How Water Warms Our Planet

The hydrological cycle. Estimates of the observed main water reservoirs (black numbers in 10^3 km3 ) and the flow of moisture through the system (red numbers, in 10^3 km3 yr À1 ). Adjusted from Trenberth et al. [2007a] for the period 2002-2008 as in Trenberth et al. [2011].

This site has long asserted that “Oceans Make Climate”. Now a recent study reveals the dynamics by which water influences temperatures over land as well. The paper is Testing the hypothesis that variations in atmospheric water vapour are the main cause of fluctuations in global temperature by Ivan R. Kennedy and Migdat Hodzic, published in Periodicals of Engineering and Natural Sciences, August 2019.  Excerpts in italics with my bolds. H/T Notrickszone.


Global warming issues have caused intensive research work in related areas, from land use, to urban environment to data science use in order to understand its effects better [25], [26], [27]. In this paper we focus on water related effects on global warming. Although water is recognised as the main cause of the greenhouse effect warming the Earth 33 oC above its black body temperature, water vapour is usually given a secondary role in global models, as a positive feedback from warming by all other causes. Despite its dominant effect in generating the weather, changes related to water are not seen as having a primary role in climate change, the focus being primarily on CO2. With positive feedback from primary warming, the effect of increasing CO2 is trebled [15] by water vapour increase. This conclusion is based on the perception that there are no significant trends in the hydrological cycle that could cause climate forcing. But this overlooks the effect of more than 3500 km3 of extra surface and ground water used annually in irrigation [17] to grow food for the human population. This quantity of extra water increases steadily year by year, well correlated with increasing atmospheric CO2, growing about 60% of world food requirements. Even so, the amount used in irrigation probably only adds about 3% to the annual hydrological cycle [9] of 113,000 km3. Is this sufficient to exert a significant extra greenhouse effect? Here we advance the hypothesis that it does and should be included in climate models.

A critical assumption of the IPCC consensus of global warming is that an increasing concentration of CO2 causes more retention of radiant heat near the top of the atmosphere, largely as a result of reduced emission of its spectral wavelengths centred on 15 microns. The radiative-convective model assumes that the lowered emissions at reduced pressure, number density and higher, colder altitudes from this GHG now provides an independent and sustained forcing exceeding 1-2 W per m2. It is assumed that once this reduction in OLR in the air column from increasing CO2 has occurred it must be compensated by increased OLR at different wavelengths elsewhere, maintaining balance with incoming radiation.

This critical assumption still lacks empirical confirmation.

Water Drives Atmospheric Warming

The importance of water in helping to keep the Earth’s atmosphere warm in the short term is beyond dispute. Table 1 summarises previously estimated rates for thermal energy flows into and out of the atmosphere [23]. As shown in the table, more than 80% of the power by which the temperature of air is maintained above the Earth’s black body temperature of -18 C is facilitated by water. Most significant of these air warming inputs from water is the greenhouse effect by which water vapour absorbs longwave radiation emitted from the surface, retaining more energy in air. However, warming from absorption of specific quanta by water vapour of incoming short wave solar radiation (ISR) and the latent heat of condensation of water vapour, exceeding the cooling effect of vertical convection, also contribute to warming of air.

Thus, the greenhouse gas (GHG) content of the atmosphere effectively provides resistance to heat flow to space increasing the transient storage of solar energy, with a warming effect analogous to resistances in an electrical circuit. By comparison to water, other polyatomic greenhouse gases like CO2 play a minor role in this process, totalling less than 20% of warming. Furthermore, the fact that the minor GHGs are relatively well-mixed by the turbulence in the troposphere, unlike water, means that we cannot expect to observe spatial variations in their effects. Furthermore, the heat capacity of non-greenhouse gases provides some 99% of the thermal inertia of the troposphere, although only greenhouse gases capable of longwave radiation by vibrational and rotational quanta can contribute to cooling by radiation through the top of the atmosphere as OLR. Figure 1 contrasts schematically the typical variation of outgoing longwave radiation (OLR) over marine and terrestrial environments.

On well-watered land such as southern China much less direct emission of OLR to space occurs, in contrast to Quetta, Pakistan, on the same latitude with similar incoming shortwave radiation (ISR). In contrast to humid atmospheres on land and tropical seas, relatively arid regions such as the Sahara, the Middle East and Australia provide heat vents effectively cooling the Earth, solely as a result of the radiant emissions from GHGs as OLR. The varying global emissions of OLR estimated for typical marine and terrestrial regions shown in Figure 2 mirror this scheme.

