Conor Clarke at The Atlantic blog, raised several issues with my study, “What to Do About Climate Change,” which Cato published last year.
One of Conor Clarke’s comments was that my analysis did not extend beyond the 21st century. He found this problematic because, as Conor put it, climate change would extend beyond 2100, and even if GDP is higher in 2100 with unfettered global warming than without, it’s not obvious that this GDP would continue to be higher “in the year 2200 or 2300 or 3758”. I addressed this portion of his argument in Part I of my response. Here I will address the second part of this argument, that “the possibility of ‘catastrophic’ climate change events — those with low probability but extremely high cost — becomes real after 2100.”
The examples of potentially catastrophic events that could be caused by anthropogenic greenhouse gas induced global warming (AGW) that have been offered to date (e.g., melting of the Greenland or West Antarctic Ice Sheets, or the shutdown of the thermohaline circulation) contain a few drops of plausibility submerged in oceans of speculation. There are no scientifically justified estimates of the probability of their occurrence by any given date. Nor are there scientifically justified estimates of the magnitude of damages such events might cause, not just in biophysical terms but also in socioeconomic terms. Therefore, to call these events “low probability” — as Mr. Clarke does — is a misnomer. They are more appropriately termed as plausible but highly speculative events.
Consider, for example, the potential collapse of the Greenland Ice Sheet (GIS). According to the IPCC’s WG I Summary for Policy Makers (p. 17), “If a negative surface mass balance were sustained for millennia, that would lead to virtually complete elimination of the Greenland Ice Sheet and a resulting contribution to sea level rise of about 7 m” (emphasis added). Presumably the same applies to the West Antarctic Ice Sheet.
But what is the probability that a negative surface mass balance can, in fact, be sustained for millennia, particularly after considering the amount of fossil fuels that can be economically extracted and the likelihood that other energy sources will not displace fossil fuels in the interim? [Remember we are told that peak oil is nigh, that renewables are almost competitive with fossil fuels, and that wind, solar and biofuels will soon pay for themselves.]
Second, for an event to be classified as a catastrophe, it should occur relatively quickly precluding efforts by man or nature to adapt or otherwise deal with it. But if it occurs over millennia, as the IPCC says, or even centuries, that gives humanity ample time to adjust, albeit at a socioeconomic cost. But it need not be prohibitively dangerous to life, limb or property if: (1) the total amount of sea level rise (SLR) and, perhaps more importantly, the rate of SLR can be predicted with some confidence, as seems likely in the next few decades considering the resources being expended on such research; (2) the rate of SLR is slow relative to how fast populations can strengthen coastal defenses and/or relocate; and (3) there are no insurmountable barriers to migration.
This would be true even had the so-called “tipping point” already been passed and ultimate disintegration of the ice sheet was inevitable, so long as it takes millennia for the disintegration to be realized. In other words, the issue isn’t just whether the tipping point is reached, rather it is how long does it actually take to tip over. Take, for example, if a hand grenade is tossed into a crowded room. Whether this results in tragedy — and the magnitude of that tragedy — depends upon how much time it takes for the grenade to go off, the reaction time of the occupants, and their ability to respond.
Lowe, et al. (2006, p. 32-33), based on a “pessimistic, but plausible, scenario in which atmospheric carbon dioxide concentrations were stabilised at four times pre-industrial levels,” estimated that a collapse of the Greenland Ice Sheet would over the next 1,000 years raise sea level by 2.3 meters (with a peak rate of 0.5 cm/yr). If one were to arbitrarily double that to account for potential melting of the West Antarctic Ice Sheet, that means a SLR of ~5 meters in 1,000 years with a peak rate (assuming the peaks coincide) of 1 meter per century.
Such a rise would not be unprecedented. Sea level has risen 120 meters in the past 18,000 years — an average of 0.67 meters/century — and as much as 4 meters/century during meltwater pulse 1A episode 14,600 years ago (Weaver et al. 2003; subscription required). Neither humanity nor, from the perspective of millennial time scales (per the above quote from the IPCC), the rest of nature seem the worse for it. Coral reefs for example, evolved and their compositions changed over millennia as new reefs grew while older ones were submerged in deeper water (e.g., Cabioch et al. 2008). So while there have been ecological changes, it is unknown whether the changes were for better or worse. For a melting of the GIS (or WAIS) to qualify as a catastrophe, one has to show, rather than assume, that the ecological consequences would, in fact, be for the worse.
Human beings can certainly cope with sea level rise of such magnitudes if they have centuries or millennia to do so. In fact, if necessary they could probably get out of the way in a matter of decades, if not years.
Can a relocation of such a magnitude be accomplished?
Consider that the global population increased from 2.5 billion in 1950 to 6.8 billion this year. Among other things, this meant creating the infrastructure for an extra 4.3 billion people in the intervening 59 years (as well as improving the infrastructure for the 2.5 billion counted in the baseline, many of whom barely had any infrastructure whatsoever in 1950). These improvements occurred at a time when everyone was significantly poorer. (Global per capita income today is more than 3.5 times greater today than it was in 1950). Therefore, while relocation will be costly, in theory, tomorrow’s much wealthier world ought to be able to relocate billions of people to higher ground over the next few centuries, if need be. In fact, once a decision is made to relocate, the cost differential of relocating, say, 10 meters higher rather than a meter higher is probably marginal. It should also be noted that over millennia the world’s infrastructure will have to be renewed or replaced dozens of times – and the world will be better for it. [For example, the ancient city of Troy, once on the coast but now a few kilometers inland, was built and rebuilt at least 9 times in 3 millennia.]
