IPCC Fourth Assessment Report: Climate Change 2007
Climate Change 2007: Working Group I: The Physical Science Basis Was Any Part of the Current Interglacial Period Warmer than the Late 20th Century?

The temperature evolution over the Holocene has been established for many different regions, often with centennial-resolution proxy records more sensitive to specific seasons (see Section 6.1). At high latitudes of the North Atlantic and adjacent Arctic, there was a tendency for summer temperature maxima to occur in the early Holocene (10 to 8 ka), pointing to the direct influence of the summer insolation maximum on sea ice extent (Kim et al., 2004; Kaplan and Wolfe, 2006). Climate reconstructions for the mid-northern latitudes exhibit a long-term decline in SST from the warmer early to mid-Holocene to the cooler pre-industrial period of the late Holocene (Johnsen et al., 2001; Marchal et al., 2002; Andersen et al., 2004; Kim et al., 2004; Kaplan and Wolfe 2006), most likely in response to annual mean and summer orbital forcings at these latitudes (Renssen et al., 2005). Near ice sheet remnants in northern Europe or North America, peak warmth was locally delayed, probably as a result of the interplay between ice elevation, albedo, atmospheric and oceanic heat transport and orbital forcing (MacDonald et al., 2000; Davis et al., 2003; Kaufman et al., 2004). The warmest period in northern Europe and north-western North America occurs from 7 to 5 ka (Davis et al., 2003; Kaufman et al., 2004). During the mid-Holocene, global pollen-based reconstructions (Prentice and Webb, 1998; Prentice et al., 2000) and macrofossils (MacDonald et al., 2000) show a northward expansion of northern temperate forest (Bigelow et al., 2003; Kaplan et al., 2003), as well as substantial glacier retreat (see Box 6.3). Warmer conditions at mid- and high latitudes of the NH in the early to mid-Holocene are consistent with deep borehole temperature profiles (Huang et al., 1997). Other early warm periods were identified in the equatorial west Pacific (Stott et al., 2004), China (He et al., 2004), New Zealand (Williams et al., 2004), southern Africa (Holmgren et al., 2003) and Antarctica (Masson et al., 2000). At high southern latitudes, the early warm period cannot be explained by a linear response to local summer insolation changes (see Box 6.1), suggesting large-scale reorganisation of latitudinal heat transport. In contrast, tropical temperature reconstructions, only available from marine records, show that Mediterranean, tropical Atlantic, Pacific and Indian Ocean SSTs exhibit a progressive warming from the beginning of the current interglacial onwards (Kim et al., 2004; Rimbu et al., 2004; Stott et al., 2004), possibly a reflection of tropical annual mean insolation increase (Box 6.1, Figure 1).

Extratropical centennial-resolution records therefore provide evidence for local multi-centennial periods warmer than the last decades by up to several degrees in the early to mid-Holocene. These local warm periods were very likely not globally synchronous and occurred at times when there is evidence that some areas of the tropical oceans were cooler than today (Figure 6.9) (Lorenz et al., 2006). When forced by 6 ka orbital parameters, state-of-the-art coupled climate models and EMICs capture reconstructed regional temperature and precipitation changes (Sections and, whereas simulated global mean temperatures remain essentially unchanged (<0.4°C; Masson-Delmotte et al., 2005b), just as expected from the seasonality of the orbital forcing (see Box 6.1). Due to different regional temperature responses from the tropics to high latitudes, as well as between hemispheres, commonly used concepts such as ‘mid-Holocene thermal optimum’, ‘altithermal’, etc. are not globally relevant and should only be applied in a well-articulated regional context. Current spatial coverage, temporal resolution and age control of available Holocene proxy data limit the ability to determine if there were multi-decadal periods of global warmth comparable to the last half of 20th century.

