Saturday, August 6, 2016

5/20/19, rev 3/9/22,9/15/22,6/25/23,8/15/23
Climate Change Drivers*
Dan Pangburn, P.E.** (ret), MSME, ASME life member
pangburndan@gmail.com

CONTENTS

1. Introduction
2. CO2 does not control climate
3. Environmental Protection Agency mistakes
4. Thermalization reduces CO2 influence on climate. How GHE works.
5. Radiance calculated by MODTRAN6
6. Water vapor
7. Effect of water vapor increase
8. Water vapor change with time
9. World Sources of Increased Water Vapor
10. Effect of TPW increase on clouds
11. Approximate effect on the planet of the net of ocean surface temperature (SST)
12. Comparison of approximation with ‘named’ ocean cycles
13. Atmospheric carbon dioxide
14. Climate sensitivity determined using MODTRAN6
15. Sunspot numbers
16. AGT measurement data set
17. The AGT Model
18. Comments regarding the method used in developing and using Equation (1)
19. Results from the AGT model
20. Attribution quantified by sunspot number anomaly proxy.
21. Hind Cast Estimate
22. Step changes in AGT are not possible
23. Caveats on Predictions
24. Conclusions
25. References










·                  *       This document is periodically updated to improve clarity and incorporate revisions, improvements, and new data.
**   P.E. stands for Professional Engineer. A Professional Engineer is a person who has been tested by and certified by a state government as one who, subject to liability, may decide the design and functional adequacy of things that, if inadequate, could endanger the public.





Summary

Thermalization and the steep decline with altitude of the population density of water vapor (WV) molecules up to the tropopause, about 8-16 km, explain why water has a dominate effect on climate. Atmospheric carbon dioxide (CO2) and other trace greenhouse gases have little if any net effect on climate. Much of upward photon emission from WV molecules goes directly to space, especially from around 4 km and wavenumber 500 and substantially at lower wavenumbers up to the tropopause, which is at about 10 km on average, 16 km at low latitudes. Above the tropopause the atmosphere is cooled mostly by outward directed radiation from CO2.

 Reported average global temperature (AGT) since before 1900 is accurately explained (96+% match with measured trend 1895-2020). Although the warming influence of WV increase is closely proportional to the increase due to the sun (which is quantified by a proxy; the time-integral of sunspot number anomalies), the temperature change 1909-2020 is probably best explained by derived attribution of ocean cycles to the average global temperature rise, sunspot number anomaly proxy and increased atmospheric water vapor . Since about 2005 the increasing water vapor has been countering the AGT decline which would otherwise be occurring. 


A note about insignificant factors:
In just about all properties and processes there are inconsequential exceptions and/or uncertainties. These are often allowed-for by claiming them to be ‘insignificant’. In all that follows, insignificant factors are excluded whether explicitly stated or not.

Meaning of terms

The words ‘water vapor’ can be misinterpreted. WV is a transparent gas. If something is visible, like steam or a cloud, it is not WV but is condensed liquid water droplets or tiny bits of ice.

 The term ‘vapor pressure’ has different meanings in different disciplines. In meteorology it means the partial pressure of WV in the atmosphere. In most other disciplines and general use, it means the pressure developed by the liquid as a result of its impetus to change phase and become a gas. This impetus depends only on the temperature of the liquid water. It is unambiguously called saturation vapor pressure. In this document, the saturation vapor pressure is sometimes referred to simply as vapor pressure (VP). The pressure of WV in the atmosphere might be identified as its partial pressure.

Another difference in term usage is the meaning of the word ‘feedback’. In engineering it usually refers to feedback factor, a dimensionless number which is the ratio minus 1 of response with feedback to the response if there were no feedback. In most science disciplines it refers to the magnitude of the response to a forcing which contributes to the cause of the forcing. In Climate Science it is measured in W/m2.

1. Introduction
The only way that energy can significantly leave earth is by thermal radiation. Only solid or liquid bodies and greenhouse gases (ghg) can significantly absorb/emit in the wavelength range of significant terrestrial radiation. Ghg significantly absorb/emit only at specific wavelengths which are characteristic for each molecule specie. They both absorb and emit at the same frequency signature. Both magnitude and frequency can be calculated using Quantum Mechanics [52] (some prefer to call it Quantum Physics). In the range of terrestrial temperatures, non-ghg must transfer energy by thermal conduction in the gas to ghg (or to liquid or solid bodies) for this energy to be radiated. Note: The expression ‘greenhouse gas’ is somewhat misleading (greenhouses actually work primarily by suppressing convection). A more correct understanding is that so-called ghg can significantly absorb/emit radiation in the wavelength range of significant infrared radiation associated with earth temperatures.

The word ‘trend’ is used here for temperatures in two different contexts. To differentiate, α-trend (period about 69.6 yr) is an approximation of the net of measured ocean surface temperature oscillations after averaging-out the year-to-year fluctuations in reported average global temperatures. The term β-trend (period = centuries) applies to the slower average energy change of the planet which is associated with change to the average temperature of the bulk volume of the material (mostly ocean water) involved.

Some ocean cycles have been named according to the particular area of the oceans where they occur. Names such as PDO (Pacific Decadal Oscillation), ENSO (el Nino Southern Oscillation), and AMO (Atlantic Multi-decadal Oscillation) might be familiar. They report the temperature of the water at or near the surface. The average temperature of the bulk water that is participating in these oscillations cannot significantly change so quickly because of high effective thermal capacitance [1]. The effective thermal capacitance of the planet is approximately equivalent to the top 110 meters of the oceans.

This high thermal capacitance absolutely prohibits the rapid (year-to-year) AGT fluctuations which have been reported, from being a result of any credible net forcing. According to one assessment [1], the time constant (time to reach 63.2% (=1-1/e) of the final change following a step change in forcing) is about 5 years. A likely explanation for much of the reported year-to-year fluctuations is that they substantially involve the roiling (at the observed tiny fluctuation magnitude) ocean surface [44]. Volcanic activity is occasionally also a temporary contributor. A simple calculation shows the standard deviation of the reported annual average surface measurements to be about ±0.09 K with respect to a smoothed trend.

The ꞵ-trend is a better indicator of the change in global energy; which is the difference between energy received and energy radiated by the planet.

Knowledge of the kinetic theory of gases, thermalization, Maxwell-Boltzmann distribution of energy among atmospheric gas molecules, some thermodynamics, the absorb/emit wavelengths of water vapor and other ghg, heat transfer analysis, and the rudiments of Quantum Mechanics provide understanding and  a rational explanation of what happens with terrestrial thermal radiation.

All gas molecules which are significantly IR-active at earth temperatures are called ghg molecules. This includes water vapor molecules.


2. CO2 does not control climate
What is meant by the statement that carbon dioxide (CO2) is a greenhouse gas (ghg)? If by that is meant CO2 absorbs electromagnetic radiation with a wave length of 15 microns (including +/- a micron or so due mostly to pressure broadening, especially near sea level), well, that was demonstrated in the lab a long time ago and remains true. But if by that is meant CO2 substantially contributes to global warming, there is multiple compelling evidence (most identified earlier [2] ) that CO2 has little, if any, controlling effect on climate:
2.1. In the late Ordovician Period (about 450 mya), the planet plunged into and warmed up from the Andean/Saharan ice age, all at about 10 times the current CO2 level [3].
2.2. Over the Phanerozoic eon (last 542 million years) there is no correlation between CO2 level and AGT [3].
2.3. Antarctic ice core data show that during the last and previous glaciations AGT trend changed direction before CO2 trend [2].
2.4. Since 2001, average temperature uptrend calculated by Global Climate Models (GCMs, aka General Circulation Models) which assume CO2 causes Anthropogenic Global Warming (AGW) is about twice what is measured. [13]. The one Russian model output essentially matches measurements but carries the usual ‘free world’ suspicions as to its authenticity. This is shown, with recent data added, on Figure 0.1.


Figure 0.1: Comparison of GCM output with measured.

2.5. Analysis of CO2 and temperature data 2002-2008 shows a close correlation between dCO2/dt and lower tropospheric temperature. This demonstrates that CO2 level follows temperature and not the reverse. [30]
2.6. Average global water vapor has been increasing faster than possible, calculated on the basis of increased vapor pressure of water resulting from temperature increase of the liquid surface water. (Section 8 here) This demonstrates that average global temperature increase has been driven by water vapor increase, not the reverse.
2.7. The data from all reporting agencies agree that at least as far back as 2002 average global temperature tracks WV not CO2 (Figure 0.2).





Figure 0.2 Average global temperature tracks water vapor not CO2.

2.8. Hitran [52] using Quantum Mechanics calculates, besides many other things, the relative absorb/emit intensity of water vapor molecules vs CO2 molecules. Comparison at zero altitude is shown in Figure 0.25. Comparison by the ratio of the summation of the products intensity times wavenumber for each transition for each molecule species is (Σ Ii * wni)WV / (Σ Ij  * wnj)CO2 ≈ 1520/46 = 33. On average at ground level, according to the comparatively low populations used by Hitran, WV molecules outnumber CO2 molecules by about 8,000/330 ≈ 24 to one. After accounting for molecule count, each WV molecule is still 33/24 ≈ 1.37 times more effective at warming (absorb/emit of thermal radiation) than a CO2 molecule.

 

The relative effectiveness of the increases of WV and CO2 over the last 35 years is calculated as follows:

CO2 increase in 3.5 decades [21], 1988 to 2023: 420 - 350 = 70 ppmv

 

Average global water vapor increase trend from Figure 3, which is a graph of NASA/RSS TPW data, is 0.03942/29 * 100 * 10 = 1.36 % per decade.

