Water Vapor #3: Updated NIR Results

Since 2000, there have been a number of very important new studies, increasing the estimated NIR absorption parameters of water vapor, stimulated by the discovery of the HITRAN-96 clerical errors, but amounting to significant increases over and above those errors. The increases seem to account for much missing atmospheric absorption. Again, the issue here is not the "greenhouse" effect of water vapor, but a type of anti-greenhouse effect of water vapor, whereby water vapor absorbs incoming solar radiation in near infra-red (NIR) and visible wavelengths, rather than absorbing outgoing IR wavelengths. This directly cuts into hypothesized positive feedback effects of increased IR absorption of water vapor by creating a negative feedback. Belmiloud et al [2000] with a variety of blue-chip authors have a compelling cut-phrase:

There is another lesson to be learned. Making sure the database is valid is necessary foundation for all modelling of atmospheric radiation transfer, especially so when theory and observation fail to agree.

Here are some references: Djedjiga Belmiloud, Roland Schermaul, Kevin M. Smith, Nikolai F.Zobov, James W. Brault, Richard C. M. Learner, David A. Newnham, and Jonathan Tennyson , New Studies of the Visible and Near-Infrared Absorption by Water Vapour and Some Problems with the HITRAN Database, GEOPHYSICAL RESEARCH LETTERS, VOL. 27, NO. 22, PAGES 3703-3706, NOVEMBER 15, 2000 http://www.tampa.phys.ucl.ac.uk/jonny/papers/262.pdf

Abstract. New laboratory measurements and theoretical calculations of integrated line intensities for water vapour bands in the near-infrared and visible (8500-15800 cmàƒ⣃ ‹’€ ”‚¬’„¢1) are summarised. Band intensities derived from the new measured data show a systematic 6 to 26% increase compared to calculations using the HITRAN-96 database. The recent corrections to the HITRAN database [Giver et al., J. Quant. Spectrosc. Radiat. Transfer, 66, 101-105, 2000] do not remove these discrepancies and the differences change to 6 to 38 %. The new data is expected to substantially increase the calculated absorption of solar energy due to water vapour

Roland Schermaul, Richard C. M. Learner, David A. Newnham, John Ballard, Nikolai F. Zobov, Djedjiga Belmiloud and Jonathan Tennysonz, The Water Vapor Spectrum in the Region 8600–15 000 cm-1: Experimental and Theoretical Studies for a New Spectral Line Database II. Linelist Construction, Journal of Molecular Spectroscopy 208, 43–50 (2001) http://www.tampa.phys.ucl.ac.uk/jonny/papers/270.pdf

The new laboratory measurements of R. Schermaul et al. (J. Mol. Spectrosc. 208), for the near-infrared and visible spectrum of water vapor, covering the 2º C ±, 3º, 3º C±, and 4º polyads, are combined with accurate calculations of weaker lines to provide a new, comprehensive linelist of water transitions for the spectral region 8600–15 000 cm¡1. The resulting ESA–WVR linelist reproduces the raw laboratory observations to a high level of agreement at all but the longest wavelengths. This linelist has been made available in a standard format for general use.

L. S. Rothman and J. Schroeder, Millennium HITRAN Compilation, Twelfth ARM Science Team Meeting Proceedings, St. Petersburg, Florida, April 8-12, 2002 http://www.arm.gov/publications/proceedings/conf12/extended_abs/rothman-ls.pdf Roman N. Tolchenova, Mizuho Tanaka, Jonathan Tennyson; Nikolai F. Zobova, Sergei V. Shirina, Oleg L. Polyansky, Ahilleas N. Maurellis, Water line intensities in the near-infrared and visible, Journal of Quantitative Spectroscopy & Radiative Transfer 82 (2003) 151–163 http://www.tampa.phys.ucl.ac.uk/jonny/papers/317.pdf

Water is the single most important molecule for models of the earth’s atmosphere but line parameters for water, particularly at shorter wavelengths, are dierent results. The experimental data used to construct the ESA-WVR linelist (J. Mol. Spectrosc. 208 (2001) 32) is re-analyzed with a focus on e>ects due to pressure determination in the cell, subtraction of the baseline and parameterization of the line pro=les. A preliminary re-analysis suggests that the line intensities given by the ESA-WVR study should be closer to those of Brown et al. (J. Mol. Spectrosc. 212 (2002) 57) used in the HITRAN. This shows the vital importance of validating the data for water by independent means.

