This is a post with big potential. A feedback the other climate models miss?
All the establishment models assume carbon dioxide warms the sky, which leads to the surface warming*, and the feedbacks then apply to the surface warming. It’s in the model architecture, the models can’t do it any other way. But what if the feedbacks don’t wait — what if the feedbacks start right away, up in the atmosphere? What if, say, CO2 warms the air, and that affects humidity and or clouds right then and there? These would be feedbacks operating on tropospheric warming, and they can reroute that energy.
Potentially, this blows everything away. If the energy blocked by increasing CO2 is merely escaping Earth through emissions from another gas in the atmosphere, like say, the dominant greenhouse gas, water-vapor, then could this explain why the effect of Co2 has been exaggerated in the conventional models?
We call this the “rerouting feedback” because when it’s harder for energy to escape to space through the CO2 pipe, this feedback would reroute it out through the water vapor pipe instead.
Put another way, as Earth emits (relatively) less energy through carbon dioxide’s favorite wavelengths, some of that blocked heat, possibly transferred through kinetic collisions, just reroutes out to space on the water vapor wavelengths instead.
No matter how many thousands of runs someone does on a conventional climate simulation, this outcome could never occur.
If this feedback is real and significant, it could explain why CO2 is not as potent as the IPCC supposes.
We will be expanding on this hypothesis in the future. It explains some anomalous observations. There are serious implications and tests we will develop.
7. The Rerouting Feedback
In post 5 we noted that the architecture of the conventional model only allows feedbacks that are responses to surface warming, thereby omitting any feedbacks that are primarily in response to climate drivers. In post 6 we discussed where outgoing longwave radiation (OLR) is emitted from, and introduced the “pipes” terminology. Now we build on both.
This post proposes the existence of the “rerouting feedback”, a feedback in response to an increase in the CO2 concentration, where the action takes place high in the atmosphere. It is omitted from the basic sensitivity calculation because it is not a response to surface warming, and it is also omitted from the large computer models (GCMs). Represented by fC in Fig. 1 of post 5, it reduces the radiation imbalance ultimately caused by an increase in CO2 and thus the warming influence of rising CO2.
For this discussion, let us suppose that all heat escapes the Earth through the four main pipes: the CO2 pipe, the water vapor pipe, the cloud top pipe, or the surface pipe (see Fig. 3 of post 6).
“How does the outgoing radiation rearrange itself among the four pipes?”
Increasing the CO2 concentration impedes the flow of OLR (or heat) through the CO2 pipe, so there is less OLR emitted on the CO2 wavelengths. The heat backs up a little, warming the atmosphere, but when steady state is resumed the total OLR is the same as it was originally because the absorbed solar radiation (ASR) is the same (ignoring the minor albedo feedbacks to surface warming).
The crucial question is: in light of the lowered OLR in the CO2 pipe, how does the OLR rearrange itself among the four pipes?
A pipe’s OLR is solely determined by the temperature of its emitting layer — the OLR in the surface pipe is determined by the surface temperature, the OLR in the water vapor pipe is determined by the average temperature of the water vapor emissions layer (WVEL) which in turn is determined by its average height and the lapse rate, and so on. Knowing the rearrangement of OLR between the pipes would allow us to know the change in OLR in the surface pipe, and thus the surface warming and the equilibrium climate sensitivity (ECS).
In the conventional model, increasing the CO2 concentration causes a sympathetic decrease in OLR in the water vapor pipe, due to amplification by water vapor feedbacks — the influence of extra CO2 is represented as a forcing, equivalent to extra ASR, which warms the surface, causing more evaporation and more water vapor, which presumably causes the WVEL to ascend because there is more water vapor in the atmosphere, whereupon the WVEL is cooler, which reduces the OLR in the water vapor pipe. So, in the conventional model, the surface and cloud top pipes must compensate for decreases in OLR in both the CO2 and water vapor pipes, by increasing their combined OLR by a matching amount. Obviously this requires much more surface warming than if the water vapor pipe also increased its OLR in response to the decreased OLR in the CO2 pipe.
