18 October 2010

The fallacy of the greenhouse effect

Part 2 Part 3 Part 4 (Index)

An object can warm through the absorption of electromagnetic radiation (EMR). However, an object passively warmed can't warm the object providing the warmth. Were this to be so energy could be multiplied for no extra input merely by having objects mutually radiate EMR. But there is no such temperature multiplication because the amount mutually exchanged cancels. It does not add as is required by greenhouse theory.

The reason that an object can be heated by EMR on earth by the sun is because the sun is warmer than the earth.  The idea of the greenhouse effect is that the cooler, upper layers of air are able to warm a warmer ground by backradiation. But this can not happen.

Greenhouse theory would even require that the backradiation from the earth to the sun warms the sun by a small (if practically imperceptible) amount. This is impossible too because a cooler object can not warm a warmer one unless work is done.  But greenhouse gas, not having an energy source, can not provide this work nor can the earth provide work to the sun.

200-plus years of thermal study must be thrown out the window if we are to believe that EMR from a colder object can warm a warmer one. If that were the case energy could be made from nothing merely by bringing two objects together mutually radiating EMR (such as the air and ground) such as the following examples show.

Let object A represent a warmer object radiating to an infinite heat sink C maintained at absolute zero.  Object B is introduced into its field of radiation and so is warmed:

Object B comes up to equilibrium temperature.    Now it has its own radiation and object A is subject to more radiation than it was before (which was zero).  If the cooler object B can warm the warmer object A through backradiation then object A will heat to a higher temperature than before for free merely because object B is passively warmed.

Object A and B now radiate more energy to the universe than when A was by itself merely by B's presence. This is clearly not possible.

Now to carry it further let the sphere B be replaced by many such spheres B on one side of A.  They all radiate as much EMR as the original object B. With eight spheres the energy is multiplied eightfold according to greenhouse theory:

Object A is eight times as warmed by backradiation as it was when there was only one sphere B.  In the case of the atmosphere this is the equivalent of putting more greenhouse gas "energy absorbers" in the atmosphere.

Now let the EMR blocking coverage continue from eight spheres B to a hemispherical shell B.  This is a cross-section through B:

Now half the radiation of A is blocked by B.   According to greenhouse theory half of this half will be radiated back to object A thus warming it.  Object A now emits 100% +  25% now re-radiated back by the hemisphere B.  An extra 25% energy gain for free!

Let the hemispherical shell B become a fully enclosing spherical shell B:

According to greenhouse theory as much of the amount emitted outside of shell B will be emitted inside.  But what happens to the amount emitted inside?  Does it add to the energy?  According to greenhouse theory yes, but how can it?  Only the amount emitted to the outside of the system is relevant.

(The outside shell B must radiate a total amount which was equal to A's original output.  Being at a larger radius it will have a lower emission temperature.  Effectively the shell B is a red shifter of the EMR spectrum.  Yet the amount of energy emitted in total will be the same.)

If adding shells as greenhouse blockers could work as an energy multiplier then Willis Eschenbach's steel greenhouse model would work.  Many commenters on that thread in defence of the greenhouse effect objected to its preposterousness but it is not warranted because the model is an accurate representation of the greenhouse effect; it's just that the greenhouse effect is preposterous.

For me the resolution to the paradox is to view heat energy like a stream that only flows downhill, from warmer to cooler, despite the presence of backradiation.

An ordinary, human blanket creates warming by blocking convection not by "backradiation". An emergency aluminium foil blanket warms by a high reflectivity/low emissivity, not by absorption and re-emission.

In the case of the earth the only way that a chemical can alter temperature is by a lowered emissivity.  But greenhouse gases being good absorbers are also good emitters as per Kirchhoff's law.

An empirical example of how a cooler object can not warm a warmer object can be seen in the operation of a vacuum furnace.
Representative vacuum furnace production run: (a) furnace and workload temperatures versus time; (b) power requirement to achieve the temperature profile.
My note:  (The lower black line is a more efficient power supply than the upper red dotted line).
In a vacuum furnace the air is removed, so the only way to convey heat is via EMR.  For the first 2 hours the power input (black line) is made higher at 40% of max power to bring the furnace walls up to temperature.  Then the power is backed off to 20% for a further 26-hour period in which the 7.5 ton load (the product to be heat treated) comes up to the desired temperature.

Notice how the power input stays constant throughout the latter 26-hour heating period despite this massive 7.5 ton load coming up to nearly the same temperature as the oven and backradiating all of that EMR to the furnace walls? 

