Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4750534 | Review of Palaeobotany and Palynology | 2011 | 14 Pages |
The aim of this investigation was to delimit pollen percentage threshold values of the major European trees as an indication of regional presence. These values can then be applied to mapping the palaeodistribution of these trees at coarse regional scales. A brief review of the most commonly applied thresholds in the published literature demonstrates the complexity of the problem and the need for this kind of study. Pollen threshold values are proposed for the presence of each of the 20 tree taxa investigated by comparing modern pollen data (2195 samples collected from a network of sites across Europe) with two sources of plant species distribution information: the modified Atlas Florae Europeae (AFE) and the map of potential natural vegetation (PNV). For 17 taxa, the modern samples frequently contain small proportions of its pollen (0.1–0.5%) regardless of the sample location with respect to the species distribution of that taxon. The best agreement between the presence of pollen and the distribution information from both AFE and the PNV map was found for Quercus, Pinus, Picea and Fagus. 12 taxa agree with AFE in as few as 9 or as many as 57% of the samples, depending upon taxon, while 13 taxa agree with the PNV map in as few as 2 or as many as 52% of the samples. Most of the pollen types reflect the AFE distribution of the parent trees better than the PNV distribution. The occurrence of pollen of Fagus, Abies, Picea and Quercus, however, relates to the tree distribution information from the PNV map as well or even better than it does to the AFE information. Based on this dataset we suggest new threshold values that minimise the presence of pollen outside the distribution, without a large reduction of the number of rightly assigned samples. We also emphasise that in heterogeneous data-sets a pollen threshold of 0.5% is a better criterion for regional presence of a tree than the often chosen continuous pollen curve.
► Although problematic pollen threshold values are needed to map past tree distribution. ► Pollen from surface samples are compared to the distributions of parent trees. ► Old thresholds are reviewed and new thresholds suggested. ► A threshold of 0.5% is preferable to presence or continuous curve criterion. ► Tree-lines are particularly difficult to describe using thresholds.