Neither of those. I am talking about a sudden feedback accelerated release of large amounts of methane on far more than a local scale, causing a quick increase in atmospheric temps sufficient to render some of the densely populated areas of the planet occasionally (and therefore essentially) uninhabitable during heat waves. This, I submit, would be catastrophic. This possibility accepts the asymptotic approach of a maximum upper temperature far below the runaway greenhouse, as apparently indicated by various articles linked here.
Good treatment of that question from a post on RealClimate:
==========================================
An Arctic methane worst-case scenario
7 January 2012
Let’s suppose that the Arctic started to degas methane 100 times faster than it is today. I just made that number up trying to come up with a blow-the-doors-off surprise, something like the ozone hole. We ran the numbers to get an idea of how the climate impact of an Arctic Methane Nasty Surprise would stack up to that from Business-as-Usual rising CO2
Walter et al (2007) says that Arctic lakes are 10% of natural global emissions, or about 5% of total emissions. I believe that was considered to be remarkably high at the time but let’s take it as a given, and representing the Arctic as a whole. If the number of lakes or their bubbling intensity suddenly increased by a factor of 100, and it persisted this way for 100 years, it would come to about 200 Gton of carbon emission, which is on the same scale as our entire fossil fuel emission so far (300 Gton C), or roughly the amount of traditional reserves of natural gas (although I’m not sure where estimates are since fracking) or petroleum. It would be a whopper of a surprise.
Scaling Walter’s Arctic lake emission rates up by a factor of 100 would increase the overall emission rate, natural and anthropogenic, by about a factor of 5 from where it is today. The weak leverage is because the high latitudes are a small source today relative to tropical wetlands and anthropogenic sources, so they have to grow a lot before they make much difference to the sum of all sources.
The steady-state methane concentration in the air scales nearly linearly with the emission rate. Actually, the concentration goes up somewhat faster than a constant times the emission rate, because the lifetime in the atmosphere gets longer (IPCC TAR). Let’s err on the side of flamboyance (great word in this context) and say the concentration of methane in the air goes up by a factor of 10 for the duration of the extra methane emission (meaning that the lifetime doubles).
Using the modtran model on line I get a radiative forcing from 10 * atmospheric methane of 3.4 Watts/m2 (the difference in the instantaneous IR flux out, labeled Iout, between cases with and without 10x methane). Using the TAR estimates of radiative forcing gives 2.7 Watts/m2.
But methane is a reactive gas and its presence leads to other greenhouse forcings, like the water vapor it decomposes into. Hansen estimates the “efficacy” of methane radiative forcing to be 1.4 (Hansen et al, 2005, Shindell et al, 2009), so that puts us to 4 or even 5 Watts/m2.
This is about twice the radiative forcing today from all anthropogenic greenhouse gases today, or (again according to Modtran) it would translate to an equivalent CO2 at today’s methane concentration of about 750 ppm. That seems significant, for sure.
Or, trying to “correct” for the different lifetimes of the gases using Global Warming Potentials, over a 100-year time horizon (which still way under-represents the lifetime of the CO2), you get that the methane would be equivalent to increasing CO2 to about 500 ppm, lower than 750 because the CO2 forcing lasts longer than the methane, which the GWP calculation tries in its own myopic way to account for.
But the methane worst case does not suddenly spell the extinction of human life on Earth. It does not lead to a runaway greenhouse. The worst-case methane scenario stands comparable to what CO2 can do. What CO2 will do, under business-as-usual, not in a wild blow-the-doors-off unpleasant surprise, but just in the absence of any pleasant surprises (like emission controls). At worst comparable to CO2 except that CO2 lasts essentially forever.
======================
So again you have a case close to the IPCC worst-case prediction (A1F1.) Which is good because then you also have their fairly exhaustive analyses of what that sort of temperature increase would do.