Search

To better understand the long-term changes of the East Asian summer monsoon precipitation (Pjja), quantitative reconstructions and model simulations are needed. Here, we develop continental-scale pollen-based transfer functions for Pjja with weighted averaging–partial least squares (WA-PLS) regression and a Bayesian multinomial regression method. We apply these transfer functions to a set of fossil pollen data from monsoonal China for quantitatively reconstructing the Pjja changes over the last 9500 years. We compare the reconstructions with Pjja simulations from a coupled atmosphere–ocean–sea ice general circulation model (the Kiel Climate Model, KCM). The results of cross-validation tests for the transfer functions show that both the WA-PLS model (r2 = 0.83, root mean square error of prediction (RMSEP) = 112.11 mm) and the Bayesian model (r2 = 0.86, RMSEP = 107.67 mm) exhibit good predictive performance. We stack all Pjja reconstructions from northern China to a summary curve. The stacked record reveals that Pjja increased since 9500 cal. yr BP, attained its highest level during the Holocene summer monsoon maximum (HSMM) at ~7000–4000 cal. yr BP and declined to present. The KCM output and the reconstructions differ in the early-Holocene (~9500–7000 cal. yr BP) where the model suggests higher Pjja than the reconstructions. Moreover, during the HSMM, the amplitude of the Pjja changes (~20–60 mm above present) in simulations is lower than the reconstructed changes (~70–110 mm above present). The rising (declining) Pjja patterns in reconstructions before (after) the HSMM are more pronounced and fluctuating than in simulations. Other palaeohydrological data such as lake-level reconstructions indicate substantial monsoon precipitation changes throughout the Holocene. Our results therefore show that the KCM underestimates the overall amplitude of the Holocene monsoon precipitation changes. [ABSTRACT FROM AUTHOR]

A modern pollen dataset from China and Mongolia (18–52°N, 74–132°E) is investigated for its potential use in climate reconstructions. The dataset includes 2559 samples, 229 terrestrial pollen taxa and four climatic variables — mean annual precipitation (P ann ): 35–2091 mm, mean annual temperature (T ann ): –12.1–25.8 °C, mean temperature in the coldest month (Mt co ): –33.8–21.7 °C, and mean temperature in the warmest month (Mt wa ): 0.3–29.8 °C. Modern pollen–climate relationships are assessed using canonical correspondence analysis (CCA), Huisman–Olff–Fresco (HOF) models, the modern analogue technique (MAT), and weighted averaging partial least squares (WA-PLS). Results indicate that P ann is the most important climatic determinant of pollen distribution and the most promising climate variable for reconstructions, as assessed by the coefficient of determination between observed and predicted environmental values (r 2 ) and root mean square error of prediction (RMSEP). Mt co and Mt wa may be reconstructed too, but with caution. Samples from different depositional environments influence the performance of cross-validation differently, with samples from lake sediment-surfaces and moss polsters having the best fit with the lowest RMSEP. The better model performances of MAT are most probably caused by spatial autocorrelation. Accordingly, the WA-PLS models of this dataset are deemed most suitable for reconstructing past climate quantitatively because of their more reliable predictive power. [ABSTRACT FROM AUTHOR]