topic, but there appears to be no simple answer, and certainly no consensus.
Dear Prof. Huff, I've only just caught up with your interesting discussion on the hazard - risk terminology question. It seems to me that none of your correspondents have actually identified the key fact that, at least for "risk analysis" and decision-making purposes, there are two main probabilities in play: the probability (in a given unit of time) of a hazard occurring (eruption; pyroclastic flow; earthquake, whatever) AND the probability that damage/loss/injury will result CONDITIONAL on that hazard occurring. This disaggregation into two probabilities makes it much easier to calculate the overall risk probability, and answers the point made by many in your discussion that a hazardous event can occur without necessarily having any risk associated with it (in human/societal terms). Of course, for decision-making there is often a requirement to be further sub-divisions of these conditional probabilities in terms of different time- and space scales. Now, Michael Fournier d'Albe (who also co-authored the UNESCO booklet) implicitly recognised this probability duality by defining a "hazard" as the probability of occurrence of the type of (natural) event in question, i.e. he lumped probability in together with the definition of the nature of the event. The question is thus really one of semantics. However, Fournier d'Albe's convention appears to be at some odds with general linguistic usage. It could be resolved, I suggest, by using the term "hazard" only to define the nature of the event, and then use "threat" to denote the probability of its occurrence (weather forecasters talk of the "threat of rain", implying a high probability in the next 24 hours, for instance). So, in volcanology, we could say that dome collapse pyroclastic flows are a HAZARD around the flanks of a particular volcano, and the pyroclastic flow THREAT at point X is the probability that one will reach there in the next N hours/days. Then, one can compute the risk of the conditional probability of casualties, damage etc. by inserting the threat element into the equation, rather than the hazard. Thank you for stimulating this interesting debate: it is actually very important in the context of promoting public understanding during a volcanic crisis, and your students' initiative has the great merit of making everyone think harder about clarifying the issues. Best regards, Willy Aspinall, willy@aspinall.demon.co.uk.
NB 12/10/01: The MTU webpage is scheduled to be revised to read: A volcanic hazard refers to any potentially dangerous volcanic process (e.g. lava flows, pyroclastic flows, unstable dome growth). A volcanic risk is any potential loss or damage that might be incurred by people, property, businesses, etc. as a result of the volcanic hazard. Risk not only includes the potential monetary and human losses, but also includes vulnerability.
Dear Prof Huff, Further to my recent email to you on the hazard-risk question, I had intended to mention also that in relation to this issue you might consider recommending to your students the book "The Mathematics of Natural Catastrophes" by Gordon Woo (Imperial College Press, 1999). Woo gives what I think is a unique, comprehensive, yet concise, précis of the contemporary state-of-the-art in nearly every aspect of hazard and risk assessment in relation to natural disasters. I'm sure your best students would find this book both stimulating and enlightening.
Dear Professor Huff, I read with interest your compilation of definitions of hazard and risk. Here's what I came up with for USGS Circular 1073, Living with Volcanoes: the U.S. Geological Survey's volcano hazards program. These make the fundamental distinction between hazard as a product of nature and risk as incurred only by human occupation. Both can be quantified without losing the fundamental distinction. HAZARD, RISK, AND ACCEPTABLE RISK The terms hazard and risk, often used interchangeably, are defined herein as follows: Hazard: an event or process that is potentially destructive. Risk: the magnitude of a potential loss-of life, property, or productive capacity-within the area subject to hazard(s). The threat posed by volcano hazards is a function of those natural processes and can be viewed as constant, whether or not lives and property are in jeopardy. Degree of risk, by contrast, is directly tied to the scale and value of human activity in the path of potential hazards. Options for controlling volcanic hazards are limited: most volcanic events cannot be modified by humans. However, much can be done to minimize volcanic risk by prudent land use, timely warnings, and community preparedness. Acceptable risk is that which individuals, businesses, or governments are willing to accept in return for perceived benefits. The level of acceptable risk is usually defined by local governments, taking into account information on volcanic hazards and combining it with economic, social, and political factors specific to the area threatened. In some situations risks will be judged acceptable without special precautions. In other situations local governments can take action to reduce the risk to acceptable levels, such as zoning to control population density, or restricting placement of critical facilities (for example, hospitals) in areas judged to be threatened.
Last modified 12/11/01