A major goal in palaeoclimatology is to reconstruct reliable, well-dated, terrestrial climate records and to correlate them with the better understood oceanic and ice-core data. It has long been recognised that as a consequence of the temperature stability and high relative humidity (~99%) of cave air, cave calcites (speleothems) can preserve a reliable palaeotemperature signal preserved as variations in calcite d18O through time. Here we present new U-Th dated stable isotope records for post-glacial (<18 ka) speleothem calcite from selected caves in S.W. Ireland. Mean annual temperatures in S.W. Ireland reflect sea-surface temperature fluctuations in the N. Atlantic, and so well-dated terrestrial records from this region can provide tests for models which ascribe to the N. Atlantic thermohaline circulation, sudden changes in the climate of north-west Europe. In detail, the
d18O ratio in speleothem calcite is controlled by two competing processes; namely (i) a surface meteoric water/precipitation effect and (ii) an in-cave temperature effect and the balance between the two processes depends on location. Thus, in S.W. Ireland, the meteoric water/precipitation effect is strong at coastal sites (positive d 18O - temperature relationship), whereas the cave-temperature effect (negative d18O - temperature relationship) dominates at inland sites. Mass-spectrometric U-Th dated d18O curves for late glacial and Holocene speleothems from the Mitchelstown
(d18O = -4.7 to -3.5) and Crag (d18O = -4.29 to -1.95) caves in S.W. Ireland will be presented and compared with other terrestrial palaeoclimate proxy data to illustrate the potential of cave calcites as palaeoclimatic recorders. The stable isotope data from the two Irish sites can be combined to derive a preliminary late-glacial/Holocene palaeo-temperature curve featuring clear evidence for Younger Dryas cooling (by more than 5 °C), and a gradual mid-Holocene warming (by 1-2 °C), followed by gradual cooling towards the present-day. The termination of the Younger Dryas inferred from the speleothem record (11,565 ± 145 years) is in good agreement with that inferred from annual layer counting in the Greenland Ice-Core Project (GRIP) core (11,640 ± 250 years). The data also provide new constraints on the rates of Younger Dryas cooling (rapid) and warming (gradual), and the implications of these observations for models of the Younger Dryas will be discussed.