The role of microhabitat for bryophyte establishment in reclamation of boreal wetlands

Minerogenous peatlands that accumulate deep deposits of organic matter (fens) were an important part of the pre-disturbance landscape across Alberta’s oil sands mining area. Bryophytes occupy 80–100% of the ground layer in these site types and form an important part of the accumulated organic matter. Bryophytes are small plants that occur in microhabitats that are an integral part of the surrounding plant community. The objective of this paper is to evaluate the establishment of bryophytes at an experimental wetland (Sandhill Wetland-SHW) by assessing the changing vegetation and substrate regimes over eight years after wet-up. Included in this analysis is an account of the ground layer flora between 2014 and 2021 along with evaluation of associated environmental factors. At SHW, a number of peatland bryophytes initially established in numerous areas across the wetland. However, in subsequent years increasing water tables with expanding Typha-dominating areas, high sedge productivity along with increasing vascular plant cover, and increasing sodicity have reduced the microhabitats available for bryophytes, with the result being a much-diminished set of peatland bryophyte species. We suggest that future reclamation designs for fens should include microhabitats for bryophytes, as bryophytes are foundational species and integral to fen development. In comparison, reclamation designs for non-peat forming ecosystems that do not have abundant bryophytes, such as marshes and saline wetlands, do not require these microhabitat conditions. Installing a recently harvested peat substrate appears to provide suitable conditions for early establishment of bryophytes.


Introduction
Surface mining activities associated with oil sands mining in Alberta, Canada, permanently changes the hydrology, substrate, and vegetation of the local landscape. Prior to mining disturbances, wetlands, especially peatlands that have a well-developed ground layer of bryophytes, were an important component of the regional landscape. To mitigate these disturbances that are projected to attain 4,800 km 2 , oil companies are required to reclaim disturbed landscapes to functioning ecosystems, including forests, wetlands, and lakes (Government of Alberta 2020). Wetlands are expected to be an important component of the reclaimed oil sands landscape as they comprised approximately half of the pre-disturbance boreal landscape of western Canada, with > 90% classified as peatlands (Rooney et al. 2012;Vitt et al. 1996). Peatlands develop a deep organic layer originating from plant material that remains in the catotelm after aerobic decomposition, a large portion of which is bryophytic in nature , and has an important role in carbon sequestration (Nwaishi et al. 2015a,b;Rooney et al. 2012) and water regulation (Waddington et al. 2015).
Within the oil Sands Administrative Area of 140,329 km 2 , bogs are estimated to cover 8,962 km 2 and fens 29,083 km 2 (Wieder et al. 2021), about 29% of the land surface (Lee and Cheng 2009). Fens, minerotrophic wetlands with peat-forming plants, were the most common wetland type on the pre-disturbed landscape (Rooney et al. 2012). Amongst the challenges for reclamation of wetlands, especially fens, are hydrological functioning, organic soil reconstruction, and the return of peatland plant communities that include a productive, species-rich ground layer of bryophytes that is similar to natural peat-forming communities (Vitt and Chee 1990).
The importance of bryophytes for future organic soil accumulation is fundamental to maintaining site water balance and development of the peat-column. Bryophyte stems and leaves (especially those of mosses) decompose more slowly than vascular plant litter across a variety of ecosystems (see references in Turetsky et al. 2008) and in most peatlands compose a majority of the remaining undecomposed plant material of the peat column (Vitt and Wieder 2009;Vitt and House 2021). Natural boreal fens, whether sedge-dominated, shrub-dominated, or wooded with an open tree canopy, all have ground layers composed of 80-100% bryophyte cover (Hartsock et al. 2021). The bryophyte-dominated ground layer is one of the fundamental characteristics of fens (and all boreal peatlands) that differentiates them from non-peatforming wetlands such as marshes (no trees) and swamps (with shrubs or trees). The development of such a bryophyte-dominated ground layer is a necessary attribute for reclamation of boreal fens.
Bryophytes are small plants that occupy habitats in close proximity to their substrate. As a result, substrate and its surrounding microenvironment are critical for individual species to survive and expand. Microhabitat characteristics include (1) substrate characteristics, (2) associated vascular plants and their abundances, and (3) local water balance.
