The Risk of Coal Production

As mentioned on the Risk page, the death numbers for coal seem far too high. In Table 1 of the paper by Kharecha and Hansen, 2013 [Archive pdf 725kb] state death rates of the order of 30 per TWhr for coal and that the use of nuclear power instead of coal power has saved 1.87 million lives. While it is correct that nuclear has proven to be safer than coal, the coal death rate seems unbelievably high. Kharecha and Hansen get the coal death numbers from their reference 16 (Markandya and Wilkinson, 2007) [Archive pdf 480kb]. That paper says that the source of the death rates is the ExternE project done some years ago. The ExternE report [Archive pdf 123kb] says
"Systematic statistical studies also indicate that there is no threshold effect for the common air pollutants from combustion of fossil fuels or biofuels."
I wanted to dig further to find out where THAT info came from. In Chapter 3 [Archive pdf 3.6Mb] of the ExternE coal publication I found this buried on page 103:
"There are now numerous studies which associate particulate air pollution with a wide range of acute health effects. These associations have been found at normal background levels in a wide range of locations. In the present study we have assumed that there is no threshold level of pollution for these effects."
Note - they assumed LNT. Reading though the pages, there seems to be no end to hedged statements and provisos. For instance a flat background death rate is assumed across the board, independent of region. They reference Schwartz and others who found increased mortality for people with lung ailments when pollution went up - which is not surprising. One review paper by Pope and Dockery, 2006, [Archive pdf 2.3Mb] states that acute (a few days or so) effects are insignificant (contradicting the quote above which states that acute effects were seen) - ie brief exposure to particulate matter has little negative effect. But chronic (long term) effects were noticed for people living next to busy streets and roadways. I can't help thinking though that income levels have a marked effect on where you live and on lifestyle in general; so data about people living next to a busy street is problematic for a number of reasons not related to pollution. The authors also admit to the usual problems associated with statistically extracting low level effects against a background of compounding effects of deaths from all other sources. It is not hard to imagine some negative effects due to pollution but extrapolating the delta in the progression of disease is error prone to say the least. Assuming a rate of 30 deaths per TWhr means that a 1 GW coal fired station leads to more than 250 excess deaths per year. For each station, year after year. Hard to believe.

You see how this goes? The data shows increased deaths in locations with more pollution, correcting as best they can for other effects - a process that is tenuous (and they admit it). The assumption of LNT is subsequently made which gets turned into an indication which gets turned into a fact. Tell me again about the robustness of peer review. This is bad science.

Apparently I am not alone in doubting these excessive claims. See this article at The Record.com. [Archive pdf 79kb] and Hormesis for Fine Particulate Matter (PM 2.5) [Archive pdf 1.5Mb]

Just to be clear, I am not defending coal. And coal does appear to kill significantly more than other energy sources. I am just against what appears to me to be a self-serving hyping of honest (but tenuous) original studies. I also acknowledge that coal power is relatively cheap and that access to relatively cheap power leads to a higher standard of living and death reduction. So one cannot dismiss coal (or any power source) out of hand.

End note: Estimates of deaths due to air pollution are statistical deaths; you won't see 'air pollution' as the cause of death listed on a death certificate. The estimates are more along the lines of the shortening of one's life particularly for those whose health is compromised via breathing ailments. See does-air-pollution-kill-40000-people-each-year-uk [Archive pdf 1.4Mb] for a good discussion of this, including the large error bars in the estimates. See also air-pollution-cause-40000-deaths-every-year-fact-check [Archive pdf 3.5Mb] where the point is made that it is 'days lost, not deaths per se.