Identification of past wetland habitats, their commonness and analogism using macrofossil indicators derived from large vegetation databases

Previous research on ecosystem and climate dynamics using fossil data has mainly focused on tracking shifts at the species or biome levels or relied exclusively on abiotic proxies. We argue there is a critical need to identify individual habitats in the fossil record. First, the palaeoclimatological reconstructions based on geochemical proxies are frequently habitat-dependent. Second, changes at the habitat level integrate not only climate, but also edaphic factors and disturbances, providing a more complex insight into succession and resilience drivers. Here, we present a novel method to identify past habitats by interconnecting plant macrofossil and vegetation databases. We utilised the recently developed FEVER index, aimed for indicating vegetation types or habitats using fossil assemblages, and calibrated it with recent cross-climate European vegetation data. When applied to the extensive supra-regional macrofossil database, the method succeeded in reconstructing changes in habitat proportion in Central European wetlands over the past 15,000 years. The reconstruction is ecologically well interpretable, with aquatic and most fen habitats dominating in the oldest periods, while waterlogged forests and bogs expanded since the Middle Holocene due to macroclimatic reasons. Treeless fens reappeared and wet grasslands initiated in the last two millennia due to anthropogenic disturbances and landscape deforestation. The method further allows for quantifying the analogism between past and recent communities of the same habitat. Most habitats, except for calcareous quaking mires and wet shrub tundra, increased analogism in the latter half of the Holocene. We share an R script and a European-scale calibration dataset for wide use.