Also, so long as we are concerned about potential geological catastrophes whose probability of occurrence and impacts have yet to be scientifically estimated, we should also consider equally low or higher probability events that might negate their impacts. Specifically, it is quite possible — in fact probable — that somewhere between now and 2100 or 2200, technologies will become available that will deal with climate change much more economically than currently available technologies for reducing GHG emissions. Such technologies may include ocean fertilization, carbon sequestration, geo-engineering options (e.g., deploying mirrors in space) or more efficient solar or photovoltaic technologies. Similarly, there is a finite, non-zero probability that new and improved adaptation technologies will become available that will substantially reduce the net adverse impacts of climate change.
The historical record shows that this has occurred over the past century for virtually every climate-sensitive sector that has been studied. For example, from 1900-1970, U.S. death rates due to various climate-sensitive water-related diseases — dysentery, typhoid, paratyphoid, other gastrointestinal disease, and malaria —declined by 99.6 to 100.0 percent. Similarly, poor agricultural productivity exacerbated by drought contributed to famines in India and China off and on through the 19th and 20th centuries killing millions of people, but such famines haven’t recurred since the 1970s despite any climate change and the fact that populations are several-fold higher today. And by the early 2000s, deaths and death rates due to extreme weather events had dropped worldwide by over 95% of their earlier 20th century peaks (Goklany 2006).
With respect to another global warming bogeyman — the shutdown of the thermohaline circulation (AKA the meridional overturning circulation), the basis for the deep freeze depicted in the movie, The Day After Tomorrow — the IPCC WG I SPM notes (p. 16), “Based on current model simulations, it is very likely that the meridional overturning circulation (MOC) of the Atlantic Ocean will slow down during the 21st century. The multi-model average reduction by 2100 is 25% (range from zero to about 50%) for SRES emission scenario A1B. Temperatures in the Atlantic region are projected to increase despite such changes due to the much larger warming associated with projected increases in greenhouse gases. It is very unlikely that the MOC will undergo a large abrupt transition during the 21st century. Longer-term changes in the MOC cannot be assessed with confidence.”
Not much has changed since then. A shut down of the MOC doesn’t look any more likely now than it did then. See here, here, and here (pp. 316-317).
If one wants to develop rational policies to address speculative catastrophic events that could conceivably occur over the next few centuries or millennia, as a start one should consider the universe of potential catastrophes and then develop criteria as to which should be addressed and which not. Rational analysis must necessarily be based on systematic analysis, and not on cherry picking one’s favorite catastrophes.
Just as one may speculate on global warming induced catastrophes, one may just as plausibly also speculate on catastrophes that may result absent global warming. Consider, for example, the possibility that absent global warming, the Little Ice Age might return. The consequences of another ice age, Little or not, could range from the severely negative to the positive (if that would buffer the negative consequences of warming). That such a recurrence is not unlikely is evident from the fact that the earth entered and, only a century and a half ago, retreated from a Little Ice Age, and that history may indeed repeat itself over centuries or millennia.
Yet another catastrophe that greenhouse gas controls may cause is that CO2 not only contributes to warming, it is also the key building block of life as we know it. All vegetation is created by the photosynthesis of CO2 in the atmosphere. In fact, according to the IPCC WG I report (2007, p. 106), net primary productivity of the global biosphere has increased in recent decades, partly due to greater warming, higher CO2 concentrations and nitrogen deposition. Thus , there is a finite probability that reducing CO2 emissions would, therefore, reduce the net primary productivity of the terrestrial biosphere with potentially severe negative consequences for the amount and diversity of wildlife that it could support, as well as agricultural and forest productivity with adverse knock on effects on hunger and health.
There is also a finite probability that costs of GHG reductions could reduce economic growth worldwide. Even if only industrialized countries sign up for emission reductions, the negative consequences could show up in developing countries because they derive a substantial share of their income from aid, trade, tourism, and remittances from the rest of the world. See, for example, Tol (2005), which examines this possibility, although the extent to which that study fully considered these factors (i.e., aid, trade, tourism, and remittances) is unclear.
Finally, one of the problems with the argument that society should address low probability high impact events (assuming a probability could be estimated rather than assumed or guessed) is that it necessarily means there is a high probability that resources expended on addressing such catastrophic events will have been squandered. This wouldn’t be a problem but for the fact that there are opportunity costs associated with this.
According to the 2007 IPCC Science Assessment’s Summary for Policy Makers (p. 10), “Most of the observed increase in global average temperatures since the mid-20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations.” In plain language, this means that the IPCC believes there is at least a 90% likelihood that anthropogenic greenhouse gas emissions (AGHG) are responsible for 50-100% of the global warming since 1950. In other words, there is an up to 10% chance that anthropogenic GHGs are not responsible for most of that warming.
This means there is an up to 10% chance that resources expended in limiting climate change would have been squandered. Since any effort to significantly reduce climate change will cost trillions of dollars (see Nordhaus 2008, p. 82), that would be an unqualified disaster, particularly since those very resources could be devoted to reducing urgent problems humanity faces here and now (e.g., hunger, malaria, safer water and sanitation) — problems we know exist for sure unlike the bogeymen that we can’t be certain about.
Spending money on speculative, even if plausible, catastrophes instead of problems we know exist for sure is like a starving man giving up a fat juicy bird in hand while hoping that we’ll catch several other birds sometime in the next few centuries even though we know those birds don’t exist today and may never exist in the future.