A 21st century greenhouse gas budget?

How much greenhouse gas can the world safely emit? Is this a useful question to ask?

One working assumption for some (footnote 1) is that a global average temperature rise of more than 2 Centigrade during the 21st century may be “dangerous” – i.e. have severe adverse impacts on human development and the ecosystems on which it depends. (It is noted, however, that even if the rise is less than 2C there may still be adverse impacts.)

It is uncertain how much greenhouse gas may be emitted consistent with a politically acceptable risk of a temperature rise of no more than 2C. One line of reasoning might appear to suggest stabilisation of atmospheric concentrations at no more than 450 ppmvCO2e (or is it 475?) by 2030 to 2050 followed by a reduction, fits the bill (2).

First question: to what extent is this line of reasoning on the right track? What caveats should be attached? If you think this is the wrong track what would you recommend instead?

Second question: if this line of reasoning is on the right track (and even if it's not), what if anything can one glean from the IPPC Fourth Assessment Report summary (AR4), published on 2 Feb 07, regarding a 21st century greenhouse gas budget?

AR4 indicates that the only emissions scenario which, by best estimate, results in a rise of less than 2C is B1 (although it should be noted that “if radiative forcing were to be stabilised in 2100 at B1...levels, a further increase in global mean temperature of about 0.5C would still be expected, mostly by 2200”).

So how many gigatonnes of carbon (GtC) are emitted under B1? The Special Report on Emissions Scenarios prepared in relation to the Third Assessment Report (see here and here -- B1 is bottom left hand box) show emissions peaking at something over 10GtC per year between about 2015 and 2055 or 2060 (black line), declining to approx half present levels by end of this century.

The AR4 notes : “The climate-carbon cycle is expected to add carbon dioxide to the atmosphere as the climate system warms, but the magnitude of this feedback is uncertain. This increases the uncertainty in the trajectory of carbon dioxide emissions required to achieve a particular stabilisation level of atmospheric carbon dioxide concentration. Based on current understanding of climate carbon cycle feedback, models suggest that to stabilise at 450ppm CO2 could require that cumulative emissions over the 21st be reduced from approximately 607 [630 to 710] GtC to approx 490 [375 to 600] GtC” (3).

490 GtC is at the bottom end of the cumulative emissions range envisaged under B1 in SRES (see here). An emission trajectory consistent with this bottom end would – I assume – be one in the area shaded green below the black line for B1 indicated here. That is, emissions peaking at no more than about 10GtC/yr by about 2020 (growing more slowly between now and that date than on the black line trajectory) and beginning to fall quite soon after 2030 to well under half (perhaps under a third) of current levels by the end of the century (4).

Third question: Is the best working assumption we can make for now that total global emissions of 490GtC are likely to be consistent with a probable temperature rise of less than a 2C? And can we characterise the range of this probability with any confidence? (5) Should we be looking at another approach altogether, and if so why? (6) If so, what is or are the best alternatives?

Notes

(1) 2C has been the EU target. Mastrandrea and Schneider identify 2.85C as their median threshold for dangerous anthropogenic influence (Avoiding Dangerous Climate Change, chapter 27).
(2) at least, according to one reading of Meinhausen (Avoiding Dangerous Climate Change, chapter 28)
(3) apparently this is CO2 only, not CO2e, and it’s not clear to me when AR4 is suggesting stabilisation at this level would occur.
(4) obviously [?] they could rise more quickly so long as they then fell more quickly, consistent with total 21st century emissions of 490GtC
(5) This question has political and operational implications. For example, Doniger et al suggest a total carbon budget for the United States of 84GtC on reasoning which I take to be along these lines. (An ambitious, centrist approach to global warming, Science, 3 November 2006)
(6) For example, that it is unrealistic. The Stern review on the economics of climate change says: 'To stabilize at 450 ppm Co2e without overshooting, global emissions would need to peak in the next ten years and then fall at more than 5% per year, reaching 70% below current levels by 2050. This is likely to be unachievable with current and foreseeable technologies.' (p 218). Writing in The Observer today, the UK government's chief scientific adviser does not indicate what a "realistic" stabilisation level might be.