Spatially extensive and temporally continuous three REVEALS pollen-based vegetation reconstructions in Europe over the Holocene.

Fossil pollen records are the best empirical data to provide natural (climate-induced) and anthropogenic (human-induced) past vegetation composition (H. H. Birks, 2008; H. J. B. Birks & Willis, 2008; Gaillard et al., 2010, 2018; Willis & Birks, 2006). Nevertheless, pollen percentages have a no-linear relationship with vegetation abundances due to plant species-specific differences in pollen production, dispersal, and deposition (Prentice, 1986; Sugita, 2007). These intertaxonomic differences are corrected by ‘Regional Estimates of VEgetation Abundance from Large Sites’ (REVEALS) model, taking into account two important parameters: relative pollen productivity and fall speed of pollen to infer plant cover. Currently REVEALS model is the most appropriate method to reconstruct plant cover at a regional spatial scale of ca. 100 x 100 km and it allows to transform pollen data from large lakes, but can also produce regional vegetation cover estimates from multiple small-sized sites (Fyfe et al., 2013; Mazier et al., 2012; Sugita, 2007; Trondman et al., 2015a). REVEALS estimates of the abundance of 31 and 46 taxa, 10 plant functional types (PFTs), and 3 land-cover types (LCTs) were produced for Europe at a 1˚x1˚ spatial scale for 25 consecutive time windows from 11.7 ka calibrated years BP (i.e. before present (1950)) to present as part of the TERRANOVA project (H2020-MSCA-ITN-2018) grant agreement No 813904. This is the third generation (version march 2022) of REVEALS reconstructions in Europe, it was used a total number of 1607 pollen records from pollen databases/archives and individual authors (see TERRANOVA_metadata.xlsx and TERRANOVA_contributors.xlsx), and for the references related to the pollen records (see TERRANOVA_metadata_list_references.pdf). The first generation was published in Trondman et al., 2015 and Gaillard, 2019, and the second generation (Fyfe, Ralph M; Githumbi, Esther; Trondmann, Anna-Kari; Mazier, Florence; Nielsen, Anne Birgitte; Poska, Anneli; Sugita, Shinya; Woodbridge, Jessie; LandClimII contributors; Gaillard, Marie-JoséFyfe, Ralph M; Githumbi, Esther; Trondmann, Anna-Kari; Mazier, 2022). The REVEALS estimates are based on three RPP-means datasets: i) RPP-means dataset from (Githumbi et al., 2022), RPPs.st1 (31 taxa); ii) a new synthesis proposed in this study inclusive of strictly entomophilous taxa RPPs.st2 (46 taxa); iii) a composed synthesis, RPPs.st3 (31 taxa) derived (Mazier et al., 2012) RPP-means dataset combined with southern RPP studies (Githumbi et al., 2022; Grindean et al., 2019) (see TERRANOVA_RPP.sts_values _FSP_taxa_PFT_LCT.xlsx). All the details information can be found in Serge et al. in prep. 25-grid based estimates at 1°x1° (ca. 100 km x 100 km) across 30°-71° N, 20°W-47°E (north-western, central Europe, Mediterranean area and part of the East until 47°E) were produced. REVEALS results are extracted by time window, producing 25 matrices of mean REVEALS estimates and 25 matrices of corresponding mean SEs for each of RPP taxa and each grid cell. As three RPP-means datasets are tested, 3 REVEALS results are produced per time window (see TERRANOVA_RVresults_march2022.zip). It was used 25 time windows i.e. (-65) to 100 BP, 100 – 350 BP, 350 – 700, 700 – 1200, 1200 -1700 BP and in 500 consecutive time windows from here to 11,700 BP (BP – before present (1950)). The consecutive of 500-year-long time windows ensures that a maximum of the REVEALS reconstructions have low SEs and it is still meaningful for the study of past land-cover changes over several millennia (Githumbi et al., 2022; Li et al., 2023). All REVEALS estimates and their SEs are given in proportions of the grid cell (the total of all REVEALS estimates sum up to 1). There are 25 files labelled TW.(number of time window).RVestimates.csv and 25 files labelled TW.(number of time window).standarderrors.csv. The RVestimates.csv files contain the REVEALS estimates for each land-cover type (LCT), plant function type (PFT) and taxa which has a unique grid ID and corresponding longitude and latitude. The standard error .csv files contain the standard errors for each land-cover type (LCT), plant function type (PFT) and taxa. The information about the quality of the REVEALS estimates in each grid cell for each time windows is provided in the TERRANOVA_GC_quality_by_TW.xlsx file. There are 3 categories i.e. 1 is high quality (1 or more large lakes, OR 2 or more sites of any site type), 2 is low quality (only 1 small lake, OR small bog OR large bog) and no data. Extra information about each site is provided in the metadata file. The third generation of pollen-based REVEALS land cover in Europe over the Holocene provides unique possibilities to explore spatial-temporal changes in past land cover and biodiversity over long time periods, to analyse the dynamic of the vegetation, for conservation ecology in the face of environmental changes. (2023-03-24)