Jai, first,
Thank you!
Now, on to the debate...
...
Andrew,
There is massive amounts of uncertainty about where we are headed.
If by "where we are headed" you mean "the global climate system", I can't disagree with that. But if you mean "September sea ice extent in the coming 20 years", I would say there are not that many uncertainties. It's trending to essentially zero (there is always a "noise floor" in any measurement, but let's disregard that and assume that 1 x 10^6km2 is well above the "noise floor" of Arctic sea ice extent). And since the anthropogenic forcing that is causing it to go to zero is getting stronger, what non-anthropogenic forcing could you invoke that would overcome the anthropogenic forcing and reverse the observed trend?
(and I am not even going to mention the accelerated decrease in aerosols which is occurring concurrently with the last stages of the demise of Arctic sea ice - because that's where you excel
)
Within that uncertainty is the near term climate response which if you look at the available data you would have to be somewhat of a lukewarmer to expect our FIRST sea ice extent below 1M km^2 in the 2040-2050 time frame or later.
Oh, I clearly believe the FIRST September sea ice extent below 1M km^2 year question will be settled soon enough, despite the models and Gompertz fit extrapolations and "slow transition hypothesis", etc.
however, there are many ideas about what will happen once this occurs that could lead to slowing of the long-term sea ice impact (i.e. see Chris' "Slow transition" thread). Some of these include increased cloudiness in summer melt season (This view has a very strong representation among the mainstream cryosphere climate modelers and scientists), leading to a rapid reduction in ice mass loss during the melt season or the freshwater lensing effect from Hansen's (et al.) exponential greenland ice melt (see ASLR's work on the "Hansen et al paper: 3+ meters SLR by 2100" also in Consequences.
Yes, there is a (relatively short) list of hypothesized negative feedbacks. The uncertainty here is over their magnitude and even if they exist at all.
On the other hand there is a (relatively longer) list of scientifically proven and easily quantifiable positive feedbacks.
So do we go with the "few unknown and not quantifiable negative feedbacks" or "many well-known and easily quantifiable positive feedbacks" to make our educated guesses?
If you assert that your understanding is the absolute truth
don't flatter me, it's not going to work... well, perhaps just a little... OK, go ahead!
then you have to have a reasoned analysis that clearly shows why these uncertainties are not going to produce the effect that others may expect. There are also many, many other uncertainties that could also produce a slowing or even a return of sea ice even after the first few years of <1M km^2).
(emphasis mine)
"many, many... could..." ? But that's the "invisible dragon in my garage" argument! Not good! I can list you the known positive feedbacks that will force September sea ice to zero, you are saying that there could be things that we don't know that would prevent September sea ice from staying at zero.
Do you see the asymmetry here?
https://en.wikipedia.org/wiki/Ad_hoc_hypothesishttps://en.wikipedia.org/wiki/The_Demon-Haunted_World#Dragon_in_the_garagewe are talking about decadal timescales in dynamic system that even the most powerful super-computer models do not yet model accurately.
Again, models can be useful, but a 10-years in the future
extremely accurate sea ice model is not going to be very useful because Arctic sea ice will be gone for good, just as you don't need a blood circulation model to tell when a person is dead, or run your model simulations to ascertain whether there is any possibility that the heart may beat again.
If you force a system beyond some thresholds, it breaks. Even if you remove the forcing, it doesn't come back.
And yes, I am aware that model simulations have shown that Arctic sea ice loss could be 100% reversible with no hysteresis - IF we could remove the CO2 forcing. The only problem is, we can't.