North Atlantic Oscillation imprints in the Central Iberian Peninsula for the last two millenniafrom ordination analyses to the Bayesian approach

Resumen

The climate variability of the Iberian Peninsula (IP) can be explained in terms of relatively few large-scale atmospheric modes of climate variability, such as the North Atlantic Oscillation (NAO), the East Atlantic (EA) and the Scandinavian (SCAND) patterns. Many studies on the present-day IP climatology clearly show that the NAO is the most prominent mode, especially in winter. However, the most recent investigations have highlighted that, in spite of this importance, other climate modes seem to play a key role in both modulating the NAO-climate relationship and controlling certain meteorological parameters, such as the precipitation and/or temperature during a given season of the year. These complex present-day climate dynamics have begun to be well characterized from the meteorological perspective, but little is known about the past evolution of these climate interactions. Furthermore, there is a reasonable understanding of the past NAO evolution in the northern and the southern latitudes of the IP, but almost no information is available on the evolution of this climate mode in the Central IP. This knowledge is crucial to accurately characterizing the past climate evolution of the entire IP. Within this framework, the main aim of this PhD thesis is to characterize the impacts of the NAO on the Central IP over the last 2,000 years. For that, three steps have been followed: 1) the establishment of a conceptual model to depict the NAO influence on the ice phenology of the Peñalara and Cimera alpine lakes (Iberian Central Range, ICR); 2) to characterize the main climate changes in the ICR using Peñalara and Cimera sediments and the spatio-temporal evolution of the NAO, as well as its relationships with other climate modes, over the last two millennia; and 3) to perform a quantitative regional NAO reconstruction using the geochemical composition of the Cimera Lake sediments and a random walk-modularised Bayesian model. The conceptual lake model formulated to understand the present-day influence of the NAO on the limnological evolution of Peñalara (2016 m asl) and Cimera (2140 m asl) lakes was established using Pearson's correlation coefficients between seasonal time-scale series of the NAO index, climatic data (i.e., precipitation, air temperature and precipitation data) and ice phenology records from both lakes. The results suggest that the effects of the NAO are only reflected in the thawing process via the air temperature and the insulating effect of snow accumulation on the ice cover. An altitude component is evident in our survey because the effects of the NAO on Peñalara Lake are restricted to winter, whereas for higher Iberian alpine lakes (i.e., Redon Lake, Pyrenees), the effects extend into spring. A latitudinal component is also clear: in northern Europe, the NAO signal is primarily reflected in lake ice phenology via the air temperature, whereas our results confirm that in southern Europe, the strong dependence of both precipitation and temperature on the NAO determines the importance of these climatic variables for lake ice cover. The past NAO impacts on the Central IP were determined by the multi-proxy characterization of the sediments of Peñalara and Cimera lakes using ordination statistical analyses. This approach was used to reconstruct the intense runoff events, the lake productivity and the soil erosion in the Cimera Lake catchment and to interpret these factors in terms of temperature and precipitation variability in the ICR for the last two millennia. The spatio-temporal integration of this reconstruction with other Iberian reconstructions was employed to identify the main climate drivers over this region. During the Roman Period (RP; 200 BC – 500 AD) and the Early Middle Ages (EMA; 500 – 900 AD), N–S and E–W humidity gradients, respectively, were dominant in the IP, whereas during the Medieval Climate Anomaly (MCA; 900 – 1300 AD) and Little Ice Age (LIA; 1300 – 1850 AD), these gradients were not evident. These differences could be ascribed to the interactions between the NAO and EA climate modes. During the RP, the generally warm conditions and the E–W humidity gradient in the IP indicate a dominant interaction between a negative NAO phase and a positive EA phase (NAO-–EA+), whereas the opposite conditions during the EMA indicate a NAO+–EA- interaction. The dominantly warm and arid conditions during the MCA and the opposite conditions during the LIA in the IP indicate the interaction of the NAO+–EA+ and NAO-–EA-, respectively. Furthermore, the higher solar irradiance and fewer tropical volcanic eruptions during the RP and MCA may support the predominance of the EA+ phase, whereas the opposite conditions during the EMA and LIA may support the predominance of the EA- phase, which would favour the occurrence of frequent and persistent blocking events in the Atlantic region. In addition, evidence of African dust inputs in these lakes could denote a coupled displacement between the Intertropical Convergence Zone and the NAO during the study period. Finally, a Bayesian random walk-modularised model was formulated to quantitatively reconstruct the evolution of the NAO impacts in the ICR (NAOICR) for the last two millennia using the raw chemical element profiles obtained from the Cimera Lake sediments using an X-Ray-Fluorescence Avaatech® core scanner. The obtained quantitative values of the NAOICR were in accordance with previously reconstructed precipitation and temperature conditions. In addition, the comparison of the NAOICR with other NAO approaches show that the local impact of the NAO can also display global aspects of this climate mode and that this impact reconstruction could therefore be considered an approximately regional index for the entire IP.