02 March, 2007

It'll Be a Cold Day in H*ll

Two weeks ago Dennis Prager noted that the ratio of cold-related to heat-related deaths in the UK in any given year is 10:1, or 20,000 versus 2,000. I haven't been able to track down Mr. Prager's original source, but he piqued my curiosity. If it's true in Britain, might it be true elsewhere? It appears that it is, as this heavily footnoted NIH document makes plain [emphases mine].

Mortality in Britain is lowest when the mean daily temperature is 17-18°C. The number of heat related deaths per year, obtained as the number of excess deaths on days hotter than this, has averaged around 800 in recent years. Most of those deaths are of people over 70 years of age, and most occur in the first day or two of a period of high temperature...

According to some predictions heat related mortality will increase drastically as global warming develops, but recent evidence is relatively reassuring. Heat related mortality is similar in hot and cold parts of western Europe and in hot and cold parts of the United States. This implies that the populations of hot regions have adjusted by physiological or other means to their hotter summers. In Britain annual heat related deaths are in any case far fewer than cold related deaths, so that the initial effect of increased temperatures all year round, [even] before such adjustment, would be to reduce net annual mortality.
Based on this, Prager's ratio may even have been conservative. The estimate for annual cold-related deaths in the UK is 20,000... versus 800 for heat. Fifteen to one. And that excludes deaths due to influenza and other seasonal illnesses that--despite what all of our mothers told us--aren't statistically correlated with climatic conditions.

Prager is at the very least stating the consensus of the medical community: all other things being equal, extremes of cold (deep, sudden and/or prolonged) kill more people than extremes of heat. It is true both short and long term, e.g., when we look at the expansion or retreat of human populations over centuries and millennia. (In those longer-term cases food scarcity issues come into play in addition to the increases in storminess and general weather volatility associated with colder climatic periods). The NIH piece continues [emphases mine]
Analysis of actual changes in heat related mortality during global warming since 1971 is even more reassuring. Despite rises in mean summer temperatures of at least 1°C in southeast England and North Carolina heat related mortality has not risen in southeast England and has virtually disappeared in North Carolina. The latter represents something more than adaptation as it could prevent the mortality rising with higher temperatures but could not make it fall. The likely explanation is the increase in air conditioning (from 57% to 72% between 1978 and 1997) that has occurred in households in that region of the United States and is in turn attributable to increasing prosperity.
Those are increases in prosperity that Gore's $553 trillion in proposed spending on global warming would bring to a halt and probably roll back. Unfortunately, after laying out these facts, the article goes off into non-sequitor never-never land saying, in essence, that just because there isn't any evidence of a problem (and in fact is strong evidence against it being a problem) doesn't mean we ought not to do something about it.

Hunh?

It's hard for me to figure out what isn't obvious about this. If deaths from cause X are ten times those from cause Y and we reduce deaths from X by 10% while increasing deaths from Y by 10% there are fewer deaths overall. A lot fewer. In fact, if one follows this simple algebra, then a general rise in average temperature would be expected to reduce overall deaths from temperature extremes of any kind whereas a general decrease in average temperature would increase deaths.

Sure the curves may be non-linear as we get to extremes, but even the pessimists agree: we're not going there. The earth has been hotter. Under those conditions, humans thrived and expanded. Sure there are probably local, temporary exceptions. But it's hard to overcome a 15:1 ratio with anecdotal evidence. (Doncha remember that day last August when uncle Bert had a coronary trying to put in his window air conditioner?)

A scan of the literature, even the global-warming-friendly stuff, seems to indicate that at worst, death rates from heat are only slightly less than those from cold. The case is still clear.

Let's make this more even concrete. You are Al Gore. (Sorry. This will be quick.) You're standing in a big lecture hall, filled with 1,000 people. Let's make it even more fun and pretend that it's the Oscars--all those anorexic actresses shivering in skimpy dresses make it even more realistic. In front of you on the podium is a thermostat. You know that if you turn it up by (say) five degrees, one person will die. If you decide to turn it down by an equal amount however, 10-15 people will die. What do you do? Well, if you're Al Gore, it's obvious: you turn it down. Why? Because half the audience is cheering wildly for you to do so.

The whole scenario is reminiscent of those psychology of authority experiments they did in the 1950s where subjects went ahead and 'electrocuted' an actor subject in the other room on the authority of a scientist telling them it would be OK.

This is a variant of the inability of policy-makers and the general public to make rational trade-offs based on actual morbidity and mortality statistics in unrelated categories. I.e., in a rational world where the goal was to prevent death, much of the sturm and drang over gun control would be directed at swimming pools. (They kill far more people each year than do guns). In this case it ought to be easier because average warming means less cooling. (My young niece got this one well before she went to pre-school.)

But we do not live in a rational world. When a scientific hypothesis is believed to be true based on the number of Oscars it garners, we have up and moved to Wonderland. Off with her head!

I ask again: What is the moral case for halting a trend likely to reduce death and famine?