It's only the amount of heat loss to the outside of the furnace that the power needs to supply.  The reverberation of EMR within a system neither adds nor subtracts from the energy content.  This is empirical evidence against the greenhouse effect.

During the daytime earth's atmosphere provides cooling.  In the daytime sun some water placed in beer bottle with the backside painted black will heat to over 64C in a vacuum sleeve.  The same bottle will without the vacuum will heat to just 41C. 

(Data at Green Power Science here.)

(Also, for reference, the equilibrium temperature of a sphere in the direct sun from ground to space here. From here.)

Without the atmosphere it would be about as hot as the surface of the moon where the daytime temperature is 107C. From 107C to 64C shows the shielding effect of earth's atmosphere.

Convection and thermal inertia (thermal mass) of the atmosphere takes the temperature down even lower than the 64C mentioned above.

CO2 absorbs all its EMR within the first 10 - 25 metres.  (1,2,3) The slight warming that this creates is meaningless by the time you move upward in the atmosphere where the vertical temperature gradient dominates.

You could argue that the slight warming of air above from CO2 absorption causes a shallower temperature gradient and so slows convection.  But observe that regardless of the temperature on or near the ground -- from 50C in a desert to -40C in Antarctica -- when you move up to a certain height the vertical temperature and gradient is the same (depending on latitude).

Moving upward from the ground there is a negative temperature gradient -- the air gets cooler with altitude by about 6.5C per km.

The real mystery to me is what creates this vertical temperature gradient.  It dominates over the slight CO2 warming that occurs with 25 metres of the ground.

Seeing as heat can not flow from a cooler object to a warmer one (even with backradiation) the upper layers of troposphere can't warm the warmer surface. The exception to this is the temperature inversion that occurs below 2km at night and occasionally during the day.

A temperature inversion is the only way there can be a greenhouse effect. Then the air above can warm the ground below. But even then it is a passive slower of heat and, while it might keep it from getting colder at night, it can't make the earth 33C warmer than it otherwise would be. And the occasional daytime temperature inversion is not going to create the 33C warming attributed to the greenhouse effect.

As mentioned above every EMR absorber is equally well an emitter by Kirchhoff's law. Carbon dioxide emits just as much as it absorbs unless it is changing temperature. But it isn't changing temperature that much as being only 380 parts in every million the warming effect is minuscule and can't significantly warm the air compared to the energy delivered by convection and water evaporation.

The following is a graph of EMR taken over the Arctic taken from the SkepticalScience website.  It seems to show a lowered emissivity for the earth due to "energy blockers":

It shows the emission and absorption at CO2 absorbing frequencies centered around a wavelength of 15 µm.   

It  looks like a mirror image in the CO2 absorption zone with energy reflected down. It's really that the atmosphere is optically thick at those frequencies of light and looking up from the ground you see the downward component of a unidirectional emission of CO2 and H2O from the warmer, lower layer.  Looking down from high above you see emission from the cooler, upper troposphere.

What we are really seeing in the above is not energy blocking but the point of emission being moved from the earth's surface where it is warmer to the upper troposphere where it is cooler.  

Yet even this vertical displacement can't make the earth warmer because emission and absorption is not an "energy blocker" like a lowered emissivity could be said to be.

An interesting thing to do with the above graph from SkepticalScience would be to turn the sensor on the ground facing down instead of up and turn the sensor at 20km facing up instead of down to demonstrate how the CO2 is emitting as much as it is absorbing at every altitude. 

This visualisation can be approximated by a handy modtran simulator at the University of Chicago.  With it you can alter CO2, humidity, clouds, ground temperature, etc. 

Enter the webpage above and change nothing in the left pane except the humidity to zero and hit "Submit the Calculation" and you get the following graph:

1. Iout, W / m2 = 345.714
Ground T, K = 299.70
Sensor altitude = 70 km
Looking down, no clouds, rel. humidity = 0, tropical latitude
CO2 375ppm:

In this graph the higher frequencies are on the right.  The divot in the middle centered on 670 wavenumber is the CO2 absorption with "wings" -- a widened absorption area.  H2O absorption also overlaps a bit with this CO2 absorption.

Now in our model let's increase the CO2 to 10,000ppm CO2:
2. Iout, W / m2 = 324.048
Ground T, K = 299.70
Sensor altitude = 70 km
Looking down, no clouds,
rel. humidity = 0, tropical latitude
CO2 10,000ppm:

It says there's 21 less W/m2 emitted into space.  The wings of absorption get greater.  And so, earth should get warmer.  But is emittancy really the only factor that cools the earth?

