Quantitative pollen-vegetation relationships in southern subtropical evergreen broadleaved forests in China and the effects of human disturbance

The quantitative relationship between modern pollen and vegetation provides a critical foundation for reconstructing past vegetation, with relative pollen productivity (RPP) serving as a key calibration parameter. However, in subtropical evergreen broadleaved forests (SEBFs) in China, RPP studies remain scarce, and the impact of human disturbances on RPP estimates has yet to be adequately assessed, limiting the accuracy of quantitative palaeovegetation reconstructions. This study was conducted in Dinghu Mountain Nature Reserve and its surrounding areas in Zhaoqing, Guangdong Province, and included 31 sampling sites. We performed pollen analysis alongside detailed vegetation surveys and utilized ERV submodel 2 and Prentice’s model to estimate the RPP of 9 common plant taxa in the southern SEBFs. There was a particular focus on evaluating the interference effects of bamboo plantations on the estimation of RPP. The results indicate that bamboo within the family Poaceae contributes minimally to surface soil Poaceae pollen because of its unique flowering characteristics, such as long flowering cycles and monocarpic reproduction. The incorporation of bamboo into the Poaceae vegetation coverage in the analysis led to excessively high RPP values for the other taxa. When bamboo coverage was removed from the Poaceae family, the recalculated RPP values aligned closely with those reported in previous studies. The RPP values, ranked from highest to lowest, were as follows: Castanopsis (12.33±0.03) > Araliaceae (1.60±0.03) > Mallotus (1.53±0.26) > Pinus (1.47±0.03) > Rosaceae (1.07±0.02) > Poaceae (1±0) > Euphorbiaceae (0.44±0.03) > Anacardiaceae (0.26±0.03) > Theaceae (0.15±0). Notably, the RPP values for Mallotus, Araliaceae, Theaceae, and Euphorbiaceae represent the first estimates for China’s subtropical region. Differences between certain RPP estimates and those of previous studies may be attributed to factors such as species composition, vegetation structure, and model selection. The findings of this study highlight that due to the widespread distribution of artificial bamboo forests in China’s subtropical regions, future RPP studies should carefully consider the influence of Poaceae. This consideration is essential for improving the accuracy of the application of fossil pollen for quantitative paleo-vegetation reconstruction in these regions.