and varved sediments (e.g., Rozanski et al., 1992) offer possibilities.
Figure 5 shows a 1200-year multi-parameter record from GISP2 subjected to a 50-year smoothing. Records of decadal Iceland sea-ice frequency and of decadal Iceland temperatures derived from the sea-ice record (Bergthorsson, 1969) have been added for comparison. The period covered includes the Medieval Warm Period (MWP), from about A.D. 1000 to between A.D. 1350 and A.D. 1450, and the Little Ice Age (LIA), from about A.D. 1450 until about A.D. 1900, as defined in northwest Europe. The Icelandic record shows clear warm and cold periods corresponding to the northwest European MWP and LIA. The GISP2 d18O record, on the other hand, shows no clear, sustained evidence of the MWP and LIA. Rather, it shows a pattern of minor, century-scale fluctuations throughout this part of the record. During the period of the LIA the warm phase of these fluctuations appears to be more subdued, giving on average a colder climate. A similar observation was made at Crête, Greenland (Dansgaard et al., 1975). Accumulation, determined by identification of individual annual layers and flow modeling (Alley et al., 1993; Meese et al., 1994b), decreases after about A.D. 1200 and is, on average, lower during the LIA period. Terrestrial source indicators such as particles, Ca, Mg, and K, as well as marine indicators such as Na and Cl, are low during the early part of this period (MWP and early LIA) and increase around A.D. 1600, indicating increased transport by stronger winds since A.D. 1600 (see Mayewski et al., 1993b). Particles can also be of volcanic origin, in which case they are accompanied by elevated levels of non-sea-salt sulfate, and may include volcanic glass shards. The non-sea-salt sulfate, which reflects primarily volcanic input as plotted here, does not indicate that volcanism was a major factor in LIA climate (Mayewski et al., 1993b). Concentrations of nitrate (from lightning and soil exhalation, for example) and methanesulfonic acid (MSA, from oxidation of ocean-produced dimethylsulfide) decrease around A.D. 1350 and A.D. 1430, respectively, coinciding with the transition from the MWP to the LIA in Europe. Ammonium, which reflects primarily biomass burning as plotted here, shows high values during periods of climate (and, presumably, vegetation) change.
The multi-parameter ice-core data thus reveal a far richer and more detailed picture of climatic changes and the accompanying environmental alterations during the MWP and ensuing LIA than was available from other sources. This highly detailed record extends back through the glacial-to-interglacial transition well into the last glacial period. The significance for society of the relatively minor climate fluctuations observed in the recent GISP2 ice core record is dramatically illustrated by the fate of the Norse colonies on Greenland that thrived during the MWP from about A.D. 1000 to A.D. 1300, but lost contact with Iceland and disappeared when climate deteriorated after A.D. 1400.
Figure 5 shows coherent patterns of change in various properties in the recent record. Yet it is difficult to study the dynamics of comparatively small century-scale climate changes in the presence of much larger short-term climate fluctuations that are not yet understood. A better situation exists across the last glacial-to-interglacial transition, which is apparent between 1670 and 1800 m of depth in the GISP2 core. There the ice has preserved in great detail a record of various climate parameters that show rapid century-scale fluctuations during this period of major climate change. A detailed time scale for the core is being constructed from the annual layering observed in visual stratigraphy, in the electrical conductivity (ECM), and in the particle concentrations of the ice (Alley et al., 1993; Meese et al., 1994a). This will provide the core with an internal time scale with annual resolution for determining the relative timing of the changes in climate-related core properties, such as stable isotopes, chemistry, dust, ice fabric, and atmospheric gas and isotope composition, during the transition from glacial to interglacial climate.
The d18O record of the glacial-to-interglacial transition (Figure 6) clearly shows the Younger Dryas as the last period of full glacial, low d18O values before the Holocene, lasting from about 11.6 to 12.9 kyr BP (Alley et al., 1993). The Younger Dryas is preceded until 14.7 kyr BP by a period of ''warmer" but not fully interglacial climate, probably corresponding to the Allerød/Bølling interstadial complex described in Europe (Mangerud et al., 1974). The interstadial is interrupted by three brief (about a century) cooler episodes, of which the middle one is fairly weak. The ECM record (Taylor et al., 1993), which is measured continuously along the core with a resolution of 1 mm per sample, shows a pattern of highs and lows quite similar to those of the d18O, except that ECM values drop to almost zero during periods of low d18O. The second, weak, cold period of the Allerød/Bølling is not clearly developed in the ECM. The ECM records the presence or absence in the ice of alkaline dust that neutralizes the acid responsible for most of the electrical conductivity of the core (Taylor et al., 1992); it is thus strongly anti-correlated with the calcium concentrations (Mayewski et al., 1993c) in the ice.
The high-Ca, high-dust/low-ECM, low d18O combination in the Younger Dryas indicates a cold climate with strong winds. This finding, which is supported by other terrestrial and marine source indicators such as Mg, Na, and Cl (Mayewski et al., 1993c), fits the general concept of increased meridional temperature gradients and atmospheric circulation during glacial times (Manabe and Hahn, 1977). Comparison of the three Allerød/Bølling cool phases shows that while d18O and Ca concentrations assume intermediate values, the ECM drops to near zero in the two colder ones and shows little change in the third. This is as expected for