We studied ptyctimous moss mites, which are characteristic of forest habitats, in Norwegian broadleaf forests considered as biodiversity hotspot areas in Fennoscandia. The study aimed to evaluate the effect of different factors (regional locality, annual precipitation, mean annual temperature, forest type, forest wetness and microhabitat) on the ptyctimous mites and on discovering their richness in broadleaf forests. Samples were collected from nine broadleaf forests in western, southern and eastern Norway, in different climatic conditions, six forest types, three forest wetness states and eight microhabitats. Overall, 3,341 ptyctimous mites were collected and their abundance differed significantly among the regions, forest types and microhabitats. Forest type turned out to be the most important factor, responsible for 24.5% of the total variation in the abundance of the ptyctimous mites. Other important factors were forest wetness and microhabitat. In total, 27 species, i.e., 87% of all ptyctimous mites known from before in Norway were found and the species richness was highest in the east and lowest in the west of the country. Atropacarus striculus was most common and most abundant; it made nearly 30% of all ptyctimous mites collected. On the other hand, a quarter of the species were represented by less than 10 specimens; most of these were new records for Norway. Among ten species discovered as new to Norway, four were also new to Fennoscandia. These findings confirm the unique character and high biological diversity of Norwegian broadleaf forests.

Study sites

Samples were collected in nine broadleaf forest (Fig. 2). Two of them were located in Western Norway (Neshalvøya − referred later as W1 and Kolneset − referred as W2), five in Southern Norway (Leirvik − S1, Søyland − S2, Verpåsen − S3, Solvang − S4, and Stamsøy − S5) and two others, in eastern part of the country (Kjeøya – E1 and Opstad − E2). They were categorized to six different forest types (Fig. 3). One of these types (a rich broadleaf forest) was represented by four forests, each located in a different vegetation zone. All study sites were characterized by an oceanic climate but differed in the temperature and precipitation (Table 1).

The sampling areas in Western Norway have high precipitation, relatively cool summer (10-15°C in July and August) and mild winter (with average temperature in the coldest month, February -2°C, Holtan 2009). The study site W1 was situated close to a stream floating along small vertical rock formations, where the bedrock consisted mainly of schist and some alkaline rocks. The most abundant trees and bushes were hazel (Corylus avellana), ash (Fraxinus excelsior), lime (Tilia cordata), grey alder (Alnus incana), birch (Betula pendula), and rowan (Sorbus aucuparia). The forest floor was overgrown by mosses, ferns and several herb species, with dominant sweetscented bedstraw (Galium odoratum). Fallen dead wood with mosses were frequent at the site. The other site, W2, was a steep and east-exposed slope with small dry ridges. The bedrock consisted mainly of different types of gneiss, with the loose masses of weathering material. The forest belonged to an unusual type of habitat, a low-herb broadleaf forest with little historical disturbance. The forest was dominated by tree species such as oak (mostly Quercus robur), lime, Scots pine (Pinus sylvestris), ash, hazel, aspen (Populus tremula), and guelder rose (Viburnum opulus). The ground vegetation was scattered and dominated by habitat-specific plants, e.g., mountain melick (Melica nutans), sticky catchfly (Lychnis viscaria), angular Solomon's seal (Polygonatum odoratum), and mosses. Moreover, the occurrence of some old trees and dead wood characterized the sampling site. The presence of some creeping soft grass (Holcus mollis) and pollarded lime trees indicated former grazing and pruning activities, respectively.

Two forests in Southern Norway (S1 and S2) had slightly warmer summer (about 15°C in July and August) and winter (around 0.1°C in the coldest month – February), comparing to Western Norway. The area was dominated by poor bedrock and consisted of acidic and hard rocks (such as granite, gneiss, amphibolite). The terrain was heavily hilly with many rivers and small valleys. One of the forests, S1, was a swamp dominated by common alder (Alnus glutinosa), with some old oaks in inner parts, and a sandy beach on the lake. The other sampling site, S2, was a rich and dry broadleaf forest. It occupied south and west-south slope overgrown by oak, hazel, ash, birch, rowan and sycamore maple (Acer pseudoplatanus).

Three other sampling sites (S3-S5)  had relatively warm summer (around 19°C in July and August) and winter (around -1.0°C in February). The sampling site S3 belonged to the southeastern Norwegian bedrock area, which consisted mainly of gneiss with a district direction southwest-northeast, parallel to the coast, but also some granite. It was situated on an amphibolite ridge with heterogeneous terrain (hilly, rock walls, stone blocks, lime rich patches) and a small valley along the stream. The geology in sampling sites S4 and S5 was very variable, spanning both nutrient-rich (easy-to-soluble) shale rocks and hard and acidic rocks such as granite and gneiss. Forest S3 was dominated by spruce (Picea abies) with enclosure oak trees and a strong mosaic patterns related to nutritional and lime-richness. Rich parts were otherwise characterized by large hazels. The herb layer was species-rich with several red-list species in high categories. Forest S4 was a large and relatively intact marsh forest with both, older trees and rich lots. The site was characterized by a mosaic of internal topography and many large and small swamps between small shrubs and rocks. Dominant tree species were common alder and spruce, but also ash, birch and pine occurred. Oak, asp, hazelnut, rowan and juniper grew on the edge. The herb layer was dominated by lady fern (Athyrium filix-femina), sedges, wood millet (Milium effusum) and several plant species that are indicators of wet habitats. Forest S5 was species-rich, intact and heterogeneous, dominated by partly very old oaks, lime, ash, maple, aspen, field elm (Ulmus minor), hazel, wild cherry (Prunus avium), willows (Salix spp.) and pine. The site consisted of east- or south-facing slopes and hilly areas cut by cracks. The bedrock was mainly gneiss bands with amphibolite passages. The herb and shrub vegetation were species-rich, both in drier and wetter parts, and included European blueberry (Vaccinium myrtillus), liverwort (Anemone hepatica), sweet-scented bedstraw (Galium odoratum), lily of the valley (Convallaria majalis), wood club-rush (Scirpus sylvaticus), meadowsweet (Filipendula ulmaria), and broad buckler-fern (Dryopteris dilatata).

