cc: "Jonathan Overpeck" , "Eystein Jansen" , "Keith Briffa" date: Fri, 5 Aug 2005 11:33:10 +0400 from: "Olga Solomina" subject: ground surface temp ch 4 and 6 coordination to:  Dear Tingjun, Please have a look at the portion of our chapter concerning the ground surface temperature reconstructions - just to be sure that it is not repeated somehow in your chapter. The text is written by Keith Briffa ([1]k.briffa@uea.ac.uk), so you might contact him if it is necessary. All the best, olga What do large-scale temperature histories from ground surface temperature measurements tell us?[2][OS1] Ground surface temperature (GST) histories reconstructed from direct measurements of subsurface temperatures in boreholes have been presented by several geothermal research groups (Pollack and Huang, 2000; Harris and Chapman, 2001; Pollack and Smerdon, 2004); see Pollack and Huang (2000) for a review of the methodology. These have all been derived using the contents of a publicly-available database of borehole temperatures (Huang and Pollack, 1998) that presently includes 695 sites in the Northern Hemisphere. These geothermal reconstructions provide independent estimates of surface temperature changes over the past five centuries, with which to compare the other proxy reconstructions, but because the Earth acts as a low-pass filter on downward-propagating temperature signals, decadal and shorter fluctuations are generally unresolved in borehole reconstructions. The coupling between above and below ground temperatures is known to be complex and affected by factors such as ground vegetation and snow cover. Within the year, summer evapotranspiration can cool the ground relative to the surface air temperature; while in winter ground surface freezing can prevent the subsurface transfer of cold air temperature signals. On longer timescales changing vegetation and snow cover are likely to result in complex and spatially varying biases in the way SAT are reflected in local GST reconstructions (Gosnold et al., 1997; Pollack and Huang, 2000; Hinkel et al., 2001; Kane et al., 2001; Sokratov and Barry, 2002; Lin et al., 2003; Mann and Schmidt, 2003; Stieglitz et al., 2003; Bartlett et al., 2004; Chapman et al., 2004; Smerdon et al., 2004, 2005). The few studies to date, using global simulations by three-dimensional coupled models, provide contradictory evidence of the likely accuracy of deep soil temperatures as an indicator of SAT on longer timescales. Gonzalez-Rouco et al. (2003) using the ECHO-G model, concluded that deep soil temperatures were indistinguishable from continental annual SAT, but in another simulation using GISS model E, Mann and Schmidt (2003) found a significant discrepancy between cold-season GST and SAT trends on multi-decadal timescales, linked to changing snow cover trends. Neither of the model simulations used in these studies included time-varying vegetation cover. Figure 6.5.1-1a includes one reconstruction of average Northern Hemisphere GST, that by Pollack and Smerdon (2004). They state that this (like all geothermal reconstructions) shows a somewhat muted estimate of the real 20th-century trend, because about half of the borehole sites at the time of measurement (which varies by up to decades) had not yet been exposed to the significant warming of the last two decades of the 20th century. The one standard deviation uncertainties for their series (not shown here) are 0.1 (in 1500), 0.5 (1800) and 0.3 (1900) ºC. These are minimum errors (associated with various regional aggregations of local records) and do not take account of site specific noise in individual local site reconstructions (Pollack and Huang, 2000). This reconstruction indicates temperatures during the 16th and 17th centuries that coincide with the lower range of multiproxy reconstructions, values in the centre of the range during the 19th and early 20th century and an overall warming near to 1.0ºC over the last 500 years. Not all hemispheric analyses of these borehole data give the same magnitude of warming (Rutherford and Mann, 2004) but all (Huang et al., 2000; Harris and Chapman, 2001; Beltrami, 2002a) are more consistent with the multi-proxy reconstructions that show the greatest warming during this period, and in less agreement with those that show least. ______________________________ [3][