 

From Figure 2, at 30 degrees latitude (area to pole = area to equator) average global WV = 10,000 ppmv. WV increase in 3 decades = .0136 * 10,000 * 3.5 = 476 ppmv

 

Therefore, WV molecules have increased 476/70 =6.8 times faster than CO2 molecules and WV increase has been 6.8 * 1.37 ≈ 9.3 times more effective at increasing ground level temperature than CO2 increase 1988-2023. (Much of the world human population has been falsely indoctrinated)

 

Well above the tropopause, radiation emitted from molecules there to space is primarily from CO2 molecules. If you ignore the increase in water vapor (big mistake), near the surface, WV averages about 10,000 ppmv. The increase in absorbers at ground level since 1900 is then about 10,410/10,295 - 1≈ 0.01 = 1%. WV above the tropopause (about 10 km) is limited because of the low temperature (~ -50 °C, saturation vapor pressure of ice 3.94 Pa [58], total pressure 26,500 Pa) to about 3.94/26,500 = 0.000149 = 149 ppmv while the CO2 fraction remains essentially constant with altitude at (in 2019) about 410 ppmv; up from about 295 ppmv in 1900. The increase in emitters to space at high altitude (~> 20 km, 0.055 atm), and accounting for the lower atmospheric pressure, is (410 + 149)/(295 + 149) * 0.055 ≈ 0.069 = 6.9%. This (increased CO2 adds more % emitters than absorbers) explains why CO2 increase does not cause significant warming (except at the poles). The result being that Climate Sensitivity is not significantly different from zero.

 The exception at the poles is because it’s cold there at ground level so WV molecule count is already low. Therefore, redirection of energy to WV molecules which radiate it to space does not occur.



Figure 0.25: At zero altitude, CO2 absorb/emit is barely discernable compared to WV.

2.9. Global average water vapor is about 10,000 ppmv (Figure 2). Based on TPW percent increase (Figure 3), WV increase since 1988 has been about 0.0431/29 * 10,000 = 14.86 ppmv per year. CO2 (348 ppmv in 1988) increase has been about (410 – 348)/30 = 2.067 ppmv per year. Therefore, about 14.86/2.067 ≈ 7 WV molecules were added for each CO2 molecule.
2.10. CO2 increased 2002 to 2018 by 40% of the increase 1800 to 2002
1800 avg. Lawdome, Neftel, Friedli = 281.6 ppmv
2002 avg. Mauna Loa/Keeling = 373.3 ppmv
Nov, 2018 Keeling = 410.0 ppmv
(410-373.3)/(373.3-281.6) = 0.40 à 40%

The recent temperature data which track water vapor, and two previous 30+ year downtrends in temperature with relentlessly rising CO2, demonstrate that CO2 has little if any effect on average global temperature. A possible explanation for the insensitivity of AGT to CO2 is that the small percent increase in the number of absorbers near the surface (water vapor and CO2) is compensated for by the large percent increase in emitters (mostly CO2) above the tropopause. [51]


3. Environmental Protection Agency mistakes
The US EPA asserts [10] Global Warming Potential (GWP) is a measure of “effects on the Earth's warming” with “Two key ways in which these [ghg] gases differ from each other are their ability to absorb energy (their "radiative efficiency"), and how long they stay in the atmosphere (also known as their "lifetime").”

The EPA calculation overlooks the very real phenomenon of thermalization. Trace ghg (all ghg except water vapor) have little if any effect on climate because absorbed energy is immediately thermalized which allows the energy to be redirected to water vapor [51]. Thermalization is a key factor in understanding the radiant energy flow by which CO2 and other ghg except water vapor (WV) have little if any net effect on climate.

The EPA calculation of the GWP of a ghg also erroneously overlooks the fact that any added outgoing radiation cooling from the increased temperature the ghg might have produced is also integrated over the “lifetime” of the gas in the atmosphere so the duration in the atmosphere ‘cancels out’. Therefore GWP, as calculated by the EPA, egregiously overestimates the influence on average global temperature of noncondensing greenhouse gases. The influence (forcing) of a ghg cannot be more than determined by its immediate concentration. 

The EPA assessment completely ignores the effect of water vapor which, by far, is the most important ghg and appears to be the only ghg that has a significant effect on climate.


4. Thermalization reduces CO2 influence on climate.
At a scale of the size of atoms, the atmosphere consists of gas molecules with empty space between them. Activity of the gas molecules determines the properties which can be measured as temperature and pressure. Imagery of the activity of the molecules making up the atmosphere is helpful. Wikipedia, in the article on kinetic theory of gases, has a pretty good 2-D animation of the 3-D activity. It shows simulated molecules bouncing elastically off each other and the walls of the container. At any point in time, the speed (and energy) of the molecules ranges from zero to high values with the highest probability being towards the low end. (This fairly simple perception works well for consideration of macroscopic properties such as pressure, temperature and viscosity. A deeper understanding involves probability, molecule configuration, electric fields and the sometimes contentious Quantum Mechanics.)

The average amount of time that passes between absorption and emission of a photon by a molecule of CO2 in the atmosphere is about a second [5, 6]. (Common sense mandates that if this elapsed time was zero there would be no way to determine absorption)

The elapsed time between collisions between gaseous molecules at sea level average temperature and pressure averages about 0.0002 µsec [7]. Each collision results in transferring at least part of the momentum and energy, which a ghg molecule acquired from the photon, to neighboring molecules. After multiple collisions, essentially all of the added photonic energy becomes distributed among other molecules. The energy in all absorbed photons is thermalized.

Energy transfer at the gas molecule level is a mechanism which contributes to the misleadingly named greenhouse effect (GHE). Radiation energy travels from molecule to molecule at the speed of light but dwells in the ghg molecules for about a second. More ghg molecules means longer cumulative dwell time. This slows passage of the energy through the atmosphere so a steeper temperature gradient (higher surface temperature) is required to maintain the energy flux.


The process of sharing the energy with other molecules is thermal conduction in the gas. The process of absorbing photons and sharing the absorbed energy with other molecules is thermalization. Thermalized energy carries no identity of the molecule that absorbed it. Ghg molecules can absorb/emit only photons with certain quanta of energy. Energy itself is quantized at the extremely fine level of 6.626E-34 J s (Planck constant). For all practical purposes, the wavelength spectrum is continuous and there are no forbidden wavelengths of photons. But only terrestrial wavelength photons at about 15 microns can be absorbed/emitted by CO2 molecules.

Emission of electromagnetic radiation from both solid and liquid surfaces of the earth complies with the Planck spectrum (with emissivity ≈0.99) and Stephan-Boltzmann (T4) law. Most particles of clouds, smoke and aerosols emit similarly because they typically contain millions of molecules. Emission of radiation from gas molecules is different (the only gas molecules which can absorb/emit significantly at earth temperatures are ghg molecules i.e. they are IR-active). Emission from ghg is restricted to specific wavelengths which are characteristic for each ghg species. Potential for emission depends on the energy levels of individual molecules. Molecule energy levels are determined probabilistically according to the Maxwell-Boltzmann distribution. The average molecule energy level of the Maxwell-Boltzmann distribution exhibits as the temperature of the gas.

A ghg molecule which is above absolute zero will emit photons (of allowed wavelength for that molecule). This radiant thermal emission from a gas which contains both ghg and non-ghg involves, for lack of a better term, reverse-thermalization, where ghg which have been cooled by photon emission are rewarmed by thermal conduction from surrounding molecules. Reverse-thermalization might alternatively be explained by collisions at adequate energy level between like-species of ghg molecules. The end result at the macro level would look about the same (emission from the planet, except at the poles, is dominated by water vapor molecules).

Significant terrestrial thermal radiation is nearly all in the wavelength range 6.5-200 microns (wavenumber 1538-50/cm). (Note: Divide 104 by either one to get the other; e.g. 104/wavenumber 667 = 15 microns, 104/15 microns = wavenumber 667). An early report, which was useful for development of guided missiles, is helpful in understanding Quantum Mechanics as applied to water vapor and carbon dioxide in the atmosphere [45].

Radiation from the surface is very close to the Planck spectrum from a black body at the specified surface temperature. Below wavenumber 550 about 95% of thermal radiation that gets absorbed from the surface is absorbed by WV within the first meter; 99% within two meters. Above about 550/cm the absorption by WV declines rapidly and absorption by CO2 increases. But the population of CO2 is so low that even at the peak of CO2 absorption at 667/cm these levels of absorption are not reached for several km altitude. The energy absorbed by the CO2 is redirected to WV via thermalization. More CO2 simply means that the energy redirection occurs at lower altitudes.

Thermalized energy carries no identity of the molecule that absorbed it. The thermalized radiation warms the air, reducing its density, causing updrafts which are exploited by soaring birds, sailplanes, and occasionally hail. Updrafts are matched by downdrafts elsewhere, usually spread out but sometimes recognized by pilots and passengers as ‘air pockets’ and micro bursts.


5. Radiance calculated by MODTRAN6
MODTRAN6 [48] is a computer program developed for the Airforce Research Laboratory which (besides other things) can calculate the radiation flux at selected elevations in the atmosphere for specified constituents and conditions. It contains default values for several environments including the tropics and the 1976 Standard Atmosphere. Values for the rate-of-change of water vapor and atmospheric temperature vary with altitude for different latitudes and seasonal conditions as shown in MODTRAN documentation [28].

Figures 0.5 and 0.6 are typical graphs calculated by MODTRAN6. They show radiation flux absorbed by ghg other than water vapor (WV) being redirected to WV with increasing altitude. The redirection is quantified by the progressively increasing depth of the ‘notches’ at the characteristic wavenumber ranges for each ghg (except WV). Note that, for CO2, the redirection at 20 km is greatest and at 50 km and higher some of the flux returns to the CO2 molecules. Somewhere in the vicinity of 20 km some of the energy is redirected back to the CO2.


The ‘notches’ are evidence of energy redirection from CO2 to WV. Redirection is possible because of continuous thermalization and reverse-thermalization at all altitudes in the gaseous atmosphere.

