Assessing the suitability of pollen dispersal and deposition model: A model-empirical data comparison study in the spruce (Picea schrenkiana) forest edge of Tian Shan Mountains, northwestern China

Abstract

Montane ecotones such as the forest-steppe transition are highly sensitive and responsive to climate change. Thus, reconstructing changing ecotone position is a valuable tool for reconstructing past climate change in mountainous regions, although identifying this from the pollen record can be challenging. Mathematical models of the pollen-vegetation relationships can be used to explore the detectability thresholds for ecotones, and the reliability of these reconstructions depends considerably on the suitability of the pollen dispersal and deposition models used. In this study, we compared pollen assemblages collected across the forest-steppe ecotone in the Tian Shan Mountains with predictions of a selection of pollen dispersal and deposition models as a test of whether they are appropriate to be used in mountainous regions. The modern pollen spectra show clear association with the local vegetation communities. Picea pollen dominate in the samples from the spruce forest, remain relatively high within hundreds of meters from the forest edge, and drop to background level (0.8% on average) by 1 km away from the forest edge. Both the log₂(z) model and the z⁻¹ model (z = distance from forest edge) performed well at simulating these changes, suggesting that although they do not incorporate topographic variations, these models can still be used with confidence to explore and to reconstruct past changes in the position of this ecotone at least over the Holocene.