Modern pollen thresholds for tree presence on the eastern Tibetan Plateau and their potential application

Abstract

Because of the issue of exogenous arboreal pollen, it is imperative to investigate modern pollen thresholds for tree presence to aid in the reliable reconstruction of past forest distribution. The influence of exogenous pollen is especially prominent on the eastern Tibetan Plateau, with its long elevational gradient and short distance between forest and meadow. Here, we compiled 1186 modern pollen samples from forest ecotones on the eastern Tibetan Plateau (evenly distributed) to investigate the quantitative pollen thresholds indicating tree presence. Canonical correlation analysis (CCA) suggests that arboreal and non-arboreal pollen can be reliably separated into different environmental gradients, implying arboreal and non-arboreal pollen could not coexist in large numbers on the eastern Tibetan Plateau. Logistic regression identifies the thresholds for selected arboreal and non-arboreal pollen that indicate tree presence (p-values <0.05). For instance, a pollen threshold of 4.1% for Abies, 3.7% for Alnus (3.7%), 11% for Picea, 14.3% for Quercus (including evergreen and deciduous type), 33.7% for Betula, and 22.1% for Pinus indicate tree presence at >0.5 probability, while a threshold of 0.2% Ephedra, 0.2% Nitraria, and 0.7% Caryophyllaceae could accurately indicate tree absence at <0.15 probability. Test cases, based on fossil pollen records going back to 18 cal ka BP from Ximen Co and Lake Naleng on the eastern Tibetan Plateau, confirmed that our estimated thresholds can be used to infer the presence of Abies around Lake Naleng during the Holocene thermal maximum (10–5 cal ka BP). This research will contribute to a better understanding of the relationships between treeline and pollen on the eastern Tibetan Plateau.