Clearly, water vapour is the most critical factor in the mechanism by which the air column of the lower troposphere is charged with heat energy. It is of interest from this figure and in Table 1 that the exact sum of the effects of all greenhouse gases in directly warming air, including conduction from the surface, charges the lower atmosphere with sufficient heat to generate the downwelling radiation from greenhouse gases directed towards the surface [12]. Water is the main source of this back radiation [18], well understood to be responsible for keeping the surface air warmer in humid atmospheres, thus raising the minimum temperature.

None of the variation in OLR in Figure 1 can be attributed to the well-mixed GHGs such as CO2.

Furthermore, unlike the greenhouse effect of CO2, which is regarded as increasing only in in a logarithmic manner as its concentration rises, the greenhouse effect of water on retaining heat in the atmosphere should vary more linearly, even in the case of absorption of surface radiation, as its vapour spreads into dryer atmospheres; this potential is illustrated in Fig.1 in the descending zones of Hadley cells at sub-tropical latitudes.

Fig. 1 Global values of mean OLR from 2003-2011 (downloaded August 2, 2017, AIRS OLR 2003-2011 average htpp:// estimated by Giorgio, G.P., June 24, 2014). The russet areas show regions of greater OLR, with outgoing radiation above the average of ca. 240 W per m2, thus tending to cool the Earth. Note how the upper troposphere above arid continental regions provides a vent for the greatest rate of cooling.

Thermal Effects from Water are Direct and Linear

An approximately linear response in increasing air temperature to changes in atmospheric water content is reasonable. Unlike the well-mixed CO2, there are marked spatial and temporal variations in atmospheric water content, with much of the Earth’s surface in significant deficit, particularly in the sub-tropical zone subject to Hadley cell recycling, emphasised over semi-arid land. To the extent that additional water vapour spills over into these dryer regions on land the greater the area of the Earth that is subject to the greenhouse effect. This response can be contrasted to the effect of increasing CO2, which has a logarithmic relationship between climate forcing and concentration in the atmosphere [14], [15], each doubling causing a similar increase in temperature. Because there is no obvious regional effect of CO2 on the weather or regional climate, the effect of any increases in its concentration can only be theoretically inferred. If additional heat is retained in the atmosphere by increasing greenhouse effects from CO2 or water, the air temperature near the surface is expected to increase to keep global values of ISR and OLR in balance. A critical assumption of the IPCC consensus for climate change is that increasing CO2 causes more retention of heat in air near the top of the troposphere, largely as reduced emission from the edges of its spectral peak centred on 15 microns. This edge effect is predicted to be visible from space as a cooling of its spectrum, providing a negative forcing of 1-2 W per m2. It is assumed that this forcing must be compensated by increased OLR at different wavelengths as a result of the increased temperature.

Fig. 3 Satellite measurements of global-zonal OLR ( NOAA website, downloaded August 20, 2017). The 1998-2000 El Nino peaked at about 1.03 C above the minimum temperature in the preceding La Nina, with zonal OLR varying approximately 4 W/m2; see also (8)

This is regarded as a result of convective elevation of the maritime atmosphere, reducing the outgoing longwave radiation (OLR) about 100 W/m2 locally and 4 W/m2 globally from an increase in global water vapour of about 4%. This suggests a linear response from greenhouse warming to increased water vapour content of the atmosphere. Note that the extra heat in the atmosphere during an El Nino is controlled by all these sources of warming, as shown in Figure 2. Whatever the source of extra heat in the ocean, by moving extra water into the atmosphere as vapour it warms the atmosphere by the resultant greenhouse effect, reducing OLR, as well as direct warming by sunlight in the air column. In Table 4, another estimate of the possible effect of irrigation on global warming by comparison with the El Nino-La Nina cycle [22] is made. Consistent with the irrigation water hypothesis the El Nino has been long known to significantly reduce the OLR over the Pacific Ocean up to 25% [3], recognised as a result of elevation of emission of the OLR from water being elevated and therefore a colder altitude. Assuming 60% of irrigation water becomes vapour in the troposphere and a longer rain-out time of 15 days in dry regions compared to less than a week over the oceans with a global average of 8.5 days [19], a steady state of about 100 km3 of extra water vapour results from irrigation.

