Journal of Molecular Spectroscopy 208, 32–42 (2001) doi:10.1006/jmsp.2001.8373, available online at http://www.idealibrary.com on The Water Vapor Spectrum in the Region 8600–15 000 cm -1 : Experimental and Theoretical Studies for a New Spectral Line Database I. Laboratory Measurements Roland Schermaul, ∗ Richard C. M. Learner, ∗,1 David A. Newnham,† R. Gary Williams,† John Ballard,† Nikolai F. Zobov,‡ ,2 Djedjiga Belmiloud,‡ and Jonathan Tennyson‡ ∗ Laser Optics and Spectroscopy, Blackett Laboratory, Imperial College of Science, Technology and Medicine, Prince Consort Road, London SW7 2BW, United Kingdom; †Atmospheric Science Division, Space Science and Technology Department, Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire OX11 0QX, United Kingdom; and ‡Department of Physics and Astronomy, University College London, Gower Street, London WC1E 6BT, United Kingdom E-mail: david.newnham@rl.ac.uk, gary.williams@rl.ac.uk, j.ballard@rl.ac.uk, j.tennyson@ucl.ac.uk Received November 30, 2000; in revised form April 26, 2001 New laboratory measurements are presented for the near-infrared and visible spectrum (8600–15 000 cm −1 ) of water vapor. Spectral line parameters, principally intensities and air-broadening coefficients, are derived from Fourier transform spectroscopic measurements at high resolution (0.03 cm −1 ), a range of optical path lengths (5–513 m), and temperatures of both 252 and 296 K. Experimental line parameters are derived for 5034 assigned transitions and thorough error analysis shows parameter errors of less than 2.5% for one-third and less than 5% for over half of the lines. Calculated spectra, derived using these line parameters, reproduce the original spectra to within 2%. A comparison of the line intensities with those in the HITRAN-96 database identifies large errors in the latter with random differences that exceed a factor of two for many lines, and systematic differences between 6 and 26% depending on the water band under consideration. The recent corrections to the HITRAN database by Giver et al. (J. Quant. Spectrosc. Radiat. Transfer 66, 101–105 (2000)) do not remove these discrepancies and the differences change to 6–38%. The new data are expected to substantially increase the calculated absorption of solar energy due to water vapor in climate models. C  2001 Academic Press Key Words: water vapor; near-infrared and visible spectrum; line intensities; air-broadening coefficients. 1. INTRODUCTION Water is a superficially simple molecule that has a very com- plex spectrum. It is one of a small number of atmospheric molecules that absorb significant amounts of visible and near- infrared radiation—radiation at wavelengths near the maximum of the solar emission spectrum. It thus has a major influence on radiation transfer in the terrestrial atmosphere and understanding its absorption properties is essential for climate studies. Model calculations of the Earth’s climate rely on molecular databases, of which HITRAN-96 (1) is the most used, to provide informa- tion about the spectral properties of molecules in the atmosphere. Recently, a problem has been identified in model calculations of atmospheric absorption in both clear and cloudy skies (2). Many climate models substantially underestimate the globally averaged short-wavelength absorption compared to atmospheric 1 Deceased. 2 Permanent address: Institute of Applied Physics, Russian Academy of Sci- ence, Uljanov Street 46, Nizhnii Novgorod, Russia 603024. observations, by as much as 30% of the total atmospheric absorp- tion in the case of clear skies. This anomaly limits our knowledge of the natural atmosphere and our ability to predict the climato- logical effect of anthropogenic perturbations on the atmosphere. Much research has been directed at identifying atmospheric ab- sorbers additional to those already in the databases to make up the discrepancy, but no critical check of the existing database, especially on water—the primary greenhouse gas and major ab- sorber of solar radiation—has been performed. The water vapor spectrum has been studied, using both ex- periment and theory, in far greater detail than that of any other molecule and has, as a consequence, an extensive literature. The main sources of information on the near-infrared and visible spectrum of H 16 2 O is the work underlying the HITRAN database, which is dominated by laboratory Fourier transform (FTS) mea- surements made at the National Solar Observatory in Tucson, Arizona, and analyzed by a number of investigators (3–7). Some of these data have been (re-)analyzed more recently (8–10) and virtually all unidentified lines have been classified, though no significant changes have been made to the line intensities. New 32 0022-2852/01 $35.00 Copyright C  2001 by Academ ic Press All rights of reproduction in any form reserved.