Advertisement

New Forests

, Volume 50, Issue 2, pp 217–239 | Cite as

Exploring drivers and dynamics of early boreal forest recovery of heavily disturbed mine sites: a case study from a reconstructed landscape

  • Morgane MerlinEmail author
  • Frances Leishman
  • Ruth C. Errington
  • Bradley D. Pinno
  • Simon M. Landhäusser
Article

Abstract

Ecological processes driving tree success in the early stages of succession are complex and often poorly understood, involving direct and indirect relationships among multiple agents modulated by legacies. Reclamation areas of open-pit mines provide a unique opportunity to study these relationships, as these sites are often homogeneous and have few ecological legacies. Our study evaluated the early performance of three boreal tree species planted at different densities over a large reclaimed landscape (57 ha) in northern Alberta, Canada, that varied in landscape topography, surface soils, and coarse woody material. A range of soil, topographical, and seedling characteristics were measured over five growing seasons. The two surface soils had stark differences in their physical and chemical properties. Overall, seedling survival was high (> 80%) for jack pine (Pinus banksiana) and white spruce (Picea glauca) and somewhat lower (60%) for trembling aspen (Populus tremuloides). All three species grew taller in the fine-textured than in the coarse-textured soils. Linear models identified simple relationships among some of the monitored variables. To further explore more complex relationships, we built a structural equation model for jack pine growth, 2 and 5 years after planting. On coarse-textured soil, the factors controlling pine growth shifted from intrinsic factors of seedling quality in the second growing season to more complex interactions in the fifth growing season, driven by soil nutrients, water availability and colonizing vegetation. We believe these models have the potential to be useful in forest reclamation, by identifying driving factors that could be monitored to indicate reclamation success.

Keywords

Forest restoration Pinus banksiana Populus tremuloides Structural equation model Seedling survival Environmental drivers of growth 

Introduction

Recognizing and understanding complex ecosystem processes and interactions is a prominent challenge in ecological research (Bazzaz 1979). Within the context of forest recovery after severe disturbance, determining possible successional trajectories is challenging, as there is a large suite of processes that need to be considered which are dependent on factors operating at different spatial and temporal scales (Turner et al. 1993; Everham and Brokaw 1996; Foster et al. 1997; Kuuluvainen 2002). For example, the presence of a canopy (tree or shrub) following a disturbance, even at an individual scale, can have a significant impact on the recovery process (Chazdon 2003; Lindenmayer et al. 2010; Baker et al. 2016). Apart from competitive effects, colonizing vegetation may also provide positive “nurse” effects for other forest species by preventing the establishment of more competitive species (Vieira et al. 1994; Carpenter et al. 2004; McKee et al. 2007; Macdonald et al. 2015b). Planting of targeted tree species is often used to accelerate forest restoration and recovery, promoting the rapid establishment of a closed canopy and the establishment of volunteer trees and understory vegetation as well as modifying the physical and biological site conditions (Parrotta et al. 1997; Omeja et al. 2011; Macdonald et al. 2015a).

The functions and processes governing the trajectories of ecosystem recovery in boreal forests are likely to be different from those of other forest ecosystems due to the unique conditions present in the boreal region. These include a short growing season and extreme climatic conditions (Kirdyanov et al. 2003; Seo et al. 2011), low tree diversity but high understory diversity (Shugart et al. 1992; Gilliam 2007), and frequent recurring fire disturbances (Bergeron et al. 2002; Turner 2010), conditions which all influence fundamental ecosystem processes. High-intensity anthropogenic disturbances in the boreal region, such as surface mining and subsequent reclamation, provide a unique opportunity to test processes in forest recovery after disturbances that are outside of the natural disturbance regime. To access the resources, surface mining involves the complete removal of vegetation, soil layers and overburden material. In Alberta, public land disturbed by industrial activities has to be returned to conditions equivalent to the land capability prior to disturbance (Government of Alberta 2017). Forests ecosystems have to be reconstructed including soil and vegetation on public forestland. The establishment of boreal tree species is thus one of the main objectives when reclaiming these boreal forest landscapes after disturbance. The redevelopment of a tree canopy is considered an essential step in the initiation and re-establishment of other ecosystem components and functions (Macdonald et al. 2015a; Jacobs et al. 2015). However, obtaining the desired canopy conditions can be challenging on newly constructed landscapes (Mackenzie and Naeth 2009; Drozdowski et al. 2012; Macdonald et al. 2015b).

Soil physical, hydrological, and mechanical properties (Hébert et al. 2006; Barr et al. 2007; Hogg et al. 2008; Hember et al. 2017) as well as nutrient availability (Tamm 1991; Manninen et al. 2009; Drozdowski et al. 2012; Fisher et al. 2012) play a critical role in dictating forest recovery and determining the potential of the site to support future forests. However, other site variables such as topographic heterogeneity at different scales, tree composition, canopy closure and coarse woody debris presence play a significant role (Beatty 1984; Cornett et al. 1997; Frouz et al. 2011; Schott et al. 2014; Melnik et al. 2017). The presence of colonizing vegetation may either hamper tree seedling establishment and growth in early stages or facilitate it (Andersen et al. 1989; Landhäusser and Lieffers 1998; Thiffault et al. 2003; McIntire and Fajardo 2014). Reclamation sites provide an opportunity to test some of these fundamental questions, as both biotic and abiotic conditions can be manipulated during site reconstruction to explore the role of specific site components.

Focusing on a few pivotal ecosystem components has been a strategy to construct an understanding of overall ecosystem function and unveil possible interactions (Beattie 1996; Pollack 2003). However, ecosystem components have complex interactions with one another and their importance can change throughout forest succession (Comita et al. 2009) and with the presence of legacies from the pre-disturbance forest conditions (Foster et al. 1998; Savage and Mast 2005). Expanding the range of variables might help to understand the complex interactions at play in forest recovery, but exploring these more complex interactions requires new tools for analysis. Structural equation modeling (SEM) aims to assess the network of direct and indirect interactions among abiotic and biotic variables with the aid of graphic modeling, thus potentially unveiling relationships not seen with traditional linear modeling. SEM has been increasingly used in ecological studies (Malaeb et al. 2000; Grace et al. 2011; Gimenez et al. 2012; Eldridge et al. 2015) and has the potential to become a tool in assessing restoration efforts by providing both fundamental knowledge of succession, as well as criteria and indicators of successful reclamation practices.

In this context, our research had two main objectives: (1) to assess individual planted tree species growth and survival in relation to soil characteristics and competing vegetation, and (2) to build a network model assessing the direct and indirect relationships linking reclamation practices, soil properties, and vegetation development with Pinus banksiana seedlings growth the second and fifth growing seasons after planting.

Materials and methods

Reclaimed landscape description

Research plots were located within a larger reclamation area of a former tailings pond in the Athabasca Oil Sands Region near Fort McMurray, Alberta, Canada (57°2″N, 111°35′W). During reconstruction, a local watershed (~ 57 ha) was created using features of dry upland hills (hummocks) and wet lowland areas between the hummocks (Fig. 1). The reconstructed watershed (Sandhill Fen Watershed) is situated on Syncrude Canada Ltd.’s Mildred Lake lease. The watershed was constructed on a nominally 10 m thick cap of tailings sand that is underlain by approximately 35 m of interbedded composite tailings and tailings sand layers created during the extraction process of bitumen from the ore (Wytrykush et al. 2012). The upland area of the watershed has eight hummocks that vary in size and height (Fig. 1). Hummocks were created by mechanically placing tailings sand, which were then capped with different salvaged soil materials to recreate soil conditions that are representative of two major forest types of this region (see below) (Pollard et al. 2012). One of the cover soils was reconstructed by placing a 40 cm subsoil layer of salvaged Pleistocene fluvial sand and then covered with 15 cm of salvaged forest floor material (FFM). The FFM material, a mixture of the organic forest floor layer and the underlying mineral soil, had been salvaged to a depth of 15 cm from P. banksiana forests on coarse-textured Brunisols (Soil Classification Working Group 1998). The second cover soil type used for some of the hummocks was a 40 cm thick layer of a fine-textured clay-rich subsoil material that was subsequently covered with a 20 cm FFM material cap salvaged to a depth of 20 cm from mesic upland forest sites dominated by Populus tremuloides on fine-textured Luvisols (Beckingham and Archibald 1996; Soil Classification Working Group 1998). All cover soil materials were directly placed without stockpiling to preserve the viability of the propagules bank within. Due to the differences between the two reconstructed soils (mostly driven by soil texture) and the site conditions created by these soils, we characterized the cover soils as either “coarse” or “fine soil” from here on to assess seedling performance. More information on the physical and chemical properties of the two soil materials can be found in Supplementary Information Table SA1. Five of the hummocks in the watershed had a coarse soil cap, while three other hummocks received a fine soil cap (Fig. 1). Following cover soil placement, coarse woody debris (mainly stems and large branches of white birch and trembling aspen) were placed on all hummocks in volumes ranging from 0.2 to 200 m3 ha−1. Construction of the reclamation area was completed in the spring of 2012.
Fig. 1

