We describe plant communities along the mesotopographic gradient in the low-elevation subcontinental mountains of NE Finland (Utsjoki region). We sampled vascular plants, bryophytes, and lichens along 18 mesotopographic ridge-snowbed transects comprising a total of 180 plots. We used non-metric multidimensional scaling (NMDS) ordination with envfit to explore the differentiation of plant communities in relation to mesotopography, elevation, rock cover, the cover of bare ground, snow bed size, and snowmelt time. The classification of communities was performed using DIANA clustering. Plant communities were differentiated along the mesotopographic gradient, snowmelt time, elevation, and rock cover. The DIANA analysis distinguished seven clusters corresponding to the following communities: Betula nana−Lichenes heath, Empetrum−Myrtillus−Stereocaulon heath, Empetrum−Pleurozium−Lichenes heath, graminoid-rich snow-protected heath, Oreojuncus trifidus–Avenella flexuosa snow-protected heath, Polytrichastrum sexangulare–liverwort snowbed, and Salix herbacea–Kiaeria starkei snowbed. Because of the strong impact of snowmelt time on plant community structure and distribution of communities, it is likely that climate change-induced changes in snow conditions are affecting tundra vegetation, and especially snow beds are threatened. Snowbed communities in the Utsjoki region roughly align with previously described vegetation associations of mountain areas in NW Europe. The assignment of the graminoid-rich snow-protected health community remains uncertain.

Vegetation data was sampled in the summer of 2019 along 18 mesotopographical transects in NE Finland (Utsjoki). GPS coordinates of the transects were recorded using Garmin 64SX GPS. Transects were designed to cover the snow depth gradient at a 15−30 m distance from the top of the ridge to the depression with snowbed vegetation. Eight 0.8 m x 0.8 m plots were placed at regular intervals along each transect. In each plot, vascular plants, bryophytes, and macro lichens were identified, and their coverage was visually estimated. The coverage of each species was estimated using a modified logarithmic Hult−Sernander−Du Rietz scale consisting of 10 classes. The identifications were made at the species level, when possible, based on collected specimens (i.e. if specimens were large enough and contained all needed distinguishing characteristics). Lophozia spp., including Barbilophozia sudetica, were treated as collectives as were some lichens such as Cladonia arbuscula and C. mitis. The coverage classes were later converted to percentage values of each class center, which correspond to cover percentages of 71.2, 35.6, 18.9, 8.9, 4.4, 2.2, 1.1, 0.5, 0.25, and 0.125. The coverage of bare soil and stones was also estimated using the same scale. Two extra plots were sampled for each snowbed, at a 1.5−5 m distance from the last plot at the snowbed end of the transect. Extra plots are marked in the data with the letter E. To render long-term monitoring of vegetation, both ends of the transects were marked in the field, and coordinates were collected using a GPS receiver.

We estimated the mesotopography for each plot (1−8, which 1 = ridge, 8 = snow bed) along with the elevation, size (m²), and snowmelt time for snow beds. Snowmelt time was evaluated from Landsat and Sentinel 2 images from 3−5 summers, depending on the availability of cloud-free images. Based on snowmelt time, snow beds were divided into three classes (1−3) which of 1 represents early-melting, 2 intermediate-melting, and 3 late-melting snowbeds. Early-melting snow beds were mostly snow-free at the end of June (range 8 June−16 July), intermediate-melting snowbeds at the beginning of July (range 24 June−28 July), and late-melting snow beds at the end of July or beginning of August (range 10 July−17 August). There was considerable year-to-year variation in snowmelt times, while the order of snowmelt appeared to be quite constant. Thus, the estimates describe relative, but not exact, snow melting time.