Figure 0.5: TOA flux for 1976 Standard Atmosphere at MODTRAN6 default values.


Figure 0.6: TOA flux for tropics atmosphere at MODTRAN6 default values.

Figures 0.5 and 0.6 are somewhat misleading; especially below about 10 km, because they are accounting for the entire energy flux including that involving latent heat and solar energy absorbed directly by atmosphere and clouds. Approximately 161 W/m2 of solar energy reaches the solid and liquid surfaces of the planet [29]. Added to this is about 16 W/m2 of broad spectrum radiation from clouds which reaches the surface through the atmospheric window. The energy leaving the surface includes about 78 W/m2 from heat of vaporization of water (annual rainfall averages about a meter and what comes down had to have gone up). Another 17 W/m2 has been added by convective heat transfer, leaving 161 + 16 – 78 – 17 = 82 W/m2 in thermal radiation. This compares favorably to up radiation minus down radiation of 345 – 271 = 74 W/m2 at zero altitude as calculated by MODTRAN6 but is only about 73% of 382 – 269 = 113 W/m2 as calculated by MODTRAN.

The non-radiant flux is replaced with radiant flux and the solar energy that was absorbed by the atmosphere and clouds are incorporated with increasing altitude. Most of this takes place by about 10 km (32,808 ft) so the graphs at 20 km and higher should be reasonably valid for radiance.

Most of the photons emitted by the water vapor molecules are at wavelengths different from the comparatively narrow band that CO2 molecules can absorb. The greatly reduced population of water vapor molecules above about 10 km means fewer molecules capable of absorbing radiation from water vapor molecules at lower elevation so most of the outward directed photons at these lower wavenumbers go directly to space. Effectively, much of the terrestrial thermal radiation energy absorbed by CO2 (and other non-condensing ghg) is thermalized, redirected to, and radiated to space from water vapor molecules.

At very high altitudes, temperature, molecule spacing and time between collisions increases to where reverse-thermalization to CO2 (and O3) molecules becomes significant as does radiation from them to space. This causes the spikes at the nominal wavenumbers of CO2 and O3.

Results from both MODTRAN and MODTRAN6 contribute to a credible approximation of the transition of the sum of convective, latent and radiative energy flux at the surface to purely radiative energy flux at top-of-atmosphere (TOA). This approximation, using the default data in the codes for standard atmosphere, is shown in Figure 0.7.
Figure 0.7: Approximate transition of surface energy flux to purely radiant flux at TOA.

Figures 1 and 1.5 are typical graphs showing TOA thermal radiation from the planet. The TOA radiation from different locations on the planet can be decidedly different, e.g. as shown in Figure 9 of Reference [8]. Figure 1, here, might be over a temperate ocean and thus typical for much of earth’s surface. The area under the black trace is about 300 W/m2 which is somewhat more than the planet average of about 240 W/m2. Figure 1.5 is similar with area under red curve ≈269 W/m2.





Figure 1: Thermal radiation from below assessed from top-of-atmosphere. Lower wavenumber photons are lower energy. (original graph is from NASA [46])

Typical TOA emission spectra such as shown in Figures 1 and 1.5 show the ‘notch’ associated with the CO2 absorb/emit band. Notch development with increasing altitude was shown in Figures 0.5 and 0.6. Existence of this notch demonstrates that terrestrial radiation in this wavelength range is absorbed by IR-active gases which include water vapor and the ghg which do not condense in the atmosphere such as CO2

Perhaps not as obvious, the presence of the 600-740 wavenumber notch (~ beige area in Fig 1) along with smaller notch at ozone (O3) also demonstrates thermalization and that the radiation energy which was absorbed by trace IR-active molecules was thermalized and substantially redirected to the absorb/emit lines of water vapor molecules. The WV molecules, progressively with altitude, emit the radiation energy to space.

Water vapor is greatly reduced at higher altitudes (>~10 km), which allows some reverse-thermalization of the radiant flux back to CO2 at the wavenumber range 600-740. The TOA notch is about 12% less deep above 50 km than it was at 20 km. The approximate 18 W/m2 (in Fig 1) which is not reverse-thermalized back to the notch explains the reduced flux at the notch.


The energy entering the atmosphere from the surface matches very closely the Planck spectrum for the temperature of the surface and emissivity about 0.99. In Figure 1, the flux through the atmospheric window indicates a surface temperature of about 293 K. For wavenumbers 600-740 /cm the power (energy rate) at TOA if no CO2 is (MODTRAN at same total flux) about 0.33 W / m^2 / cm^-1 * 140 cm^-1 ≈ 46 W/m^2. The power radiating at the notch from the tropopause and above is about 0.2 W /m^2 / cm^-1 * 140 cm^-1 = 28 W/m^2. The 46 – 28 = 18 W/m^2 that is not emitted at the wavenumber range 600-740 has to be emitted at other wavenumbers. The redirected power which is emitted in the wavenumber range 600-740 (28 W/m^2) could be from both CO2 and water vapor. 



Figure 1.5: Typical TOA radiant emission.


An ‘experiment’ demonstrating the effect of reduced water vapor in the atmosphere already exists. Near the poles, the extremely low temperatures result in very low water vapor content while the CO2 level is about the same as everywhere else. With few water vapor molecules available to emit radiation, more of the TOA emission is from CO2 molecules near 15 microns as shown in Figure 9 of Ref [8]. Because the GHE at the poles is dominated by CO2 , an increase in CO2 causes the slight warming at the poles which has been observed.

Approximately 98% of dry atmospheric molecules are non-ghg; nearly all nitrogen and oxygen with about 1% argon. Near the surface, they are substantially warmed by thermalization of the photonic energy absorbed by the ghg molecules and, at higher altitudes, cooled by reverse-thermalization back to the ghg molecules which radiate the energy to space.


6. Water vapor
Average measured global atmospheric water vapor (total from surface to TOA) over the years is provided here at Figure 3. Measured CO2 level is at Figure 7.

WV increase is a cause of warming (average global temperature increasing) because it is a ghg. Part of WV increase is a result of surface water warming because its saturation vapor pressure increases with temperature. The saturation vapor pressure increase causes an increase in the rate of WV molecules being forced into the atmosphere (when the atmosphere at ground level is less than saturated with WV which is usually the case). An additional source of WV increase is human activity, especially irrigation. This is discussed further in Section 9.

 In the atmosphere, condensed water can exist as water, ice or super-cooled water [56] (super-cooled water is liquid water below 0.0 °C). Accurate numerical values for saturation vapor pressure of liquid water [57] and ice [58] are graphed in Figure 1.7. Saturation vapor pressure for super-cooled water can be calculated using the Bolton equation [55]. The Bolton equation for saturation vapor pressure in kPa vs temperature in C is

 p = 0.6112 * e^(17.67 * T / (T+243.5))                      (z1)

As shown in Figure 1.7, saturation vapor pressure increases progressively with temperature. Of interest is the % increase in saturation vapor pressure per degree increase in temperature. This is readily calculated from the numerical data for both liquid water and ice from:

1/1 increase/Tave = (pj – p(j-1)/(Tj – T(j-1))/Tave                        (z2)

Where,

1/1 = %/100

j and (j – 1) are adjacent values in the table

Tave = average temperature of the adjacent values.

 

The same thing for super-cooled water is obtained using the first derivative of the Bolton equation which is

dp/dT = p * 17.67 * 243.5/(T+243.5)^2                     (z3).

This, divided by p to get the 1/1 value curve, is shown in the bottom graph at Fig. 1.7.

Slightly more accurate formulae for calculating saturation vapor pressure were derived by Jianhua Huang [62].

Saturation vapor pressure depends ONLY on the temperature of the ice or liquid water. The 1/1 change in saturation vapor pressure per Celsius degree for water, ice and super-cooled water are shown in the lower graph of Figure 1.7. 

Figure 1.7: Saturation vapor pressure of ice &water and fractional rate of change per C degree change vs temperature.

The atmospheric temperature decreases with altitude so the accommodation for WV increases with altitude to about 12%/C° at the tropopause (°C is a temperature, C° is a temperature difference and is used interchangeably with K for Kelvin degree). This can result in the perhaps counterintuitive condition that as surface temperature increases, WV (specific humidity) increases but accommodation for WV in the atmosphere increases even more, so relative humidity decreases.

 Based on ocean temperatures from [59], the area-weighted change in saturation vapor pressure per C degree at sea level is about 0.0633 / C°. The amount of compounding is unknown but cannot be greater than 0.0633+0.0633^2+0.0633^3+… = 0.0676/C°. It is conservatively estimated to be about 0.067/C° = 6.7%/C°

If there is no additional source of WV, the percent change of atmospheric water vapor is assumed to vary directly with the percent change in liquid water saturation vapor pressure.

Figure 2: Water vapor declines with latitude and rapidly with altitude. [9] (original from NASA)

All absorbed radiation is thermalized i.e. the absorbed energy is shared with surrounding molecules. Only WV molecules can absorb/emit radiation in the wavenumber range 25 to 550/cm.

At the tropopause it is very cold, about 50 degrees Celsius below zero (-58 °F). At this low temperature nearly all WV molecules have condensed (or cooled by radiation and descended) leaving only about 149 ppmv WV above this altitude compared to a global average of about 10,000 ppmv at ground level. Accounting also for reduced pressure to about 0.26 atm results in a WV population per unit volume above the tropopause of 149/10000 * 0.26 ≈ 1/258 of what it is at ground level. This huge gradation in WV molecules favors outward radiation; with increasing amounts escaping directly to space with increasing altitude.

The ‘hash’ in graphs of TOA radiation flux vs wavenumbers, e.g. Figure 1 or 1.5, is evidence of radiation originating from a range of altitudes (temperatures).