The availability of a variety of substrate types (e.g., soil, rocks, logs) is a critical aspect of bryophyte species richness (Slack 1977;Söderström 1988;Vitt and Belland 1997), and rare species are rare largely because their microhabitats are rare (Heinlen and Vitt 2003). On wetland reclamation sites, the presence of organic soil may provide an important source for diaspores and enhance the chances of bryophyte survival (Lemmer et al. 2022). Although microhabitats are usually thought of as small local areas (such as crevices in cliff faces), they also may be quite expansive, for example-conifer forest floors or peatland lawns and carpets. Each of these 'microhabitats' no matter their extent, has similar environmental characteristics allowing specific bryophyte species to successfully inhabit the habitat.
Associated vascular plants provide the structural component of the microhabitat, influencing the amount of sunlight, intercepting precipitation, and in wetlands and forests, producing a litter layer that competes with bryophytes for substrate space (Vitt 2019). Structural heterogeneity is one of the most important influences on the diversity of bryophytes (Vitt et al. 1995;Newmaster et al. 2003;Cole et al. 2008).
Local water balance is critical for bryophytes as they have no roots, and most species are poikilohydric. As a result, bryophytes are unable to access water from beneath the soil surface and rely on surface water, precipitation, and favorable evaporative environments. For example, one factor influencing high species richness in old growth forests (compared to young forests) is moist local microclimate conditions (Rambo and Muir 1998;Newmaster et al. 2003). Favorable local water balance may be achieved by high humidity in rock crevices, streamsides, hummock-hollow development in ombrotrophic peatlands, or water tables near or just below the soil surface.
Suitable microhabitats for bryophytes in wetlands are a balance between (1) vascular plant cover providing some shade and appropriate levels of humidity as opposed to too much cover producing copious litter and dense shade, (2) relatively stable water tables just below the surface of the substate as opposed to flooded surfaces or periodically dry substrates, and (3) water and soils with suitable base cation concentrations similar to those of natural peat-forming wetlands of the region as opposed to saline or brackish wetlands.
Sandhill Wetland (SHW) is one of two landscapescale projects constructed to test methods for creating peat-forming wetlands similar in biotic composition and function to regional fens within the constraints of a post-mining landscape with potentially high concentrations of sodium (Ketcheson et al. 2016). The objective of this paper is to evaluate the developing bryophyte flora as a measure of initial success of Sandhill Wetland by assessing how changing vegetation and substrate regimes have affected the establishment of peatland bryophytes. Specifically, seven years of vegetation surveys over eight years after wet-up that included detailed accounts of the ground layer flora, along with assessments of associated environmental factors are evaluated. This work advances upon that presented in Vitt and House (2015) that assessed the bryophyte flora two years after the beginning of the reclamation process, while this study evaluates how bryophytes have responded to longer-term environmental change over the seven subsequent years. In this paper we address the following questions: (1) where on the wetland did bryophytes establish, (2) has the bryophyte flora changed over the eight years after wet-up of Sandhill Wetland and, (3) how did changing vascular plant communities and associated chemical (especially sodium) concentrations and water table fluctuations affect bryophyte establishment.

Climate
The climate of the study site is characterized as subhumid with multi-year water deficits (Devito et al. 2012Smerdon et al. 2005Smerdon et al. , 2007. Mean June 1-Oct 1 air temperatures varied from 12.1 °C (2016) to 16.2 °C (2021). Average annual precipitation (Oct. 1-Sep. 30 -2013 was 376 mm and for the growing season (May 1-Sep 30) 326 mm. Growing season precipitation varied from 251 mm in 2021 to 436 mm in 2020; 4 years were above average and five below average. Especially noteworthy for the changing water tables on the wetland were three significant rain events-a 92 mm event on Jul 21, 2018, two events of 54 mm (on Jul 9) and 58 mm (on Sep 3) in 2016. Data from 2013 to 2018 from meteorological station on site-from Biagi et al. (2019 and 2019-2021 data from McMurray Airport-https:// acis. alber ta. ca/ weath er-data-viewer. jsp accessed on November 22, 2021.

