Temporal variability in central European climate reconstructions, 1501–2020 CE, and its attribution

Abstract

<p>Despite their pivotal role in climate research, direct instrumental records of meteorological variables are only available for the most recent part of climate history. Even in regions with longest tradition of weather measurements, such as central Europe, the existing series rarely comprise more than two centuries of reliable data. However, documentary sources, both quantitative and qualitative, can be employed to substantially extend the available records. Using the resulting multi-centennial data, previously unexplored features of climate system’s evolution can then be studied.</p><p>In this analysis, temporal variability in annual and seasonal series of temperature, precipitation and drought indices (Standardized Precipitation Index - SPI, Standardized Precipitation Evapotranspiration Index - SPEI, Z-index), pertaining to the territory of contemporary Czechia, has been studied over the 1501–2020 CE period. The series under investigation were reconstructed from multitude of Czech documentary data sources, combined with instrumental observations. Phenoclimatic temperature and SPEI reconstructions, derived from historical records of cereal and grape harvest dates, were also employed and compared to their documentary-based counterparts.</p><p>Statistical attribution analysis, utilizing multiple linear regression, confirmed the influence of covariates related to volcanic activity (prompting temporary temperature decreases, especially during summer) and the North Atlantic Oscillation (influential in all seasons except summer for all target variables) in the Czech climate reconstructions. Statistically significant components correlated with multidecadal variability in the northern Atlantic and northern Pacific (represented by multiproxy-reconstructed AMO and PDO indices) were identified in the Czech temperature and precipitation series as well as in all drought indices. Additionally, using wavelet and cross-wavelet analysis, notable oscillations shared by the AMO/PDO variations and the Czech climate series were found, particularly at periods of approximately 70–100 years.</p>