Overview of the reclaimed landscape showing upland hummocks capped with either a fine soil (solid outline) or a coarse soil (dotted outline). Two seedling planting density prescriptions were applied on the landscape and represented here: a medium planting density (5000 stems ha−1, light grey) and a high density planting treatment (10,000 stems ha−1, dark grey). Individual research plots are presented on the map as black squares

In the first week of June 2012, upland areas were planted with 1-year-old container grown tree seedlings. Seedlings were grown commercially at nurseries from open pollinated local seed sources. Trembling aspen and white spruce seedlings were grown in 515A (250 ml, 5.1 cm diameter, 15.2 cm deep) styroblock containers, while jack pine seedlings were grown in 412A (125 ml, 4.6 cm diameter, 11.7 cm deep) styroblock containers (BeaverPlastic, Edmonton, AB).To assess the impact of planting density on forest recovery, each hummock was partitioned into three roughly equal sections and planted to either high density (10,000 stems ha−1), medium density (5000 stems ha−1) or stayed unplanted as a control (Fig. 1). Since this study assessed planted seedling performance only, we report data only from the planted areas. The composition of planted tree species on the different hummocks was based on the two dominant forest types that naturally can be found in this region. Hummocks with coarse soil were planted predominantly with jack pine (Pinus banksiana Lamb.) (80%) and a mixture of aspen (Populus tremuloides Michx.) (10%) and white spruce (Picea glauca (Moench) Voss.) (10%). The fine soil hummocks on the other hand were planted predominately with P. tremuloides (80%) and 10% of P. glauca and P. banksiana. Upland areas other than the hummocks that were not part of our measurements were planted at an operational density of 2000 stems ha−1 and included various mixtures of tree and shrub species: aspen, white birch (Betula papyrifera Marsh.), white spruce, jack pine, dogwood (Cornus sericea L.), and green alder (Alnus viridis ssp. crispa (Chaix.) D.C.).

Plot layout and measurements

A total of 52 permanent research plots (15 × 15 m) were established across the hummocks of the upland research area of the watershed in the medium and high planting density treatments (Fig. 1). The plots were spatially arranged across the landscape to capture a range of cover soils, planting prescription, topographical positions, i.e. aspect and slope position. As such, 38 plots were established on coarse soil on five hummocks on level, north and south-facing areas, and 14 plots were established on fine soil on the three hummocks on level and west-facing slopes (Fig. 1). Based on the topographical position a heat load index was calculated for each research plot following McCune and Keon (2002), to convert the factorial variable “aspect” into a continuous variable combining both aspect and slope. To minimize disturbance impacts through foot traffic, a measurement plot (7 × 7 m) was nested within the center of each research plot. All seedlings were measured in each measurement plot and vegetation recovery was assessed in four 1 m2 vegetation plots established in each corner of a measurement plot.

Seedling performance

Planted seedling performance was assessed in all high and medium density planting plots. Survival was assessed on all planted seedlings in each measurement plot for each species at the end of the fifth growing season (i.e. at the end of the 2016 growing season). Ten individuals of the dominant tree species in each plot and all seedlings of the other two planted tree species (up to ten) were selected for growth measurements (a seedling was considered a subsample). Mortality across the site contributed to lowering the number of seedlings available for measurement for white spruce (n = 18 plots across the landscape). Height was measured yearly on these seedlings at the end of each growing season. For all further analysis, survival and height were averaged per measurement plot. As part of the planted seedling assessment, foliar nutrient concentrations were determined in August 2013 and 2016 on foliar samples of trembling aspen and jack pine collected from each measurement plot. Nitrogen concentration was determined using the Dumas Combustion Method (Dumas 1831) on a TruSpec CN Carbon Nitrogen Determinator (©2010 LECO Corporation). Concentrations of P and K were determined by inductively coupled plasma optical emission spectrometry (ICP-OES) after microwave-assisted nitric acid digestion. To compare relative changes in leaf nutrition between years, leaf nutrient concentrations were used. Leaf nutrient concentration rather than content was used, as leaf samples are routinely collected during performance surveys (Alberta Environment and Sustainable Resource Development 2013) and do not require a knowledge of the total size of the pool, needed to estimate nutrient content.

Ground cover

Coarse woody debris presence and volume was assessed in each measurement plot using a line intercept method adapted from Van Wagner (1982), Marshall et al. (2000) and Natural Resources Canada (2008). For the structural equation models for pine on the coarse soils (see below), the distribution of the coarse woody debris volume was strongly bimodal, thus coarse woody debris was used as a categorical variable with two levels: “presence” (volume > 20 m3 ha−1) and “absence” (volume < 20 m3 ha−1). On the fine soil (see below), coarse woody debris volume was much more evenly distributed, and consequently it was used as continuous variable “coarse woody debris volume” for assessing trembling aspen growth. Ocular estimation of plant cover—distinguished into vascular vegetation and bryophytes—was done in each vegetation plot during July of 2013 and 2015. Bryophyte cover was separated from other colonizing vegetation cover to take into account the development of vertical structure in the colonizing vegetation community. Percent cover was estimated to the closest 1% below 10%, and to the closest 5% for greater cover. If cover was considered to be less than 1%, it was assigned a cover of either 0.5% or trace (0.05%).

Edaphic variables

Precipitation data was obtained from climate stations present on site (McMaster University, Hamilton, ON, Canada). Soil water availability was measured in the center of each measurement plot using MPS-2 matric water potential sensors (Decagon Devices Inc. Pullman, WA, USA). Data was recorded by an EM50 datalogger (Decagon Devices Inc. Pullman, WA, USA). One sensor was installed at a depth of 10 cm (midway through the forest floor material, hereafter referred to as “Water potential at 10 cm”) and the other sensor was installed at a depth of 15 or 20 cm (at the interface of the FFM cover and the subsoil, hereafter referred to as “Water potential at 20 cm”). Collection of soil matric water potential data began only in May 2013 in order to allow sensors to calibrate to the soil conditions throughout the winter since installation in August 2012. Soil water potential was averaged over the growing season (May 1st–September 31st) for each year.

Soil available nutrient supply rate was assessed using PRSRTM probes (Plant Root Simulator probes, Western Ag. Innovations, Saskatoon, Canada), which are ion exchange resin membranes measuring cation and anion supply rates in situ. The quantity of soil ions adsorbed is a function of soil properties (physical, chemical and biological), soil moisture, and burial time, which control the soil nutrient availability to plants. Four pairs of anion and cation probes were installed in the top 10 cm soil layer from mid-July to end of August in each measurement plot for 42 days during each of the 2013 and 2016 growing seasons. The four anion and four cation probes were pooled for one analysis per measurement plot. Inorganic nitrogen supply rate was determined colorimetrically using an automated flow injection analysis system. Inductively-coupled plasma spectrometry was used for the other nutrient ion supply rates (K+, H2PO4).

To accurately represent the soil environment fine roots experience, both the soil water potential sensors and the PRS probes were installed at soil depths where the majority of fine roots from the seedling and the colonizing vegetation can be found (Landhäusser et al. 2012b; Laclau et al. 2013; Bockstette et al. 2017). Additionally, sensors and probes were located within or at the boundary between the cover soil layer and the subsoil, providing relevant information on the cover soil characteristics.

Statistical analyses

Linear models

All of the following analyses were conducted using the statistical R software (R Core Team 2015). All models were mixed effects models with Hummock as a random effect. Each variable was checked for normality assumptions using the Shapiro–Wilk test and appropriate transformations were done for any non-normal variable. Heteroscedasticity was checked for using the Breusch–Pagan test and when significant, variance was allowed to change per hummock to fit the models assumptions. Plots were treated as independent replicates within each hummock. Non-significant interactions were dropped from the model when necessary. Post-hoc tests were conducted using the package emmeans using estimated marginal means and the Tukey method for p value adjustment. The effects of soil, planting density and year on soil water potential and nutrients supply rates were assessed using a three-way ANOVA. An additional model assessing the effects of Aspect and Year in a two-way ANOVA was used on the coarse soil only for soil water potential. Survival of each planted species was analyzed using binomial generalized linear mixed effects models to assess for the effects of soil, planting density and aspect in 2016 (survival at the end of the fifth growing season).

Cumulative yearly growth for each planted tree species was analyzed for the effect of soil type using linear mixed effects models, as was final height at the end of the growing season in 2016. Mean planted seedling height in 2013 and 2016 was analyzed for each species on their target soil (jack pine on coarse soil, trembling aspen on fine soil and white spruce across the landscape) separately using linear mixed effects models on each species to assess for the effects planting density and aspect, coarse woody debris placement and/or volume, heat load index, soil water potential at 20 cm, foliar nutrition (seedling nitrogen, potassium and phosphorus), colonizing vegetation cover and initial seedling characteristics. The interactions between coarse woody debris and heat load index as well as soil water potential at 20 cm were assessed and removed if not significant.

