We all see the changes outside so we need facts to stop this trend not only words.

http://www.alternative-energies.net/a-few-solutions-to-fight-climate-change-in-2015/

It is time to learn a few tips that can help us stop the temperature to reach higher values. ]]>

[2]Environ. Res. Lett. 9 (2014) 031003 (7pp) doi:10.1088/1748-9326/9/3/031003

Perspective Implications of potentially lower climate sensitivity on climate projections and policy

Joeri Rogelj, Malte Meinshausen, Jan Sedláˇcek and Reto Knutti

These try to explain the possibility of lower effects from CO2 by increased water vapor – kept in atmosphere with a longer residence time than 9 days [Trenberth used 8 days in his ’97 GEB, – to fix latent heat at 78 or 80 w/m2 – non-moving value.. while Graeme Stephens calls it 88+/-10 W.m2 [2012 article on GEB in Nature]..

]]>Hi Paul,

I too am unsure about the assumption…primarily I wonder whether the feedbacks are indeed time invariant for each zone, which is the primary assumption of Armour et al. I have been planning on running a few quick tests on this soon…

]]>Troy,

Thanks for your response. Having considered the matter further (in light of your response), I withdraw my original objection.

My main reservation (now) is that the approach rests on the assumption that local feedbacks are invariant between model cases for the same model. I do not know how robust this assumption is.

Thanks

Paul

Hi Paul,

Thanks for you comments, and sorry I’ve been so slow to respond. I have wondered a similar thing myself, but I am not sure that horizontal heat transfer must explicitly be taken into account in order to assess the bias I am describing here.

First, I think we are in agreement that theoretically (using the same assumptions as Armour et al) one can diagnose λi for each zone without consideration of Q, right? After all, while Q affects the temperature increase, λi is only concerned with the change in outgoing flux normalized by that temperature increase. I don’t believe I am assuming that there is radiative equilibrium within each zone at TOA, as we are primarily dealing with the change in that outgoing flux, not its value in absolute terms. Moreover, it seems to me that any change in Q (vs, for example, the shorter timescale vs. equilibrium case) would implicitly be included in those different spatial warming patterns that we use for weighting the zonal TOA feedbacks.

“In other words, if you want to calibrate two models against each other to assess the degree of bias in lambda, I don’t think that you can do so by just looking at the TOA net radiative flux in each zone.”

I agree with this sentence, but isn’t this the point of then weighting by the normalized temperature increase in each zone for different states? Essentially I am saying “if we take from model X the spatial pattern of T at some point (either equilibrium or CO2-only idealized scenarios) and the spatial pattern of TOA feedbacks, what does the observed spatial warming pattern imply about the bias in lambda according that model?”. I suppose I don’t understand what assumptions I am making here that are not present in Armour et al (2013)?

Regardless, I will try to whip up some simplified demo with only 2 zones and some value of Q that either changes or remains constant, to see if I can satisfy myself whether failing to account for Q is significant or not, unless you have something similar already handy?

Thanks,

-Troy

]]>I’ve just realised that part of what I wrote above doesn’t scan. When I wrote Qi-1 = Qi+1 = 0, what I should have said was that Qi-1 = 0 when i = 1 and Qi+1 = 0 when i = n; these are just the end-member boundary conditions. Sorry for any confusion caused. ]]>

You might find your temperature-weighting approach runs into some problems because of the influence of meridonial flux. The individual zonal energy balance equations need to include these terms in order to make sense of the final steady-state solution, since locally, the net radiative flux balance does not go to zero at steady state. Consider n latitudinal zones, say. The local equations can be expressed as:-

dHi/dt = Net radiative flux (for ith zone) plus heat flow in from zone( i-1) less heat flow out to zone (i+1)

= Fi – λi * Ti + Qi-1 – Qi+1 for i = 1 to n

Qi-1 = Qi+1 = 0

When these equations are summed up, all of the Q terms (i.e. the meridonial heat transfer terms) sum to zero. However, they do have a serious impact on the averaging of local properties. Although the net heat flux, dHi/dt, must go to zero at steady state, the local energy balance is achieved by a balance of meridonial and radiative flux which can be both substantially non-zero.

The average temperature gain at time t is the areal weighted sum of terms like (Fi + Qi-1 – Qi+1 – dHi/dt)/λi .

At steady state, dHi/dt -> 0 and the effective equilibrium temperature is given by the areal weighted sum of (Fi + Qi-1 – Qi+1)/λi (Equation A)

The total change in TOA net flux is equal to the total areal weighted forcing, F. Hence the overall aggregate feedback shown by the model will be equal to the total areal-weighted forcing divided by the temperature given by Equation A.

In other words, if you want to calibrate two models against each other to assess the degree of bias in lambda, I don’t think that you can do so by just looking at the TOA net radiative flux in each zone.

Regards

Paul