Forest canopy cover composition and landscape influence over bryophytes communities in Nothofagus forests of southern Patagonia

Understanding the influence environmental drivers on understory vegetation is important for conservation efforts under climate change. Bryophytes are one of the most diverse groups in temperate forests but also the least known. In addition, the environmental drivers (e.g., forest structure, microclimate, soil conditions or substrate) influencing over bryophyte community among Nothofagus forest types are poorly known. The aim of this study was to evaluate the influence of forest canopy-layer composition on the structure (cover) and the composition (richness and diversity) of bryophyte communities (mosses and liverworts) in two contrast landscape types (coast and mountain) in southern Patagonia. Three natural Nothofagus forest types (pure deciduous, pure evergreen, and mixed deciduous-evergreen) in two landscapes (coast < 100 m.a.s.l.; mountain > 400 m.a.s.l.) were selected (N = 60 plots). In each forest plot, we established one linear transect (10 m length) to measure bryophyte cover (point-intercept method). The data were evaluated using ANOVAs, Chi-square test and multivariate analyses. The mosses were mostly austral-antarctic origin, and the liverworts were all endemics. The principal substrates for the bryophytes development in the forest floor were litter and decaying woods. Moreover, many bryophytes species act as a substrate for natural tree regeneration. The forest structure was the main driver of bryophytes community in the coast landscape, while the slope was the principal driver of bryophytes in the mountain landscape. These differences were mainly explained for the microclimate into the forests (e.g., soil moisture and air temperature), and for the regional climate in the landscapes (e.g., air temperature and soil conditions). Notably, the mixed forest, mainly in the coast, presented exclusive species that were not present in the deciduous and evergreen pure forests. The conservation efforts should include management considerations both the stand and landscape levels based on the potential climate-change impact over bryophyte communities.

Introduction forests e.g., [20,21]. Among them, Lencinas et al. [20] described high variability in the ground-bryophyte photos of the canopy were taken. We used a 8 mm fish-eye lens (Sigma Ex-AF4, Japan) mounted on a 35 125 mm digital camera (Nikon D50, Japan) with a tripod and level, which was oriented to the North. The camera 126 was set 1 m above the forest floor, which is enough to avoid registering understory or shrub cover in these  Europe; E = endemic; PAN = pantropical-type Podocarpus; A = austral-antarctic. We also analysed the 168 occurrence frequency (OF, %) and the main taxonomic groups (Ms = mosses and Li = liverworts) (S1 Table).

169
Finally, the substrate (understood as microhabitat) where the specimens were growing was registered at each 170 sampling point in the transect: litter (LT), decaying wood (DW), bare soil (BS), stone (St), and epiphytic on 171 branches and/or bark in the forest floor (EP) whose development was prior to falling to forest floor. At each point, the presence of tree seedlings growing on bryophytes was also recorded. We then calculated the frequency of tree regeneration on bryophytes for each forest and landscape types.

208
Forest structure, forest floor cover and microclimate 209 The forest structure showed significant differences among forest types and landscapes types, except for 210 canopy cover (%) ( Table 1). The basal area was higher in evergreen forests (84.5 m 2 ha -1 ) than in mixed forests 211 (77.1 m 2 ha -1 ) and deciduous forests (66.5 m 2 ha -1 ). Likewise, the basal area was higher in the mountain (80.7 m 2 212 ha -1 ) than in the coast (71.4 m 2 ha -1 ). Dominant height and diameter at breast height presented differences along a 213 gradient: deciduous forests > mixed forests > evergreen forests. RLAI was higher and similar in deciduous and 214 mixed forests than in evergreen forests. Likewise, RLAI was significantly higher on the coast (2.5) than in the 215 mountain (2.3). Climate drivers showed significant differences among forest types (e.g., soil temperature and 216 relative air humidity) and landscape types (e.g., soil moisture and relative air humidity). The soil temperature 217 was higher in deciduous (6.2°C) > evergreen (5.7°C) > mixed forests (4.3°C), while relative air humidity was 218 higher in mixed (62%) > evergreen (55%) > deciduous forests (39%). In addition, landscape type showed 219 differences for soil moisture, which were higher in the mountain than in the coast, and relative air humidity was 220 higher in the coast than in the mountain. The soil and forest floor conditions also showed significant differences 221 among forest types (e.g., slope, pH, bare soil, vascular plants cover) and between landscape types (e.g., bare soil, resistance to penetration, bare soil and lichen cover were higher on the coast than in mountain. There were 226 significant interactions for basal area, relative air humidity, lichen cover, pH, and resistance to penetration 227 ( Table 1; Fig 2). On the other hand, basal area, dominant height and diameter at breast height did not differ 228 significantly between N. pumilio and N. betuloides species in mixed forests (S2 Table). Neither of these 229 differences was found between landscapes type, only in diameter at breast height (coast > mountain).

230
Interactions were not significant either (S2 Table). The interactions in basal area occurred for all forest types had 231 similar values, which of deciduous forests had significantly lower basal area than mixed and evergreen forests in 232 the mountain. The basal area in the coast and mountain landscapes differed only in mixed forests (Fig 2). The 233 relative air humidity was similar in the deciduous forests in landscape types. However, the mixed and evergreen 234 forests, presented higher relative air humidity in the coast (Fig 2). Interaction in lichen cover occurred by 235 significantly higher cover in coast than in mountain in mixed and evergreen forests (Fig 2), but not in the 236 deciduous. Statistical differences were not detected among forest types for each landscape type. In pH, a 237 significant interaction was detected due to all forest types had significantly values on the coast, but mixed and 238 evergreen forests had similar pH values in the mountains (Fig 2). The resistance to penetration was similar in all 239 forest types in the mountain, but on the coast it was significantly higher in deciduous than in mixed and 240 evergreen forests. However, the resistance to penetration was significantly higher in the coast than in the Table 1