Structural equation models

To further explore relationships among initial seedling characteristics, environmental and site variables, seedling height and year since planting, structural equation modeling (SEM) was applied to the multivariate data following the guidelines by Grace et al. (2010, 2012). Initially the study aimed to use the SEM technique to assess seedlings performance for each tree species on their target soil. However, the low number of measured seedlings per plot across the landscape and measurement plots on fine soil (for trembling aspen, n = 14, for white spruce, n = 18) prevented the use of the SEMs for these two species. This modelling technique was thus only applied to jack pine on the coarse soil. The working meta-model presented in Fig. 2 shows the underlying conceptual structure used to build the models following the example of Grace et al. (2010).
Fig. 2

Theoretical structural equation model used for years 2013 and 2016. The theoretical constructs are represented in dotted boxes and are as follows: (1) “Abiotic conditions” encompasses environmental conditions of the system limiting biological productivity, (2) “Reclamation practices” describes the degree to which reclamation practices have modified the site after cover soil placement, (3) “Colonizing vegetation” refers to colonizing vegetation covers and (4) “Seedling performance” refers to measures of foliar nutrition and seedling growth. The hypothesized pathways linking the constructs are labelled with the expected processes and their direction (+: positive, −: negative). The pathway directions and labels demonstrate the expected processes between constructs based on theory and previous observations

Extrinsic or environmental factors are represented by the abiotic conditions, reclamation practices and colonizing vegetation concepts in Fig. 2. Intrinsic factors are represented by seedling characteristics in Fig. 2. The observed variables behind the theoretical constructs in the model are presented in Table 1. Links between the observed variables in the structural equation model were based on existing knowledge of relationship between them, and thus are referred to as “drivers” for the seedling responses. The measurement plot was used as the experimental unit, with the observed variables averaged when multiple values per measurement plot were available. Although the use of latent variables—composite variables encompassing observed variables of similar nature—is often recommended for ecological studies (Grace and Bollen 2007; Grace et al. 2010), the low number of samples in our study as well as our goal to identify specific drivers motivated our choice to keep observed variables as individual variables in the model. As our sample size was small (36 measurement plots—experimental units—sampled per year as two plots had to be excluded from all 38 coarse-textured soil plots), a Bayesian approach using a Monte Carlo Markov Chain (MCMC) procedure (Lee 2007) was selected instead of common maximum likelihood procedures that rely on large sample sizes. All links between the observed variables were represented with linear relationships to limit model complexity, resulting initially in 17 pathways and a ratio of 2.2 of samples per pathways. The use of a Bayesian approach was supported due to the low ratio of samples to pathways. Estimation was performed using WinBUGS (Lunn et al. 2000) implemented through the R package R2winbugs (Sturtz et al. 2005). Three independent chains were used in the MCMC estimation process, with 500,000 iterations per chain, with a burn-in of 100,000. A thinning of 50 was used to ensure no autocorrelation among estimates. Results reported are repeatable based on independent estimation runs and all parameters were evaluated to determine whether there was basis for model simplification based on assessments of 95% credible intervals and no loss of explanatory power. Links that could be ignored were then removed and a new simplified model was built and assessed. Pearson’s correlations were calculated for relationships between residuals for the endogenous variables and the exogenous variables to quantify linear relationships. The ultimate decision to include additional linkages in a revised model was based on evaluation of parameter significance for included links. We used a query approach to estimate quantities equivalent to classical standardized coefficients (Grace et al. 2012). Each year was modelled separately to investigate the changes in the potential drivers of P. banksiana height. The equation specifications for the final structural equation models for each year are available in the Supplementary Information B.
Table 1

Theoretical constructs (see Fig. 2) and observed variables related to those constructs and their properties

Theoretical construct

Observed variables related to construct

Properties of variables

Abiotic conditions

Soil water potential at 20 cm (“WP20”)

Very low to close to zero (− 2500 to 0); continuous

Heat Load Index (“HLI”)

Continuous

Soil N supply rate (“Total N”)

Continuous

Soil K supply rate (“K”)

Continuous

Soil P supply rate (“P”)

Continuous

Reclamation practice

Presence of woody debris (“CWD”)

Categorical with two levels: presence-absence

Density of planted jack pine seedlings (“Density”)

Categorical with two levels: medium/high

Colonizing vegetation

Vascular vegetation cover (“Vegcover”)

0–100%; semi-continuous

Bryophyte cover (“moss”)

0–100%; semi-continuous

Seedling performance

Height

Continuous

Leaf N concentration (“Seedling N”)

Continuous

Leaf K concentration (“Seedling K”)

Continuous

Leaf P concentration (“Seedling P”)

Continuous

The variables labels used in the results are shown in parenthesis and between quotes in the second column

Results

Edaphic conditions

Average soil water potential decreased from 2013 to 2016 across the whole landscape with a more pronounced drop in the fine-textured soil (Fig. 3a). Soil water potential on the fine soil was lower than on the coarse soil in 2015 for the 10 cm depth and in both 2015 and 2016 for the 20 cm depth (p < 0.03). High density plots had lower soil water potential in 2016 than medium density plots on the fine soil for both depths (p < 0.03) whereas no difference was found on the coarse soil (Fig. 3b). No difference was found between plot aspect for either depth or soil type. For more detailed information on the annual patterns of soil water potential and precipitation refer to Supplementary Information Figs. SA1 and SA2.
Fig. 3

Average soil water potential (MPa) over the growing season (May 1st–September 30th) from 2013 to 2016 for two soils (coarse and fine) (a) and two seedling density treatments in 2016 (b): medium (M, 5000 stems ha−1), and high (H, 10,000 stems ha−1). Significant statistical differences between soils and years for each year are represented by letters. The dotted line at − 1.5 MPa represents the universally accepted wilting point for plants

Soil nitrogen supply rate decreased between 2013 and 2016 and was not different between the two soil types in either year (Fig. 4). Soil potassium and phosphorus supply rates were lower on the fine soil than on the coarse soil in both 2013 and 2016. No differences among planting density treatments were found for the supply rate of any nutrients.
Fig. 4

Soil nutrient availability (µg 10 cm2 per burial length) for 2013 (white) and 2016 (dark grey) on both coarse and fine soils as determined with PRSRTM probes [a total nitrogen (N), b potassium (K), c phosphorus (P)]. The statistical differences labelled here are from log-transformed data. Labels: “*”: p value < 0.05, “**”: p value < 0.01, “***”, p value < 0.001, ns not significant. Cover soil was a main effect across years for K and P

Seedling survival

Survival at the end of the fifth growing season was variable across species: trembling aspen: 62.5%, jack pine: 86.3%, white spruce: 94.5%. Soil type had no impact on survival of jack pine, but trembling aspen and white spruce survival rates were lower when planted at high density on the coarse soil only (trembling aspen: p < 0.05, mortality odds ratio of high versus medium planting density = 3.3, white spruce: p < 0.06, mortality odds ratio of high versus medium planting density = 10.5). Jack pine mortality was marginally higher in the high density planting treatment (p < 0.06, mortality odds ratio of high versus medium planting density treatment = 2.6) and on the south-facing plots (p < 0.05, mortality odds ratio of south versus north-facing plots = 9.6).

Seedling growth and foliar nutrition

For all three planted species, total height in 2016 was significantly greater on the fine soil compared to the coarse soil (all species, p < 0.001, Fig. 5). Trembling aspen seedlings growing on the fine soil did not show any differences between planting densities and aspects. Jack pine seedlings growing on the coarse soil had greater total height on north-facing slopes compared to other aspects (p < 0.01). Planting density did not affect jack pine total height. Density and aspect did not affect white spruce growth on either soil.
Fig. 5

Mean annual total height of the three planted species (a trembling aspen, b jack pine and c white spruce) on the coarse (black points, solid line) and fine soil (black triangles, dotted line). Statistical differences between years and soil have been evaluated for each species separately. The letters indicate a statistical difference for which p value < 0.05 between the years and soil. Error bars represent 95% confidence intervals

Changes in total height between soil types followed different patterns between species (Fig. 5): the coarse and fine soils diverged in 2013 for trembling aspen and not until 2014 for the conifers, with seedlings on the fine soil approximately twice as tall as seedlings on the coarse soil. For all species, lowest annual height growth was observed in 2015 (results not shown).

