ALBERT

Description: Environmental reconstructions based on fossil pollen rely on the understanding of modern pollen distribution along climatic and biogeographic gradients. This study analyses the modern pollen spectra of Central America using three basic approaches: (1) the evaluation of using modern pollen spectra to differentiate the main vegetation types of the region, (2) the usage of non-linear regression to predict individual pollen abundances as a function of climate, and (3) the construction of pollen—climate transfer functions. Standard pollen analysis was carried out on mud—water interface samples from 81 lakes in the Yucatan Peninsula and adjacent mountains of Guatemala and Mexico. Detrended correspondence and cluster analyses were used to evaluate the biogeographic patterns revealed by this modern pollen data set. Non-parametric locally weighted scatterplot smoothing (LOESS) regression was used to construct pollen—climate functional relationships. Five modern vegetation types were clearly identifiable through their associated pollen spectra: Pinus forest, Quercus forest, mountain mesophyllous forest, tropical rainforest, and tropical seasonal forest. The last group includes three subcategories (evergreen seasonal, tropical semi-deciduous, and tropical deciduous forests), which were not separable via this analysis. Precipitation and temperature trends were consistent and robust for at least 28 and 30 taxa, respectively, in the LOESS regression. While floristic patterns driven by temperature were clearly reflected by the pollen spectra, those driven by precipitation were less sharply defined. Nevertheless, pollen data from the study area offered good resolution to identify broad biogeographic patterns. Furthermore, individual taxa showed high predictability along precipitation and temperature gradients, allowing the theoretical construction of pollen—climate transfer functions. This study provided valuable tools for the interpretation of fossil pollen sequences from Central America.

Description: In general, a moderate drying trend is observed in mid-latitude arid Central Asia since the Mid-Holocene, attributed to the progressively weakening influence of the mid-latitude Westerlies on regional climate. However, as the spatio-temporal pattern of this development and the underlying climatic mechanisms are yet not fully understood, new high-resolution paleoclimate records from this region are needed. Within this study, a sediment core from Lake Son Kol (Central Kyrgyzstan) was investigated using sedimentological, (bio)geochemical, isotopic, and palynological analyses, aiming at reconstructing regional climate development during the last 6000 years. Biogeochemical data, mainly reflecting summer moisture conditions, indicate predominantly wet conditions until 4950 cal. yr BP, succeeded by a pronounced dry interval between 4950 and 3900 cal. yr BP. In the following, a return to wet conditions and a subsequent moderate drying trend until present times are observed. This is consistent with other regional paleoclimate records and likely reflects the gradual Late Holocene diminishment of the amount of summer moisture provided by the mid-latitude Westerlies. However, climate impact of the Westerlies was apparently not only restricted to the summer season but also significant during winter as indicated by recurrent episodes of enhanced allochthonous input through snowmelt, occurring before 6000 cal. yr BP and at 5100–4350, 3450–2850, and 1900–1500 cal. yr BP. The distinct ~1500-year periodicity of these episodes of increased winter precipitation in Central Kyrgyzstan resembles similar cyclicities observed in paleoclimate records around the North Atlantic, likely indicating a hemispheric-scale climatic teleconnection and an impact of North Atlantic Oscillation (NAO) variability in Central Asia.