Although WV is reduced to 149 ppmv above the tropopause, CO2 remains at its ground level ppmv value. Therefore, above the tropopause most of the remaining thermalized energy is reverse-thermalized to CO2 molecules which emit it to space.

At very high altitudes, increased molecule spacing and greatly diminished water vapor molecules result in the sharp peaks at nominal absorb/emit wavelengths of non-condensing ghg (See Figure 1 & 1.5).

A common observation which demonstrates that water vapor increase contributes to global warming is cloudless nights cool faster and farther when absolute water vapor content of the atmosphere is lower. This simple observation also demonstrates the existence of the misleadingly named greenhouse effect (GHE) and that water vapor is a ghg.


7. Effect of water vapor increase

Energy goes from ghg molecule to ghg molecule by electromagnetic radiation (EMR) at the speed of light. But the photon energy absorbed by a molecule spends an average of about a second in the molecule. More ghg molecules means more time spent absorbed and less time at the speed of light. That is why more water vapor (and at the poles more CO2) slows the rate of energy flow from the surface to space so the surface temperature must increase to maintain the energy flux.

The average global water vapor increase since 1895-2018 is about 10% (Fig 3). The average global temperature increase 1895 to 2018 attributable to WV increase according to Equation (1) and temperature data thru 2018 is about 0.70 K (Table 1).

The procedure to determine the effect of water vapor increase using MODTRAN [47] is to determine a total radiant flux for a particular base condition (scale factor = 1), apply the scale on water vapor and then by trial and error change to temperature offset, determine the surface temperature which produces the same total radiant flux for the condition of study as for the base condition. The difference in surface temperature between the base condition and the condition under study is the effect on surface temperature of the change to WV. The default levels for all noncondensing ghg (e.g. 400 ppmv for CO2) were used for all cases to prevent any calculated effect from a change to them.

Water vapor change is investigated at three scales. All available conditions were evaluated with vapor pressure (VP) held constant and also with relative humidity (RH) held constant. In all conditions except 10% increase for tropical atmosphere, the temperature change was greater with RH held constant. The base condition always gave exactly the same results for RH as VP except Flux for tropical atmosphere.


Table 0.5 Temperature change as determined by MODTRAN [47] for several conditions. All are clear sky except one with light rain and one using the std cirrus model as noted.
Condition
Flux, W/m2
Held constant
WV scale
Surface temp, K
∆T, K
Tropical atmosphere 
(up to 20 deg latitude, 34.2% of area)
298.52
VP
1
299.7
base
1.1
300.59
0.89
1.2
301.07
1.37
296.824
RH
1
299.7
base
1.1
300.53
0.83
1.2
301.38
1.68
Mid-latitude summer 
(20 to 60 deg latitude, 52.4% of area)
288.064
VP
1
294.2
base
1.1
294.595
0.395
1.2
294.97
0.77
RH
1.1
294.78
0.58
1.2
295.35
1.15
Mid-latitude winter
234.118
VP
1
272.2
base
1.1
272.43
0.23
1.2
272.645
0.445
RH
1.1
272.49
0.29
1.2
272.765
0.565
Subarctic summer 
(above 60 deg latitude, 13.4% of area)
269.538
VP
1
287.2
base
1.1
287.55
0.35
1.2
287.88
0.68
RH
1.1
287.69
0.49
1.2
288.165
0.965
Subarctic winter
201.305
VP
1
257.2
Base
1.1
257.37
0.17
1.2
257.52
0.32
RH
1.1
257.4
0.2
1.2
257.59
0.39
1976 Std atmosphere
266.272
VP
1
288.2
base
1.1
288.55
0.35
1.2
288.88
0.68
RH
1.1
288.7
0.5
1.2
289.19
0.99
Mid-latitude summer, light rain & nimbo-stratus
283.542
VP
1
294.2
base
1.1
294.55
0.35
1.2
294.88
0.68
RH
1.1
294.69
0.49
1.2
295.17
0.97
Mid-latitude summer, Std cirrus model
270.448
VP
1
294.2
base
1.1
294.56
0.36
1.2
294.9
0.7
RH
1.1
294.71
0.51
1.2
295.2
1.0


MODTRAN6 [48] provides similar capability but imposes predefined WV vs altitude profiles for each condition in place of selecting to hold either VP or RH constant. The WV profiles incorporate the constraint that humidity cannot exceed 100%. The already high WV in tropical areas is apparently not compatible with increasing it by 10%. Similar plots to those made by MODTRAN are obtained by setting MODTRAN6 to radiance, sensor altitude to 99 km, spectral range wavenumbers 250 to 1500, and resolution to 1.2/cm.

Table 0.6: Summary of results using MODTRAN6.

Condition
Flux, W/m2
WC mult
Water column, atm-cm
Temperature, K
∆T, K
Mid-latitude summer
(52.4 %)
269.594
1
3635.9
294.2
base
1.1
3999.49
295.636
1.436
Mid-latitude winter
249.969
1
1059.7
272.2
Base
1.1
1165.67
272.34
0.14
Tropics (34.2 %)
268.546
1
5119.4
299.7
base
1.1
5631.34
304.325
4.625
Subarctic summer
(13.4 %)
265.441
1
2589.4
287.2
base
1.1
2848.34
287.81
0.61
Subarctic winter
231.737
1
517.73
257.2
base
1.1
569.503
257.196
-0.004
US standard, 1976
269.339
1
1762.3
288.15
base
1.1
1938.53
288.69
0.54


Area weighted, winter/summer average results in ∆T of 0.5 K for MODTRAN and 0.75 K (excluding tropics) for MODTRAN6 for 10% water vapor increase. These compare to 0.70 K for Equation (1) (Table 1).


8. Water vapor change with time
Water vapor is the ghg which makes earth warm enough for life as we know it. Increased atmospheric water vapor contributes to planet warming. Water vapor molecules are more effective at absorbing terrestrial thermal radiation than CO2 molecules. Humanity’s contribution to atmospheric water vapor increase is primarily (≈ 90%) as a result of increased irrigation (Figure 3.5), with comparatively small contribution from cooling towers at electricity generating facilities. Fossil fuels make an insignificant contribution to water vapor. Switching to ‘renewables’ will have little, if any, effect on climate.

Because water vapor is a ghg, increased water vapor causes the planet to warm, which further increases saturation vapor pressure of liquid water and therefore further increases water vapor so there is a cumulative effect. In control system analysis and electric circuit analysis as done by engineers, this is called positive feedback and is quantified by a dimensionless number which is the ratio minus one of the change with feedback to the change if there was no feedback. This cumulative effect also increases the rate of cooldown. (The term ‘feedback’ has a different meaning to Climate Scientists and is quantified in units of W/m2).

 Planet warming increases the saturation vapor pressure of water (Figure 1.7) contributing to the water vapor increase. At present water vapor is increasing faster than possible based on AGT increase alone. Global temperature increase Dec 1978 – Feb 2022 from the UAH [4] trend is about 0.135 K per decade (this automatically includes any feedback effect). The assessment in Sect 6 resulted in the calculation of average saturation vapor pressure increase with temperature of 6.7% per degree C including compounding. The percent increase in water vapor is assumed to be about the same as the percent increase in saturation vapor pressure. Percent increase in water vapor due to temperature increase = 0.135 * 6.7% = 0.9045% per decade.

 Measured % increase of Total Precipitable Water (TPW) from Figure 3 is 0.04313/29 = 1.49% per decade. Thus measured increase in WV is about 1.49/0.9045 ≈ 1.65 times the amount for liquid water temperature increase alone..

 The comparison of the measured WV change with the WV change calculated from temperature change over the full time period since Jan 1988 can be calculated incrementally and shown graphically. The file for calculated change is generated in EXCEL where each row contains:

 

WVn = WV(n-1) + (Tn – T(n-1))* R * (WV(n-1) +F)                        (z4)

 

Where:

WVn = calculated WV in month n, kg/m^2

Tn = temperature anomaly in month n, K

R = effective rate of WV increase resulting from feedback of temperature increase, 0.067/K (= 6.7 %/K)

F = added to avoid circular reference of (WV(n-1)+WVn)/2. F is calculated as an increase to each month equal to half a month at the final slope. This requires iteration.

For GISS [23] thru Dec 2021, F = 0.0241234/24 = 0.001005 kg/m^2/month. Slope at F = 0 is 0.0241226. Effect over 34 yr = (0.0241234 - 0.0241226)*34 = 0.0000027 kg/m^2

 The starting calculated WV is adjusted to make the starting trends the same.

 The results of this algorithm are shown along with the actual WV anomaly measured and reported by NASA/RSS [11] (plus 28.73). The calculated WV increase since Jan 1988 for GISS reported temperatures [23] is shown in Figure 2.8 and for UAH reported temperatures in Figure 2.9. Figure 2.8 also shows the linear trends assuming three different values for average relative humidity remaining constant with increasing air temperature (as calculated in GCMs).



Figure 2.8: Measured WV vs calculated WV based on GISS reported average global temperatures [23]. (R = 0.067/K) and trends assuming constant relative humidity.


Figure 2.9: Measured WV vs calculated WV based on UAH reported average global temperatures [4]. (R = 0.067/K

The average ratio of measured WV/calculated WV is obtained by the ratio of the slopes of the regression lines. The ratio for GISS with R = 0.067 is 0.041882/0.0241234= 1.736. The ratio for UAH is 0.041882/0.015051 = 2.78. The observation that measured WV increased faster than the determination using temperature demonstrates that there has to be an additional source of WV above that resulting from just temperature increase.

A corroboration of the long-term temperature trend is as follows: Assume that at the beginning of the warm up the temperature increase was caused by something else. Then the WV increase can be calculated from that temperature increase using the vapor pressure vs temperature for water and the assumption that % increase in WV = % increase in vapor pressure. But the WV has increased more than that so there has to be an additional source of WV. The additional source of WV (nearly all from irrigation) is the something else that produced the initial warming. 

