Aim The summits of mountain ranges at mid-latitude in the Northern Hemisphere and the Arctic share many ecological properties including comparable climate and similar floras. We hypothesise that the orogeny during the Oligocene-Miocene combined with global cooling allowed the origin and early diversification of cold-adapted plant lineages in these regions. Before establishment of the Arctic cryosphere, adaptation and speciation in high elevation areas of these mountains ranges may have led to higher species richness when compared to the Arctic. Subsequent colonisation from mid-latitude mountain ranges to the Arctic may explain similar but poorer flora. Location Arctic-Alpine regions of the Northern Hemisphere. Methods We mapped the cold climate in the Northern Hemisphere for most of the Cenozoic (60 Ma until present) based on paleoclimate proxies coupled with paleoelevations. We generated species distribution maps from occurrences and regional atlases for 5464 cold-adapted plant species from 756 genera occupying cold climates. We fitted a generalised linear model to evaluate the association between cold-adapted plant species richness and environmental as well as geographic variables. We performed a meta-analysis of studies, which inferred and dated the ancestral geographic origin of cold-adapted lineages using phylogenies. Results We found that the subalpine-alpine areas of the mid-latitude mountain ranges comprise higher cold-adapted plant species richness than the Palearctic and Nearctic polar regions. The topo-climatic reconstructions indicated that the cold climatic niche occurred first in mid-latitude mountain ranges (42-38 Ma), specifically in the Himalayan region, and only later in the Arctic (22-18 Ma). The meta-analysis of the dating of the origin of cold-adapted lineages indicated that most clades originated in central Asia between 39 and 7 Ma. Main conclusions Our results support the hypothesis that the orogeny and the progressive cooling in the Oligocene-Miocene generated cold climates in mid-latitude mountain ranges, before the appearance of cold climates in most of the Arctic. Early, cold mountainous regions likely allowed for the evolution and diversification of cold-adapted plant lineages followed by the subsequent colonisation of the Arctic. Our results are in line with Humboldt’s vision of integrating biological and geological context in order to better understand the processes underlying the origin of arctic-alpine plant assemblages.
Paleoclimate Reconstructions
Cenozoic estimated air surface temperature (EAST) in the Northern Hemisphere at one degree resolution with coordinate reference system (CRS) "+proj=longlat +ellps=WGS84 +datum=WGS84 +no_defs +towgs84=0,0,0".
The file ‘EAST_60to0Ma_Hagen_etal_2019.csv’ contains the x and y coordinates followed by the EAST at the given coordinate and time, starting at 60 Ma and ending at 0 Ma (i.e. present) as indicated by the column names. The file ‘read_paleoclimate.R’ contains code on how to load and convert ‘EAST_60to0Ma_Hagen_etal_2019.csv’ into a raster with the correct coordinate reference system and an example on plotting the Northern Hemisphere estimated air surface temperature at 60Ma.
Richness Patterns
Data of species and genus richness for the Northern Hemisphere for the present, masked for cold regions criteria and aggregated at two degree resolution with CRS “+proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0”
The raster ‘cold-adapted_species.tif’ and ‘cold-adapted_genera.tif’ contains the richness of cold-adapted species and genera respectively, and is the data set used to produce Figure 3. The six files ‘cold-adapted_family_[family name].tif’ contain the richness for the Compositae, Poaceae, Leguminosae, Brassicaceae, Cyperaceae, and Rosaceae families. These are the six most common cold-adapted families for the Northern Hemisphere and is the data used to produce Figure 4.
The script ‘read_plor_richness.patterns.R’ loads and plots one simple example of richness patterns in a Artic Polar Stereographic projection (EPSG:3995). The folder ‘ocean_layer’ contains an ocean layer shapefile used to facilitate data visualization.
Species Range Mapping
In a simplified way, the four main process from data to species ranges are:
1. Extract occurrence data from GBIF species that are pre-selected as cold-adapted or that complement Hulten’s ranges maps and the Panarctic Flora checklist. For that, data manipulation and a pre-resolution of species taxonomic names is necessary (see ‘1.extract_pre-select_cold_plants_gbif.R’).
2. After resolving taxonomic names and synonyms within datasets, merge distribution data (see ‘2.merge_dataset.R’).
3. From all merged points, create range maps using ecoregions and climatic layers information (see ‘3.create_range.R’).
4. After having clean datasets, merge all datasets (see ‘4.merge_datasets.R’).
Additional steps as the removal of plants not belonging to the Angiosperms clade, masking of the cold areas and the manual check of species names are not provided here, given their simplicity, specificity to a cluster computer and large quantity of intermediate individual files.
Inside the folder ‘support_functions’, there are four functions called by the scripts mentioned above. All functions have a short description and parameters are shortly described.
The file (i) ‘range_from_points.R’ contains the function used to create range maps from occurrence data. The file (ii) ‘resolve_taxonomic_names.R’ contains the function used to automatically resolve taxonomic names and look for synonyms while deleting clear mistakes. The taxonomic name resolution procedure involved a manual check of all species. The file (iii) ‘merge_occurance_data.R’ merges occurrence data from a same dataset considering duplicates and corrected species names. This function requires a species names resolution matrix and a specific structure for datasets (i.e. occurrences inside a folder named ‘points’ with a txt file for each species name containing x and y coordinates). The final occurrence data is stored in a folder named ‘merged_point’. The file (iv) ‘compress_housekeeping.R’ contains a function conveniently called after generating the merged points. It compresses all occurrence data inside ‘point’ folder and deletes the folder while keeping a compressed zip file for reference.