Atmospheric controls on rainfall and soil water 18O; implications for proxy-based NAO reconstruction in Europe

Principal Supervisor: Prof. P. Frank McDermott


The North Atlantic Oscillation (NAO) is de ned as the normalised sea level pressure di erence between the Azores and Iceland, and is the leading mode of European winter climate variability. The NAO is responsible for much of the inter-annual variability in mean wind speed and direction, heat and moisture transport between the Atlantic and the surrounding areas, as well as the intensity, number and trajectory of storm tracks (Hurrell et al., 1995). The instrumental NAO index is, however, too short (1821-present) to make de nitive statements about possible changes in its behaviour that might be linked to anthropogenic climate change. Consequently, there have been several e orts to use a range of climate sensitive proxies to reconstruct the NAO back through the last few millennia (Goodkin et al., 2008; Trouet et al., 2009; Moreno et al., 2012). However, the only available reconstruction for this period suggested a persistent, extreme and positive NAO state during the Medieval Climate Anomaly (MCA; 1050-1250) implying a fundamental change in atmospheric dynamics that is not captured by Ocean-Atmosphere General Circulation Models (OAGCMs; Lehner et al., 2012).

Identi cation of regions where the proxy-NAO relationship is robust during the instrumental period is essential to nd a suitable location for paleoclimate studies seeking to reconstruct the NAO. For this reason, my study focuses partly on the causes of temporal and spatial nonstationarity of the NAO-climate and NAO-rainfall 18O relationships over Europe. In particular, this thesis documents for the rst time that di erent NAO=EA (i.e. East Atlantic pattern) and NAO=SCA (i.e. Scandinavian pattern) combinations systematically influence winter climate conditions in Europe as a consequence of NAO dipole migrations. These migrations influence the non-stationarity of NAO-climate relationships, yielding a complex NAO-climate link in some regions. Likewise, the NAO-rainfall 18O relationship is also shown to be a ected. This is especially important, as 18O is a natural proxy commonly used in paleoclimatology. These findings could, for instance, explain some of the discrepancies between previously published NAO reconstructions.

In addition, since speleothems precipitate from water derived originally from meteoric precipitation, a good understanding of the processes influencing 18O before capture in speleothem carbonate is required to successfully attribute changes in stalagmite 18O to climate parameters. Of particular concern is how the 18O of in ltrated water may be modi ed in the soil zone. Here, a new soil water model which successfully captures the 18O variability observed at drip sites in La Garma cave (N. Spain) is presented to provide a better framework with which to evaluate the relationship between GAR-02 stalagmite from La Garma cave and rainfall 18O.
The obtained results clearly demonstrate the rationale for the choice of N. Spain as a suitable location for NAO reconstruction. Hence, a new 1,200-year 18O record from stalagmite GAR-02 from La Garma cave is presented as a new NAO proxy. GAR-02 preserves clear annual laminations which have been sampled for oxygen isotope analysis and its chronology is based on a Bayesian age-depth model using lamina counts in combination with 14C measurements. Crucially, while GAR-02 18O con rms an overall positive NAO circulation state during the MCA, it is not as extreme as that of Trouet et al. (2009). Also, several excursions towards negative NAO states,
consistent with documentary data, are reconstructed from GAR-02 18O. These results have important implications for OAGCMs, because these appear to produce too little long-term NAO variability during the last millennium.