 The increased water vapor also eventually causes increased cloud cover and/or lower average cloud altitude which counters temperature increase and will eventually limit it. Sustained increase of only about 1.7% of cloud area would result in an eventual temperature decline of 0.5 °C [22]. 

More water vapor in the atmosphere means more warming. Water vapor exhibits a logarithmic decline in absorption effect for equal added increments of water vapor (Fig. 3 of Ref. [12]).

Sect 9 of [61] includes a graph which shows the average global temperature increase vs average global WV increase.

Essentially all of the ghg effect on earth comes from water vapor. Clear sky water vapor measurements over the non-ice-covered oceans in the form of total precipitable water (TPW) have been made since 1988 by Remote Sensing Systems (NASA/RSS) [11]. A graph of this measured ‘global’ average anomaly data, with a reference value of 28.73 added, is shown in the left graph of Figure 3. This data is extrapolated earlier using CO2 level as a proxy, with the expression kg/m^2 TPW = 4.5247 * ppmvCO2^0.31286. This produced a match with the magnitude and slope of the trend at the intersection of the extrapolation with the measured values. The result is the right-hand graph of Figure 3 which shows approximately 7% increase 1960-2005.


Figure 3: Average clear air Total Precipitable Water over all non-ice-covered oceans as measured by NASA/RSS using satellite based instrumentation and with extrapolation by me. (Left graph is by month, right graph is by year average.). Estimated future assumes continuation of the slope of the monthly trend.

Clouds (average emissivity about 0.5) consist of solid and/or liquid water particles that radiate approximately according to Planck spectrum and Stephan-Boltzmann (S-B) law (each particle contains millions of molecules).


9. World Sources of Increased Water Vapor
Irrigation, industrialization, and, increasing population are causing the rise in atmospheric water vapor (WV) above that from feedback (engineering definition of feedback) due to liquid water temperature increase. A survey of available on-line literature provides direct and indirect quantification of significant global sources of the extra increase.

Transportation fuel, linearly interpolated to 2017, amounts to 113E15 BTU/y [31]. Energy content of a typical liquid fuel is 115,000 BTU/gal [32]. Liquid fuels weigh about 6.073 lb/gal = 2.75 kg/gal. Therefore transportation fuels amount to
113E15 * 2.75/115000 = 2.7E12 kg fuel/y                 (a)

About 1.42 kg of WV is produced for each kg of liquid fuel [32] so the amount of WV produced by transportation is
2.7E12 * 1.42 = 3.8E12 kg WV/y                  (b)

World electricity generation is now about 25,000 TWH/y [33]. At an average efficiency of 50% this requires a thermal input of 50,000 TWH/yr. Fuel source fractions of energy [34] in 2017 are approximately 0.38 coal, 0.36 natural gas and 0.26 non fossil fuel.

Coal combustion produces about 0.4 kg WV/kg coal [35]. Energy content of bituminous coal is about 8200Wh/kg [36]. The amount of WV resulting from burning coal to generate electricity is then
50E15 * 0.38 * 0.4/8200 = 0.93E12 kg WV/y                       (c)

The amount of WV produced by natural gas (nearly all methane, CH4) is readily calculated from the dominant chemical reaction
CH4 + 2O2 => CO2 + 2H2O                (d)

Where a mole of methane weighs about 16 g and the two moles of WV weigh about 18 g each.
Natural gas energy content is about 15,400 Wh/kg [36]. The amount of WV resulting from burning natural gas to generate electricity is then
50E15 * 0.36 *36/16/15400 = 2.6E12 kg WV/y                    (e)

The total WV from all fossil fuel used to generate electricity is then
0.9E12 + 2.6E12 = 3.5E12 kg WV/y                         (f)

Waste energy during electricity generation can be approximately accounted for by evaporation of water in cooling towers, etc. At 50% efficiency the waste energy is the same as the energy in the electricity produced, 25,000 TWH/yr = 25E12 kWh/y.
Latent heat of water = 2257 kJ/kg = 0.627 kWh/kg = 1.594 kg/kWh.
The amount of WV from waste heat (cooling tower, etc.) during electricity generation is then
25E12 * 1.594 = 39.8E12 kg WV/y                           (g)


Irrigation is by far the largest source of WV. The increase in irrigation is indicated by the increase in withdrawal for agriculture as shown in Figure 3.5 [37].
Figure 3.5: Global water withdrawal includes both ground water and surface water [37]

The total agricultural area equipped for irrigation in 2009 was 311E10 m2 of which 84% were actually being irrigated [38]. Estimating an increase of 2% to 2017, the total area being irrigated is now about
311E10 * 0.84 * 1.02 = 266E10 m2                            (h)
This is more than 4 times the area of France.

Total annual fresh water withdrawal (both ground and surface) is now 3,986 km3 = 3.986E15 kg/y [39]. Of this, about 70% is for agricultural use [40]. This works out to
3.986E15 * 0.7/266E10 = 1052 kg/m2/y ≈ 1 m/y                     (i)
which appears reasonable because average rainfall for the planet is about 1 m/y.

Evapotranspiration, WV from plants and landscape, is discussed in the ‘thematic discussion’ of Aquastat [37]. From there, the amount of precipitation on land is 110,000 km3 of which the fraction evapotranspirated is 0.56 + 0.05 = 0.61. Given the planet surface area of 510.1E6 km2, and land fraction of 0.29 this results in the equivalent liquid depth of the total amount of water leaving the surface as WV as
110,000 * 0.61/0.29/510.1E6 = 0.00045 km = 0.45 m                       (j)

Water weighs 1000 kg/m3 so evapotranspiration amounts to 450 kg/m2

Worldwide about 86% of irrigated area is flood irrigated [54]. To simplify calculation, assume all irrigation is flood irrigation approximated as furrow type [41]. Optimum frequency is to flood the furrows about every 10 days [42]. Thus about half the area is covered by water 10% of the time where evaporation from the water is about one meter per year [43] and the rest of the time, the additional evaporation is assumed to be according to the calculated evapotranspiration. Evapotranspiration prior to irrigation must have been low or irrigation would not be done. Evapotranspiration with irrigation, to be cost effective, is most likely to be much more than calculated. These two uncertainties are assumed to approximately cancel each other. A further assumption is that, on average, irrigation is applied for about 1/3 year. The total amount of WV resulting from irrigation is then

[(0.1 * (1 + 0.45)/2 + 0.9 * 0.45) * 266E10]/3 = 42.3E10 m3 = 42.3E13 kg/y           (k)

 

These calculations are summarized in Table 1

 

Water vapor source

E13 kg/y

%

Irrigation

42.3

90.0 %

Transportation fuel

0.4

0.8 %

Fossil fuel for electricity generation

0.3

0.7 %

Cooling towers, etc. for electricity generation

4.0

8.5%

Total

47.0

100 %

Table 0: Summary of contributions to atmospheric water vapor.

 

From Table 0, approximately 42.3/47 = 0.90 or 90% of atmospheric WV increase above that due to feedback (engineering definition of feedback) from liquid water temperature increase results from irrigation. WV added by irrigation might be particularly influential because it is added at locations where natural WV is low aiding evaporation and water is shallow so it would warm quickly.

The assessment here that irrigation accounts for 90% of WV corroborates what Shiklomanov 1997 determined as mentioned by Doll in 2002 [60].

Given the earth area of 510E12 m^2 and average annual precipitation of about a meter or 1000 kg/m^2 the increased water use, mostly for irrigation, results in 47E13/5.1E17 = 0.0009 = 0.09 % equivalent increase in global precipitation.


10. Effect of TPW increase on clouds (added 9/17/18)

As TPW is increased, condensation (cloud formation) could occur at lower average cloud altitude and thus higher temperature except that increased atmospheric temperature increases the accommodation for WV which would reduce the formation of clouds. If lower altitude, higher temperature clouds occur, they would radiate more energy to the cosmic background and partially counter the warming from increased water vapor.


11. Approximate effect on the planet of the net of ocean surface temperature (SST)

The average global ocean surface temperature oscillation is only about ±1/7 C°. It is defined to not significantly add or remove planet energy. The net influence of SST oscillation on reported AGT is defined as α-trend. In the decades immediately prior to 1941 the amplitude range of the trends was not significantly influenced by change to any candidate internal forcing effect; so the observed amplitude of the effect on AGT of the net ocean surface temperature trend anomaly then, must be approximately the same as the amplitude of the part of the AGT trend anomaly due to ocean oscillations since then. This part is A ≈ 0.3 C° total highest-to-lowest extent with a period of approximately 69.6 years (verified by high R2 in Table 1).

 The measured AGT trajectory (Figure 9) suggests that the least-biased simple wave form of the net effect of average surface temperature oscillation of all of the oceans on average global temperature is approximated by the sine wave:

 Toceans = A/2 * sin((y – 1925) * 2 * π/69.6)                  (z5)

 Where:

Toceans is the effective temperature anomaly of the planet surface as a result of the average ocean surface temperature oscillation, C°.

A = Effective peak to peak amplitude of the ocean oscillation average, C°.

 This puts the most recent maximum in 2012. Maximum R2 in equation (1) was obtained with A=0.298. Equation (z5) is graphed in Figure 4 for A=0.298.

Figure 4: Effect on planet surface temperature of ocean surface temperature oscillations (α-trend). They do not on the long term significantly affect the bulk energy of the planet.

In previous releases of this analysis, the effect of ocean oscillations on planet surface temperature was approximated with a saw-tooth wave form of period 64 years.


12. Comparison of approximation with ‘named’ ocean cycles
Named ocean cycles include, in the Pacific north of 20N, Pacific Decadal Oscillation (PDO); in the equatorial Pacific, El Nino Southern Oscillation (ENSO); and in the north Atlantic, Atlantic Multidecadal Oscillation (AMO).

