Resolving ‘ghost’ diversity in pollen records
Increasing our ability to resolve under-represented plant taxa in pollen records
Pollen analysis is a key technique for examining past changes in vegetation communities. However, the pollen grains of different plant species within some genera or families are largely indistinguishable from each other (e.g. Poaceae, Fuscospora, Coprosma) leading to a large amount of dark diversity that cannot be resolved using traditional microscopic techniques. In New Zealand, for example, pollen is only able to detect at most c. 12.4% of the seed-plant diversity (236 pollen types can be identified compared with 1896 native seed-plant species).
We are currently developing techniques for extracting and sequencing ancient DNA from pollen grains. By using primers developed for each broad pollen type we are able to identify pollen grains with a higher degree of taxonomic resolution, allowing deeper insights into past changes in plant community composition. This approach also allows ancient DNA from plants to be assessed in a secure stratigraphic context, overcoming potential issues around DNA leaching through sediment profiles which can afflict bulk-sediment analyses in temperate regions.
In addition to the indistinguishable pollen types found in some plant Families, another common problem with pollen analysis is the under-representation of plant taxa in the pollen record -such as Lauraceae - these are species that rarely get recorded. We are trying to reconstruct the Holocene history of some key "ghost taxa' using modelling approaches.
Janet Wilmshurst