as interpreted by Enfield and Cid. To these speculations may be added our suggestion, above, of a possible association of spectral temperature oscillations on El Niño time scales with a bidecadal oscillation.
In conclusion, we suggest that the atmospheric CO2 record may be of some use in the study of decadal variability in climate. Conclusive evidence of decadal periodicities in climatic parameters will probably be established, however, only if and when a mechanism is found that is physically reasonable and successfully predicts changes in climate.
In the oral presentation of this paper, we presented an hypothesis that forcing of ocean temperatures by oceanic tidal dissipation might explain the observed decadal oscillation in global temperature and its disappearance in the early part of this century. We were attracted to the coincidence of the times of prominent decadal temperature signal with the times when the strongest equilibrium oceanic tidal forcing was produced by the gravitational forces of the sun and moon. At these times, within 10 years of 1881 and 1974, episodes of unusually strong tidal forcing occurred at times of the strongest tides of the year, because the sun and moon tended then to be both in unusually close alignment with the Earth and at nearly closest approach. In contrast, in the 1920s, when the decadal signal was weakest, only weaker tidal forcing occurred at times of the strongest tides. Because strong tidal forcing occurs at intervals of 9 years when the alignments of the sun, moon, and earth are nearly optimal for strong tidal forcing, and otherwise at 3- and 6-year intervals, the decadal signal itself and the 6-year signal in the middle of the record period also correspond to the periodicities of tidal forcing.
It was not possible to prepare an appropriate presentation of the tidal discussion for this volume. An article describing the tidal hypothesis is in preparation.
We are grateful to scientists who gave generously of their time to discuss the subjects of solar phenomena and climatic variation with us in the course of preparation of this article. We specifically thank Philip Jones and Thomas Wigley, Michael Ghil, James Hansen, Harry van Loon and Gerry Meehl, David Parker, and Henry Diaz and Thomas Karl. We are also grateful for discussions with Tim Barnett, Robert Bacastow, Daniel Cayan, and Hugh Hudson, colleagues of ours at the University of California at San Diego. We further thank Drs. Jones and Hansen for supplying us with their temperature data sets, and David Parker for supplying us with extensive global and regional sea surface temperature data in graphic form. Computer time was provided by the San Diego Supercomputer Center. Financial support came from the National Science Foundation via Grant ATM-91-21986 and from the U.S. Department of Energy via Grant FG03-90ER-60940.
Table A1 presents the global-average anomalies of air temperature used in our study. The anomalies are listed for each calendar month, followed in the last column by the annual average. They were obtained by averaging Northern and Southern Hemisphere anomalies. The hemispheric data for 1855 through 1989, expressed both over land and in surface seawater, were supplied to us by P.D. Jones (personal communication of April 1990). These data were supplied to us as monthly averages, whereas for 1987 through 1989 only annual averages were available with respect to temperatures in surface seawater. For these final three years of our data set we estimated global averages for each month from the monthly anomalies supplied to us for temperatures over land.
For the Northern Hemisphere we multiplied the monthly land values for each given year by the ratio of the annual global mean for that year to the mean for land. For the Southern Hemisphere we assumed that no intra-annual variations existed over the oceans, and that temperatures over land contributed 30% to the hemispheric averages. For the full data set we accepted the premise of Jones et al. (1986c) that sea surface and marine air temperatures follow each other closely on interannual time scales, so that combined land and surface sea-water-temperature data portray global-average interannual variations in surface air temperature.
Recently P.D. Jones (personal communication of April 1993) kindly supplied us with updated monthly data for all years of our study. We have repeated our calculations with these data and find only very small changes in spectral amplitudes and frequencies, and thus in the time plots shown in this article.