A modern pollen–temperature calibration data set from Korea and quantitative temperature reconstructions for the Holocene

Modern surface pollen samples from the mountains along the east coast of Korea were used to derive pollen–temperature transfer functions. Detrended correspondence analysis (DCA) and detrended canonical correspondence analysis (DCCA) were performed to test the robustness of the relationship between the modern pollen assemblages and temperatures. The relationship exhibited a high correlation (DCA, r = −0.887; DCCA, r = −0.908). The performance of the best weighted averaging partial least squares (WA-PLS) transfer function was statistically good ($r_{\text{jack}}^2$ = 0.74; RMSEP = 1.79°C). In order to quantitatively reconstruct the Holocene temperature changes, the best model was applied to five fossil pollen records produced from four coastal lagoons of the east coast and one high-altitude peat bog. Anomalies calculated from reconstructed paleotemperature data were combined to generate a synthesis temperature reconstruction for the east coast of Korea, in which the ‘Medieval Warm Period’, ‘Little Ice Age’, and ‘Migration Period’ were clearly shown. This study demonstrated the validity of the quantitative reconstruction of paleotemperature using the pollen–climate transfer function, even in heavily human-impacted areas such as the Korean Peninsula.