Back to normal CO2, let us now add 100% relative humidity water vapour to it:
3. Iout, W / m2 = 287.844
Ground T, K = 299.70
Sensor altitude = 70 km
Looking down, no clouds, rel. humidity = 1, tropical latitude
CO2 375ppm:

A 58 W/m2 less emittancy for 0 to 100% humidity compared to a 21 W/m2 difference from 375 to 10,000ppm CO2.  So,  water vapour is much more powerful than CO2.  But while no one is suggesting increases in CO2 to 10,000ppm, changes in relative humidity  from zero to 100%, or parts in between, happen regularly on earth and yet it does not boil up.

Water vapour does change adiabatic cooling rates but not because of the greenhouse effect but the latent heat of condensation.  Charts of adiabatic cooling do not take into account greenhouse gases.

If EMR was the only way that heat was dissipated from the surface then clouds, which block EMR, would cause tremendous warming.

At night clouds keeps us warm through a slightly lowered emissivity/increased reflectivity and by blocking convection but this won't make the planet warmer than it otherwise would be.   During the day clouds make us cooler, not warmer, by reflecting sunlight up. 

At the very periphery of earth's atmosphere EMR is the only heat output.  It is only at the periphery of an object where emissivity can make a difference to the object as a whole.  But the emissivity at the edge says nothing of the heat flow within the material.

Let us explore this with the modtran simulator by adding a thick layer of clouds:
4. Iout, W / m2 = 290.858
Ground T, K = 299.70
Sensor altitude = 70 km
Looking down, cumulus clouds base 0.66km top 2.7km, rel. humidity = 0, tropical latitude
CO2 375ppm

Now the emittancy has gone down by a whopping 55 W/m2 compared to run 1 above.  According to Hansen et al 1 W/m2 = 3/4 of a degree C.  So 55 W/m2 change will cause a 41C temperature increase!?  Ask yourself does this happen when there's thick cloud over a large area, day or night?

So far every modtran graph I have provided has been at the sensor height of 70km looking down over the tropics.  To prove my point that the lowered emittance is due to the height at which the emission takes place we can view the emission spectrum from varying altitudes.

First it's useful to note in the right hand pane along with the emission curve there is another graph underneath with the air temp and the concentration of some gases.
5.  Iout, W / m2 = 395.012
Ground T, K = 299.70
Sensor altitude = 5 km
Looking down, no clouds, no humidity, tropical latitude
CO2 375ppm:

Wavelengths of the window frequencies are affected by atmospheric conditions like clouds and water vapour. Although most of the EMR comes from the ground at window frequencies it still doesn't mean that that EMR is the main mode of transmission through the atmosphere.  For example, the top of the clouds will merely replace the ground as the blackbody radiator.  This is because convection and water evaporation are  valves that offset CO2-global-warming-by-absorption.

Where the absorption valley bottoms out in the graph of greenhouse gas absorption represents the height that the emission takes place.  Even with 10,000ppm of CO2 in the air the temperature line that the valley bottoms out at is the same because there is saturation for CO2 absorption/emission even at 375ppm; the valley just broadens with more greenhouse gas it doesn't deepen.

What you get is the bottom of the valley, no matter the greenhouse gas concentration, follows the temperature gradient for that height of emission (green line above).

5 km looking down:
6.  Iout, W / m2 = 395.012
Ground T, K = 299.70
Sensor altitude = 5 km
Looking down, no clouds, no humidity, tropical latitude
CO2 375ppm:

At 5km the width of the valley is already at what it is at 70 km looking down (test run 1) while the bottom of the absorption valley is near the temperature for 5km altitude.

The knee of the temperature graph is about at 17 km in the modtran model.  This gives the lowest, deepest value for the valley because the air is at its lowest temperature at that height.
7.  Iout, W / m2 = 347.284
Ground T, K = 299.70
Sensor altitude = 17 km
Looking down, no clouds, no humidity, tropical latitude
CO2 375ppm:

Moving up again in altitude raises the bottom of that valley to the temp at the height above 17km:
8.  Iout, W / m2 = 345.4
Ground T, K = 299.70
Sensor altitude = 40 km
Looking down, no clouds, no humidity, tropical latitude
CO2 375ppm:

With the atmosphere thinning out above that height the EMR pretty much stays constant with height after about 35 km.

But it's interesting that above the 17km temperature knee in the modtran model despite the presence of CO2 the emission increases with altitude for about 18 km (from 17 to 35 km altitude).  Thus showing that CO2 is a good emitter.

Above the altitude at which convection and adiabatic cooling dominates (which is the troposphere) CO2 assists the emission of EMR to space.