Ocean cycles are perceived to contribute to AGT in two ways: The first is the direct measurement of sea surface temperature (SST). The second is warmer SST increases atmospheric water vapor which acts as a forcing and therefore has a time-integral effect on temperature. The approximation Equation (z5) accounts for both ways.

 Measured SST data is available for three named cycles: PDO index [17], ENSO 3.4 index [18] and AMO index. The poor correlation between the sine approximation (z5) and measured PDO is shown in Figure 5.

Figure 5: Comparison of idealized approximation of ocean cycle effect and the calculated effect from PDO and ENSO.

The AMO index [19] is formed from area-weighted and de-trended SST data. It is shown with two different amounts of smoothing in Figure 6 along with the good match with the sine approximation, (z5) with A = 0.298.

Figure 6: Comparison of idealized approximation of ocean cycle effect and the AMO index through 2020.


The high Coefficient of Determination in Table 1 and the comparison in Figure 6 corroborate the assumption that the sine profile defined by Equation (z5) provides adequate approximation of the net effect of all named and unnamed ocean cycles in the calculated AGT anomalies. The AMO index, which dominates the effect of ocean cycles, has not yet started the decline indicated by the approximation.


13. Atmospheric carbon dioxide
The level of atmospheric carbon dioxide (CO2) has been widely measured over the years. Values from ancient times were determined by measurements on gas bubbles which had been trapped in ice cores extracted from Antarctic glaciers [20]. Spatial variations between sources have been found to be inconsequential [2]. The best current source for atmospheric carbon dioxide level [21] is Mauna Loa, Hawaii. The left graph in Figure 7 provides insight as to the fraction of atmospheric CO2 for various times and conditions. The planet came perilously close to extinction of all land plants and animals due to the low level of CO2 at the end of the last glaciation. For plant growth, even at the current level the atmosphere is impoverished for CO2.
Figure 7: Atmospheric carbon dioxide levels.


The trajectory shape, including data back to 1700 from Law Dome (275 ppmv), was used as a proxy to extrapolate TPW back to 1700.


14. Climate sensitivity determined using MODTRAN6
MODTRAN6 calculates the radiation flux given inputs of water vapor, CO2, and surface temperature. The procedure to determine Climate Sensitivity (CS) to CO2 is simply to find the flux change separately from CO2 change and temperature change holding water vapor constant and then apply:

T/CO2 = T/Flux *Flux/CO2 * CO2/T

This gets the temperature change as a ratio to CO2 change. Multiply by the CO2 ppmv change to get the temperature change in K.

For the 1976 standard atmosphere;
Water column, atm, cm
CO2
Temperature, K
Flux
Partial derivative
1762.3
400
288.15
269.339

400
288.35
269.611
T/F = 0.2/0.272 = 0.735
407
288.15
269.320
ꝺF/CO2 = -0.019/7 = -0.00271
T/CO2 = 0.735 * -0.00271 * 400/288.15 = -0.00276  K/ppmv CO2
Because the temperature is reduced by added CO2 the temperature must be increased by the same amount to obtain the required flux. Climate Sensitivity, increasing pre-industrial CO2 = 275 ppmv by 275 ppmv, for this calculation is then CS = 275 * 0.00276 = +0.76 K.


The same procedure was applied to the other conditions which have default values available in MODTRAN6. Like zones were averaged winter and summer and zones weighted 34.2% tropics (to ±20 deg latitude), 52.4% mid latitude (±20 to ±60 deg latitude), and 13.4% subarctic (±>60 deg latitude). This calculation resulted in CS = 1.07 K for doubling CO2 to 550 ppmv from 275 ppmv. 

These calculations of CS use the default values in MODTRAN6. Assessments in Section 2 have determined that CS is not significantly different from zero.


15. Sunspot numbers
Sunspots have been regularly recorded since 1610. In 2015 historical (V1) SSN were reevaluated in light of current perceptions and more sensitive instruments and are designated as V2. The V2 SSN data set is used throughout this assessment. V2 SSN [15] are shown in Figure 8.

Sunspot numbers (SSN) are seen to be in cycles each lasting approximately 11 years. The current cycle, called 24, has been comparatively low, peaked in 2014, and is now (Aug 2018) near zero.

The Maunder Minimum (1645-1700), an era of extremely low SSN, was associated with the Little Ice Age. The Dalton Minimum (1790-1820) was a period of low SSN and low temperatures. An unnamed period of low SSN (1880-1930) was also accompanied by comparatively low temperatures.

An assessment of this is that sunspots are somehow related to the net energy retained by the planet, as indicated by coincident changes to the average global temperature trend. Fewer sunspots are associated with cooling, and more sunspots are associated with warming. Thus the hypothesis is made that SSN are proxies for the rate at which the planet accumulates (or loses) radiant energy over time. Therefore the time-integral of the SSN anomalies is a proxy for some of the amount of energy retained by the planet above or below breakeven.

Also, a lower solar cycle over a longer period might result in the same increase in energy retained by the planet as a higher solar cycle over a shorter period. Both amplitude and duration are inherently accounted for in Equation (1).

SSN change correlates with change to Total Solar Irradiance (TSI) so SSN anomaly could also be acting as proxy for TSI anomaly. Because AGT change has been found to correlate with SSN change, the SSN change might also act as a catalyst on some other factor (perhaps clouds [22]) which have a substantial effect on AGT. Considered forcings are a power thing so each yearly average power quantity is divided by the effective thermal capacitance to obtain the temperature increment (an energy thing). Ocean cycles are considered to contribute temperature directly; any accompanying forcing averages to zero over time. 
Figure 8: V2 SSN [15] thru 2020


16. AGT measurement data set
In the last few years, reported temperature data, especially land temperature data, have been changed by the reporting agencies. This detracts from their applicability in any correlation.

Rapid year-to-year changes in reported temperature anomalies are not physically possible for true energy change of the planet. The sharp peak in 2015-2016, which coincides with an extreme El Nino, is especially distorting. 

 A further bit of confusion is introduced by satellite determinations. Anomalies they report as AGT anomalies are actually for the lower troposphere (LT), have a different reference temperature (reported anomalies determined using satellite data are about 0.2 K lower), and appear to be somewhat more volatile (about 0.15 K further extremes than surface measurements) to changes in forcing.

The data set used for this assessment is the current (5/27/16) HadCRUT4 data set [23] extended through 2020. This set is shown in Figure 9.


Figure 9: HadCRUT4 data set as used here.


17. The AGT Model
Most modeling of global climate has been with Global Climate Models (GCMs) where physical laws are applied to a three-dimensional grid consisting of hundreds of thousands of discrete blocks (elements). Interactions between the discrete blocks are analyzed using super computers with an end result being calculation of the AGT trajectory. This might be described as a ‘bottom up’ approach. Although theoretically promising, multiple issues currently exist with this approach. Reference [13] discloses that essentially all of the more than 100 current GCMs are obviously faulty. The growing separation between calculated and measured AGT as shown at Figure 9 in Ref. [14] also suggests some factor is missing.

The approach in the analysis presented here is ‘top down’. This type of approach has been called ‘emergent structures analysis’. As described by Dr. Roy Spencer in his book The great global warming Blunder, “Rather than model the system from the bottom up with many building blocks, one looks at how the system as a whole behaves.” That approach is used here with strict compliance with physical laws.

The basis for the algorithm for assessment of AGT is the first law of thermodynamics, conservation of energy, applied to the entire planet as a single entity. Much of the available data are forcings or proxies for forcings which must be integrated (mathematically as in calculus, i.e. accumulated over time) to compute energy change. Energy change divided by effective thermal capacitance is temperature change. Temperature change is expressed as anomalies which are the differences between annual averages of measured temperatures and some baseline reference temperature; usually the average over a previous multiple-year time period. (Monthly anomalies, which are not used here, are referenced to previous average for the same month to account for seasonal norms.)

With CO2 rejected as a significant contributor to warming, the remaining factors to be explicitly considered are:
            Ocean cycles
            SSN
            WV
            the tiny net feedback from the degree or so increase in AGT which has occurred since the depths of the Little Ice Age.

These factors are combined in an EXCEL file, with coefficients (A,B,C,D,F) assigned to each factor. The file is set up for numerical integration in yearly steps, where any EXCEL row is as follows:

Tanom = Tanom(n-1) + A/2 * sin((yn – 1925) * 2 * π/69.6)– A/2 * sin((y(n-1) – 1925) * 2 * π/69.6)+ thcap-1 * {B*[Sy-Sref] + C*(lnTPWn - lnTPW(n-1)] – F * [(Ty/T1700)4 – 1]}         (1)

Where:
Tanom = Calculated average global temperature anomaly at the end of year y with respect to the baseline of the anomaly for the measured temperature data set, K
A = highest-to-lowest extent in the sine approximation of the net effect on planet AGT of all ocean cycles, K
y = year for which the net gain or loss in temperature (energy) is being calculated
n = any year
thcap = effective  thermal capacitance [1] of the planet = 17±7 W yr m-2 K-1
1895 = Selected beginning year of acceptably accurate worldwide temperature anomaly measurements to be compared to calculated anomalies.
B = combined proxy factor and influence coefficient for energy change due to sunspot number ‘anomaly’ change, W yr m-2
Sy = average daily v2 sunspot numbers [15,16] in year y
Sref = baseline for referencing SSN anomalies. Average v2 SSN 1640-2018.
C = combined proxy factor & influence coefficient for energy change due to TPW change, W yr m-2
TPWn = total precipitable water in year y, kg m-2 (from measurements and extrapolation for Fig 3)
F = 1 to account for change to S-B radiation from earth due to AGT change, W yr m-2
Ty = AGT calculated by adding the HadCRUT4 reference temperature to the reported anomaly, K
T1700 = AGT in 1700 = 288.05 K
D = starting value for Tanom (Tanom in 1699) which shifts the calculated trajectory vertically on the graph, without changing its shape, to best match the measured data, K.