Jack pine foliar nitrogen concentration was higher in 2013 than in 2016 for both soil types (Table 2, p < 0.001) and was higher on the fine soil than the coarse soil in 2013 but not in 2016 (year × soil interaction, p < 0.05). Trembling aspen foliar nitrogen concentration was similar between soil types and years. Jack pine foliar potassium concentration was higher on the fine than on the coarse soil in 2013 but not in 2016 (year × soil interaction, p < 0.05) while there were no differences in trembling aspen foliar potassium between soils. Foliar phosphorus concentrations were similar between soil types for both species. For both soils and species, foliar potassium and phosphorus concentrations were higher in 2013 compared to 2016 (p < 0.05).
Table 2

Foliar nutrient (N, K and P) concentrations of planted trembling aspen and jack pine seedlings planted on coarse or fine soils in 2013 and 2016

Year

Soil

Nitrogen (mg g−1)

Potassium (mg g−1)

Phosphorus (mg g−1)

Mean

SD

Mean

SD

Mean

SD

Trembling aspen

 2013

Coarse

19.36 (ns)

2.81

6.64 (a)

1.49

1.71 (a)

0.28

Fine

19.94 (ns)

2.35

7.33 (a)

1.63

1.62 (a)

0.21

 2016

Coarse

19.06 (ns)

2.45

5.50 (b)

1.64

1.49 (b)

0.26

Fine

19.25 (ns)

1.99

4.42 (b)

0.72

1.37 (b)

0.15

Jack pine

 2013

Coarse

17.32 (a)

2.47

4.66 (a)

0.85

1.32 (a)

0.15

Fine

19.50 (b)

2.09

3.73 (b)

0.48

1.31 (a)

0.12

 2016

Coarse

13.92 (c)

1.15

3.09 (c)

0.41

1.12 (b)

0.12

Fine

13.53 (c)

0.80

2.90 (c)

0.38

1.08 (b)

0.07

Mean values and standard deviation for each species, year and soil are presented for foliar nitrogen, potassium and phosphorus. All values are presented in mg g−1. The letters between parentheses in the mean columns for each foliar nutrient indicate a statistical difference between the years and soil

ns not significant

Environmental drivers of seedling growth

Trembling aspen height in 2013 was negatively related to foliar nitrogen concentration and coarse woody debris volume, but positively related to foliar potassium and phosphorus concentrations (Table 3). Trembling aspen height in 2016 was positively related to foliar potassium concentrations only. There were no environmental or seedling-related variables related to white spruce height in 2013, but initial seedling height negatively influenced white spruce height on the fine soil in 2016.
Table 3

Linear models of the environmental and seedling-specific variables that affected height of trembling aspen on the fine soil and white spruce on both soils for the time periods of 2012–2013 and 2012–2016

Species

Response variable

Significant variables

R2

2012–2013

 Trembling aspen

Total height

Coarse woody debris volume (slope = − 0.52, p = 0.002)

Foliar K (slope = 5.44, p = 0.012)

Foliar P (slope = 54.61, p = 0.018)

Foliar N (slope = − 23.94, p = 0.059)

0.82

 White spruce

Total height

No model

 

2012–2016

 Trembling aspen

Total height

Foliar K (high vs. low concentration = 26.88, p = 0.012)

0.49

 White spruce

Total height

Soil × initial height (slope fine soil = -2.07 p = 0.019, coarse soil ns)

0.66

For each significant variable or interaction of variables, the slope (linear relationship), or difference (factor levels differences) is presented with its associated p value. The marginal R2 for each model is presented

ns non-significant

Structural equation models for jack pine on the coarse soil were developed with total height in 2013 and 2016 as the response variables. These models could not be applied to trembling aspen and white spruce due to the limited number of replicates. Comparing the SEMs for jack pine in 2013 and 2016 indicates a significant shift in factors that influenced seedling height from intrinsic seedling-based variables to extrinsic environmental-based variables (Fig. 6). In 2013 (Fig. 6a), foliar nitrogen concentration reflective of initial planting stock characteristics, was the only factor that could be directly linked to seedling total height. Soil nitrogen and potassium indirectly influenced pine height through a positive effect on foliar nitrogen concentration. Presence of coarse woody debris was positively associated with higher levels of foliar N and K whereas heat load index was negatively associated with foliar potassium. The cover of colonizing vegetation, including the bryophyte cover, did not affect any of the measured pine performance variables in 2013.
Fig. 6

Final structural equation models for jack pine total height assessed in 2013 (a) and 2016 (b) on coarse soils. R2 values for each of the response variables are displayed. Black solid and dashed arrows represent links between response and predictor variables. The numbers associated with the arrows are queries; they represent the percentage of variation of the response variable relative to its range when the predictor variable varies between its minimum and maximum value. The thickness of the arrow represents the strength of the link, with thicker arrows for queries above 0.5 and thinner arrows for queries below 0.2. The dotted line represents residual correlation between response variables, and its associated Pearson’s moment of correlation. These correlations could not be included in the models as they were either not the focus of the study or their estimation was not strong enough. Please refer to Table 1 for the abbreviations of variables

Environmental factors had higher significance for explaining seedling height in 2016. Jack pine height was directly and positively influenced by soil potassium supply rates, foliar potassium and to a lesser extent coarse woody debris presence (Fig. 6b). Presence of coarse woody debris was positively associated with higher levels of foliar phosphorus and potassium whereas heat load index was negatively associated with foliar potassium. Similarly, vegetation cover was positively linked with soil potassium supply rate as well as coarse woody debris presence. Soil water potential was positively influenced by vascular vegetation and negatively influenced by bryophyte cover. Bryophyte cover was dominated by the early successional mosses Ceratodon purpureus (Hedw.) Brid. and Bryum argenteum Hedw. Jack pine height was associated with decreased soil water potential. Linear models confirmed the findings of the structural models on total height for jack pine, but failed to identify the network relationships involving soil water potential and vegetation.

Discussion

Planted seedling performance showed different responses among species and large differences between the two soil types across the reconstructed landscape. The use of both linear mixed effects models and structural equation models uncovered both direct and indirect relationships among seedling characteristics, environmental variables, and growth. These relationships and their strengths changed over time, moving away from a strong influence of initial seedling characteristics after the second growing season to complex interactions with environmental variables after the fifth growing season.

Overall, survival of the planted conifer seedlings across the landscape after five growing seasons was high, above 80% for both jack pine and white spruce seedlings, while trembling aspen survival was lower, around 60% and was somewhat unexpected. The reduced outplanting performance of aspen might be related to the quality of the aspen seedling stock used in this study, as seedlings had low root to stem ratio (1.34 g g−1), low RCD (2.83 mm) and small terminal buds, all characteristics indicating lower quality seedling stock with a greater potential for poor outplanting performance (Landhäusser et al. 2012a, b). Models on seedling height confirm the crucial importance of seedling quality and their initial characteristics in the early establishment and performance of seedlings after outplanting (Grossnickle 2012; Ivetić et al. 2016). Foliar nitrogen concentrations had a strong effect on seedling height [as previously reported for nursery-grown seedlings (Salifu et al. 2009; Schott et al. 2016)], and so did initial seedling height. Impairment of root system development and the associated functions following planting is considered a major cause of transplanting shock, and leads to added stress of seedlings which can be exacerbated by dry conditions (Haase and Rose 1993). This early in forest regeneration, density-dependent intra- and inter-specific competition likely played a minimal role in seedling mortality (Franklin et al. 2002; Zhao et al. 2007), but may have contributed to higher mortality of aspen seedlings on plots with high densities of jack pine seedlings on the coarse soil. On the fine-textured soil where competition from the colonizing vegetation was much more intense, aspen seedlings may have had a greater disadvantage due to above and below-ground competition (Midoko-Iponga et al. 2005; Messier et al. 2009; Bockstette et al. 2017), especially when considering the potential for reduced seedling stock outplanting performance. White spruce survival was high, and seedling height was solely related to soil material and initial seedling characteristics in 2016, highlighting its tolerance to stress early in development compared to pine and aspen (Reich et al. 1998).

Seedling growth was about twice as much on the fine soil for the three tree species compared to the coarse soil. However, nutrient supply rates and soil water potential indicated reduced availability of these resources on the fine soil in 2016. The larger trees and the associated development of leaf area, including the colonizing vegetation, on these soils created a higher demand for water and nutrients, thus lowered the availability of soil water and nutrients compared to the coarse-textured soil (Knoop and Walker 1985; Singh et al. 1998; Haase 2008; Bedford and Small 2008). The seasonal dynamics of the soil water potential in 2016 clearly show a greater draw on soil water on the fine-textured soil, especially after precipitation events (Supplementary Information Figs. SA1 and SA2). Although not limiting now, water stress may become a greater factor affecting seedling performance in the future as seen during the dry summer of 2015 (25% less precipitation) when soil water potential approached the wilting point (− 1.5 MPa), periodically limiting seedling growth. The importance of sporadic precipitation events may be exacerbated in the future as canopy and thus water demand continue to expand on the landscape. When considering the coarse soil, areas with taller seedlings had greater water demand and resulted in decreased soil water potential, while controlling colonizing vegetation through its shade. Where the pine seedlings were small, colonizing vegetation had a greater foothold, but overall soil moisture availability was higher potentially due to lower demand of the colonizing vegetation compared to the planted seedlings. These responses point towards a complex relationship between competition (above and below-ground) and the abiotic environment, simultaneously affecting moisture and light availability. What exactly initiated and drove these differences is not clear, but could be related to a heterogeneous distribution of biological legacies (i.e. soil propagule bank) and soil characteristics [i.e. vertical textural distribution of the sandy soil material (Zettl et al. 2011)]. Interestingly, the high bryophyte cover was associated with lower soil water potentials. Bryophyte cover was in some areas up to 100%, and dominated by species that are common on disturbed, dry, and exposed sites (Økland et al. 2008). They can intercept and quickly evaporate precipitation, particularly during small rainfall events, with their large surface area (Pypker et al. 2006; Weber et al. 2016) potentially contributing to lower soil water potentials.

The models uncovered a significant shift in the drivers determining trembling aspen and most specifically jack pine seedling height on their respective target soils. Intrinsic seedling characteristics explained most of the differences in seedling height a year after planting, but extrinsic environmental variables and their feedbacks with abiotic soil conditions became more prominent when assessing seedling performance at the end of the fifth growing season. Similar shifts have also been described in other reclamation sites (Pinto et al. 2011; Schott et al. 2016). Both the linear and structural equation models indicated a change in the linkage between initial foliar nutrients and seedling growth with potassium replacing nitrogen in importance over time. Foliar potassium and phosphorus concentrations significantly decreased between 2013 and 2016, while foliar nitrogen stayed constant. The greater importance of potassium in seedling growth over nitrogen for trembling aspen and jack pine seedlings can be considered unusual for this region, as most boreal ecosystems are considered to be nitrogen limited (Larsen 1980; McGuire et al. 1992; Menge et al. 2008; Näsholm et al. 2013; Sigurdsson et al. 2013; Franklin et al. 2014; Ostertag and DiManno 2016), although this view has recently been challenged (Ste-Marie and Houle 2006; Manninen et al. 2009). The importance of potassium for pine species has been shown previously (Stone 1953; Stone and Leaf 1967; Stone and Kszystyniak 1977) as well as for aspen seedlings (Pinno et al. 2012). Soil potassium availability measured with PRS probes in our coarse soil was similar to those measured in a natural jack pine stand on similar soils in the region with PRS probes in 2013 (F. Leishman, personal communication); however, potassium availability on the site as assessed through conventional soil extractions in Price (2014) showed much lower availability than in other studies from both natural (Huang and Schoenau 1996; Brais et al. 2000) and reclaimed sites (Brown and Naeth 2014). Reduced mycorrhizal colonization of these seedlings in this landscape reconstructed soil may also have contributed to the potassium limitation of seedlings, as K nutrition in plants is improved by mycorrhizal associations, and especially under limiting K conditions (Garcia and Zimmermann 2014; Hankin et al. 2015). The PRS probes are highly sensitive to biological, physical and chemical processes operating in the soil environment, and therefore they only provide a small snapshot of nutrient availability over a determined period of time and thus are not able to capture nutrient availability dynamics over a whole growing season.

Reclamation practices, such as the placement of coarse woody debris, can alleviate some nutrient deficiencies by providing a long-term source of nutrients (Hafner and Groffman 2005), create favorable microsite conditions (Haskell et al. 2012; Brown and Naeth 2014; Kwak et al. 2015), and control competitive colonizing vegetation cover (Hoffman 2016). The placement of CWD has recently become a more common practice in forest reclamation (Macdonald et al. 2015a). Although it is too early in our study to see any long-term benefits of CWD placement on soil nutrient availability, the shelter provided by CWD could have relieved some other environmental stressors, which would explain the positive relationship with seedling foliar nutrient concentrations in our study. While CWD can reduce soil water evaporation and provide cooler soil temperature, with reduced competition, excessive amounts of CWD could have also negative effects through the occupation of growing space, increased water interception, and nitrogen immobilization (Gilman and Grabosky 2004; Homyak et al. 2008; Miller and Seastedt 2009; Hoffman 2016).

Structural equation models offer the potential to untangle network relationships common in ecology and plant sciences. In our study, the SEMs provided a unique opportunity to explore a complex data set of biotic and abiotic variables for interactions among drivers of early jack pine seedling establishment on a reclaimed landscape over time. This modeling approach allowed for a more detailed exploration of the data, exposing relationships that linear modeling alone would not be able to detect, particularly highlighting indirect causal pathways. However, SEMs have to be built with the aid of underlying theoretical knowledge and expert opinion to inform the inclusion of certain pathways and their direction. Sampling strategy and number of replicates greatly limit the ability to apply the SEMs to ecological data. This was one main limitation to our study, as the data allowed us to investigate only one of the six species by soil type combination in this study (i.e. jack pine on coarse soil) due to the limited number of observations for the other possible combinations. Nonetheless, the SEMs allowed us to delve deeper into the complex relationships, linking biotic and abiotic variables within a developing ecosystem. We believe that SEM can aid in the accumulation of new knowledge of the ecosystem processes involved and can provide guidance for the collection of new information necessary to explore other relationships and improve future SEMs for ecosystem recovery.

By intensively monitoring this reclaimed landscape, we uncovered temporal shifts in the variables governing planted seedling performance as well as explored the role of heterogeneous soil properties in survival and early growth of planted boreal tree species. Our results warrant a more detailed focus on the role of nutrient uptake and availability in reclamation, especially on potassium and phosphorus availability. While soil water availability was not limiting at this stage of forest recovery, water supply might play a larger role in later stand development and landscape fluxes (Ketcheson et al. 2016; Devito et al. 2017). As these forests mature in combination with a changing climate, we can expect that some of these relationships become exacerbated, thus highlighting the importance of integrating the ecological processes at different spatial and temporal scales in heterogeneous landscapes.

Notes

Acknowledgements

We would like to acknowledge the funding support provided by the National Science and Engineering Research Council (NSERC), Syncrude Canada Ltd. and the Canadian Oil Sands Innovation Alliance (COSIA). We thank Carla Wytrykush and Jessica Piercey for their logistical support and all those who provided field support for this project over the years (Lynnette Allemand, Alexander Goeppel, Jessica Grenke, Ashley Hart, Elizabeth Hoffman, Robert Hetmanski, Caren Jones, Adam Kraft, Shaun Kulbaba, Angeline Letourneau, Mika Little-Devito, Michelle McCutcheon, Katharine Melnik, Sarah Thacker, and Alison Wilson). We also thank two anonymous reviewers for their comments on the manuscript.

Supplementary material

11056_2018_9649_MOESM1_ESM.docx (1.1 mb)
Supplementary material 1 (DOCX 1128 kb)

References

  1. Alberta Environment and Sustainable Resource Development (2013) Criteria and indicators framework for oil sands mine reclamation certification. Fort McMurray, AB. September 18, 2012Google Scholar
  2. Andersen CP, Bussler BH, Chaney WR et al (1989) Concurrent establishment of ground cover and hardwood trees on reclaimed mined land and unmined reference sites. For Ecol Manag 28:81–99CrossRefGoogle Scholar
  3. Baker TP, Jordan GJ, Baker SC (2016) Microclimatic edge effects in a recently harvested forest: do remnant forest patches create the same impact as large forest areas? For Ecol Manag 365:128–136.  https://doi.org/10.1016/J.FORECO.2016.01.022 CrossRefGoogle Scholar
  4. Barr AG, Black TA, Hogg EH et al (2007) Climatic controls on the carbon and water balances of a boreal aspen forest, 1994–2003. Glob Change Biol 13:561–576.  https://doi.org/10.1111/j.1365-2486.2006.01220.x CrossRefGoogle Scholar
  5. Bazzaz FA (1979) The physiological ecology of plant succession. Annu Rev Ecol Syst 10:351–371.  https://doi.org/10.1146/annurev.es.10.110179.002031 CrossRefGoogle Scholar
  6. Beattie M (1996) An ecosystem approach to fish and wildlife conservation. Ecol Appl 6:696–699.  https://doi.org/10.2307/2269463 CrossRefGoogle Scholar
  7. Beatty SW (1984) Influence of microtopography and canopy species on spatial patterns of forest understory plants. Ecology 65:1406–1419.  https://doi.org/10.2307/1939121 CrossRefGoogle Scholar
  8. Beckingham JH, Archibald JD (1996) Field guide to ecosites of Northern Alberta. Natural Resources Canada, Canadian Forest Service, Northwest Region, Northern Forestry Centre, Edmonton, Alberta. Special Report 5Google Scholar
  9. Bedford DR, Small EE (2008) Spatial patterns of ecohydrologic properties on a hillslope-alluvial fan transect, central New Mexico. CATENA 73:34–48.  https://doi.org/10.1016/J.CATENA.2007.08.005 CrossRefGoogle Scholar
  10. Bergeron Y, Leduc A, Harvey B, Gauthier S (2002) Natural fire regime: a guide for sustainable management of the Canadian boreal forest. Silva Fenn.  https://doi.org/10.14214/sf.553 CrossRefGoogle Scholar
  11. Bockstette SW, Pinno BD, Dyck MF, Landhäusser SM (2017) Root competition, not soil compaction, restricts access to soil resources for aspen on a reclaimed mine soil. Botany 95:685–695.  https://doi.org/10.1139/cjb-2016-0301 CrossRefGoogle Scholar
  12. Brais S, David P, Ouimet R (2000) Impacts of wild fire severity and salvage harvesting on the nutrient balance of jack pine and black spruce boreal stands. For Ecol Manag 137:231–243.  https://doi.org/10.1016/S0378-1127(99)00331-X CrossRefGoogle Scholar
  13. Brown RL, Naeth MA (2014) Woody debris amendment enhances reclamation after oil sands mining in Alberta, Canada. Restor Ecol 22:40–48.  https://doi.org/10.1111/rec.12029 CrossRefGoogle Scholar
  14. Carpenter FL, Nichols JD, Pratt RT, Young KC (2004) Methods of facilitating reforestation of tropical degraded land with the native timber tree, Terminalia amazonia. For Ecol Manag 202:281–291.  https://doi.org/10.1016/J.FORECO.2004.07.040 CrossRefGoogle Scholar
  15. Chazdon RL (2003) Tropical forest recovery: legacies of human impact and natural disturbances. Perspect Plant Ecol Evol Syst 6:51–71.  https://doi.org/10.1078/1433-8319-00042 CrossRefGoogle Scholar
  16. Comita LS, Uriarte M, Thompson J et al (2009) Abiotic and biotic drivers of seedling survival in a hurricane-impacted tropical forest. J Ecol 97:1346–1359.  https://doi.org/10.1111/j.1365-2745.2009.01551.x CrossRefGoogle Scholar
  17. Cornett MW, Reich PB, Puettmann KJ (1997) Canopy feedbacks and microtopography regulate conifer seedling distribution in two Minnesota conifer-deciduous forests. Écoscience 4:353–364.  https://doi.org/10.1080/11956860.1997.11682414 CrossRefGoogle Scholar
  18. Devito KJ, Hokanson KJ, Moore PA et al (2017) Landscape controls on long-term runoff in subhumid heterogeneous Boreal Plains catchments. Hydrol Process 31:2737–2751.  https://doi.org/10.1002/hyp.11213 CrossRefGoogle Scholar
  19. Drozdowski BL, Naeth AM, Wilkinson SR (2012) Evaluation of substrate and amendment materials for soil reclamation at a diamond mine in the Northwest Territories, Canada. Can J Soil Sci 92:77–88.  https://doi.org/10.4141/cjss2011-029 CrossRefGoogle Scholar
  20. Dumas JBA (1831) Procédés de l’Analyse Organique. Ann Chim Phys 247:198–213Google Scholar
  21. Eldridge DJ, Beecham G, Grace JB (2015) Do shrubs reduce the adverse effects of grazing on soil properties? Ecohydrology 8:1503–1513.  https://doi.org/10.1002/eco.1600 CrossRefGoogle Scholar
  22. Everham EM, Brokaw NVL (1996) Forest damage and recovery from catastrophic wind. Bot Rev 62:113–185.  https://doi.org/10.1007/BF02857920 CrossRefGoogle Scholar
  23. Fisher JB, Badgley G, Blyth E (2012) Global nutrient limitation in terrestrial vegetation. Glob Biogeochem Cycles.  https://doi.org/10.1029/2011GB004252 CrossRefGoogle Scholar
  24. Foster DR, Aber JD, Melillo JM et al (1997) Forest response to disturbance and anthropogenic stress. Bioscience 47:437–445.  https://doi.org/10.2307/1313059 CrossRefGoogle Scholar
  25. Foster DR, Motzkin G, Slater B (1998) Land-use history as long-term broad-scale disturbance: regional forest dynamics in central New England. Ecosystems 1:96–119.  https://doi.org/10.1007/s100219900008 CrossRefGoogle Scholar
  26. Franklin JF, Spies TA, Van Pelt R et al (2002) Disturbances and structural development of natural forest ecosystems with silvicultural implications, using Douglas-fir forests as an example. For Ecol Manag 155:399–423.  https://doi.org/10.1016/S0378-1127(01)00575-8 CrossRefGoogle Scholar
  27. Franklin O, Näsholm T, Högberg P, Högberg MN (2014) Forests trapped in nitrogen limitation—an ecological market perspective on ectomycorrhizal symbiosis. New Phytol 203:657–666.  https://doi.org/10.1111/nph.12840 CrossRefPubMedPubMedCentralGoogle Scholar
  28. Frouz J, Kalčík J, Velichová V (2011) Factors causing spatial heterogeneity in soil properties, plant cover, and soil fauna in a non-reclaimed post-mining site. Ecol Eng 37:1910–1913.  https://doi.org/10.1016/J.ECOLENG.2011.06.039 CrossRefGoogle Scholar
  29. Garcia K, Zimmermann SD (2014) The role of mycorrhizal associations in plant potassium nutrition. Front Plant Sci 5:337.  https://doi.org/10.3389/fpls.2014.00337 CrossRefPubMedPubMedCentralGoogle Scholar
  30. Gilliam FS (2007) The ecological significance of the herbaceous layer in temperate forest ecosystems. Bioscience 57:845–858.  https://doi.org/10.1641/B571007 CrossRefGoogle Scholar
  31. Gilman EF, Grabosky J (2004) Mulch and planting depth affect live oak (Quercus virginiana Mill.) establishment. J Arboric 30:311–317Google Scholar
  32. Gimenez O, Anker-Nilssen T, Grosbois V (2012) Exploring causal pathways in demographic parameter variation: path analysis of mark-recapture data. Methods Ecol Evol 3:427–432.  https://doi.org/10.1111/j.2041-210X.2011.00150.x CrossRefGoogle Scholar
  33. Government of Alberta (2017) Environmental protection and enhancement act. Revised Statutes of Alberta 2000 Chapter E-12. http://www.qp.alberta.ca/documents/acts/e12.pdf. Accessed 20 April 2018
  34. Grace JB, Bollen KA (2007) Representing general theoretical concepts in structural equation models: the role of composite variables. Environ Ecol Stat 15:191–213.  https://doi.org/10.1007/s10651-007-0047-7 CrossRefGoogle Scholar
  35. Grace JB, Anderson TM, Olff H, Scheiner SM (2010) On the specification of structural equation models for ecological systems. Ecol Monogr 80:67–87.  https://doi.org/10.1890/09-0464.1 CrossRefGoogle Scholar
  36. Grace JB, Harrison S, Damschen EI (2011) Local richness along gradients in the Siskiyou herb flora: R. H. Whittaker revisited. Ecology 92:108–120.  https://doi.org/10.1890/09-2137.1 CrossRefPubMedGoogle Scholar
  37. Grace JB, Schoolmaster DR, Guntenspergen GR et al (2012) Guidelines for a graph-theoretic implementation of structural equation modeling. Ecosphere 3:art73.  https://doi.org/10.1890/es12-00048.1 CrossRefGoogle Scholar
  38. Grossnickle SC (2012) Why seedlings survive: influence of plant attributes. New For 43:711–738.  https://doi.org/10.1007/s11056-012-9336-6 CrossRefGoogle Scholar
  39. Haase DL (2008) Understanding forest seedling quality: measurements and interpretation. Tree Plant Notes 52:24–30Google Scholar
  40. Haase DL, Rose R (1993) Soil moisture stress induces transplant shock in stored and unstored 2+ 0 Douglas-Fir seedlings of varying root volumes. For Sci 39:275–294Google Scholar
  41. Hafner SD, Groffman PM (2005) Soil nitrogen cycling under litter and coarse woody debris in a mixed forest in New York State. Soil Biol Biochem 37:2159–2162.  https://doi.org/10.1016/J.SOILBIO.2005.03.006 CrossRefGoogle Scholar
  42. Hankin SL, Karst J, Landhäusser SM (2015) Influence of tree species and salvaged soils on the recovery of ectomycorrhizal fungi in upland boreal forest restoration after surface mining. Botany 93:267–277.  https://doi.org/10.1139/cjb-2014-0132 CrossRefGoogle Scholar
  43. Haskell DE, Flaspohler DJ, Webster CR, Meyer MW (2012) Variation in soil temperature, moisture, and plant growth with the addition of downed woody material on lakeshore restoration sites. Restor Ecol 20:113–121.  https://doi.org/10.1111/j.1526-100X.2010.00730.x CrossRefGoogle Scholar
  44. Hébert F, Boucher J-F, Bernier PY, Lord D (2006) Growth response and water relations of 3-year-old planted black spruce and jack pine seedlings in site prepared lichen woodlands. For Ecol Manag 223:226–236.  https://doi.org/10.1016/j.foreco.2005.11.005 CrossRefGoogle Scholar
  45. Hember RA, Kurz WA, Coops NC (2017) Relationships between individual-tree mortality and water-balance variables indicate positive trends in water stress-induced tree mortality across North America. Glob Change Biol 23:1691–1710.  https://doi.org/10.1111/gcb.13428 CrossRefGoogle Scholar
  46. Hoffman EH (2016) Influence of environmental and site factors and biotic interactions on vegetation development following surface mine reclamation using coversoil salvaged from forest sites. University of AlbertaGoogle Scholar
  47. Hogg EH, Brandt JP, Michaelian M (2008) Impacts of a regional drought on the productivity, dieback, and biomass of western Canadian aspen forests. Can J For Res 38:1373–1384.  https://doi.org/10.1139/x08-001 CrossRefGoogle Scholar
  48. Homyak PM, Yanai RD, Burns DA et al (2008) Nitrogen immobilization by wood-chip application: protecting water quality in a northern hardwood forest. For Ecol Manag 255:2589–2601.  https://doi.org/10.1016/J.FORECO.2008.01.018 CrossRefGoogle Scholar
  49. Huang WZ, Schoenau JJ (1996) Microsite assessment of forest soil nitrogen, phosphorus, and potassium supply rates in-field using ion exchange membranes. Commun Soil Sci Plant Anal 27:2895–2908.  https://doi.org/10.1080/00103629609369748 CrossRefGoogle Scholar
  50. Ivetić V, Grossnickle S, Škorić M (2016) Forecasting the field performance of Austrian pine seedlings using morphological attributes. iForest 10:99.  https://doi.org/10.3832/ifor1722-009 CrossRefGoogle Scholar
  51. Jacobs DF, Oliet JA, Aronson J et al (2015) Restoring forests: what constitutes success in the twenty-first century? New For 46:601–614.  https://doi.org/10.1007/s11056-015-9513-5 CrossRefGoogle Scholar
  52. Ketcheson SJ, Price JS, Carey SK et al (2016) Constructing fen peatlands in post-mining oil sands landscapes: challenges and opportunities from a hydrological perspective. Earth Sci Rev 161:130–139.  https://doi.org/10.1016/J.EARSCIREV.2016.08.007 CrossRefGoogle Scholar
  53. Kirdyanov A, Hughes M, Vaganov E et al (2003) The importance of early summer temperature and date of snow melt for tree growth in the Siberian Subarctic. Trees 17:61–69.  https://doi.org/10.1007/s00468-002-0209-z CrossRefGoogle Scholar
  54. Knoop WT, Walker BH (1985) Interactions of woody and herbaceous vegetation in a southern African savanna. J Ecol 73:235.  https://doi.org/10.2307/2259780 CrossRefGoogle Scholar
  55. Kuuluvainen T (2002) Natural variability of forests as a reference for restoring and managing biological diversity in boreal Fennoscandia. Silva Fenn.  https://doi.org/10.14214/sf.552 CrossRefGoogle Scholar
  56. Kwak J-H, Chang SX, Naeth MA, Schaaf W (2015) Coarse woody debris increases microbial community functional diversity but not enzyme activities in reclaimed oil sands soils. PLoS ONE 10:e0143857.  https://doi.org/10.1371/journal.pone.0143857 CrossRefPubMedPubMedCentralGoogle Scholar
  57. Laclau J-P, da Silva EA, Rodrigues Lambais G et al (2013) Dynamics of soil exploration by fine roots down to a depth of 10 m throughout the entire rotation in Eucalyptus grandis plantations. Front Plant Sci 4:243.  https://doi.org/10.3389/fpls.2013.00243 CrossRefPubMedPubMedCentralGoogle Scholar
  58. Landhäusser SM, Lieffers VJ (1998) Growth of Populus tremuloides in association with Calamagrostis canadensis. Can J For Res 28:396–401CrossRefGoogle Scholar
  59. Landhäusser SM, Pinno BD, Lieffers VJ, Chow PS (2012a) Partitioning of carbon allocation to reserves or growth determines future performance of aspen seedlings. For Ecol Manag 275:43–51.  https://doi.org/10.1016/J.FORECO.2012.03.010 CrossRefGoogle Scholar
  60. Landhäusser SM, Rodriguez-Alvarez J, Marenholtz EH, Lieffers VJ (2012b) Effect of stock type characteristics and time of planting on field performance of aspen (Populus tremuloides Michx.) seedlings on boreal reclamation sites. New For 43:679–693.  https://doi.org/10.1007/s11056-012-9346-4 CrossRefGoogle Scholar
  61. Larsen JA (1980) The boreal ecosystem, 1st edn. Academic Press, New YorkGoogle Scholar
  62. Lee S-Y (2007) Structural equation modeling: a Bayesian approach, 1st edn. Wiley, ChichesterCrossRefGoogle Scholar
  63. Lindenmayer DB, Knight E, McBurney L et al (2010) Small mammals and retention islands: an experimental study of animal response to alternative logging practices. For Ecol Manag 260:2070–2078.  https://doi.org/10.1016/J.FORECO.2010.08.047 CrossRefGoogle Scholar
  64. Lunn DJ, Thomas A, Best N, Spiegelhalter D (2000) WinBUGS - A Bayesian modelling framework: concepts, structure, and extensibility. Stat Comput 10:325–337.  https://doi.org/10.1023/A:1008929526011 CrossRefGoogle Scholar
  65. Macdonald SE, Landhäusser SM, Skousen J et al (2015a) Forest restoration following surface mining disturbance: challenges and solutions. New For 46:703–732.  https://doi.org/10.1007/s11056-015-9506-4 CrossRefGoogle Scholar
  66. Macdonald SE, Snively AEK, Fair JM, Landhäusser SM (2015b) Early trajectories of forest understory development on reclamation sites: influence of forest floor placement and a cover crop. Restor Ecol 23:698–706.  https://doi.org/10.1111/rec.12217 CrossRefGoogle Scholar
  67. Mackenzie DD, Naeth MA (2009) The role of the forest soil propagule bank in assisted natural recovery after oil sands mining. Restor Ecol 18:418–427.  https://doi.org/10.1111/j.1526-100X.2008.00500.x CrossRefGoogle Scholar
  68. Malaeb ZA, Summers JK, Pugesek BH (2000) Using structural equation modeling to investigate relationships among ecological variables. Environ Ecol Stat 7:93–111.  https://doi.org/10.1023/A:1009662930292 CrossRefGoogle Scholar
  69. Manninen OH, Stark S, Kytöviita M-M et al (2009) Understorey plant and soil responses to disturbance and increased nitrogen in boreal forests. J Veg Sci 20:311–322.  https://doi.org/10.1111/j.1654-1103.2009.05529.x CrossRefGoogle Scholar
  70. Marshall PL, Davis G, LeMay VM (2000) Using line intersect sampling for coarse woody debris. In: Forest Research Technical Report, TR-003. Research Section, Vancouver Forest Region, BCMOFGoogle Scholar
  71. McCune B, Keon D (2002) Equations for potential annual direct incident radiation and heat load. J Veg Sci 13:603–606.  https://doi.org/10.1111/j.1654-1103.2002.tb02087.x CrossRefGoogle Scholar
  72. McGuire AD, Melillo JM, Joyce LA et al (1992) Interactions between carbon and nitrogen dynamics in estimating net primary productivity for potential vegetation in North America. Glob Biogeochem Cycles 6:101–124.  https://doi.org/10.1029/92GB00219 CrossRefGoogle Scholar
  73. McIntire EJB, Fajardo A (2014) Facilitation as a ubiquitous driver of biodiversity. New Phytol 201:403–416.  https://doi.org/10.1111/nph.12478 CrossRefPubMedGoogle Scholar
  74. McKee KL, Rooth JE, Feller IC (2007) Mangrove recruitment after forest disturbance is facilitated by herbaceous species in the Caribbean. Ecol Appl 17:1678–1693.  https://doi.org/10.1890/06-1614.1 CrossRefPubMedGoogle Scholar
  75. Melnik K, Landhäusser SM, Devito K (2017) Role of microtopography in the expression of soil propagule banks on reclamation sites. Restor Ecol.  https://doi.org/10.1111/rec.12587 CrossRefGoogle Scholar
  76. Menge DNL, Levin SA, Hedin LO (2008) Evolutionary tradeoffs can select against nitrogen fixation and thereby maintain nitrogen limitation. Proc Natl Acad Sci USA 105:1573–1578.  https://doi.org/10.2307/25451324 CrossRefPubMedGoogle Scholar
  77. Messier C, Coll L, Poitras-Larivière A et al (2009) Resource and non-resource root competition effects of grasses on early-versus late-successional trees. J Ecol 97:548–554.  https://doi.org/10.1111/j.1365-2745.2009.01500.x CrossRefGoogle Scholar
  78. Midoko-Iponga D, Krug CB, Milton SJ (2005) Competition and herbivory influence growth and survival of shrubs on old fields: Implications for restoration of renosterveld shrubland. J Veg Sci 16:685–692.  https://doi.org/10.1111/j.1654-1103.2005.tb02411.x CrossRefGoogle Scholar
  79. Miller EM, Seastedt TR (2009) Impacts of woodchip amendments and soil nutrient availability on understory vegetation establishment following thinning of a ponderosa pine forest. For Ecol Manag 258:263–272.  https://doi.org/10.1016/J.FORECO.2009.04.011 CrossRefGoogle Scholar
  80. Näsholm T, Högberg P, Franklin O et al (2013) Are ectomycorrhizal fungi alleviating or aggravating nitrogen limitation of tree growth in boreal forests? New Phytol 198:214–221.  https://doi.org/10.1111/nph.12139 CrossRefPubMedGoogle Scholar
  81. Natural Resources Canada (2008) Canada’s National Forest Inventory ground sampling guidelines: specifications for ongoing measurement. Canadian Forest Service, Pacific Forestry Centre, VictoriaGoogle Scholar
  82. Økland RH, Rydgren K, Økland T (2008) Species richness in boreal swamp forests of SE Norway: the role of surface microtopography. J Veg Sci 19:67–74.  https://doi.org/10.3170/2007-8-18330 CrossRefGoogle Scholar
  83. Omeja PA, Chapman CA, Obua J et al (2011) Intensive tree planting facilitates tropical forest biodiversity and biomass accumulation in Kibale National Park, Uganda. For Ecol Manag 261:703–709.  https://doi.org/10.1016/J.FORECO.2010.11.029 CrossRefGoogle Scholar
  84. Ostertag R, DiManno NM (2016) Detecting terrestrial nutrient limitation: a global meta-analysis of foliar nutrient concentrations after fertilization. Front Earth Sci 4:23.  https://doi.org/10.3389/feart.2016.00023 CrossRefGoogle Scholar
  85. Parrotta JA, Turnbull JW, Jones N (1997) Catalyzing native forest regeneration on degraded tropical lands. For Ecol Manag 99:1–7.  https://doi.org/10.1016/S0378-1127(97)00190-4 CrossRefGoogle Scholar
  86. Pinno BD, Landhäusser SM, MacKenzie MD et al (2012) Trembling aspen seedling establishment, growth and response to fertilization on contrasting soils used in oil sands reclamation. Can J Soil Sci 92:143–151CrossRefGoogle Scholar
  87. Pinto JR, Marshall JD, Dumroese RK et al (2011) Establishment and growth of container seedlings for reforestation: a function of stocktype and edaphic conditions. For Ecol Manage 261:1876–1884.  https://doi.org/10.1016/J.FORECO.2011.02.010 CrossRefGoogle Scholar
  88. Pollack HN (2003) Uncertain science—uncertain world. Cambridge University Press, CambridgeCrossRefGoogle Scholar
  89. Pollard J, McKenna GT, Fair J et al (2012) Design aspects of two fen wetlands constructed for reclamation research in the Athabasca oil sands. In: Fourie AB, Tibbett M (eds) Mine closure 2012. Australian Centre for Geomechanics, Brisbane, pp 815–829Google Scholar
  90. Price R (2014) 2013 Sandhill Fen soil sampling program. NorthWind Land Resources Inc., EdmontonGoogle Scholar
  91. Pypker TG, Unsworth MH, Bond BJ (2006) The role of epiphytes in rainfall interception by forests in the Pacific Northwest. II. Field measurements at the branch and canopy scale. Can J For Res 36:819–832.  https://doi.org/10.1139/x05-286 CrossRefGoogle Scholar
  92. R Core Team (2015) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R-project.org/
  93. Reich PB, Tjoelker MG, Walters MB et al (1998) Close association of RGR, leaf and root morphology, seed mass and shade tolerance in seedlings of nine boreal tree species grown in high and low light. Funct Ecol 12:327–338.  https://doi.org/10.1046/j.1365-2435.1998.00208.x CrossRefGoogle Scholar
  94. Salifu KF, Jacobs DF, Birge ZKD (2009) Nursery Nitrogen loading improves field performance of bareroot oak seedlings planted on abandoned mine lands. Restor Ecol 17:339–349.  https://doi.org/10.1111/j.1526-100X.2008.00373.x CrossRefGoogle Scholar
  95. Savage M, Mast JN (2005) How resilient are southwestern ponderosa pine forests after crown fires? Can J For Res 35:967–977.  https://doi.org/10.1139/x05-028 CrossRefGoogle Scholar
  96. Schott KM, Karst J, Landhäusser SM (2014) The role of microsite conditions in restoring trembling aspen (Populus tremuloides Michx) from seed. Restor Ecol 22:292–295.  https://doi.org/10.1111/rec.12082 CrossRefGoogle Scholar
  97. Schott KM, Snively AEK, Landhäusser SM, Pinno BD (2016) Nutrient loaded seedlings reduce the need for field fertilization and vegetation management on boreal forest reclamation sites. New For 47:393–410.  https://doi.org/10.1007/s11056-015-9522-4 CrossRefGoogle Scholar
  98. Seo J-W, Eckstein D, Jalkanen R, Schmitt U (2011) Climatic control of intra- and inter-annual wood-formation dynamics of Scots pine in northern Finland. Environ Exp Bot 72:422–431.  https://doi.org/10.1016/J.ENVEXPBOT.2011.01.003 CrossRefGoogle Scholar
  99. Shugart HH, Leemans R, Bonan GB (1992) Introduction. In: Shugart HH, Leemans R, Bonan GB (eds) A systems analysis of the global boreal forest, 1st edn. Cambridge University Press, Cambridge, pp 1–8CrossRefGoogle Scholar
  100. Sigurdsson BD, Medhurst JL, Wallin G et al (2013) Growth of mature boreal Norway spruce was not affected by elevated [CO2] and/or air temperature unless nutrient availability was improved. Tree Physiol 33:1192–1205.  https://doi.org/10.1093/treephys/tpt043 CrossRefPubMedGoogle Scholar
  101. Singh JS, Milchunas DG, Lauenroth WK (1998) Soil water dynamics and vegetation patterns in a semiarid grassland. Plant Ecol 134:77–89.  https://doi.org/10.1023/A:1009769620488 CrossRefGoogle Scholar
  102. Soil Classification Working Group (1998) The Canadian system of soil classification, 3rd edn. Agriculture and Agri-Food Canada Publication 1646Google Scholar
  103. Ste-Marie C, Houle D (2006) Forest floor gross and net nitrogen mineralization in three forest types in Quebec, Canada. Soil Biol Biochem 38:2135–2143.  https://doi.org/10.1016/j.soilbio.2006.01.017 CrossRefGoogle Scholar
  104. Stone EL (1953) Magnesium deficiency of some northeastern pines. Soil Sci Soc Am J 17:297–300.  https://doi.org/10.2136/sssaj1953.03615995001700030029x CrossRefGoogle Scholar
  105. Stone EL, Kszystyniak R (1977) Conservation of potassium in the Pinus resinosa ecosystem. Science 198:192–194.  https://doi.org/10.2307/1744944 CrossRefPubMedGoogle Scholar
  106. Stone EL, Leaf AL (1967) Potassium deficiency and response in young conifer forests in eastern North America. In: Forest Fertilization, Proceedings of the 5th Colloquium of the International Potash Institute. Jyväskilä, Finland, pp 219–229Google Scholar
  107. Sturtz S, Ligges U, Gelman A (2005) R2WinBUGS: a package for running WinBUGS from R. J Stat Softw 12:1–16.  https://doi.org/10.18637/jss.v012.i03 CrossRefGoogle Scholar
  108. Tamm CO (1991) Nitrogen in terrestrial ecosystems, 1st edn. Springer, BerlinCrossRefGoogle Scholar
  109. Thiffault N, Jobidon R, Munson AD (2003) Performance and physiology of large containerized and bare-root spruce seedlings in relation to scarification and competition in Québec (Canada). Ann For Sci 60:645–655.  https://doi.org/10.1051/forest:2003057 CrossRefGoogle Scholar
  110. Turner MG (2010) Disturbance and landscape dynamics in a changing world. Ecology 91:2833–2849.  https://doi.org/10.1890/10-0097.1 CrossRefGoogle Scholar
  111. Turner MG, Romme WH, Gardner RH et al (1993) A revised concept of landscape equilibrium: disturbance and stability on scaled landscapes. Landsc Ecol 8:213–227.  https://doi.org/10.1007/BF00125352 CrossRefGoogle Scholar
  112. Van Wagner CE (1982) Practical aspects of the line intersect method. Petawawa National Forestry Institute, Canadian Forest Service, Chalk RiverGoogle Scholar
  113. Vieira ICG, Uhl C, Nepstad D (1994) The role of the shrub Cordia multispicata Cham. as a ‘succession facilitator’ in an abandoned pasture, Paragominas, Amazônia. Vegetatio 115:91–99.  https://doi.org/10.1007/bf00044863 CrossRefGoogle Scholar
  114. Weber B, Berkemeier T, Ruckteschler N et al (2016) Development and calibration of a novel sensor to quantify the water content of surface soils and biological soil crusts. Methods Ecol Evol 7:14–22.  https://doi.org/10.1111/2041-210X.12459 CrossRefGoogle Scholar
  115. Wytrykush C, Vitt DH, McKenna G, Vassov R (2012) Designing landscapes to support peatland development on soft tailings deposits: Syncrude Canada Ltd’s Sandhill Fen Research Watershed initiative. In: Vitt D, Bhatti J (eds) Restoration and reclamation of Boreal ecosystems—attaining sustainable development. Cambridge University Press, Cambridge, pp 161–178CrossRefGoogle Scholar
  116. Zettl J, Lee Barbour S, Huang M et al (2011) Influence of textural layering on field capacity of coarse soils. Can J Soil Sci 91:133–147.  https://doi.org/10.4141/cjss09117 CrossRefGoogle Scholar
  117. Zhao D, Borders B, Wang M, Kane M (2007) Modeling mortality of second-rotation loblolly pine plantations in the Piedmont/Upper Coastal Plain and Lower Coastal Plain of the southern United States. For Ecol Manag 252:132–143.  https://doi.org/10.1016/J.FORECO.2007.06.030 CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V., part of Springer Nature 2018

Authors and Affiliations

  1. 1.Department of Renewable Resources, 4-42 Earth Sciences BuildingUniversity of AlbertaEdmontonCanada
  2. 2.Natural Resources Canada, Canadian Forest ServiceNorthern Forestry CentreEdmontonCanada

Personalised recommendations