Site development and characteristics
Sandhill Watershed is located approximately 40 km north of Fort McMurray, Alberta on Syncrude Canada's oil sands lease at 57°02′16.10″N,111°36′06.09″W and 57°02′29.03″N, 111°35′06.37″W. A 52-ha study site that contains 35 ha of uplands and a 17 ha central wetland (SHW) was constructed in the northwestern corner of a previously mined area  referred to as East-in-pit. The watershed was built over four years (2009)(2010)(2011)(2012), when the entire East-inpit was first filled with 35 m of inter-bedded composite tailings and tailings sand layers followed by a 10 m of a tailings sand structural cap. On the wetland, the 10 m tailings sand cap was overlain by 0.5 m of clay till and 0.5 m of fresh peat harvested from a nearby peatland slated for mining with mixed characteristics, including microsites having bog and rich fen vegetation. Peat placement was completed in January 2011 (Wytrykush et al. 2012). In November of 2011, a native seed mix was spread across the wetland. About 80% of this mixture was composed of seeds of Carex aquatilis (Vitt et al. 2016). In spring of 2012, plant introduction plots (16) were established and a variety of vascular plants introduced to these plots. Additional species were introduced in spring of 2013. The wetland site remained dry until August 2012 when water was introduced from a natural lake through a series of pipes and a water storage pond (Wytrykush et al. 2012), with continued additional water inputs in 2013 (Wytrykush et al. 2012;). Thus, 2013 was the first full year of adequate water available for plant growth. In 2013, vascular plant establishment was limited, with only scattered individuals of graminoids, annual upland weeds, and cattails (Typha latifolia) present on the site. By 2014, plant abundances increased, with mean vascular plant cover reaching 65% (Vitt and House 2015); however, much of this cover was provided by weedy annuals and a variety of forbs. Vascular plant cover increased in 2015 to between 90 and 100% cover, with three distinct vegetation types zonally distributed on the wetland (Vitt et al. 2016). The main three zones were each characterized by different dominant plant species and different water tables. The wettest zone, with persistent standing water was dominated by Typha latifolia; the second zone with water tables close to or at the peat surface for most of the growing season was dominated by Carex aquatilis, and the third zone with water tables below the peat surface was dominated by Calamagrostis canadensis (Vitt et al. 2016;House et al. 2022). Additionally, some areas continued to have a variety of abundant forbs and graminoids (i.e., Equisetum arvense, Lotus corniculatus, Melilotus alba, Carex atherodes, Carex utriculata, and Rubus idaeus).

Methods
We used a grid of 91 permanent plots (78-87 surveyed annually depending on access) centered at 40 m intervals across SHW. We conducted annual surveys in early August from 2014 to 2019 and 2021, representing years 2 through 7 and 9 following the introduction of water to the wetland. At each grid point we identified all bryophytes that occurred within a circular 8 m 2 plot. We used vascular plant abundance data, water table data, and water chemistry values as reported in House et al. (2022). Georeferenced points were created from coordinates recorded at the surveyed locations using a handheld Garmin Montana 680 GPS unit. Maps were generated in Arc-Map version 10.6 with results presented for 2014, and for alternate years (2015,2017,2019,2021) over the eight-year time span. We divided the 33 bryophyte species into those occurring in peatlands and those widespread in a number of habitats (non-peatland species-Environment and Parks 2017). We examined the association of peatland bryophyte occurrence in plots located in four vegetation-types as mapped in House et al. 2022): (1) plots with 30% or greater cover of Typha latifolia, (2) plots with 50% or greater cover of Carex aquatilis, (3) plots with 50% or greater cover of Calamagrostis canadensis, and (4) plots with one or more abundant forb and graminoid species and with less than 30% of Typha latifolia, 50% Calamagrostis canadensis, or 50% Carex aquatilis. Vascular plant cover in this group varied from 53 ± 4% [S. E.] in 2014 to between 69 ± 6% in 2017, and 101 ± 8% in 2019.
We used X 2 Goodness of Fit to determine if the frequency distribution of plots with peatland bryophytes and total number of plots differed between the four vegetation groups in 2104, 2015, 2017, 2019, and 2021. We examined whether vascular plant species richness differed between years for plots in the four vegetation types and for plots with peatland bryophytes and without peatland species (alpha diversity, Whittaker 1972) using two-way ANOVA (F-statistic) and Tukey's pairwise comparisons when data passed a Shapiro-Wilks test for normality (p < 0.05 failure). If data were non-normal, a Kruskal-Wallis one-way ANOVA (H-statistic) on ranks was used and Tukey's was used for pairwise comparisons (SigmaPlot v. 11.0, 2008).

