Hmmm, entering a cooling phase?

I thought I would briefly comment on a recent paper by McCarthy et al. called Ocean impact on decadal Atlantic climate variability revealed by sea-level observations. The paper seems to be suggesting that the Atlantic MultiDecadal Oscillation (AMO) may be entering a negative phase. Their Press release says

the global climate is on the verge of broad-scale change that could last for a number of decades.

…..Since this new climatic phase could be half a degree cooler.

Their Conversation article says

This is known as the Atlantic Multidecadal Oscillation (AMO), and the transition between its positive and negative phases can be very rapid. For example, Atlantic temperatures declined by 0.1ºC per decade from the 1940s to the 1970s. By comparison, global surface warming is estimated at 0.5ºC per century – a rate twice as slow.

So, they seem to be suggesting that we could be entering a phase of slower warming, or – possibly – even cooling. This is rather odd as the recent Steinman, Mann & Miller paper says

This natural cooling trend appears to reflect a combination of a relatively flat, modestly positive AMO and a sharply negative-trending PMO. Given the pattern of past historical variation, this trend will likely reverse with internal variability instead, adding to anthropogenic warming in the coming decades.

which seems to be suggesting we’re about to enter a period where internal variabililty will amplify anthropogenically driven warming.

So, why does this recent paper suggest the opposite? Well, Sou has already largely explained this. The McCarthy et al. paper simply uses linear detrending to estimate the AMO. However, this is only reasonable if the forced response is linear. It almost certainly is not, and the Steinman, Mann & Miller paper uses actual forcing data to determine the underlying, internally forced variability, which is why their result is different – and probably more physically plausible – than that of McCarthy et al.

However, you can also do some basic sanity checks as to whether or not entering a cooling phase makes physical sense. The NOAA Ocean Heat Content (OHC) data suggests that we’re accruing energy at a rate of around 10²³J per decade. About one-third of this goes into the upper 300m of the ocean, which we can reasonably assume is coupled to the atmosphere (not quite right, as it is probably the upper 100m, but this won’t change things much). The upper 300m of the oceans has a heat capacity of around 4 x 10²³ J K^-1. If this region is accruing energy at the rate of 3.3 x 10²² J/decade, that would suggest it should be warming at just below 0.1^oC per decade. Not far off what is observed if we correct for coverage bias (Cowtan & Way 2014).

So, could we enter a global cooling phase? Well, we’re increasing anthropogenic forcings by about 0.35 Wm^-2 per decade. If we were to cool, or not warm, the planetary energy imbalance would increase, the energy going into the upper ocean would presumably also increase, and we’d have to warm. That seems logically inconsistent, so a multi-decade cooling phase just seems physically implausible. We can’t really increase the energy going into the system without it also warming.

What about continuing to warm slowly, say at around 0.1^oC per decade? Well, we’re increasing anthropogenic forcings at around 0.35Wm^-2/decade. The feedback response is about 1.2Wm^-2K^-1. If we’re accruing energy at 10²³J/decade, that implies a planetary energy imbalance of around 0.7Wm^-2. If we continue to warm at 0.1^oC per decade, then the net non-Planck feedback will 0.12Wm^-2/decade, which will amplify the anthropogenic forcings by 0.12Wm^-2/decade. The Planck response, on the other hand, is 3.2Wm^-2K^-1, so would provide a negative feedback of 0.32Wm^-2/decade. So, overall, we’d be increasing the planetary energy imbalance by around 0.35 + 0.12 – 0.32 = 0.15Wm^-2/decade.

So, again, we’d be increasing the rate at which we’d be accruing energy, which would seem to imply that accelerated warming is more likely than reduced warming, or cooling. This also ignores that we could increase our emissions and increase the rate at which the anthropogenic forcing is rising. Of course, these are all just ballpark figures, but it is still the case that it is very difficult to see how we can enter a phase of slower warming, or cooling, if we continue to maintain, or increase, anthropogenic emissions. Basic energy balance arguments suggest that we have to continue warming and that – at some stage – this will have to accelerate in order to bring the system back towards energy balance.

So, it’s certainly my view that if someone is going to write a paper suggesting that we may be entering a phase of slower warming, or even cooling, they should probably do some basic sanity checks to see if it makes physical sense. That doesn’t mean that they shouldn’t publish their work, but it would seem worthwhile discussing this basic issue. It may even be possible, but that would seem to require some mechanism for continuing to increase the rate at which we sequester energy in the deeper parts of the ocean. One should – in my opinion – bear in mind that no amount of curve fitting, or complicated statistics, can trump the basic laws of physics.

Update: The University of Southampton has now updated its Press Release to say

Previously transitions to a negative AMO saw a relative cooling of half a degree in the Atlantic, if this occurs again, it may well offer a brief respite from rising temperatures, as well as resulting in fewer hurricanes hitting the United States.

which is clearer. I still think that continued slow warming for the next few decades is unlikely if we continue to increase our emissions, but this is – at least – a less ambiguous description of the possible implications of this work.