An increasing number of studies report stable-isotope and trace-element records in speleothems. This variation is controlled by diverse environmental variables including the climatically important variables, temperature and rainfall. There is, however, a paucity of laboratory studies attempting to understand the influence of these environmental controls on stalagmite geochemistry. Quantitative data from such studies would dramatically improve our ability to reconstruct palaeoclimate from stalagmites. This study produces a new calcite growth setup, which closely mimics natural processes (e.g. precipitation driven by C02 degassing, low ionic strength solution, thin solution film) but with a tight control on growth conditions (temperature, pC02, drip rate, calcite saturation index and the composition of the initial solution). Results of a series of experiments (12 experiments), which investigates synthetic calcite growth at four different temperatures (7, 15, 25 and 35°C) and three different drip rates (1.6,5.8 and 10.4 drips/min) provide important, new information on speleothem chemistry and how it relates to environmental conditions. A relationship between temperature, drip rate and growth rate is derived and shows that temperature is the dominant control on growth rate, although the effect of saturation index is not yet known. Varying degrees of δ180calcite equilibrium/disequilibrium are observed throughout the range of temperatures and drip rates and demonstrate that faster drip rates are more likely to produce 8180calcite in equilibrium with the drip solution. Trace metal results provide cave-analogue partition coefficient values for Mg, Co, Sr, Cd, Ba and U. In agreement with previous studies, Mg/Ca is shown to respond to both temperature and the percentage of calcite precipitated (%cp). D(Cd), not previously used in stalagmite studies, shows a more sensitive response to %cp and therefore has the potential to act as a proxy for effective rainfall with increases in Cd/Ca corresponding to wetter conditions. The near unity value of D(Co) limits the response of Co to %cp and therefore variation in Co/Ca should be indicative of a different kind of environmental change. In addition to the synthetic calcite experiments, early results were presented from palaeoclimate reconstruction studies in Southern Chile and Northern Morocco. U-Th chronology on the two Chilean stalagmites indicate growth during the second half of the Holocene period. The stable isotope records of both samples are comparable at 4ka BP but the records diverge after that period. Robust palaeoclimate interpretations are hampered by this divergence, by a lack of modern drip water data and of a duplicate zero-age stalagmite record. Results from Northern Morocco represent an inItial investigation of the suitability of Moroccan caves for future palaeoclimate work. Results indicate that Ghar Cahal, located close to Tetouan, may be suitable for retrieving palaeoclimate data that will be directly comparable with archaeological records from that same location.

Climatic conditions throughout the Paleocene (55 - 65.5 Ma) are believed to have remained relatively stable (Shackleton and Hall, 1984a,b; Zachos et al., 1994; Zachos et al., 2001). The Paleocene is generally considered to be time of global warmth and reduced latitudinal temperature gradients compared with the present day (e.g. Zachos et al. 1993, 1994), although generally experienced temperatures relatively lower than those of the late Cretaceous and early-mid Eocene (Zachos et al., 2001; Pagani et al., 2005). It also experienced elevated levels of atmospheric CO2 compared to the present day. Some evidence however suggests that the Paleocene may not have been as stable as is generally accepted; the devastation of marine biota at the Cretaceous/Tertiary boundary (K/Pg) may have disrupted biogeochemical cycles for up to 1 - 3 My (D'Hondt et al., 1996a; 1998), as evidenced by the sustained collapse or suppression of the benthic-planktic carbon isotope gradient (Keller and Lindinger, 1989; Zachos et aI., 1989; Zachos et al, 1992; D'Hondt et al., 1998; Coxall et al., 2006). In the late Paleocene, a decrease in global benthic temperatures (Zachos et al., 2001; Cramer et al., 2009) is recorded, coinciding with a positive carbon isotope excursion in marine carbonates. The Paleocene may therefore represent an epoch with a relatively dynamic carbon cycle and climate, and as such provides an ideal period to study the relationship between carbon cycling and climate change in a high-Co- Earth system. The central aim of this thesis was to determine whether significant climate and carbon cycle instability occurred through the Paleocene using organic biomarker approaches, including the TEX86 palaeothermometer and compound specific carbon isotope analysis of algal and terrestrial biomarkers. Transient climate change and ecological disruption occurred at the K/Pg in the southwest Pacific. Climate instability persisted for c. 1 - 1.2 My, and then proceeded to stabilise, with most climate and ecological parameters returning to pre-K/Pg values; a notable exception is the algal biomarker distributions, which reflect a restructured algal community, in keeping with the suggestion that algal community restructuring, and thus the restructuring of tropic levels, was responsible for the long-term term recovery of the benthic-planktic carbon isotope gradient. Late Paleocene climate reconstructions indicate cooling coeval with l3C-enrichment of terrestrial and marine reservoirs in the Southern Ocean (SO), perhaps associated with enhanced marine productivity and a drawdown of CO2, indicating that this period, referred to as the Paleocene Carbon Isotope Maximum (PCIM) reflects a period of cooling associated with carbon cycle changes. A tentatively dated low-latitude north Atlantic SST III record does not indicate cooling. Assuming a correct age assignment, this could indicate that SST cooling was a more regionally restricted phenomenon, and that the global benthic carbonate ISO-enrichment reflects Southern Ocean (SO) cooling and strengthening of SO sourced bottom waters. Tentatively, drawdown of CO2 and Southern Ocean cooling could have brought about early Antarctic glaciations and an associated drop in sea level. Although such an inference is contentious and evidence presented here is circumstantial, such an event provides a mechanism for the pronounced oceanographic changes, including enhanced cooling and sedimentary anoxia, occurring at the neritic mid- Waipara River site from 58.3 - 58 Ma. Furthermore, these findings suggest that the Paleocene SO climate may have been more sensitive to carbon cycle dynamics than the classic ocean heat transport models suggest; this is in agreement with Paleogene climate models which predict a stronger influence of CO2 on Southern Ocean climate than the thermal isolation brought about by circum-Antarctic circulation (Hub er and Sloan, 2001; DeConto and Pollard, 2003a; 2003b; Huber and Nof, 2004; Huber et al., 2006; DeConto et al., 2008).

The climate of the Neoproterozoic Snowball Earth is tested in the UKMO Unified Model, specifically the HadCM3 climate model. The model is largely left unchanged, but the boundary conditions, both external and initial, are adjusted to create experiments based on the Snowball Earth hypothesis. The model can reproduce multiple equilibrium climates, as have been seen in energy balance models of the Earth's climate. The modelled present day and Neoproterozoic versions of Earth can both reproduce both ice capped and ice covered climate states. Neither can reproduce a climate which remains ice free throughout the year, even with an equilibrated ocean or elevated levels of C02. In all cases the ice free climate reverts toward the ice capped climate after the first polar winter. The modelled Neoproterozoic ice covered climate, that is the climate of Snowball Earth, has a climate very different from the present day. These changes are mostly driven by the lower thermal inertia, latitudinal temperature differences and the changed meridional circulation that results. The weather of the modelled Snowball Earth climate is also very different, dom- inated by a strong diurnal variation due to solar heating, as opposed to the more varied weather in the present day. The model responds well to the conditions of the Snowball Earth climate, with temperatures similar to those predicted by a simple physical model. The model responds less well to high levels of C02 in the Snowball Earth climate. The ice model also allows excessive heat and moisture to escape from the ocean into the atmosphere compared to that that would be predicted from solid ice coverage of the ocean. The exit from a Snowball Earth state was also tested within the model. Neither an decrease in albedo nor an increase in CO2 is unable to increase the temperature of the climate system sufficiently to exit the Snowball Earth state.