This estimate also suggests an increase in temperature near 0.2C from 0.84 W/m2 of forcing based on the data given in Figure 3. This is consistent with the total effect of water vapour on global warming exceeding 25 C.

It should be noted that this dynamic effect of water on warming air includes heat pumping by evapotranspiration as well as significant warming by direct absorption of short wave solar radiation (see Fig. 2), also contributing to a more linear effect by water on warming. Since this increase estimates a primary forcing effect of new water, a positive feedback is also anticipated from increased evaporation of the ocean, suggesting that the total increase from irrigation could be of the order of 0.5 oC in the 20th century.

These global results may have more accuracy than the results obtained from the numerous grid points in global circulation models, given the additivity of errors.

Empirical Proof Comparing Dry and Irrigated Land

In Figure 4, using the same modelling as in Figure 2, the predicted steady state greenhouse effect of adding irrigation water in a comparison between dryland and irrigated land. In fact the effect of water on heat transfer to the atmospheric column is not only a result of the greenhouse effect given in the equation in the figure but also from direct absorption by water of short wave ISR and evapotranspiration, similar in total magnitude. These latter effects will be a linear function of the water vapour involved. The evaporative effect cools the surface but must transfer a similar amount of heat to the atmosphere as infrared radiation (ca. 6 microns) associated with condensation of water vapour into droplets under convective cooling as in [21]. Paradoxically, the modelling paper in [6] failed to account for any of these effects, specifically dismissing significant transfer of water vapour into the atmosphere from growth of irrigated crop growth as noted above. This provides a clue to the possible flaw in their models. Except for environments already very humid where evapotranspiration is limited, this cannot be true.

Fig. 4 Comparison of dryland and irrigated land for effect of water on heat retention in the atmosphere as an enhanced greenhouse effect. The El Nino condition of enhanced evaporation from the ocean known to strongly reduce OLR In [3] is shown as an analogue.

NCEP/ NCAR Reanalyses Coincident with the Periodic Flooding of Lake Eyre

Fig. 5 Variation in OLR from flooding of lake Eyre using NCEP-NCAR reanalysis datasets. a.Difference in OLR values between 1978 and 1974, dry and wet years. b. Difference in OLR values between 1978 and 1973, two dry years.

Rarely, during the La Nina phase of the climate cycle, the dry interior of northern Australia overlying the Great Artesian Basin may flood. Lacking riverine exits to the ocean, the massive runoff caused flows southwards, mainly accumulating in the depression below sea level in central South Australia known as Lake Eyre. In late January and February in the early months of 1974 Lake Eyre filled to a depth of six metres, its surface only returning to its hot, dry state three years later in 1977-78. This was the greatest flood ever recorded. The hypothesis in [4] suggests that this flooding should also lead to persistent elevated water vapour content of the atmosphere, predominantly downwind from the Lake Eyre basin. Using the NCEP-NCAR reanalysis datasets, which are informed by Nimbus and other satellite observations since 1970, the OLR emissions to space and the variation in humidity from this region comparing 12 months of 1974 with the same period in 1978 by subtraction of one year from the other. A significant elevation of OLR when the lake was dry by more than 10 W/m2 was observed for the 12-month period (Figure 5). This result is accompanied by increases in specific humidity consistent with an elevated greenhouse effect such as would be experienced in semi-arid areas when irrigated. The area affected downwind also showing elevated humidity is estimated as 35 times the flooded area, showing that the magnitude of this regional greenhouse effect was indeed significant.

Conclusion:  Thankfully, A Wet World is a Warm World

The neglect of the possible effect of irrigation as a significant source of anthropogenic climate change may have been a result of reluctance to consider the relatively small amount of irrigation in the hydrological cycle. Because water has been considered as providing positive feedback to warming primarily from CO2 its possible forcing effect has been overlooked. But as shown here by several different means, the more potent effect of applying water previously in the ocean or deep in the ground to dry surfaces with air in strong water deficit can be sufficient to affect global temperature. Clearly, the water vapour content of the troposphere is the major cause of the natural greenhouse effect, contributing up to two-thirds of the 33 oC warming.

Spatial and temporal variations in soil moisture and relative humidity of the atmosphere are the main factors controlling the regional outgoing longwave radiation (OLR), in contrast to the more even effects from well-mixed greenhouse gases such as CO2.