The attribution of each factor is determined by maximizing the Coefficient of Determination, R2. Wikipedia has a good example for how to set up the file to determine R2.


18. Comments regarding the method used in developing and using Equation (1)
Selection of causal factors must be rational; projections before and after the matched period must be reasonable; and R2 must be high for results to be useful.

Equation (1) is calibrated against measured AGT to determine the partial attribution of each preselected main causal factor of climate change. The attribution of each factor is revealed by the derived coefficients which are adjusted by trial-and-error to obtain the highest R2. Best estimate for the coefficients is the set which produces the best match of calculated to measured AGT 1895-2020.


Several runs with different assumed values for Sref showed low sensitivity to this parameter and also indicated lower sensitivity to SSN than previously determined. The average SSN value, including 60 years of zero SSN preceding the start of the warmup in 1700, is 66. Setting Sref = 66 resulted in estimated AGT at the depths of the Little Ice Age (LIA) to be about 1.1 K below current measured AGT. This determination of LIA AGT is consistent with estimates of AGT at the depths of the LIA by others.

19. Results from the AGT model
By definition, energy change (the average power or forcing) for each year divided by effective thermal capacitance is temperature change for that year.

In all cases in this document coefficients (A, B, C, & D) are reported which achieved maximum R2. F=1 for all cases which accounts for increased temperature driving absorb/emit as well as the radiation which goes through the ‘atmospheric window’. Incremental convergence to maximum R2 is accomplished by sequentially and repeatedly adjusting the coefficients.

Measured temperature anomalies in Figure 10 are data as shown in Figure 9. The excellent match of the up and down trends since before 1900 of calculated and measured temperature anomalies, shown here in Figure 10, demonstrate the usefulness and validity of the calculations. All reported values since before 1900 are within the range ±2.5 sigma (±0.225 K) from the calculated trend. Note: The variation is not in the method, or the measuring instruments themselves, but results from the effectively roiling (at this tiny magnitude of temperature change) of the object of the measurements. [44]

Sunspot numbers for the years 2017-2020 are from Australian Bureau of Meteorology [16]. After 2020 the assumptions for the calculations are listed on the graph.

Some noteworthy volcanoes and the year they occurred are also shown on Figure 10. Prior to 1988 only an approximation of the WV trend is available so no consistent AGT response is observed. The slight reduction of measured temperature compared to calculated temperature in 1991 might have been contributed to by mount Pinatubo, Volcano Explosivity Index (VEI) = 6.

Much larger volcanoes, VEI 7 or more (each VEI number is ten times the intensity of the next lower number), can cause significant temporary global cooling from the added reflectivity of aerosols and airborne particulates. The Tambora eruption, VEI 7, which started on April 10, 1815 and continued to erupt for at least 6 months, was approximately ten times the magnitude of the next largest in recorded history and led to 1816 which has been referred to as ‘the year without a summer’. The cooling effect of that volcano exacerbated the already cool temperatures associated with the Dalton Minimum.

Figure 10: Measured average global temperature anomalies with calculated past and future trends using Sref = 66 and with V2 SSN. R2 = 0.924140.

Notice also in Figure 10 the small difference in prediction with and without sunspots. This is because AGT is dominated by WV and, if the WV continues to increase at the trend rate since 1988, it might prevent the temperature decline many are predicting as a result of the quiet sun.

Coefficients in Equation (1) which were determined by maximizing R2 identify contributions for each of the factors explicitly considered. Factors not explicitly considered (such as unaccounted for residual (apparently random) variation in reported annual measured temperature anomalies, aerosols, CO2, other non-condensing ghg, volcanoes, ice change, etc.) must find room in the unexplained residual, and/or by occupying a fraction of the effect otherwise accounted for by each of the factors explicitly considered. 

Values of the derived coefficients and other results are summarized in Table 1. Note that a coefficient of determination, R2 = 0.92414 means a correlation coefficient of 0.96. Each % cause includes the added amount from its effect on AGT.

Some of the variation of measured temperatures is removed by smoothing using a 5-year running average. The coefficients are unchanged but results in a higher R2. The resulting graph is shown in Figure 10.6 and the associated data as file # Aintsm in Table 1.

Figure 10.6: Measured average global temperature anomalies with calculated past and future trends using Sref = 66, V2 SSN and measured temperatures 5-year smoothed prior to maximizing R2. R2 = 0.969013. 

The effect of different values for thcap have also been investigated [53].


Table 1: A, B, C, D refer to coefficients in Equation 1. The column headed # is a code identifying the particular EXCEL file used.

#

Fig

Period end

OCEAN

A

SUN

B

TPW

C

Starting anomaly

D

R2

1895-end ΔTTPW K

% cause of 1909-end AGT change

sun

sea

TPW

Aint

10

2020

.298

.00102

117.3

-.382

.92414

.70

13.2

23.7

63.1

Aintsm

10.6

2020

.298

.00102

117.3

-.382

.969013

.70

13.2

23.7

63.1


As shown in Figure 11, the AGT increase trend referred to as Global Warming or Climate Change has been profoundly dominated by water vapor increase trend. The future AGT trend depends primarily on WV. Long term WV increase correlates with irrigation increase and is inherently self-limiting. AGT increase is therefore also inherently self-limiting. Short term WV fluctuation appears to vary randomly with ocean surface temperature roiling and occasional transients associated with el Nino/la Nina. The attribution to WV assumes continuation of prior TPW increase trend, 0.051 kg/m2/yr, which might not occur.
Figure 11: Attribution of each factor to temperature change as of 2020 for both cases # Aint & Aintsm.

The high attribution of WV to AGT is explained by the population gradient from global average of about 10,000 ppmv (1%) WV at surface (4% in the tropics) to 149 ppmv at top of tropopause. Most of the outward directed photons from WV molecules at high altitude but below the tropopause make it all the way to space. This is demonstrated by the ‘hash’ in radiation flux from the WV as shown in TOA graphs of radiation flux vs wavenumber, Figures 1 and 1.5.


20. Attribution quantified by sunspot number anomaly proxy.
Although the comparatively weak connection between AGT and SSN (which acts as a proxy for the influence of solar change on AGT change) is demonstrated, the mechanism by which this takes place remains somewhat speculative.

Various papers have been written that indicate how the solar magnetic field associated with sunspots can influence climate on earth. These papers posit that decreased sunspots are associated with decreased solar magnetic field which decreases the deflection of and therefore increases the flow of galactic cosmic rays on earth.

Henrik Svensmark, a Danish physicist, found that increased flow of galactic cosmic rays on earth caused increased low altitude (<3 km) clouds and planet cooling. An abstract of his 2000 paper is at [24]. Marsden and Lingenfelter also report this in the summary of their 2003 paper [25] where they make the statement “…solar activity increases…providing more shielding…less low-level cloud cover… increase surface air temperature.” These findings have been further corroborated by the cloud nucleation experiments [26] at CERN.

These papers [24, 25] associated the increased low-altitude clouds with increased albedo leading to lower temperatures. Increased low altitude clouds would also result in lower average cloud altitude and therefore higher average cloud temperature. Although clouds are commonly acknowledged to increase albedo, they also radiate energy to space so increasing their temperature increases S-B radiation to space which would cause the planet to cool. Increased albedo reduces the energy received by the planet and increased radiation to space reduces the energy of the planet. Thus the two effects work together to contribute to AGT change of the planet.

A contributing or possibly alternate speculation is that clouds might also be affected by solar wind. End result is the same: Sunspot number anomalies are proxies for attribution of solar change to AGT change.

Simple analyses [22] indicate that either an increase of approximately 186 meters in average cloud altitude or a decrease of average albedo from 0.3 to the very slightly reduced value of 0.2928 would account for all of the 20th century increase in AGT of 0.74 K. Because the cloud effects work together, most of AGT change is from WV change, and part of the temperature change is due to ocean oscillation (low in 1901, 0.17 higher in 2000), substantially less cloud change would suffice.


21. Hind Cast Estimate
Average global temperatures were not directly measured in 1700 (accurate thermometers had not been invented yet). Recent estimates, using proxies, are few. The temperature anomaly trend that Equation (1) calculates for that time is consistent with other estimates.

As a possibility, the period and amplitude of oscillations attributed to ocean cycles demonstrated to be valid after 1895 are assumed to maintain back to 1700. Equation (1) begins integration in 1700 with D as the beginning of the combined value.

Temperature anomalies thus calculated, estimate likely trends since 1700 and actual trends of reported temperatures since they have been accurately measured world wide.  This assessment is shown in Figure 12.
Figure 12: Calculated temperature anomalies using Equation (1) with the same coefficients as for Figure 10.6 and V2 SSN. Measured temperature anomalies from Figure 9 and anomaly range estimates determined by Loehle are superimposed. 

A survey [27] of non-tree-ring global temperature estimates was conducted by Loehle including some for a period after 1700. Simplifications of the 95% limits found by Loehle are also shown on Figure 12. The spread between the upper and lower 95% limits are fixed, but, since the anomaly reference temperatures might be different, the limits are adjusted vertically to approximately bracket the values calculated using Equation (1). The fit appears reasonable considering the uncertainty of all values.

Calculated temperature anomalies look reasonable back to 1700. They qualitatively agree with Vostok, Antarctica ice core data but decidedly differ from Sargasso Sea estimates during that time (see the graph for the last 1000 years in Reference [2]). Worldwide assessments of average global temperature, that far back, are sparse and speculative. Ocean oscillations might also have been different from assumed.


22. Step changes in AGT are not possible
Interpretation of a reported sudden AGT increase (or decrease) as planet energy increase (or decrease) is physically impossible because of the huge effective thermal capacitance which results in a 5-year time constant [1] for thermal response of the planet to a step change in forcing.



