Close agreement between pollen-based and forest inventory-based models of vegetation turnover

摘要: Aim Fossil records are being increasingly used to help understand the consequences of climate change for biodiversity. Pollen from late Quaternary among most commonly fossil data, but pollen-based inferences biodiversity can potentially be confounded by spatial and taxonomic uncertainties influence non-climatic abiotic factors such as soils on vegetation–climate relationships. Using paired pollen vegetation inventories, we assess fidelity estimates compositional turnover along environmental gradients given various sources uncertainty. Location Eastern United States. Methods We modern forest composition data Forest Inventory Analysis (FIA) plots fit generalized dissimilarity models. To address how in affect turnover, coarsened spatially individual 10- 30-arcmin resolution taxonomically species genus. determine whether soil properties influenced deviance partitioning between models including or variables versus with a combination both. Results Pollen-based were highly correlated those based FIA tended lower, mainly due differences secondarily resolution. Neither nor uncertainty substantially reduced correlation pollen- FIA-based turnover. best matched when they aggregated genus Vegetation–climate relationships similar across datasets, although sometimes differed. The was negligible compared did not improve model fit. thresholds greatly form strength pollen–vegetation relationships. Main conclusions Pollen act robust proxy thereby supporting use predict temporal changes vegetation.