Two sampling sites in Eastern Norway (E1 and E2) had relatively mild summer, with average temperatures between 16.0-16.7°C in July and August. Winter was slightly colder than in the other study sites; in the coldest months (January and February) the average temperatures were around -2.9 and -3.8°C. The sampling sites were located in the south-eastern Norwegian bedrock area, which consisted mainly of gneiss and granite rocks of different composition. Forest E1 was characterized by old and large oak and lime trees. Other tree species were hazel, ash, Norway maple (Acer platanoides), and with some spruce and pine addition. The herb vegetation was partly sparse. The site E2 was a mixture of spring swamps and swamp forests. The natural basis were marine deposits, silt and clay with source springs that keep the forest floor moist. The dominant tree species was common alder, the shrub and herb layer had barberry (Berberis vulgaris), common cottongrass (Eriophorum angustifolium), rough horsetail (Equisetum hyemale) and different sedge species.

Sampling and identification

A total of 147 samples, each of a volume of 500 cm³, were collected by hand between 8^th-12^th June 2017 from the nine broadleaf forests and several microhabitats: 1) Sphagnum mosses on ground (18 samples), 2) other mosses on ground (49 samples), 3) lichens on tree twigs lying on ground (3 samples), 4) mosses on tree trunks on ground level (18 samples), 4) mosses on tree trunks 1.5 m above ground (14 samples), 6) mosses on stumps (13 samples), 7) mosses on dead wood (23 samples), and 8) dead wood (9 samples). Different number of samples from microhabitats was caused by the fact, that not in all forests all microhabitats could be found.

Adult mites were extracted using Tullgren funnels for 14 days and preserved in 90% ethanol. Ptyctimous mites were sorted out from the samples under stereomicroscope, mounted on slides in lactic acid and identified mainly by W. Niedbała according to the keys (Niedbała 2008, 2011). All species are deposited at the University Museum of Bergen, Norway, while some duplicates are donated to W. Niedbała (Adam Mickiewicz University, Poznań, Poland). Data on other oribatid mites present in the samples will be published later.

The systematics, nomenclature and global distribution of mites follow Niedbała and Liu (2018). Habitat preferences of species (Table 2) are based on Beck et al. (2014), Schatz (2015) and Weigmann et al. (2015). They include the following types: alpine, subalpine, hygrophilous (living in wet places), mesohygrophilous (species which prefer high moisture but not wet places), xerophilous (living in dry places), arboricolous (living on trees), epilithic (living on rocks, stones, walls), geophilous (living in soil), lichenicolous (living on lichens), muscicolous (living in mosses), praticolous (meadow species), silvicolous (forest species), tyrphophilous (bog species), xylophilous (living in wood) and eurytopic (occurring in more than three habitat types). Full names of species are given in Table 2, while in other tables and figures abbreviations are used. The new records of Oribatida for Fennoscandia are based on Niedbała and Liu (2018) and those for Norway are based on Niedbała and Liu (2018), and Seniczak et al. (2019a, c).

The classes of dominance (D), i.e., percentage of specimens of a particular species among ptyctimous mites in particular forest follow Seniczak (1978), including superdominants which are characteristic of extreme microhabitats (Table 3), and classes of frequency (C), i.e., percentage of samples in particular forest where the species was found (Table 4) follow Górny and Grüm (1981). Categories summarizing the status of occurrence of ptyctimous species (Table 6) follow Niedbała et al. (2020).

Statistical analyses

The ptyctimous communities were characterized by the abundance index (ind.·500 cm³) and species richness (number of species per sample). Normality of the distribution was tested with Kolmogorov–Smirnov test, while equality of variance in different samples with Levene test. The assumption of normality or equality of variance was not met, so the non-parametric Kruskal-Wallis test by ranks was used, and in case of significant differences between medians, a multiple comparison test between mean ranks was applied. These calculations were carried out with STATISTICA12.5 software.

All multivariate analyses were performed using CANOCO software (Microcomputer Power, Ithaca, NY, USA; Ter Braak 1988; Jongman et al. 1995). Response data or dependent variables (abundance of ptyctimous species) were log-transformed, log (x+1) (Łomnicki 2010), considering the down-weighting of rare species. Independent or explanatory variables were: region (Western, Southern, Eastern), annual precipitation (820-2300 mm), mean annual temperature (5.8-7.4°C), forest (W1, W2, S1, S2, S3, S4, S5, E1, E2), forest wetness (dry, medium, wet), microhabitat (Sphagnum mosses on ground, other mosses on ground, lichens on tree twigs lying on ground, mosses on trees at ground level, mosses on trees 1.5m above ground, mosses on stumps, mosses on dead wood, and dead wood), all taken as factors or dummies.

First, we checked using canonical correspondence analysis (CCA) if any of the independent variables, when treated separately, explained the variation of ptyctimous communities (simple effects). Next, they were compared to conditional effects of the same independent variables to detect any correlation or collinearity among independent variables. Further analyses were focused only on variables that significantly explained variation in the ptyctimous communities in the conditional effects space.