Reconstruction of winter temperature of southwest China between 1700 and 1949 based on a Bayesian approach

Using the period 1950–1999 as the reference, winter temperatures in southwest China from 1700 to 1949 were reconstructed with a Bayesian approach. Documentary data were used to extract weather and climate information for generating Cold Winter Index (CWI). Two ensembles of paleo-simulations (ModE-Sim and CESM-LME) served as priors. The uncertainty inherent in documentary data, together with the climate dependence of observations, informed the likelihood estimation. Three reconstructed series were produced: CWI-ModE-Sim, CWI-LME, and a time-independent CWI-ModE-Clim. This dataset provides the CWI, prior, likelihood, posterior, and the three final reconstructions.