P. Albert, K.M. Smith, R. Bennartz, D.A. Newnham, J. Fischer, Satellite- and ground-based observations of atmospheric water vapor absorption in the 940 nm region, Journal of Quantitative Spectroscopy & Radiative Transfer 84 (2004) 181–193 http://naftali.meteor.wisc.edu/~ralf/jqsrt_albert_line_absorption_2003.pdf

ments of direct absorption of solar radiation between 9000 and 13; 000 cmàƒ⣃ ‹’€ ”‚¬’„¢1 (770–1100 nm) with a spectral resolution of 0:05 cmàƒ⣃ ‹’€ ”‚¬’„¢1 are compared with line-by-line simulations of atmospheric absorption based on di:erent molecular databases (HITRAN 2000, HITRAN 99, HITRAN 96 and ESA-WVR). Di:erences between measurements and simulations can be reduced to a great amount by scaling the individual line intensities with spectral and database dependent scaling factors. Scaling factors are calculated for the selected databases using a Marquardt non-linear least-squares @t together with a forward model for 100 cmàƒ⣃ ‹’€ ”‚¬’„¢1 wide intervals between 10,150 and 11; 250 cmàƒ⣃ ‹’€ ”‚¬’„¢1 as well as for the water vapor absorption channels of the Medium Resolution Imaging Spectrometer (MERIS) onboard the European Space Agency’s (ESA) ENVISAT platform and the Modular Optoelectronic Scanner (MOS) on the Indian IRSP-3platform, developed by the German Aerospace Centre (DLR). For the latter, the scaling coeEcients are converted into correction factors for retrieved total columnar water vapor content and used for a comparison of MOS-based retrievals of total columnar atmospheric water vapor above cloud-free land surfaces with radio soundings. The scaling factors determined for 100 cmàƒ⣃ ‹’€ ”‚¬’„¢1 wide intervals range from 0.85 for the ESA-WVR molecular database to 1.15 for HITRAN 96. The best agreement between measurements and simulations is achieved with HITRAN 99 and HITRAN 2000, respectively, using scaling factors between 0.9 and 1. The e:ects on the satellite-based retrievals of columnar atmospheric water vapor range from 2% (HITRAN 2000) to 12% (ESA-WVR).

Maurellis, A. N., Lang, R., Williams, J. E., van der Zande, W. J. Smith, K. Newnham, D. A.,Tennyson, J., Tolchenov, R. N., Weakly Interacting Molecular Pairs: Unconventional Absorbers of Radiation in the Atmosphere, Proceedings of the NATO Advanced Research Workshop, p. 259-272, 2003


  1. Dave Dardinger
    Posted Apr 12, 2005 at 10:44 PM | Permalink

    Let me see if I’ve understood this correctly. This extra absorption works counter to the greenhouse effect since the absorption occurs relatively high in the troposphere and thus warms those layers directly, allowing the energy to be emitted as long-wavelength IR to space without first having to be absorbed by the surface, thus warming it and letting the surface emit the long-wavelength IR. [Usual caveats — about this being on the margin and not all IR going directly to space, etc. — applying, of course.]

  2. Steve McIntyre
    Posted Apr 13, 2005 at 5:50 AM | Permalink

    Dave, that seems like the implication to me. If the radiation had got to the surface, then it would have to be re-emitted to space, working its way back up through the atmospheric gases. But that’s my interpretation rather than an observation that I can quote from a specialist. I’m planning a couple more posts on this, one on GCMs and one with a comment by Alan Arking when the issue was posed to him.

    Now the impact of water vapor on outgoing IR is a bigger effect than the impact of water vapor on incoming NIR and visible (NIR is between mid-IR and visible and is short-wavelength), but that’s not the issue: if you under-estimate one side of the equation, you would certainly screw up your feedback calculations. Steve

  3. Dave Dardinger
    Posted Apr 13, 2005 at 7:42 AM | Permalink

    BTW, in the somewhat parallel discussion of water vapor on RealClimate presently, John Finn in comment #55 indicates a tenfold discrepency between observed and predicted H20 increases with warming (3ppm/K vs 30ppm/K). Does anyone here know if this is correct or if he’s made a mistake in his analysis?

  4. John Finn
    Posted Apr 13, 2005 at 9:43 AM | Permalink


    I was using the 30 ppm figure as a bait in the hope they would ‘correct’ me, but I have actually seen figures such as 25 ppm and 30 ppm quoted. I’m fairly sure there is a roughly tenfold difference between the observations in the Minschwaner paper and model results.

    Regarding another post on realclimate (same discussion). Is the John A who posts regularly on here the same John A who upset Gavin Schmidt. If it is he really should read the IPCC SUMMARY FOR POLICYMAKERS properly as there is a graph which clearly shows that solar forcings had little or no impact on global climate for around 900 years – and it’s only at the point when atmospheric CO2 levels rose from 280 ppm to 295 ppm did global temperatures begin the “unprecedented” rise. 🙂

  5. John A
    Posted Apr 13, 2005 at 12:52 PM | Permalink

    Re #4

    Yes, I am the same person who “upset” Gavin Schmidt.

    If it is he really should read the IPCC SUMMARY FOR POLICYMAKERS properly as there is a graph which clearly shows that solar forcings had little or no impact on global climate for around 900 years – and it’s only at the point when atmospheric CO2 levels rose from 280 ppm to 295 ppm did global temperatures begin the “unprecedented” rise.

    Fortunately you added the emoticon so I knew you weren’t being serious. Welcome to the funny farm.

