FIGURE 2-4 Smoothed time series of normalized annual mean surface air temperature averaged over China and the entire Northern Hemisphere. SOURCE: Bradley et al. (1988). Reprinted with kind permission of Springer Science and Business Media; copyright 1988.

dominated by variations in global mean temperature that occur in response to changes in the global energy balance. If this premise is correct, it follows that temperature time series at points on Earth should be more strongly correlated with the time series of the global mean temperature on these longer timescales than they are on the year-to-year timescale. Several papers offer support for this view (e.g., Leung and North 1991, Shen et al. 1994), as does the time series shown in Figure 2-4 (Bradley et al. 1988). The strong correspondence between the Northern Hemisphere and China curves indicates that much of the decade-to-decade and century-to-century variability in the mean temperature of the Northern Hemisphere since 1880 can be captured using data from the Chinese station network alone. Of course, temperature time series at individual sites within China are not as highly correlated with hemispheric mean time series as the China-mean time series in Figure 2-4, and proxy time series do not perfectly represent the true time series of surface temperature variations. Much remains to be done to place the spatial sampling requirements on a firm footing.

Another issue that arises when interpreting proxy records of surface temperature over the last 2,000 years is the degree to which temperature time series in various latitude belts are representative of the globally averaged temperature. The instrumental record of surface temperature shown in Figure 2-5 is instructive in this respect. The rise in surface air temperature that occurred during the 1920s and the slight decline during the 1950s were much more pronounced over high latitudes of the Northern Hemisphere than at lower latitudes. In contrast, the warming of the last few decades has been much more latitudinally uniform. The latitudinally dependent features in Figure 2-5 serve as a reminder that not all the variability over high latitudes, as recorded in ice core measurements and high-latitude proxies, is necessarily representative of variations in global mean temperature.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement