Aim: Support for different underlying mechanisms of species occupancy is inconsistent, yet this could be related to spatial scale. Since abiotic filtering typically acts at broader scales than biotic interactions, we hypothesise that occupancy could be more driven by species’ abiotic niche (i.e., tolerance and preference of abiotic conditions) at broad scales, whereas species’ traits affecting competitive ability could be more important at fine scales. Here we test these hypotheses by assessing relationships of occupancy to niche and trait metrics across spatial scales.

Location: Four study areas located north of Arctic Circle.

Taxon: Vascular plants.

Methods: We derived occupancy for 106 species at four spatial scales (Micro-scale with plot size of 0.04 m ² and extent of 2 km, Local-scale with plot size of 4 m ² and extent of 40 km, Regional-scale with plot size of 4 ha and extent of 800 km, and Polar-scale with plot size of 4 km ² and extent of 5200 km). We then assessed using generalized additive models whether the relationships between occupancy and species’ niche breadth, niche marginality, intraspecific trait variability (ITV) and trait distinctiveness vary across the scales.

Results: At the finer scales, ITV (especially of specific leaf area) had the highest contribution with positive relationship in explaining occupancy. At the broader scales, occupancy was better explained by niche metrics. Especially at the broadest scale, the occupancy had a positive relationship with species’ climatic tolerance.

Main conclusions: Abiotic filtering, especially related to macro-climate, drives species occupancy at broader spatial scales while biotic interactions are relatively more important at local scales. This scale-dependency of factors behind species occupancy should be accounted for when, for example, planning conservation of rare species, forecasting invasions, or anticipating the effects of changing climate on biota at local versus global scales.

spOccu_atMicroScale consists of 1920 plots of size 20 x 20 cm where all vascular plants were identified. Plots are organised in 20 x 8 m grids (12 grids) and the plots of 20 x 20 cm were randomly placed inside the 1 m2 plots of the grids (one 20 x 20 cm plot per one 1m2 plot). Plots are situated in Northern Finland on the slopes of mount Saana.

spOccu_atLocalScale  consists of  573 sites were species occurrences from four 1 m2 plots are pooled to a site. Site consists of four plots five meters from the center of the site to each direction (N, E, S, W). From each plot all vascular plants were identified. Data was collected from Northern Finland and Norway.

data_Regio_dryad consists of 296223 occurrences (coordinates) of 211 species from mainland areas of Finland, Sweden, and Norway north of Arctic Circle. Occurrences were derived from GBIF database (GBIF.org (29 March 2020) GBIF Occurrence Download https://doi.org/10.15468/dl.nolkaq), complemented with entries from national databases of Finland (https://laji.fi/en), Sweden (https://www.artportalen.se/) and Norway (https://www.artsdatabanken.no/). Only occurrences surveyed 1990–2019 and location accuracy of at least 100 m are included. Further, occurrences with duplicate coordinates per species and the occurrences falling to sea or on a glacier were removed. In the files uploaded to Dryad, the coordinates of vulnerable species are rounded to two decimals. 

data_Polar_dryad consists of 330659 occurrences of 211 species from non-glaciated terrestrial area north of Arctic Circle. Occurrences were derived from GBIF database (GBIF.org (14 April 2021) GBIF Occurrence Download https://doi.org/10.15468/dl.hvb7xm), including only occurrences surveyed 1990–2020 and with location accuracy of at least 1 km. Further, occurrences with duplicate coordinates per species and the occurrences falling to sea or on a glacier are removed. In the files uploaded to Dryad, the coordinates of vulnerable species are rounded to two decimals.