Figure 6.9

Figure 6.9. Timing and intensity of maximum temperature deviation from pre-industrial levels, as a function of latitude (vertical axis) and time (horizontal axis, in thousands of years before present ). Temperatures above pre-industrial levels by 0.5°C to 2°C appear in orange (above 2°C in red). Temperatures below pre-industrial levels by 0.5°C to 2°C appear in blue. References for data sets are: Barents Sea (Duplessy et al., 2001), Greenland (Johnsen et al., 2001), Europe (Davis et al., 2003), northwest and northeast America (MacDonald et al., 2000; Kaufman et al., 2004), China (He et al., 2004), tropical oceans (Rimbu et al., 2004; Stott et al., 2004; Lorentz et al., 2006), north Atlantic (Marchal et al., 2002; Kim et al., 2004), Tasmania (Xia et al., 2001), East Antarctica (Masson et al., 2000), southern Africa (Holmgren et al., 2003) and New Zealand (Williams et al., 2004).

Box 6.3: Holocene Glacier Variability

The near-global retreat of mountain glaciers is among the most visible evidence of 20th- and 21st-century climate change (see Chapter 4), and the question arises as to the significance of this current retreat within a longer time perspective. The climatic conditions that cause an advance or a retreat may be different for glaciers located in different climate regimes (see Chapter 4). This distinction is crucial if reconstructions of past glacier activity are to be understood properly.

Records of Holocene glacier fluctuations provide a necessary backdrop for evaluating the current global retreat. However, in most mountain regions, records documenting past glacier variations exist as discontinuous low-resolution series (see Box 6.3, Figure 1), whereas continuous records providing the most coherent information for the whole Holocene are available so far only in Scandinavia (e.g., Nesje et al., 2005; see Box 6.3, Figure 1).

What do glaciers reveal about climate change during the Holocene?

Most archives from the NH and the tropics indicate short, or in places even absent, glaciers between 11 and 5 ka, whereas during the second half of the Holocene, glaciers reformed and expanded. This tendency is most probably related to changes in summer insolation due to the configuration of orbital forcing (see Box 6.1). Long-term changes in solar insolation, however, cannot explain the shorter, regionally diverse glacier responses, driven by complex glacier and climate (mainly precipitation and temperature) interactions. On these shorter time scales, climate phenomena such as the North Atlantic Oscillation (NAO) and ENSO affected glaciers’ mass balance, explaining some of the discrepancies found between regions. This is exemplified in the anti-phasing between glacier mass balance variations from the Alps and Scandinavia (Reichert et al., 2001; Six et al., 2001). Comparing the ongoing retreat of glaciers with the reconstruction of glacier variations during the Holocene, no period analogous to the present with a globally homogenous trend of retreating glaciers over centennial and shorter time scales could be identified in the past, although account must be taken of the large gaps in the data coverage on retreated glaciers in most regions. This is in line with model experiments suggesting that present-day glacier retreat exceed any variations simulated by the GCM control experiments and must have an external cause, with anthropogenic forcing the most likely candidate (Reichert et al., 2002).

Box 6.3 Figure 1

Box 6.3, Figure 1. Timing and relative scale of selected glacier records from both hemispheres. The different records show that Holocene glacier patterns are complex and that they should be interpreted regionally in terms of precipitation and temperature. In most cases, the scale of glacier retreat is unknown and indicated on a relative scale. Lines upper the horizontal line indicate glaciers smaller than at the end of the 20th century and lines below the horizontal line denote periods with larger glaciers than at the end of the 20th century. The radiocarbon dates are calibrated and all curves are presented in calendar years. Franz Josef Land (Lubinski et al., 1999), Svalbard from Svendsen and Mangerud (1997) corrected with Humlum et al. (2005), Northern Scandinavia (Bakke et al., 2005a,b; Nesje et al., 2005), Southern Scandinavia (Dahl and Nesje, 1996; Matthews et al., 2000, 2005; Lie et al., 2004), Brooks Range (Ellis and Calkin, 1984), Canadian Cordillera (Luckman and Kearney, 1986; Osborn and Luckman, 1988; Koch et al., 2004; Menounos et al., 2004), Alps (Holzhauser et al., 2005; Jörin et al., 2006), Himalaya and Karakorum (Röthlisberger and Geyh, 1985; Bao et al., 2003), Mt. Kenya (Karlén et al., 1999), New Zealand (Gellatly et al., 1988).