23. Caveats on Predictions
1. A candidate cause must not have already been ruled out. CO2 was ruled out in Section 2.

2. Sunspot numbers are a proxy for both the influence of TSI and also, as found by Svensmark [24], the influence of clouds. Records for neither TSI nor cloud cover extend back to the LIA.

3. AGT is very sensitive to total cloud cover [22] which in turn varies with water vapor content (Fig 3), sunspots, surface temperature and atmosphere temperature. 

4. Global average ocean cycle surface temperatures (α trend) are believed to vary in period and intensity in response to as yet uncertain cause. The approximation here is working well since 1895. Future measurement might indicate a different period and/or better wave form or better yet an ocean cycle prediction based on planetary cycles or other predictable natural phenomenon. In any event, their attribution is cyclic, i.e. no significant net attribution on the long term.

24. Conclusions
Three factors explain essentially all of AGT change since before 1900. They are ocean cycles, accounted for with an approximation, solar influence quantified by a proxy which is the SSN anomaly and, the gain in atmospheric water vapor measured since Jan, 1988 and extrapolated earlier using measured CO2 as a proxy.

Others have looked at only amplitude or only duration factors for solar cycles and got poor correlations with average global temperature. As sometimes done, this procedure violates the relation between math and the physical world. The excellent (and computationally valid) correlation comes by accounting for both amplitude and duration, which is done in Equation (1). Prediction of future sunspot numbers more than a decade or so into the future has not yet been confidently done.

The β-trend is the estimated true average global temperature trend (the net average global energy trend) during the planet warm up from the depths of the Little Ice Age.

The net effect of ocean oscillations is to cause the surface temperature α-trend to oscillate above and below the β-trend. Equation (1) accounts for both trends. Measured ocean cycles appear to have not yet started the decline assumed in the approximation.

Warming attributed to increasing water vapor explains most of past warming and the possibly still rising measured AGT trend in spite of declining sunspot and ocean cycle forcings. Increasing WV might delay or even prevent global cooling.

The measured water vapor increase has been markedly more than expected from liquid surface water temperature increase alone. The extra increase appears to be mostly (about 90%) from increased irrigation. WV increase is ultimately self-limiting.

Long term (decades) prediction of average global temperatures depends weakly on long term prediction of sunspot numbers. Longer term (centuries) will depend partly on prediction of longer term variation in solar output.

Humanity has wasted over a trillion dollars in failed attempts using super computers to demonstrate that added atmospheric CO2 is a primary cause of global warming and in misguided activities to try to do something about it. An unfunded engineer, using only a desk top computer, applying a little science and some engineering, discovered a simple equation that unveils the mystery of global warming and describes what actually drives average global temperature. The refined version of the equation is presented herein as Equation (1).


25. References:

1. Effective thermal capacitance & time constant: Schwartz, Stephen E., (2007) Heat capacity, time constant, and sensitivity of earth’s climate system, J. Geophys. Res., vol. 113, Issue D15102, doi:10.1029/2007JD009373 

2. 2008 assessment of non-condensing ghg  https://www.researchgate.net/publication/350530743_Historical_Data_on_Global_Warming_provided_by_US_Government_Agencies

3. Phanerozoic AGT & CO2: http://www.geocraft.com/WVFossils/Carboniferous_climate.html

4. UAH v 6.0 numerical data http://www.nsstc.uah.edu/data/msu/v6.0/tlt/uahncdc_lt_6.0.txt

5. Average elapsed time to emit a photon https://sealevel.info/Happer_UNC_2014-09-08/Another_question.html

6. Average elapsed time to emit a photon http://rabett.blogspot.com/2013/04/this-is-where-eli-came-in.html.      

7. Time between molecule collisions: http://hyperphysics.phy-astr.gsu.edu/hbase/kinetic/frecol.html

        8. Barrett TOA radiation http://www.warwickhughes.com/papers/barrett_ee05.pdf

9. Water vapor vs altitude http://homeclimateanalysis.blogspot.com/2010/01/earth-radiator.html

10. EPA GWP  https://www3.epa.gov/climatechange/ghgemissions/gwps.html

11. NASA/RSS TPW through Dec 2022 https://data.remss.com/vapor/monthly_1deg/tpw_v07r02_198801_202212.time_series.txt 

12. Willis’ TPW graph: https://wattsupwiththat.com/2016/07/25/precipitable-water

13. Epic fail of ‘consensus’ method http://www.drroyspencer.com/2013/06/still-epic-fail-73-climate-models-vs-measurements-running-5-year-means

14. Analysis with V1 SSN sans water vapor: Pangburn 2014, Energy & Environment V25, No. 8 1455-1471

15. V2 sunspot numbers http://www.sidc.be/silso/datafiles

        16. Australian Bureau of Meteorology: http://www.sws.bom.gov.au/Solar/1/6

17. PDO https://www.ncdc.noaa.gov/teleconnections/pdo/

18. El Nino 3.4 index http://www.esrl.noaa.gov/psd/gcos_wgsp/Timeseries/Data/nino34.long.data (Linked from http://www.cgd.ucar.edu/cas/catalog/climind/TNI_N34 )

        19. AMO index http://www.esrl.noaa.gov/psd/data/correlation/amon.us.long.data

20. CO2 level at Law Dome, Antarctica: http://cdiac.ornl.gov/ftp/trends/co2/lawdome.combined.dat

21. Mauna Loa CO2: https://www.co2.earth/monthly-co2

22. Sensitivity of AGT to clouds http://lowaltitudeclouds.blogspot.com

23. Current GISS data set: http://data.giss.nasa.gov/gistemp/tabledata_v4/GLB.Ts+dSST.txt

24. Svensmark paper: Phys. Rev. Lett. 85, 5004–5007 (2000) http://prl.aps.org/abstract/PRL/v85/i23/p5004_1

25. Marsden & Lingenfelter 2003, Journal of the Atmospheric Sciences 60: 626-636 http://www.co2science.org/articles/V6/N16/C1.php

26. CLOUD experiment at CERN http://indico.cern.ch/event/197799/session/9/contribution/42/material/slides/0.pdf

27. Loehle non-tree-ring AGT http://www.econ.ohio-state.edu/jhm/AGW/Loehle/Loehle_McC_E&E_2008.pdf

        28. MODTRAN6 defaults http://modtran.spectral.com/modtran_faq

29. K & T chart http://chriscolose.wordpress.com/2008/12/10/an-update-to-kiehl-and-trenberth-1997/

30. MacRae assessment of dCO2/dT https://wattsupwiththat.com/2015/06/13/presentation-of-evidence-suggesting-temperature-drives-atmospheric-co2-more-than-co2-drives-temperature/

31. Transportation fuel https://www.eia.gov/outlooks/ieo/pdf/transportation.pdf

32. Fuel properties https://en.wikipedia.org/wiki/Gasoline

33. World electricity generation https://yearbook.enerdata.net/electricity/world-electricity-production-statistics.html

34. Fuel sources for electricity generation https://www.eia.gov/outlooks/ieo/exec_summ.php

35. WV from coal combustion http://energyeducation.ca/encyclopedia/Water_vapour

36. Energy content of bituminous coal https://en.wikipedia.org/wiki/Energy_density

37. Global water withdrawal http://www.fao.org/nr/water/aquastat/water_use/index.stm

38. Irrigated agricultural area http://www.worldwatch.org/global-irrigated-area-record-levels-expansion-slowing-0

39. Annual fresh water withdrawal http://www.fao.org/nr/water/aquastat/didyouknow/index2.stm

40. 70% of withdrawal is for agriculture http://www.fao.org/nr/water/aquastat/infographics/Withdrawal_eng.pdf

41. Surface irrigation https://water.usgs.gov/edu/irfurrow.html

42. Frequency of furrow irrigation https://naldc.nal.usda.gov/download/54786/PDF

43. Pond evaporation rate http://www.nws.noaa.gov/oh/hdsc/Technical_papers/TP13.pdf

44. Animation of roiling SST https://www.youtube.com/watch?v=aKMY4JRN0kk

45. Q-M applied to water vapor and carbon dioxide in the atmosphere (loads slowly): http://www.dtic.mil/dtic/tr/fulltext/u2/477312.pdf

46. NASA/GISS TOA graph source https://www.giss.nasa.gov/research/briefs/2010_schmidt_05/

47. MODTRAN calculator http://climatemodels.uchicago.edu/modtran/

48. MODTRAN6 calculator http://modtran.spectral.com/modtran_home#plot

49. Deleted.

50. Vapor pressure of water https://en.wikipedia.org/wiki/Vapour_pressure_of_water

51. Theory of redirected energy: http://energyredirect3.blogspot.com

52. HITRAN data base calculator http://www.spectralcalc.com/spectral_browser/db_intensity.php

53. DIY climate change analysis http://diyclimateanalysis.blogspot.com

54. Surface irrigation 86%: http://www.fao.org/3/I9253EN/i9253en.pdf

55. Bolton equation for water saturation p T https://glossary.ametsoc.org/wiki/Clausius-clapeyron_equation

56. Ice and mixed phase clouds: http://www.cas.manchester.ac.uk/resactivities/cloudphysics/background/ice/

57. Wexler, vapor pressure of water: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5312760/

58. Wexler, thermodynamic calculations for the vapor pressure of ice: https://nvlpubs.nist.gov/nistpubs/jres/81A/jresv81An1p5_A1b.pdf

59. Ocean temperatures: https://rwu.pressbooks.pub/webboceanography/chapter/6-2-temperature/

60. Water vapor generated by humanity is 90% from irrigation: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2001WR000355

61. Water vapor vs CO2 for planet warming: https://watervaporandwarming.blogspot.com

62. Accurate formulae for calculating saturation vapor pressure: https://journals.ametsoc.org/view/journals/apme/57/6/jamc-d-17-0334.1.xml