Bryophyte species richness
Over the 8 years post wet-up, our seven annual vegetation surveys yielded a total of 33 bryophyte species, 20 of these considered species that inhabit natural peat-forming plant communities of fens and bogs of the region. Thirteen species are either species characteristic of early succession or upland species (non-peatland species) (Table 1) Between 2014 and 2015, non-peatland species occurrences dominated over peatland species, with between 76 and 66% of the occurrences nonpeatland species; however, non-peatland species declined in 2016 and 2017 while peatland species either declined less or remained steady, so that between 2016 and 2019 peatland species occurrences made up between 58 and 51% of the total; in 2021 peatland species made up 68% of the bryophyte occurrences (Fig. 1).
Of the 20 peatland species, Aulacomnium palustre was the most frequent peatland species on the wetland, occurring in 81% of the plots in which peatland species occurred. The high frequency that this easily recognized species occurred in plots with suitable microhabitats makes it a key indicator of conditions appropriate for bryophyte colonization of reclaimed surfaces. In 2014, only eight plots contained peatland species without the occurrence of A. palustre (similarly the same is true for 2015-8 plots, 2017-2 plots, 2019-5 plots, and 2021-3 plots). Campylium stellatum and Tomentypnum nitens also occurred on the wetland in lower frequencies.

Spatial distributions of peatland bryophytes
In 2014, one year after wet-up, 72% of the plots were dominated by forbs and graminoids and 28% of the plots were dominated by the three vascular plant species that would dominate in 2015 and onwards (Fig. 2a). In 2014, of the 27 plots (out of 85 possible plots) in which peatland bryophytes were found, only 11 plots were in those dominated by Typha latifolia, Carex aquatilis, or Calamagrostis canadensis, while 16 plots were in the forb-graminoid vegetation plots. In 2015, the number of plots in the forb-graminoid vegetation type decreased to 37% and to 14% in 2021. In comparison, the number of peatland bryophytes occurring in the forbgraminoid vegetation type increased to 19 in 2015, and then decreased to 5-9 in 2017-2021. Overall, the number of plots with peatland bryophyte species decreased from 51 in 2015 to 29 in 2017 19 in 2019, and 12 in 2021 (Table 2).
Typha latifolia-dominated areas: Plots dominated by Typha increased from 2 in 2014 to 9 in 2015, 11 in 2017, 16 in 2019, and 29 in 2021. No bryophytes were found in any of these plots (Table 2, Fig. 2a-e).
Carex aquatilis-dominated areas: Seven plots were dominated by C. aquatilis in 2014, increasing to 17 in 2015; and varying from 23-24 plots in 2017-2021. In 2014, only 2 of the 7 plots dominated by C. aquatilis had peatland bryophytes. In 2015, peatland bryophyte occurrence increased, occurring in seven (41%) plots, decreasing to between 2-3 plots Table 1 List of species found on Sandhill Wetland from surveys between 2014-2021 (no survey in 2020). Species are listed as 1) Peatland species -those occurring in boreal peatlands or 2) species widespread in a variety of habitats. Habitats summarized from a variety of literature sources including BFNA (2007BFNA ( , 2014, Gignac et al. (1991) Fig. 2a-e). In all years, 63% of the plots in C. canadensis-dominated plots contained peatland bryophytes.
Sodium concentrations in the surface water of SHW have increased over the eight years since wetup (Biagi et al. 2019;House et al. 2022); however, the increases have been most pronounced along the southern boundary of the wetland, with 17 plots having concentration greater than 400 mg L −1 (Fig. 2f).

Vascular plant species richness and cover
Over all of the years, mean vascular plant species richness [alpha] in plots with mosses was 10.0 (± 0.8 S.E.) compared to 7.5 (± 0.9 S.E.) in plots without mosses (Typha-dominated plots excluded). When analyzed by vegetation type, vascular plant richness was greater in plots with mosses compared to those without only in the forb-graminoid vegetation type (f = 29.35, p = 0.001, Online Resource 1), and there differing only in 2014 (p = 0.002) and 2017 (p = 0.001). In all three vegetation types with mosses, vascular plant richness decreased over the eight years (f = 7.97, 12.58, 15.58 all p = 0.001).
Similarly, vascular plant cover (here equated to a surrogate for shade) differed between plots with and without mosses in the graminoid-forb group (p = 0.001) with plots with mosses having 92.8% ± 6.8 (S.E.), and plots without mosses had 74.3% ± 9.2. The other groups (dominated by Typha, Carex, and Calamagrostis) did not differ between plots with and without mosses (all p > 0.07 -106.0% [± 8.7 S.E] in plots with mosses compared to 89.4% [± 5.9 S.E.] in plots without mosses). Vascular plant cover increased over the eight years of sampling in each of the three vegetation types with mosses (f = 5.01 p = 0.001, f = 38.92 p < 0.001, 11.56, p = < 0.001).

Discussion
Suitable microhabitat availability for bryophytes is an important attribute for reclamations designed to recreate peatland landforms in the recovery from oil sands open pit mining. Species of bryophytes that decompose slowly (Turetsky et al. 2008) and make up a large portion of the organic matter present in the peat column are essential plant components for peatland ecosystems . On the other hand, non-peat forming ecosystems, such as marshes, often characterized by water tables above the substrate surface, and dominated by Typha spp. or broad-leaved sedges, have little or no bryophyte cover and a ground layer that does not function in the limited peat accumulation taking place in these site types (Bayley and Mewhort 2004;Trites and Bayley 2009). At SHW, there were no occurrences of bryophytes in areas dominated with high water tables and dominated by Typha latifolia, indicating the lack of suitable microhabitats in these areas for peatland bryophytes. Over the 8 years of surveys, Typha-dominated areas expanded, reducing the available habitat space for bryophytes. Also, beginning in 2016 and evident in 2017, Carex aquatilis-dominated areas increased, expanding toward the margins (House et al. 2022). High productivity of C. aquatilis in 2017 (mean  312 g m 2 - Vitt et al. 2020), compared to natural sites averaging 204 g m 2 (Thormann and Bayley 1977) produced high biomass that increased litter at the ground level, prohibiting suitable microhabitats for bryophytes, and reducing the overall vascular plant species richness (House et al. 2022). This dense litter layer, along with water management of the site, influenced the expansion of C. aquatilis-dominated areas (from 20% in 2015 to 31% in 2017 of the plots on the wetland), replacing areas previously occupied by Calamagrostis canadensis and forb-graminoid vegetation. Plots in Calamagrostis areas decreased from 29 in 2015 to 20 in 2017 and from 32 to 23 for forbgraminoid vegetation. In 2016, a wet year with two growing season rain events over 50 mm and outflow pumps off, water tables increased in the late season. An unusually wet spring in 2017 with several rain events between 20 and 40 mm also increased water tables . Throughout this period, outflow pumps remained off, creating extremely high water tables for a short time throughout the wetland, especially evident in Typha and Carex aquatilis zones in 2017 (House et al. 2022). In contrast, C. canadensis-dominated and forb-graminoid areas, with high overall species richness (House et al. 2022) were little affected by the high water tables in 2016-17. In 2017, suitable microhabitats remained intact in these areas and peatland bryophyte occurrences remained high. For example, in 2015, peatland bryophytes occurred in 41% of the Carex aquatilis-dominated plots, while in 2017 they occurred in only 10%. Comparatively in 2015, peatland bryophytes occurred in 79% of the Calamagrostis-dominated plots and 66% of the forbgraminoid plots, while in 2017, they occurred in 90% of Calamagrostis plots and 39% of the forb-graminoid plots. Bryophytes are especially sensitive to high levels of sodicity and few species occur in areas with high concentrations of sodium (Sabovljevic and Sabovljevic 2007). As a result, microhabitats high in sodium are largely unsuitable for bryophytes. Material utilized to infill oil sands in-pits have high salinity, especially high in sodium and calcium dissolved in the associated water . Although SHW was designed to partially remove these high levels of cations (Wytrykush et al. 2012), upward movement from the deposited tailings has continued to be a problem (Biagi et al. 2019;. Between 2015 and 2021, Na + concentrations increased just under two times (House et al. 2022). This increase was especially strong in 2019 with Na + concentrations in 2019, increasing to over 400 mg L −1 in many areas. The effect of these high levels of Na + on bryophyte distributions and suitable microhabitats is seen in plots in which bryophytes that occurred in plots along the southern boundary of the wetland where in 2017, 73% of the plots had peatland bryophytes (Fig. 2c), compared to 32% in 2019 and 2021 (Fig. 3d-e).
By mid-growing season in 2014 (the second growing season after wet-up in 2012-2013) vegetation was largely dominated by weedy annuals, with only 28% of the plots being dominated by one of the three species that in 2015 would provide the dominant species currently on the wetland. In 2014, water introduced in 2012-13 was not present in all areas of the wetland, leaving some scattered areas dry. Also, 2014 was a relatively dry year, with only 299 mm of precipitation during the growing season . We reported (Vitt and House 2015) 21 bryophyte species on the wetland in 2014; however, these were scattered across the wetland, with most occurring in plots not dominated by Typha, Carex, or Calamagrostis. By 2015, bryophytes occurred in 41% of the C. aquatilis, 79% of the C. canadensis plots and 66% of the forbgraminoid plots, suggesting that suitable microhabitats were available on the wetland and bryophytes had colonized from diaspores present in the peat substrate placed on the site in 2012.
The upper peat column of both bogs and fens has been shown to contain a wide diversity of bryophyte diaspores that germinate frequently under appropriate conditions (Vellak et al. 2021). The early establishment success of a large number of bryophyte species across Sandhill Wetland indicates that the initial design of placing a recently harvested peat substrate on the reclaimed surface created beneficial conditions for bryophyte colonization (Vitt and House 2015). This approach follows on to similar approaches used in restoring cut-over peatlands (Rochefort et al. 2003(Rochefort et al. , 2016 and enhancing post-fire recovery (Sheppard et al. 2023). In both cases, appropriate water table position is a critical factor for successful bryophyte establishment. However, at SHW by 2017, owing to the high water table expanding over the wetland enabling increases in Typha and Carex, suitable microhabitats only remained in areas dominated by C. canadensis and forb-graminoids. The increase in Na + concentrations in 2019 further reduced suitable microhabitats along the southern boundary of the wetland, mostly in areas dominated by C. canadensis and forb-graminoids.
Although vascular plant species richness decreased over the eight years of sampling, no association with plots having mosses vs. those without mosses is evident. Likewise, an increase in vascular plant cover over this time frame is present, but without any association with plots in which mosses occurred. However, the increase in vascular plant cover, mostly from dominant species in each of the vegetation types, over the entire wetland may have a negative influence on moss persistence.

Conclusion
In summary, suitable microhabitat conditions for peatland bryophytes are not present in Typha-dominated areas with high water tables. Likewise, Carex aquatilis-dominated areas with high productivity and a fluctuating water table do not provide suitable microhabitats for peatland bryophytes. High litter production by sedges creates conditions that are unsuitable for peatland bryophytes. Finally, high sodicity of surface water cannot be tolerated by bryophytes and suitable microhabitats are not present under these conditions.
Reclamation designs for fens must include microhabitats for bryophytes, as bryophytes are foundational species for peat accumulation and integral to fen development (Graf and Rochefort 2010). In comparison, reclamation designs for non-peat forming ecosystems that do not have abundant bryophytes, such as marshes and saline wetlands, do not require these microhabitat conditions. Installing a recently harvested peat substrate appears to provide suitable conditions for early establishment of bryophytes. At SHW, a number of peatland bryophytes successfully established in numerous areas across the wetland. However, in subsequent years increasing water tables with expanding Typha-dominating areas, high sedge productivity and fluctuating water tables, and increasing sodicity have reduced the microhabitats suitable for peatland bryophytes, with the result being a much diminished set of bryophyte species. Concomitantly, vascular plant species richness decreased, with an associated increase in cover of the dominant species. As a result, and in agreement with , SHW has shifted from having the potential to develop into a wetland with peatland characteristics to a wetland with marsh or saline wetland features. Unfortunately, these latter site types were not a common feature of the pre-disturbance landscape.
As open pits become available for reclamation, designs should consider the importance of the bryophyte layer in reconstruction of peatlands. Conditions for a successful integration of a rich peatland bryophyte assemblage into future reclamation designs for peatland site types should incorporate (1) recently harvested peat substates as has been tested on mineral oil pads (Gauthier et al. 2018), (2) controls on an unconstrained water table, (3) a more diverse set of wetland vascular plant species, (4) controls on Typha invasions, and (5) control of sodicity.