dation and the National Oceanographic and Atmospheric Administration. Climate modeling at GISS is supported by the NASA Climate Program Office. We thank R. Healy and P. Lonergan for help with model experiments; J. Hansen, R. Bradley, M. Hughes, G. Jacoby, and L. Thompson for discussions or critical comments on the manuscript; and J. Lean for providing estimates of solar variability. J.M. Wallace provided substantial comments in review.
JOHN M. WALLACE
University of Washington
Drs. Rind and Overpeck have given us a very interesting and informative summary of a coordinated series of climate sensitivity experiments designed to investigate the potential importance of five sources of climate variability: (1) inherent (random) variability of the climate system, not taking into account variability in the ocean circulation, (2) coupled atmosphere-ocean interactions, including phenomena such as the ENSO cycle and the thermohaline circulation, (3) variability in solar output, (4) forcing due to episodic injections of volcanic aerosols, and (5) increases in greenhouse gases. I was struck by the fact that they had independently chosen nearly the same categories that I had used in interpreting the observational record in my paper. This coincidence serves to illustrate that many of us share a common view as to the nature and probable causes of interdecadal climate variability.
Climate modeling is such a broad topic that no single paper of this length can hope to be comprehensive, but I think that the numerical experiments discussed in their review are well chosen and provide a great deal of food for thought. They may be viewed as a first step toward a quantitative assessment of the relative importance of various proposed mechanisms that contribute to climate variability on the interdecadal time scale.
The authors conclude that none of the five interpretations of interdecadal climate variability that they considered should be categorically discounted at this point. I agree that we need to be open to the possibility that any of these interpretations might ultimately prove to be relevant, but I also agree with the earlier suggestion that we have a duty to try to assess the relative importance of the various climate forcing mechanisms. I believe the modeling evidence reviewed by Drs. Rind and Overpeck does, in fact, suggest that some mechanisms are more important than others.
To start with, I think it can be argued that solar variability should be near the bottom of the list. In order to elicit a significant response to solar variability on the time scale of the sunspot cycle in the experiments that they described, it was necessary to artificially inflate the solar forcing of the troposphere by treating the observed temporal variability in the solar output as if it were independent of wavelength, whereas in reality it involves only ultraviolet radiation, which is absorbed in the upper atmosphere. I would also place aerosols injected by volcanic eruptions near the bottom of the list because, in order to elicit a significant interdecadal signal, it is necessary to invoke changes in aerosol loading on time scales much longer than the typical 1- to 2-year residence times observed in association with recent eruptions.
Another mechanism that I would be inclined to dismiss is decadal-scale changes in the ENSO cycle. I would do so not on the basis of the numerical experiments but on the basis of the observational evidence presented in my paper, which indicates that there is very little temporal correlation between the ENSO signal in the tropics and the observed interdecadal variability and mean temperature of the extratropical regions of the globe. And it is mainly in extratropical latitudes that one finds the large interdecadal variability in paleoclimate records.
That leaves us with greenhouse gases, the final topic discussed in Dr. Rind's paper, and inherent (random) variability of the climate system, including variations in North Atlantic deep-water formation, which might account for the dip in Northern Hemisphere temperatures during the 1960s and 1970s. I suppose that aerosols generated by sources other than volcanic eruptions (e.g., industry) might also be important. I propose this priority list for our discussion.