Sketch of the Mechanism
Increased CO2 causes a decrease in OLR in the CO2 pipe. Now consider how it might also trigger a feedback that increases OLR in the water vapor pipe, by way of partial compensation (as if fC in Fig. 1 of post 5 was negative).
From the point of view of heat in the upper troposphere, increased CO2 makes it harder to escape to space in photons fired from CO2 molecules, and therefore relatively easier to escape in photons fired from water vapor molecules. Increased CO2 thus increases the relative propensity of OLR to come from water vapor molecules. The energy has to escape to space somehow, the relative attractiveness of the CO2 pipe has decreased compared to the water vapor pipe, and the heat is essentially available to all molecules because they swap energy back and forth by thermal collisions. Furthermore, the changes in the CO2 spectrum with increased CO2 occur in the wings of the CO2 “well” (see for instance the last diagram on this page of Barrett Bellamy), at heights around 8 km, which is about the average height of the WVEL.
“…when increasing CO2 makes it more difficult for heat to radiate to space on the wavelengths at which carbon dioxide absorbs and emits, some of the blocked heat simply reroutes out to space on the water vapor wavelengths instead.”
If more OLR comes from the water vapor molecules, the population of water vapor molecules would be less energetic and would thus tend not to ascend quite so high in the Earth’s gravitational field — so the WVEL would descend slightly (which would be compatible with the non-observation of the “hotspot”; more on that in later posts). Although the population is less energetic, the top of the population is in a lower and therefore warmer place compared to it where was before the increased CO2 caused it to descend. Thus the WVEL is warmer, emitting more OLR.*
Note that it is possible for the WVEL to descend despite increased evaporation from the surface, if the extra water vapor is mainly confined to the lower troposphere and the consequent greater stability at low altitudes leads to less overturning and less transport of water vapor to the upper troposphere — indeed this seems to be happening, as reported by Paltridge et. al in 2009 , from study of the better radiosonde data from 1973.
We call it the “rerouting feedback” because some fraction of the OLR that is blocked from escaping to space out the CO2 pipe by rising CO2 levels is instead rerouted out the water vapor pipe.
“It is not a response to surface warming, but to CO2 enrichment.”
In other words, when increasing CO2 makes it more difficult for heat to radiate to space on the wavelengths at which carbon dioxide absorbs and emits, some of the blocked heat simply reroutes out to space on the water vapor wavelengths instead. This feedback takes place high in the atmosphere, far from the surface, so there is no place for it in the conventional climate model — which only contains feedbacks in response to surface warming.
This proposed feedback is contained within fC in Fig. 1 of post 5. It is not a response to surface warming, but to CO2 enrichment. It all occurs within the higher atmosphere, so it responds more strongly to variables describing the upper atmosphere and radiation than to the surface temperature. (Perhaps a suitable variable to describe the strength of the feedback is the height of the CO2 emission layer plus the height of the WVEL.)
The rerouting feedback might offset a substantial portion of the reduction in OLR in the CO2 pipe due to an increasing CO2 concentration. If it exists, the rerouting feedback would lower our estimates of the sensitivity of surface temperature to rising CO2 levels.
A Negative Feedback?
The rerouting feedback reduces the ultimate radiation imbalance due to extra CO2, so it is a negative feedback in terms of its effect on the CO2 forcing, so fC is negative. Applying the feedback diagram in Fig. 1 of post 3 with a equal to DR,2X and b to fC, the rerouting feedback changes the radiation imbalance due to increasing CO2
For example, if fC was −0.6 then
and the influence of increasing the CO2 concentration would be reduced by 70%.
Semantic point: Although the rerouting feedback reduces the sensitivity of the surface temperature to changes in CO2, and although fC is negative, it is not a feedback in response to surface warming so it is not a “negative feedback” as that term is understood in the conventional paradigm.
