Geophysical Hazards

Yesterday I started reading the Springer publication Geophysical Hazards: Minimizing Risk, Maximizing Awareness. I didn’t get very far because the book is really bad, and I’ve now decided I’m not going to finish this book – if you want to know the short version of what I thought about the book, read my review on goodreads. In this post I want to include a few observations I made along the way while reading this book, in order to at least somehow justify having read the parts of it I did; there was a little bit of interesting stuff, but you had to go through a lot of crap to get to it and it’s just not worth it.

I’ve decided to mostly just talk about the book in this post, rather than quote from it, however I do want to start out with a quote from the book:

“the Budapest Manifesto on Risk Science and Sustainability[2] provided a generic framework suitable for environmental risk management across a variety of disciplines, including hazards studies. […] It can be summarised by the following list of items that need to be examined:

• Consultation
• Concerns
• Consequences
• Calculations
• Certainties, uncertainties, probabilities
• Comparing against pre-determined criteria
• Control, mitigate and adapt
• Communicate
• Monitor
• Review.

There are 10 items listed above. Risk management praxis consists in undertaking steps 2–8 in sequential order, and then re-checking to make sure that any residual risk has been dealt with adequately. […] certain terms have not been incorporated into the Budapest Manifesto framework. In particular, the idea that one needs to determine the context within which a risk management activity takes place. The first step in risk management is always to establish the context in which to operate.”

…

Okay, on to my own observations. The first one is that mobile phones are actually really neat things which in theory could be/become excellent tools to use in a disaster setting – for example a government agency could notify/warn people of a disaster in progress via a text message (according to the book Finland decided that this was the way to go already in 2005), or, say, victims trapped under debris in a disaster setting might be identified/found through their phone signals. One major issue is that most phone networks have little excess capacity, and widespread disasters are therefore likely to interrupt such lines of communication completely (lots of people calling their loved ones all at the same time, or calling relevant local governmental agencies), making mobile phones much better suited for small-scale disasters than for large-scale disasters. The ‘phones may not be all that helpful in a big disaster setting’-point I have encountered before, e.g. here and here, but I hadn’t really thought about the fact that these sorts of dynamics don’t really play any role in small-scale disasters, where these tools may actually be immensely useful.

There’s a lot of geographic variation in the pattern of environmental hazard losses, which is a result of the interplay of a large number of variables. It turns out that we don’t know nearly as much about those variables as I’d have thought. Anyway, in some areas floods are most relevant, in others seismic activity is important, in some areas droughts are key. Half of the human fatalities from loss-causing geophysical hazards affecting the United States during the period 1960-2007 according to SHELDUS (see link below) were losses due to ‘severe weather’ (33.2%) and ‘winter weather’ (15.5%); most of the remaining fatalities were due to ‘tornados’ (14.5%), ‘flooding’ (12.2%) and ‘heat and droughts’ (15.8%) – the website is here, if you’re interested in having a closer look at this stuff. My own opinion would be that when a tornado forms it’s accurate to say that the weather is acting up a bit (if that’s not ‘severe weather’ then…) – but of course this is all just a matter of coding and figuring out how best to categorize these things. Although the above might indicate that we have a lot of data on this type of stuff we really don’t, and there are huge data problems in this area which make deciding upon proper courses of action, in terms of the proper risk reduction strategies to engage in, very difficult. Of course data problems tend to be most severe in the areas where fatalities are most likely to be high, because the populations at risk those places are particularly vulnerable. It’s important to note that missing data and vulnerability variables are not independent as such, as missing decision-relevant data may be conceptualized as factors impacting the vulnerability of the people affected by the disaster in an undesirable manner; if rescue workers have poor information, matching aid supplies to needs may be difficult, which can cost lives. Getting the distribution of disaster victims wrong can lead to misallocation of resources, and misallocation of resources often costs lives in disaster settings (“It is not uncommon for relief teams to be deployed to a disaster area without full knowledge of how many people will need aid or where they’re located relative to the impact area, let alone have information on age and gender – characteristics that are vital in the delivery of food, water, shelter, and other assistance needs.”). Simply knowing where people live is in some contexts difficult; a great majority of people on Earth have been ‘accounted for’ in a relatively recent census (“More than 85% of the world’s population has been enumerated within a national census since 2000 (NRC 2007).”), but people move around, have children and die all the time, and some countries are much better at tracking that kind of stuff than are others (which is both a good thing and a bad thing – in this context keeping track of people is a good thing). One fundamental problem, in terms of optimizing potential relief efforts, is that the more fine-grained information you have, the less likely your data is to be up-to-date; there’s a tradeoff between how often you can gather the relevant information from people and how much information you gather from each person, as obtaining each bit of information takes time and money.

Geographical variation in loss profiles is partly due to different physical processes playing a role in different places, but many other variables play a role in mediating how severe the consequences of a given event may be to a given population of people, and how likely a hazard is to become a disaster. Having good data on stuff like gender and age isn’t just helpful because it makes relief efforts easier – it’s also helpful because not all people are equally vulnerable, and variables such as these actually have an impact on the pre-disaster risk assesssment as well. Population density (if more people live in an area, the likelihood of someone (or many someones) dying from a given cause goes up – megacities are particularly vulnearable), age profile (old and infirm people as well as children may be more likely to die in disaster contexts, because of the limited mobility – which translates into a limited ability to get out of harm’s way – and care-taking requirements of these individuals), socioeconomic factors (rich people are generally better able to handle/absorb the consequences of a disaster; this is both the case at the community level and at the international level – disasters in poor countries tend to cause more fatalities and lower economic losses than disasters in rich countries. However they note in the book that it seems that the relative economic losses (disaster losses compared to national GDP) are higher in poor countries than in rich countries), ‘disaster preparedness’, infrastructure, etc. all play important roles, but little seems to be known about precisely how they mediate risks in different contexts. As one author puts it, “At present, most of the vulnerability metrics are descriptive, not predictive indices and are mainly used as a representation of multi-dimensional phenomena, such as those pre-existing conditions in communities that make them susceptible to harm.” It’s incidentally an assumption taken as a given throughout the part of the book I read that governments have important roles to play in assessing such risks and trying to counter them in various ways, although at no point do the authors even attempt to justify this assumption – one of many reasons why I dislike the book. From a political economy point of view I think it’s safe to assume that justifying public involvement in handling/preventing/… large-scale disasters which are hard to insure against is conceptually much easier than is justifying government efforts aimed at saving 10 people from dying of heat-stroke during a warm summer, and when your disaster data includes data of the latter type, cost-effectiveness considerations are bound to come up during the discussion. At least if you ask people who aren’t a member of some four-letter UN organization.