SCALE-DEPENDENT EFFECTS OF DEPOSITIONAL ENVIRONMENTS – LESSONS FROM BRITISH COLUMBIA POLLEN SPECTRA

Pollen spectra obtained from soil samples may differ from lacustrine samples in proportions of key species. Differences in pollen representation due to depositional environments (sample substrates) may be due to differential preservation as well as effective catchment size of the basin. Paleo-reconstructions may be affected by substrate effects if differences involve important pollen types. The British Columbia Modern Pollen Database allows us to examine these differences by measuring the effect of sample substrate on pollen representation.

Differences in pollen representation between substrates (soil or lacustrine) are not consistent with ecozone or spatial scale. The Coastal Western Hemlock (CWH) biogeoclimatic ecosystem (BEC) zone is most strongly affected by the effects of depositional environments, while Engelmann Spruce, Sub-Alpine Fir (ESSF) sites show differences in only a single pollen type.

Increasing the spatial extent of the Coastal ecozones to include both Mountain Hemlock and CWH zones increases the number of pollen types that appear to show differences between sample depositional environments, likely as a result of the bias in distribution of soil samples towards more southern sites. Multivariate differences between soil and lacustrine samples are not significant when tested with the Multi-Response Permutation Procedure. This pattern is seen in the other BEC zones tested; increasing the spatial scale by uniting groups of affiliated BEC zones changes the pollen types affected, while multivariate comparisons between substrates in these same grouped BEC zones are not significant.

Reconstructions of climate and vegetation in regions of high relief with strong climatic gradients are often difficult, requiring large sample sizes for proper interpretation. We find that although differences in pollen representation between substrates exist, multivariate differences between substrate types are not significant if proper pollen-type selection is used. This indicates that in regions with poor sample coverage pooling pollen spectra obtained from differing substrates will provide reliable results for reconstructions.