Consider how the climate might adjust to a decrease in OLR in the CO2 pipe. The blocked OLR has to find its way to space somehow. The resistance of the surface pipe to carrying more OLR is exceptionally high in the tropics, where most of the heat is, because heat loss from the surface via evaporation rises exponentially with surface temperature (Kininmonth 2010  elaborates on this). The resistance of the water vapor pipe to carrying more OLR might be relatively low, because it requires only that the average height of the WVEL (~8 km) ascends or descends by a few tens of meters. Like the WVEL, the cloud tops might ascend or descend slightly with little apparent energy requirement.
The energy required to warm the surface on a sustained basis, with the ocean warming that would entail, might be much greater than the energy required to change the average height of the WVEL or cloud tops sufficiently to change OLR by the same amount. (More research is needed to get the figures to assess this.) This would suggest that the bulk of the response to the decrease in OLR escaping via the CO2 pipe would come as more OLR from the WVEL or cloud tops, rather than from the surface — which is consistent with the proposed rerouting feedback and with a lower ECS.
Figure 1: Electrical analogy for heat escaping to space. The zig-zags are electrical resistors; the current (a la heat) mainly flows through the paths of least resistance — the current in a resistor is inversely proportional to its resistance. Increasing CO2 increases RC, so some current reroutes from flowing through RC to flowing through the other resistors, mainly through the other resistor with the lowest resistance.
* Update 4 Oct 2015: The mechanism sketch needs more details, as discussed in the comments (thank you Stephen Wilde and Joe Born).
If less heat is escaping to space from CO2 molecules, the upper atmosphere must warm, especially the upper troposphere where the most significant changes in the emissions spectrum of CO2 are occurring, in the wings of its 15 micron “well”. Presumably this heat warms the WVEL, also in the upper troposphere, and so more OLR is emitted from water vapor molecules. Hence the rerouting of some OLR from the CO2 pipe to the water vapor pipe, as CO2 increases.
How would this affect the height of the WVEL? If the lapse rate remained unchanged, it would have to descend in order to be warmer, so presumably it would descend — somehow (knowing the initial and final states doesn’t necessarily tell us how it did it).
However an offsetting mechanism could be as follows. There is a net loss of emission of heat to space as OLR (the increase from water vapor could never compensate for the decrease from CO2 100%, though it might be close-ish), which would change the local lapse rate, in the upper troposphere. OLR accelerates the cooling of ascending air, so the net decrease in OLR would decrease the lapse rate a little (that is, less cooling per kilometer). A decrease in the local lapse rate means it’s warmer at the same altitude, working upward from a surface at a constant temperature, so perhaps a warmer WVEL could be found at an unchanged altitude.
(The warmer WVEL would mean the water vapor population is more energetic, more able to work against the Earth’s gravity — even though it is losing energy via increased emission of OLR, there is a net increase in energy in the water vapor. So this factor would tend to increase the WVEL height slightly. However this is presumably a much smaller force than lapse rate and humidity changes.)
Paltridge et al.  note that the last few decades has seen a drier upper troposphere, which they explain as the extra water vapor due to surface warming creating greater stability at low altitudes, leading to less overturning and less transport of water vapor to the upper troposphere. Perhaps increased CO2 is also leading to a slightly drier upper troposphere, which would lower the WVEL.
[1^] Paltridge, G., Arking, A., & Pook, M. (2009). Trends in middle- and upper-level tropospheric humidity from NCEP reanalysis data. Theoretical and Applied Climatology, 98:351-359.
[2^] Kininmonth, W. (2010). A Natural Constraint to Anthropogenic Global Warming. Energy and Environment, Vol. 21, No. 4, 225 – 236.
*Yes, this is shorthand. In a technical sense, CO2 is not “warming” the ground, merely slowing heat loss while the Sun sends in more energy. Ultimately the ground ends up warmer than it would have been.