274
The liverworts were found in all microhabitats, and they not presented significant differences in the 275 association among microhabitats or forest-landscape types (Chi-square = 22.9; df = 20; p = 0.289). In contrast, 276 mosses presented significant differences for its association (Chi-square = 77.08; df = 25; p <0.001). Acrocladium 277 auriculatum was the moss species with the major variety of microhabitats, it was not found only on stones. they were not found only on stones (S3 Table). 280 The most common moss species, Acrocladium auriculatum, was significantly associated (Chi-square =  Fig 2). Interactions in 301 liverwort richness were due to in coast and mountain landscapes all forest types have significant differences, but 302 presenting low values in deciduous forests. At the same time, liverwort richness of the coast and the mountain 303 differed only in mixed forests (Fig 2). The mosses cover was similar in deciduous forests at both landscape 304 types. However, the mosses cover in mixed and evergreen forests was significantly higher in the mountain 305 compared with the coast (Fig 2).

306
Several shared species were observed among the different forest types (33% of richness = 9 species), 307 which were found both in deciduous and evergreen forests (Fig 3). Influence of forest structure, microclimate and floor cover on bryophyte 339 distribution previously selected for its statistically significance according to Pearson correlation coefficient (Table 4; S5   343  Table). Axis 1 was influenced by annual precipitation and RLAI and Axis 2 was influenced by relative air 344 humidity, air temperature and slope. When species were analyzed alone (Fig 4A), both axes showed a close 345 correlation and association between mosses and liverworts according to microclimatic drivers. Moreover, the 346 CCA separated the sampling plots in two main groups, defined by its landscape type (coast and mountain), with 347 few differences between the different forests types (Fig 4B). The environmental drivers were the most related to 348 the coasts forests (e.g., RLAI, annual precipitation, relative air humidity, air temperature), while slope influenced 349 in the mountain forests.  conditions, which may vary according to the forest community in which they are found.

399
Specific environmental drivers could influence the richness, cover and diversity of bryophytes. In our 400 study, air conditions seem to be highly important for bryophyte composition in the two gradients under 401 analysis (canopy composition and landscape), where differences were important for each forest type and 402 landscape type. Precipitation was also crucial to water storage in soils and thus keeping environmental 403 moisture in the forest sites. However, the water storage capacity in these forest soils vary greatly, and 404 generates differences in the soil moisture, and therefore differences in the bryophyte abundance as suggested 405 by Raabe et al. [40]. In this context, the regional differences (altitudinal gradients) for the cover, richness and 406 diversity of bryophyte species are evidently determined by climatic differences, like relative air humidity and 407 soil moisture, but also geological drivers such as landscape texture and soil types, which offer a wide range 408 of microhabitats for bryophytes [38]. Here, the landscape types had a strong effect on the soil physical-409 chemical conditions differing noticeably between the coast and the mountain [14]. It was observed that the slope differentiates the coast of the mountain, mainly due to the altitude in which the forest types develop. In addition, air temperature is considered a key driver structuring the bryophyte community in different forest 412 types [40]. Variations in temperature, together with higher relative air humidity were considered favourable 413 conditions for bryophytes abundance [38]. However, high temperatures were also considered to limit the 414 abundance of bryophytes in the forest [41]. Notably, this can be one reason for the differences between the 415 bryophytes between the coast and mountain because the coast had higher temperatures generating greater 416 drying of the soils in forest in the coast. Contrary, mountain soils had greater moisture (air and soil) 417 maintaining favourable microclimatic conditions for bryophytes development.

418
Previous studies indicate that the availability of atmospheric water is more important for bryophytes 419 than the level of soil moisture since bryophytes absorb water throughout their entire plant structure [42]. 420 Therefore, in forest ecosystems a high level of air humidity surrounding the understorey is important for the 421 bryophytes, as well as shaded conditions [43]. Our study showed that the relative air humidity was higher on 422 the coast (influenced by the proximity to the sea), this driver was the one that probably best influences the 423 composition of bryophytes (mainly in mixed forests and evergreen forests), while the soil moisture explains 424 better the specific composition in the mountains. Raabe et al. [40] considered that soil moisture is the main 425 driver of bryophyte diversity. In our study, although the higher soil moisture in mountain forests presented a 426 greater coverage of bryophytes, the less humid soils on the coast for the three forest types could be important to 427 conserve rare species, even if they are present in low abundance.

428
Other specific drivers can also influence richness, cover and diversity of bryophytes along with the

505
The forest types and landscape types influence differentially the richness, cover and diversity of 506 bryophytes communities. This was conditioned by the microclimate into the forests (i.e., soil moisture, relative 507 air moisture and air temperature), as well as by characteristics of the regional climate influenced by the 508 landscape type. The diversity of microhabitats was essential for mosses and liverworts development, such as 509 those generated by litter and decaying wood. Here, the evergreen forest presented greater cover of most 510 favorable microhabitats along with the mixed forests. However, the mixed forest presenting exclusive species of 511 bryophytes absents in deciduous and evergreen forests (mainly in the coast landscape). Therefore, the mixed and cover of bryophytes in all forest types. According to distribution patterns, bryophytes of Nothofagus forests 515 could be considered endemic and highly specific to these southern areas. This study constitutes an important 516 basis for continuing studies of bryophytes in the southernmost forests, investigating their role in forest dynamics 517 and as a key component of local and regional biodiversity. Conservation efforts should not only consider the 518 forest type, but also the landscape type and therefore the microclimatic conditions. Our study highlights the 519 crucial effect of climatic conditions on the composition of bryophyte community, thus this group could be more 520 sensitive to the expected changes of temperature and precipitations. This knowledge could improve ground-