  6. Posted Apr 13, 2005 at 1:03 PM | Permalink

    Astronomers have always known that water vapor absorps radiation in visible as well as infrared. Large telescopes are built at high altitude, either in a dry climate (such a desert regions in Chile) or steep, extinct volcanoes on islands (such as Hawaii or Canary Islands) to avoud as much water vapor, clouds and atmospheric extinction as possible. This knowledge has evidently not reached the climate researchers.

  7. Steve McIntyre
    Posted Apr 13, 2005 at 2:24 PM | Permalink

    Lars, I think that’s over-stating the matter. The models do consider NIR absorption. However, modeling NIR absorption by water vapor (and IR absorption) is not easy since there are thousands of lines that come into play. Because the direct impact of 2xCO2 is say 3 wm-2 or so, achieving a water vapor model that is accurate to a precision of 3 wm-2 is probably pretty hard (even before one worries about changing concentrations). I would have thought that you’d want to model it more accurately than 3wm-2 in order to assess 3 wm-2 in CO2, but this is a topic that I haven’t looked at in detail. Steve

  8. Buck Smith
    Posted Apr 23, 2005 at 9:38 AM | Permalink

    Is it not correct to compare the 3wm-2 of CO2 effect to say, a 1% change in albedo which is 13.7 wm-2, and conclude CO2 is just not likely to be a big factor.

    Great Blog! It is really good to see the academics blogging. I hope it will improve public debate and lead to saner environmental policies. We might see some physics, engineering, chemistry etc. profs do the same for public debate about energy.

  9. Roger Bell
    Posted Jun 13, 2006 at 9:09 AM | Permalink

    Re your post number 7.
    Before I retired I was modelling stellar spectra using hundreds of thousands of spectral lines of different elements, in both atomic and molecular form. Many of these were molecular lines e.g. TiO, H2O, CO,….. others were atomic.

    I was working on astronomy problems. However, the work leads me to say that someone should use water vapour and CO2 line data to calculate the effect of the earth’s atmosphere on the solar spectrum in the 12-18 micron region and compare it with real data. Are we getting agreement? If not, what has to be done? This would be a crucial test of the accuracy of the modelling.

  10. Steve McIntyre
    Posted Jun 13, 2006 at 10:07 AM | Permalink

    #9. riger, I have no idea on what the status is in this field. It’s something I look at only sporadically. I’d love to know the current thinking on “anomalous atmospheric absorption” where Ramanathan had results that were 20-25 wm-2 different than canonical Trenberth results, which he attributed to water vapor. This was a live topic about 5-8 years ago, but seems to have gone cold. I haven’t seen an article which ties up the loose ends; if anyone knows one, I’d be intereseted.

  11. John G. Bell
    Posted Jun 13, 2006 at 11:23 AM | Permalink

    Roger, Is this of any use? The related? question of water dimers came up before and at that time no one could point to an article.

    International Journal of Infrared and Millimeter Waves
    Issue: Volume 7, Number 11
    Date: November 1986
    Pages: 1795 – 1803
    Anomalous atmospheric absorption spectra due to water dimer

    Cai Peipei1, Zhang Hansheng1, Shen Shanxiong1 and I-Shan Cheng1
    (1) Department of Physics, East China Normal University, 200062 Shanghai, China

    Abstract The anomalous atmospheric absorption spectra in the window wavelength region of 8–14 um have been suggested due to the water dimer. Based on our laboratory measurements, water continuum CO2 laser absorption spectra and a resonance absorption line due to the weak local water vapor pure rotational transition have been reported. The equilibrium concentration of water dimers in the atmosphere, the electronic binding energy and the theoretical calculations for absorption attenuation have been obtained in agreement with published data.

  12. Steve McIntyre
    Posted Jun 13, 2006 at 11:52 AM | Permalink

    This seems to be a fairly slight effect (as I understand it). But it could be that the answer consists of 8 different small effects all adding up to 20-25 wm-2.

    I don’t know whether the current view is that “anomalous absorpotion” even exists or not. For a while, the measurements were argued about. IPCC TAR uses Trenberth;s values and doesn’t discuss Ramanathan in any detail, although Ramanathan is an eminent scientist (and arguably the first person to cite the value 4wm-2). So if anyone could enlighten me on that issue, I’d appreciate it.

    Ramanathan and Vogelmann 1997 thought that the effect is due to poorly modeled NIR absorption by water vapor associated with clouds . I get the impression – but this is not a researched impression – that the modelers have “dealt” with the topic either by denying the measurements or allocating the effect to aerosols. It seems to me to be a large issue, since if the anomalous NIR absorption is attributable to water vapor, that is a very large negative feedback and would have an impact on climate projections. If it is attributable to aerosols, then that is just one more parameter.

  13. Posted Aug 23, 2009 at 8:38 PM | Permalink

    Came across this looking for references on water vapor overtones. You can get some indication of how well the then current models did in 1999 from

    Click to access mlawer_ej.pdf

    The major outstanding issue was the water vapor continuum.

    A useful name to search under for this sort of thing is Chris Benner. RTFR

%d bloggers like this: