Early ecosystem establishment using forest floor and peat cover soils in oil sands reclamation

ABSTRACT Early trends of plant community development provide the basis of ecosystem function and reclamation success of oil sand extraction sites. However, few studies have explicitly investigated species-level interactions with different cover soil types, placement depths, and time since reclamation during early plant community development in boreal forests. We investigated effectiveness of forest floor mineral mix (FMM) and peat mineral mix (PMM) cover soils and placement depths (10 and 20 cm) at four research sites 4 to 13 years after reclamation. Outcomes of this study indicate FMM had a more positive influence on woody plant densities, vegetation cover, and species richness than PMM. Species assemblage, composition, dominance, and types (successional stages, habitat types, competitive-stress tolerant-ruderal strategies) also showed FMM cover soil performed better than PMM. Greater vegetation cover and richness on deeper (20 cm) cover soil placements were evident. However, this effect of cover soil depth would likely decrease with time. Dominant and subdominant species on FMM were native and early to late successional, thus trajectory community development on FMM followed typical early succession of boreal forests (from ruderal and annual to perennial communities), while PMM was dominated by non-native and annual forbs which could slow succession and ecosystem recovery.


Introduction
Understory plant communities are products of plant diversity and drivers of overstory succession, nutrient cycling, stand productivity, and wildlife communities in the boreal forest (Hart and Chen 2006;Dhar et al. 2018Dhar et al. , 2020a. Natural disturbances such as wildfires, wind throws, and insect outbreaks are common ways to dominated by shade-intolerant species (e.g. Epilobium angustifolium L., Rubus idaeus L) (De Grandpré et al. 1993;Bergeron et al. 2002;Hart and Chen 2006;Dhar et al. 2018). Boreal forests generally recover well after natural disturbances and can be resilient to human disturbances such as harvesting. However, anthropogenic disturbance regimes such as oil sands mining have more severe impacts as they alter the ecosystem in different ways than natural disturbances (Hart and Chen 2006;Lecomte et al. 2006;Maynard et al. 2014;Dhar et al. 2018). Thus, recovery of such disturbed sites can be aided by understanding the patterns that regulate ecosystem progression and processes that recreate landforms and plant community development to support a successional trajectory for ecosystem resilience .
Oil sands mining in Alberta creates one of the world's largest and most intense industrial disturbances and has removed 1055.4 km 2 of boreal forests (Alberta Environment and Parks 2023). Oil sands operators are required to reclaim disturbed land to equivalent land capability ('ability of a land to support various land uses after conservation and reclamation is similar to the ability that existed prior to an activity being conducted on the land, but that the individual land uses will not necessarily be identical'), as mandated by the Alberta Environmental Protection and Enhancement Act (EPEA) (Alberta Government 2005). Mining removes organic material, mineral soil, and overburden overlying oil sand deposits (Mackenzie and Naeth 2010;Errington and Pinno 2015). Thus, reclamation challenges include reconstruction of landforms and soil including development of organic matter, microbial and soil fauna activities, and conditions for re-establishment of resilient native vegetation Chen et al. 2017).
Reclamation outcomes depend on multiple factors including biotic and abiotic conditions and time since reclamation Chen et al. 2017;Dhar et al. 2018;Mackenzie et al. 2019). Abiotic site conditions such as cover soil type can play significant roles in plant community development (Chen et al. 2017;Mackenzie et al. 2019). In contrast, biotic factors such as microbial activity help soil formation and nutrient cycling . Two types of cover soils are commonly used in reclamation, forest floor mineral mix (FMM, upper layer of forest floor materials mixed with underlying mineral soil 1:1 to 1:5), and peat mineral mix (PMM, peat to mineral soil mix 3:2 to 3:4 by volume) Errington and Pinno 2015;Dhar et al. 2018). Due to readily available and abundant supplies, PMM is most commonly used as cover soil (Rowland et al. 2009). However, FMM contains a larger native seed and propagule bank, which produces more diverse plant communities similar to those of target upland forest endpoints (Mackenzie and Naeth 2010;Dhar et al. 2018). Undisturbed, fine-textured, upland forest soils in the oil sands region are a rich source of plant propagules with over 9,100 emergents m −2 relative to 3,600 emergents m −2 from peatland soils (Mackenzie and Naeth 2010). In contrast, greater establishment of trembling aspen on PMM due to increased surface roughness and soil water content, and decreased competition, is favourable for dispersed seed catchment, germination, and growth (Pinno and Hawkes 2015). Recently FMM has been replacing PMM as a cover soil due to its greater native species cover, richness, and abundance under laboratory and field conditions (Mackenzie and Naeth 2010;Errington and Pinno 2015;Dhar et al. 2018). However, this has not been fully validated on well-replicated large-scale research sites.
Depth of cover soil application is an important factor influencing plant community development, as depth of buried seed affects germination, emergence, and establishment (Qi and Scarratt 1998;Rokich et al. 2000;. Howell and MacKenzie (2017) suggested placement depths exceeding 10 cm should be maintained to reduce potential contaminants. Dhar et al. (2018) concluded cover soil depth should range between 10 and 20 cm, although depth can be changed based on substrate, cover soil material, site conditions, and desired species.
Considerable reclamation research has been conducted in the oil sands, focusing on overstory and understory development, stockpiling, cover soil types and depth, soils, and microbial communities (Naeth et al. 2013;Pinno and Hawkes 2015;Chen et al. 2017;Mackenzie et al. 2019;Dhar et al. 2019Dhar et al. , 2020a. However, none of this research has explicitly investigated species-level interaction with different cover soil types, placement depths, and time since reclamation during early plant community development in boreal forests. This understanding is critical, as early trends of plant community development can provide the basis of ecosystem function and reclamation success . Our objectives were to determine whether plant community abundance, diversity, composition, and dominance differ among sites reclaimed with different cover soil types and placement depths, and whether these responses vary with time since reclamation.

Study area
Research sites were located north of Fort McMurray, Alberta, Canada, in the central mixedwood natural subregion of the boreal forest ( Figure S1). Mean annual temperature is 0.7°C, with maximum 23.7°C in July and minimum -22.5°C in January (Environment Canada 2022). Mean annual precipitation is 418 mm, with 316.3 mm as rain and 134 cm as snow. Soils are mainly Gray Luvisols with fine textured glaciofluvial or medium to fine textured till parent materials. Eutric and Dystric Brunisols have developed on drier sandy sites, and organic and peaty Gleysols dominate lowlying areas. Mixedwood forests with varying proportions of Populus tremuloides (trembling aspen), Populus balsamifera L. (balsam poplar), and Picea glauca Moench (Voss) (white spruce) are the main upland vegetation types, with some inclusions of Abies balsamea (L.) Mill (balsam fir) and Betula papyrifera (paper birch). Pinus banksiana (jack pine) forests occur in drier areas. Wetland vegetation typically consists of Picea mariana (Mill.) BSP. (black spruce), Larix laricina (Du Roi) K. Koch (tamarack) and Salix spp. L. (willow).

Experimental design
The study was conducted on four oil sands mine reclamation sites with reclamation age of 4-13 years (4, 5, 7, and 13 years for Sites 1, 2, 3, and 4, respectively). All sites were constructed using a variety of substrate materials where forest floor mineral soil mix (hereafter FMM) and peat mineral soil mix (hereafter PMM) were used as cover soils. Site 1 was~2.5 ha and reclaimed using two cover soils FMM and PMM (Brown 2010). FMM was salvaged, stockpiled for 3 months, then applied at 20 cm over 30 cm of B and C horizon mixed subsoil. PMM was applied on clean overburden at 30 cm depth. Site 2 was~3.5 ha and reclaimed with two cover soils (FMM, PMM); FMM had two placement depths (10 and 20 cm), and PMM had 100 cm peat-sand (50% sand, 50% peat from a fen) (Mackenzie 2012). FMM was salvaged and stockpiled for 6 months prior to placement. Site 3 was~5 ha with two cover soils (FMM, PMM) and two placement depths (10 and 20 cm) (Mackenzie and Naeth 2010). Cover soils were applied on 35 cm secondary substrate. Site 4 was~3.5 ha with two cover soils (FMM, PMM) and one placement depth (18 cm) (Lanoue and Qualizza 2000). Cover soils were placed over 23 to 35 cm of secondary substrate. Detailed salvaged FMM and PMM donor sites characteristics can be found in Table S1. The study was conducted across the sites in systematic, random block designs with 3 to 7 replicates. Depending on the size, each plot consisted of 7 to 18 quadrats and each quadrat was 1 × 1 m with 6-10 m spacing. Total number of quadrats varied with site and were based on species area curves.

Vegetation assessment
Vegetation was assessed in August using rectangular 0.5 m 2 (1 m × 0.5 m) quadrats. At each site three transects were spaced approximately 7 m apart. Quadrats were assessed along these transects with number determined by species area curves; 84 at Site 1, 235 at Site 2, 168 at Site 3, and 120 at Site 4. In each quadrat, species cover by vegetation growth form, litter cover, bare ground, woody debris, rocks (>2 cm diameter) moss, and woody plant density (shrub, tree) including native and non-native species was determined. Systematic walk-throughs in each experimental unit were performed to locate species not found in the quadrats. Species nomenclature followed Moss (1994). The slope and aspect of each experimental unit were measured. Historical data from each site were used for assessment of response over time, as all data were collected using the same procedures.
The competitive-stress tolerant-ruderal (C-S-R) system which classifies plant species by adult strategy of response to stress and disturbance (Grime 1974;Hodgson et al. 1999;Hunt et al. 2004) was applied to determine dominant species on each treatment. Under low stress and high disturbance species tend to ruderality (R); under high stress and low disturbance species tend to tolerance (S); under low stress and disturbance species tend to be competitive (C). Under intermediate stress and disturbance species are classified as intermediate types (C/CR or R/CR) representing the relative importance of the three attributes. Other combinations can be read based on C-S-R criteria; for example, C/SC or SC can be read as competitive, competitive-stresstolerant, or C/CSR can be read as competitive, competitive-stress-tolerant ruderal. Leaf area was used to determine C-S-R functional types . Plant samples were collected from late June to early July from research sites and natural forests from the surrounding areas. Samples of approximately 10 robust, unshaded, undamaged leaves were taken from multiple flowering plants and multiple locations per plant. For dry weight, leaves were oven dried individually at 70°C for 48 hours. The WinFolia software program (Regent Instruments Inc.) was used to measure leaf area of scanned leaves. Information on lateral spread and canopy height was mainly from Moss (1994) and Tannas (2003aTannas ( , 2003bTannas ( , 2003c. Flowering start and period references were from Wilkinson (1999), Royer and Dickinson (2007).
Species with greatest cover were considered dominant and two or three with similar cover were considered co-dominant. Subdominant species were those that occupied more space in quadrats than most, > 5% of total cover (proportion based). Rank abundance curves were used with 5% criterion to determine subdominant species. A rank abundance curve was generated for each treatment. Dominant and subdominant species were classified by successional stage (early, early to late, late), C-S-R functional type (competitor, stress tolerant, ruderal, intermediates) and habitat (natural, natural and disturbed, disturbed) to determine if reclaimed sites were dominated by early successional or disturbance type species or whether late successional and natural type species were establishing.

Statistical analyses
Vegetation cover, density, richness, Shannon diversity, evenness, and species composition were used to compare cover soil types and application depth. Each site was analyzed separately using SAS statistical software at p ≤ 0.05 (version 9.3, SAS Statistical Institute), with cover soil type, application depth, and age as fixed factors. Data were tested for normality and homogeneity of variance prior to testing with analysis of variance (ANOVA). If nonnormal permutational analysis of variance (perANOVA) was used with Bray Curtis distance; if non-homogeneous Proc mixed for heterogeneous variances was used. The non-parametric Scheirer-Ray-Hare extension of the Kruskal Wallis test was run in R version 4.0.3 (R Core Team 2020) R Development Team) when data were not normal and variances were not homogeneous. Post hoc pairwise comparisons were conducted after significant ANOVAs and perANOVAs using a SAS macro Dunn's test for multiple comparisons. Permutational analysis of multivariate dispersion (PERMDISP) showed age of all data sets that failed the homogeneity test. Bonferroni-corrected pairwise comparisons were conducted after significant perMANOVAs. Further nonmetric multidimensional scaling ordination (NMDS) ordinations were conducted to verify significant perMANOVA results using ecodist library in R where Bray-Curtis dissimilarity was used. NMDS ordination examined plant community composition and species assemblages as they related to physical environmental attributes (McCune and Grace 2002).

Woody plant density
Woody plant density was significantly greater on FMM than PMM or peat-sand cover soil across all sites and years since reclamation (Table 1). Woody plant density generally increased steadily over time; sometimes with a small decrease at the beginning, a dramatic increase in year 3 or 4, and a decline by year 7. Cover soil placement depth generally had a significant effect on woody plant density. FMM cover soils at placement depth 20 cm had greater woody plant density than at 10 cm depths at all ages, whereas no distinct trend was observed for PMM depth.

Vegetation cover
Vegetation cover was significantly higher on FMM than PMM at all sites (Table 2), whereas time since reclamation was mostly significant on cover across years. For example, Site 1 cover was significantly greater on FMM than PMM in years 2 and 4; Site 2 cover in years 1, 3, and  5 was significantly greater on FMM than PMM. Vegetation cover on 20 cm FMM was significantly greater than on 10 cm depth across all sites and time since reclamation, which was not the case for PMM. Native species cover on FMM was significantly greater than on PMM; both cover soils showed an increasing trend over time. Non-native species cover had a decreasing trend with time on FMM, whereas in PMM it increased up to 3 years after reclamation then decreased in some sites. Site 4 non-native cover (21.2%) was greater than native cover (19%) 13 years after reclamation on PMM. Native species cover was greater on 20 cm than 10 cm FMM, with the opposite trend on PMM.

Species richness and diversity
Total species richness was generally significantly different on FMM and PMM, with an increasing trend with time across sites (Table 3). In the year after reclamation, FMM had greater species richness than PMM, with differences decreasing with time. FMM had significantly Table 2. Mean (±SE) total, native and non-native species cover at sites and years by cover soils and placement depth. Lower case letters indicate significant differences between treatment combinations within years and upper case letters indicate significant differences between years within treatment at p < 0.05. *PMM cover soil placement depth 100 cm.
61.6 (9.5) a/A 61.6 (9.5) a/A 0.0 Site 3 1 10  greater native species richness than PMM across all sites and time since reclamation (Table 3); with an opposite trend for non-native species (data not shown). FMM 20 cm had greatest native species richness (Table 3); PMM 10 cm had greatest non-native species richness. Shannon diversity index was significantly different between FMM and PMM at the beginning of reclamation, but with time became insignificant (Table 3). There were significant fluctuations over time with PMM diversity significantly greater in year 7 than in year 1, and FMM equivalent in years 1 and 7 at Site 3; PMM had lower diversity than all other treatments at Site 4 ( Table 3). Application depth had no effect on diversity.
In most cases, evenness was not significantly different between cover soils, although there was a significant difference for time since reclamation across all sites (data not shown). Evenness was lower on PMM than FMM and cover soil depth did not impact evenness.

Community composition
The total number of vascular and non-vascular species varied across sites and time since reclamation (Table S2).
In total 180 species were found in all four sites; 37 graminoids, 109 forbs, 28 shrubs, and 6 trees. In total 146 native and 34 non-native species were found across sites. After 13 years of reclamation at Site 4 FMM had 84 species, PMM had 66. Site 4 differences were less noticeable. After 7 years Site 3 FMM had 100 species on 20 cm and 93 on 10 cm; PMM had 87 species on 20 cm and 90 on 10 cm. At Site 1 the total number of species was 86 in FMM and 85 on PMM 4 years after reclamation. At Site 2 there were 66 species on FMM and 48 on PMM 5 years after reclamation. Species composition by site showed variable response to cover soils and placement depths. Site 1 cover soils had significantly different species composition in years 2, 3, and 4, but not year 1. This was apparent in the X1 vs X3 NMDS ordination (Figure 1(a)). The X2 vs X3 ordination showed species associations most effectively (Figure 1(b)). In both ordinations, PMM was located upper right and FMM bottom left. Associated species in PMM were Chenopodium album L. (non-native species) in year 2 and Equisetum arvense L. (common horsetail) in years 3 and 4; on FMM Fragaria virginiana, Agropyron trachycaulum, Aster ciliolatus, and Achillea millefolium L. were most associated with years 3 and 4. The X2 vs X3 ordination showed progression through time, with younger plots in the upper left corner and older plots in the lower right corner. An annual species Corydalis aurea Willd. was most associated with year 1 for both cover soils, while perennial species Epilobium angustifolium, Sonchus arvensis L., and Rubus idaeus were associated with years 3 and 4. Site 2 × 1 vs X2 ordinations showed PMM loosely clustered on the left and separated from FMM except in year 1 (Figure 1(c)). The X1 vs X2 ordination showed a temporal progression with time (bottom to top) and the X2 vs X3 ordination (Figure 1(d)) showed a temporal left to right progression. Strong species associations on the X1 vs X2 ordination (Figure 1(c)) were early to late successional perennials (Elymus innovatus, Epilobium angustifolium, Vaccinium myrtilloides, Carex spp., and Rosa acicularis with years 3 and 5 on FMM; early successional perennial species Sonchus arvensis with years 3 and 5 and Urtica dioica L. with year 1 on PMM. Site 3 NMDS ordinations showed cover soils had differences in species composition in year 1 (Figures 2(a,b)).
Only in year 2, species composition differed significantly between 10 and 20 cm depths for FMM and PMM. In year 3 composition differed significantly between cover soils. Ordinations supported this, showing independent clusters of cover soils, with no differentiation between 10 and 20 cm. The X1 vs X2 ordinations (Figure 2(a)) showed a temporal change from years 2 to 7, and X2 vs X3 ordination (Figure 2(b)) showed different rates of change in species composition for cover soils. Early to late successional perennial species Carex spp., Taraxacum officinalis Weber, and Equisetum sylvaticum were strongly associated with years 3 and 7 on PMM. Early successional species Rubus idaeus, Epilobium angustifolium, Agropyron trachycaulum, Fragaria virginiana, Sonchus arvensis, and early to late successional species Mertensia paniculata, Rosa acicularis, Achillea millefolium, Elymus innovatus, Aster ciliolatus, Vicia americana Muhl., and moss were associated with years 2, 3, and 7 on FMM. Site 4 ordinations (X1 vs X2 and X1 vs X3; Figures 2(c,d)) showed compositional differentiation between cover soil types 13 years after reclamation. In total 22 species associated with FMM and 12 with PMM, of which two non-native (Bromus inermis, Galeopsis tetrahit) associated with FMM and eight non-native species (Agropyron repens, Erucastrum gallicum, Lotus corniculatus, Medicago sativa, Melilotus alba, Poa pratensis, Sonchus arvensis, Taraxacum officinalis) associated with PMM.

Dominant and subdominant species
Site 1 dominant species differed with cover soils in years 1 and 4 after reclamation; subdominants differed in year 1 (Table 4). Year 1 dominants and subdominants on both cover soils were competitive-ruderals, with one intermediate subdominant (Achillea millefolium); year 4 cover soils had stress-tolerant-competitive, competitive, and competitive-ruderal dominants and subdominants (Table S3). Site 2 FMM and PMM were dominated by a native, early successional species (Epilobium angustifolium) in year 1. FMM subdominants were native early to late successionals; PMM had native and non-native early successionals (Table 4). Year 5 early to late successional moss dominated FMM and PMM with a non-native (Sonchus arvensis). All subdominant species were early and early to late successionals. Differentiating cover soils at Site 2 on C-S-R functional type was difficult; year 1 was dominated by a competitive, with competitive, competitive ruderal or C/CSR subdominants; only PMM had an R/ CR species (Table S3). Year 5 FMM sub-dominants were competitives, competitive-stress-tolerants, C/CSR species, and competitive-ruderals; PMM subdominants were competitive and stress-tolerant. Site 3 dominant species on FMM and PMM were almost similar in year 1 but differed in year 7 (Table 4). Subdominants in year 1 and 7 differed by soil covers. Competitive or competitive-ruderals dominated in year 1, whereas in year 7 all treatments continued to have competitive, competitive-ruderal and stresstolerant-competitive species, and moss; only PMM had an R/CSR species (Table S3). After 13 years since reclamation Site 4 FMM was dominated by native and PMM by nonnative species (Table 4). FMM dominant species were stress-tolerant-competitive; PMM dominants were competitive-ruderals, R/CSR, or S/CSR species. Differentiation of subdominant species was difficult.

Discussion
Overall, outcomes of this study indicate FMM had a positive influence on woody plant densities, cover, and species richness, and a negative influence on non-native species relative to PMM. Greater understory richness and vegetation cover with FMM was found in other oil sands reclamation studies in the laboratory and field (Mackenzie and Naeth 2010;Errington and Pinno 2015;Dhar et al. 2020). Cover soil differences in species assemblage, composition, dominance, and types (successional stages, habitat types, CSR strategies) support better performance with FMM than PMM which might be influenced by several factors. The propagule bank of donor soil and how it is transferred to the receiver site, including stockpiling, can influence species richness of emergent vegetation, and number of propagules of each species will influence its density and cover (Rokich et al. 2000;Mackenzie and Naeth 2010;Dhar et al. 2018). Condition of cover soils and how their propagules respond to those conditions affect species that arrive at the receiver site through wind and animal dispersal. Although seed dispersal from surrounding areas plays Table 4. Dominant and subdominant species initial and final year after reclamation at sites. R: low stress and high disturbance species (ruderality), S: high stress and low disturbance (tolerance), C: low stress and disturbance (competitive), C/CR or R/CR: intermediate stress and disturbance species (intermediate), C/SC or SC: competitive, competitive-stress-tolerant, C/CSR: competitive, competitivestress-tolerant ruderal, CR: competitive ruderal. an important role in revegetation (Baker et al. 2013), success of species establishment specifically for trees on disturbed sites is dependent on dispersal distance from potential seed sources, seed size, and types of seeds and species (Asselin et al. 2001). Several studies reported that biotic interactions that occur among species drive plant community development in early successional sites post fire or post logging (Dhar et al. 2018;2020a), although reclamation sites cannot be expected to follow similar developmental patterns to natural disturbances. Typically plant community development at reclamation sites is a function of biotic factors (soil propagule bank, seed dispersal, wildlife impacts), abiotic factors (environmental conditions, soil properties, landform, topography), and time since reclamation (Rokich et al. 2000;Burger and Fannon 2009;Hobbs et al. 2009;Tropek et al. 2013;Dhar et al. 2018). Differences in cover soil propagule banks may be the most obvious factor affecting plant community development as there was some variability in donor site species composition (Table S1), but was not most decisive for some plant groups. Mackenzie and Naeth (2010) found Site 3 donor FMM had significantly more grass, sedge, rush, forb, and native propagules in the upper 10 cm than donor PMM, although materials did not differ significantly in number of woody plant propagules. This was supported by our findings of total and native emergents being significantly higher in FMM than PMM. Mackenzie and Naeth (2010) found forest floor and peat mineral donor soil propagule banks contained 9000 and 3600 seeds m −2 , respectively; 8300 seeds from native species and 100 from non-native species in forest floor soil, and 3300 from native species and 17 from non-native species in peat soil. Seed densities in productive boreal forest soils were 1273 to 9108 seeds m −2 (Moore and Wein 1977;Hills and Morris 1992). At Site 3 Mackenzie and Naeth (2010) found large losses of propagules during transfer of salvaged FMM material to the receiver site. Number and type of emergents from propagule banks from cover soils at the receiver site were similar, and more related to application depth than cover soil. Woody plants and non-native species had lowest emergents. Differences in propagule bank size at the donor sites did not translate into differences at the receiver site, although the propagule bank of FMM at the receiver site had almost double the number of species of PMM. A lack of propagules emergent was observed on FMM at Site 2 as propagule densities in seed banks from receiver sites were low, with cover soils having similar numbers of grass and woody propagules (Mackenzie 2012). Despite low propagule densities, our study found FMM at the receiver site had greater species richness than PMM. No information was available about propagule banks of Sites 1 and 4 donor materials, although similar losses likely occurred through salvage and placement operations, resulting in both cover soils having propagule banks more similar to each other than that of the original donor soils, at least in growth form assemblage.
Stockpiling may have a direct impact on soil propagule vitality (Mackenzie et al. 2019;Dhar et al. 2019;Shaughnessy et al. 2022), as most of our cover soils were stockpiled 3 to 6 months except Sites 2 and 4 where stockpiling times for PMM were not well documented. Mackenzie et al. (2019) found stockpiling can lead to significant losses in viability within 8 months.
Although there is no information on causes of reduced propagule viability during stockpiling, some studies suggest in-situ germination, predation, physical and mechanical damage, seed decomposition, and loss of microbial associates as the major reasons (Rokich et al. 2000;MacKenzie et al. 2012;Naeth et al. 2013;Dhar et al. 2019).
Emergence from propagule banks was directly influenced by soil conditions, which could be driving differences seen in cover soils. In similar conditions FMM could be more favourable for seed germination and growth than PMM. Propagules in PMM prefer to grow in subhydric to hydric conditions, as peat soil is derived from bogs and fens (Beckingham and Archibald 1996). PMM placed on upland sites during reclamation tends to dry, making it difficult for propagules to establish, grow, and produce native vegetation cover; FMM cover soil was better adapted to conditions as the memory of biological legacies was carried from the upland boreal forest (Dhar et al. 2018(Dhar et al. , 2020a. Therefore FMM can provide a more suitable environment for germination and growth, particularly with available nutrients and microbial associations. Vegetation parameters such as cover and richness showed more affinity to deeper than shallow placement depths. Our study revealed greater vegetation cover on 20 cm placements for both cover soils at most sites, although this effect likely decreases with time as space becomes more restricted and is confounded by some species occupying more space than others. Mackenzie and Naeth (2010) found total, forb, woody, grass, native and non-native cover higher with 20 cm than 10 cm placement on FMM; no significant difference was found for species richness and diversity.  suggest deeper cover soil would be better for establishing some less competitive native forest species, as shallow cover soil would potentially reduce propagule viability due to exposure and desiccation. Application depth for cover soil should be within 10 to 20 cm for better vegetation growth, which is comparable to the study by Dhar et al. (2018).
Non-native cover remained low or was decreasing over time on FMM, but increasing on PMM in some sites early after reclamation. Some studies showed number and cover of non-native species was greater on PMM than FMM (Mackenzie and Naeth 2010;Errington and Pinno 2015;Dhar et al. 2020). This might be due to better access to native propagules at early stages on FMM than on PMM and possible competition. Non-native species on FMM were annuals or biennials, likely reproducing by seed, and may have difficulty establishing with competing native species. Native species may have an advantage on FMM due to interactions with microorganisms and soil conditions carried from upland soils. Non-native species on FMM declined across sites, although the initial increase might be due to exploiting resources available after reclamation and decreasing as native species developed. Increasing non-native species cover on PMM in some sites early after reclamation is comparable to other long-term studies (Pinno and Hawkes 2015;Dhar et al. 2020a). Pinno and Hawkes (2015) found nonnative species cover in PMM increased early in reclamation and was 10% after 20 years. Dhar et al. (2020a) found non-native species cover declined from 32 to 10% 24 years after reclamation. Lower native species cover and corresponding lower competition on PMM at an early stage likely allowed more non-natives to establish. While these studies indicate non-native species on PMM will decline with time, ongoing research with continuous monitoring is needed to document the potential long-term impacts of non-native species on ecosystem recovery of reclaimed sites if PMM is used as a cover soil.
Native perennial shrub species such as Rubus idaeus, a dominant species at FMM reclamation sites, are often found in recently disturbed boreal forest communities (Hart and Chen 2006). The presence of such shrub species is a good indicator of reclamation success, as shrubs accelerate forest floor development, facilitate tree establishment through snow retention, reduce drought and temperature stress, and reduce seedling mortality from browsing (Rowland et al. 2009). Two nitrogen-fixing shrubs, Alnus spp. and Shepherdia canadensis (L.) Nutt., found across sites can increase soil nitrogen pools and soil organic matter (Mummey et al. 2002), which could improve overall soil chemical and biological properties and influence plant community development in reclamation sites (Rowland et al. 2009;Dhar et al. 2018).
Dominance of functional groups is an important aspect of characterizing plant communities when describing the successional trajectory for a disturbed site (Dhar et al. 2018). In boreal forest, trees and shrubs shade ruderal species (Messier et al. 1998;Lieffers et al. 1999), which reduces competition and promotes development of forest understory species. FMM dominated by perennials with an early and early to late successional community and PMM dominated by non-natives, ruderal or early successional species shows that the trajectory of community development in FMM appears to be consistent with typical successional progress of boreal forests from ruderal and annual communities to perennial communities (Messier et al. 1998;Lieffers et al. 1999;Hart and Chen 2006;Dhar et al. 2018). Dominance by annual, non-native ruderal or early successional species in PMM could slow succession and ecosystem recovery, and increase upland species with time, indicating propagule dispersal plays a significant role in upland plant community development and ecosystem recovery which requires time.

Conclusion
Results support preferential use of FMM over PMM as woody plant density, species richness, and cover were greater in most sites. Cover soil placement depth of 20 cm performed better than 10 cm in both FMM and PMM. Non-native species cover declined on FMM and continuously increased on PMM up to a few years after reclamation. Species composition and assemblage differed between cover soil types at most sites. Dominant and subdominant species on FMM were always native and mixed early and early to late successional species, while non-native and annual forb dominated on PMM.