This is well illustrated in the 4-6 year El Nino cycles, resulting in a global mean temperature variation approaching 1 oC compared with La Nina years. Longer term, the proposed Milankovitch glaciations of paleoclimates result in declines of atmospheric temperature around 10 oC, consistent with the major reduction in tropospheric water vapour approaching 50%. Weather conditions and climate as illustrated in the greenhouse effect are clearly demonstrated in the distribution of water, particularly on land. The apparently linear relationship between the water content of the atmosphere is direct verification of the greenhouse warming effect of this greenhouse gas. By contrast, other than by correlation, there is no such direct verification possible for the greenhouse effect of CO2. We rely on the forcing equation of 5.3ln[(CO2)t /(CO2)o] to estimate the climate sensitivity with respect to varying concentration (ppmv) of this greenhouse gas. Early hopes that a clear spectral signal was available showing significantly reduced OLR from increasing CO2, proving the hypothesis of climate forcing by permanent GHGs, have not been realised [5]. A focus using new satellites on the longer wavelength OLR associated with rotations of water might help resolve this question. Up till now, OLR is estimated for this region based on shorter wavelengths. The natural experiment provided by the flooding of Lake Eyre of the greenhouse effect by significantly reducing the OLR provides confirmation that irrigation water typically applied to dry land will have a measurable greenhouse effect.

One year time lapse of precipitable water (amount of water in the atmosphere) from Jan 1, 2016 to Dec 31, 2016, as modeled by the GFS. The Pacific ocean rotates into view just as the tropical cyclone season picks up steam.


  1. Hifast · January 7, 2020

    Reblogged this on Climate Collections.


  2. cognog2 · January 7, 2020

    To me there is a joker in the pack running through this article; being the omission of any consideration of the thermodynamics relating to the phase change of water which IMO has a major influence.
    The steam tables tell us that for every kilogram of water evaporated from the surface and returning to earth in the Hydro Cycle some 694 Watthrs of energy gets dissipated into the atmosphere and beyond. This being substantially the Latent Heat, which is pumped up against gravity with a proportion attaining the height of the Tropopause in the form of cirrus clouds, where energy is transmitted to space as the ice crystals grow.
    I submit that this process is a major factor and influence in the climate budget and is largely independent of radiation matters, albeit complicated by them. It requires a detailed analysis and explanation which appears absent in this post.

    The process itself may best be described as a Rankine Cycle which is well understood and does go a long way to explain what is happening in the clouds in the journey of water through the cycle.
    The main inconvenient fact being that at phase change the energy absorbed by water does not result in a temperature increase but is used to alter the configuration of the water molecule instead, resulting in a marked increase in volume. This cuts right across the assumption that water always acts as a greenhouse gas which is only true for the stable states of water.
    Due to the large energy involved wrt the greenhouse energies it is very likely that water in net terms provides a negative feedback to any CO2 forcing and evidence for this may be found in our steam generating plants which operate at constant temperature in spite of variations in energy input.
    A further comment here is that the term convection should not be applied to rising water vapor as this occurs by way of its buoyancy wrt to dry air with no temperature differential involved. This confabulation of buoyancy with convection is endemic in the literature these days but the two are very different.

    Otherwise a most interesting article giving much food for thought.


    • Ron Clutz · January 7, 2020

      Thanks for those reflections cognog. The article does acknowledge that latent heat warms the atmosphere when water vapor condenses. But I agree that they are overly focused on radiative effects. Other scientists share your view that much of the energy dissipated as OLR at top of atmosphere was transmitted from the surface by water.
      See a recent post


      • cognog2 · January 8, 2020

        Yes there will be energy transfer with the atmosphere but not necessarily much warming if any; for again the condensation process takes place at constant temperature, although water/ice returning to earth on the last leg of the Rankine Cycle will be warmed. The temperature being determined by the prevailing local pressure. ( equivalence to the feed heater in earthbound steam plants). It is a complex situation and well beyond me to explain with evidence.
        This raises the issue of the use of the radiative Planck equation dF = S*dT in this situation where the sensitivity coefficient “S” is zero and unless this is taken into account in the calculations would lead to an overestimate of the Global Sensitivity value.
        Thanks for the link which I have already read courtesy of WUWT and your good self. Overall I get the impression that the prevailing radiative groupthink fails to appreciate that temperature is not the ONLY factor of state by which to determine energy matters.
        My regards and thanks,


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s