Analyses of a large number of continental boreholes have yielded temperature reconstructions for the last 500 years (Lachenbruch and Marshall 1986, Pollack and Huang 2000, Harris and Chapman 2001, Majorowicz et al. 2006). These reconstructions have particular value because they do not have to be calibrated against the instrumental record and because temperature itself is being measured, not a proxy for temperature. Important quantitative uncertainties exist, but general trends in these reconstructions are likely robust.
Borehole-based temperature reconstructions averaged for broad regions (including eastern North America, western North America, Europe, Australia, and South Africa) show warming from the 19th century to the present and persistently cool conditions for the preceding few centuries (Gosnold et al. 1997, Pollack and Huang 2000 and references therein, Huang et al. 2000, Harris and Chapman 2001, Pollack and Smerdon 2004, Majorowicz et al. 2006). Temperature changes for earlier times are not resolvable. Estimates of the magnitude of recent warming as a global (continental) average, or an average over the middle latitudes, are approximately 0.7–0.9°C, from the mid-19th century to the late 20th century (Huang et al. 2000, Pollack and Huang 2000, Harris and Chapman 2001), which is similar to the temperature increase estimated from the instrumental record discussed in Chapter 2 (Figures 8-1 and 8-2). Some