Over Antarctica on a very cold day (I tweaked it to the record -89C!) you can see that even if the ground gets cold CO2 is still emitting more warmly than the ground:
9.  Iout, W / m2 = 73.036
Ground T, C = -89
Sensor altitude = 70 km
Looking down, no clouds, no humidity, arctic latitude
CO2 375ppm:

Notice how the top of the hill which replaces the valley in previous graphs is near that same yellow 220K blackbody line that the valley bottomed out at even though the ground is at -89C?  The emission temperature of the CO2 is the same regardless of the ground temperature.

(One could argue that I am pushing the limit of the modtran model above with this, although I did find this graph (from here) with a similar emission shape to test run 9.)

And, as well as its increased emissivity, due to the increased adiabatic cooling rate of CO2 compared to O2 and N2 it may offset the slight warming of the lower atmosphere to provide a net cooling effect.  At least one model in agreement with this is here: THE “GREENHOUSE” EFFECT.

With the cooler atmosphere backradiating less longwave radiation than is outbound you could say that heat loss is slowed down by a lower flux of EMR.  The IPCC predicts a modest warming of 1C for a doubling of CO2.  

But  daily fluctuations of 50C or more in some places make no difference to the sign of, or above a certain height the magnitude of, the vertical temperature gradient.  It shows that convection and evaporation of water convey more heat in the troposphere than EMR and therefore neutralise the slight greenhouse warming in the very lower portion.

In conclusion, whatever greenhouse warming there is it still won't make the earth as a whole warmer than it would be otherwise because the greenhouse effect can't can't generate extra energy or a higher temperature like a lowered emissivity can.  Whatever slight warming there is all washes out in the mix compared to daily and seasonal temperatures which vary widely. 

If this negative temperature gradient didn't dominate so over greenhouse warming there would be a hot spot such as explained on the JoNova website.

Energy blocking and reabsorbing is not a mechanism of storing energy nor of increased temperature.  The reason for what creates the negative vertical temperature gradient and the supposed 33C warming is to be explored in a follow-up post.


  1. ever heard of the atmospheric window?


  2. Yeah well if you look at that SkepticalScience pic above: it is the absorption and re-emission from CO2 that creates the window of which you speak.

    Looking up from the ground there is a strong re-radiation of IR down from CO2 at the relevant absorption/emission frequencies. Meanwhile looking down from 80km there is an emission from the cooler upper atmosphere. It makes it look like a reflection but it isn't. It's just normal EMR absorption and re-emission which occurs in every object to some degree.

    And the absorbed EMR at the side of this window still doesn't warm the Earth up more than it otherwise would be.

    I believe there are other mechanisms for the temperature regulation of Earth than just EMR from the Sun and emission of the Earth.

  3. In your first example you've missed a couple of things. Firstly, you've counted some radiation twice. If B absorbs some radiation from A, then the radiation to space from A is reduced by that amount. Radiation to space will remain the same as for a lone A. Secondly, if object A is radiating, then it is cooling down; the small amount of radiation from B can't "heat" A to a "higher temperature than before". The energy returned to A reduces its rate of cooling slightly.

    Your fourth illustration with the spherical shell B can only represent the Earth/atmosphere system if it's at equilibrium, that is the temperature of A is maintained by a heat source. In your example, A is cooling down; there is no heat source, so no equilibrium. Of course any "back radiation" has to have come from A in the first place, and there's no energy "added for free".

    If we assume a constant input of heat to A, and equilibrium, any heat energy that shell B has acquired to achieve its constant temperature has come from the period BEFORE equilibrium was attained.

    NOW your diagram will represent a simple version of the Earth/Atmosphere system, where the outward radiation from B equals the energy input from the source of heat, presumably internal to A, or in reality radiation from the sun through shell B.

    Your interpretation of the EMR plots is a little askew, Most of the radiation seen in the upper plot is that which goes through the "atmospheric window" (800 to 1300 cm−1) from the surface, and not emission from the upper atmosphere.

    Incidentally, the upper plot can be used to show that the calculation for the temperature of the "non-greenhouse" Earth as -18°C, with therefore a "greenhouse" warming of 33°C to the present 15°C is wrong. The calculation is made using the Stefan-Boltzmann equation and the radiation to space of 240 W/m², which indeed give an approximate -18°C, but the equation assumes a complete Planck curve for the radiation. It's easy to see that there are large chunks missing from the ideal envelope, and the calculation is erroneous, as the total radiation is related to the area under the curve. Your later MODTRAN plots show this better for the entire curve, I've estimated a better value by "filiing in" the gaps, of -5°C, which rather knocks a large chunk off the claimed "greenhouse" warming.