Towards an improved assessment of global plant diversity in deep time: further insights from the Palaeozoic fossil record

An accurate characterisation of plant diversity patterns is critical to further understanding the development of early terrestrial ecosystems. However, biases inherent to the nature of the fossil record, along with methodological constraints, are known to distort deep-time diversity estimates. To further investigate their influence, we first evaluate whether genus-level diversity serves as an accurate predictor of species diversity, using a new, comprehensive dataset containing the temporal distribution of late Silurian to earliest Carboniferous plant macrofossil species. Results demonstrate that genus-level diversity reliably represents species diversity dynamics and that including taxa in open nomenclature does not affect diversity trajectories. Secondly, a relatively unexplored bias related to time discretisation of data is assessed using a randomised time-continuous approach. Results show that no major discrepancies in diversity estimates are recovered between substage-binned and randomised time-continuous approaches, except during the Late Devonian interval. This difference is interpreted to reflect the longer duration of geological substages at that time, as well as a lower stratigraphic resolution in taxon ranges. These factors lead to increased uncertainty regarding the evolution of plant diversity across this period compared to earlier time intervals, highlighting the necessity to better chronostratigraphically constrain these floras.