A
Direct and indirect effects of terrain, snow, and shrubs on the structure of an alpine herbaceous community in the Canadian Rocky Mountains.T6G 2H1
¹Department of Renewable Resources, University of Alberta, Edmonton, Alberta, Canada
Zacharaih J. Madsen¹ (https://orcid.org/0009-0009-7621-7568) and Scott E. Nielsen1 (https://orcid.org/0000-0002-9754-0630)
Corresponding author information
Scott E. Nielsen (scottn@ualberta.ca)
University of Alberta Campus
751 General Services Building
Edmonton, Alberta, Canada
Keywords
Alpine Terrain
Snowmelt Dynamics
Shrub–Herb Interactions
Diversity
Structural Equation Modelling
A
Acknowledgement
A
We acknowledge the local Indigenous communities and Treaty 6 First Nations for supporting research on their traditional lands near Cardinal Divide, Whitehorse Provincial Park, Alberta. We thank Whitehorse Creek Provincial Park for permission to conduct research following the research permit #23-395. We thank Alexa MacDonald and Natalya Klutz for their assistance with data collection.Abstract
Alpine ecosystems have extreme environments that limit plant growth, favoring stress-tolerant species. Climate change has resulted in the expansion of shrubs at the expense of poor competitors, including herbaceous species. Here, we explore the pathways linking local terrain features, snowmelt dynamics, and shrub cover on herbaceous plant cover and species richness at Cardinal Divide in the Canadian Rocky Mountain Front Ranges using a series of structural equation models (SEMs) that isolate direct from indirect effects among factors. We measured the biotic and abiotic conditions within 338 quadrats (0.25-m2) across a gradient of environments typical of the open and expanding krummholz alpine sites on the divide. Herbaceous cover and richness were negatively related to shrub cover but positively associated with late-snow persistence and terrain features that supported higher winter snow accumulation and greater soil depth. Dwarf deciduous shrubs were minimally affected by differences in the timing of snowmelt. Tall evergreen shrubs were uncommon but abundant where present, substantially reducing herbaceous cover and richness. The indirect effects of shrub cover tempered the positive association of snow on herbaceous cover and richness. Concave terrain surfaces and cooler-facing slopes increased herbaceous composition, while soil depth supported taller shrubs and herbaceous communities. This study underscores the complex relationships, both direct and indirect, between the abiotic and biotic environments of the alpine plant community.
Introduction
Arctic and alpine ecosystems are rapidly changing, including alterations in plant composition and diversity (Post et al. 2009, Myers-Smith and Hik 2018). Recent expansions of shrubs are associated with decades-long increases in temperatures across the Arctic and alpine tundra (Myers-Smith et al. 2019, Zong et al. 2022). Shrubs may benefit from recent warming due to extended growing seasons, particularly with warmer temperatures early in the growing season (Bjorkman et al. 2020). The expansion of shrubs has resulted in substantial shifts in the remaining plant community (Scharnagl et al. 2019). This shift is facilitated mainly by earlier snowmelt, a key phenological driver of seasonal environments (Hallinger et al. 2010, Zong et al. 2022). The implications of earlier snowmelt on alpine plant communities are essential for understanding past and future changes, yet unclear at local scales, since snowmelt patterns are themselves influenced by local terrain features that affect snow accumulation and solar exposure (Abudu et al. 2016).
Temporal patterns in snow cover and snowmelt in the alpine often vary less with altitude and more with local variations in slope-aspect and surface roughness, which influence snow accumulation or loss during winter and spring (Anderton, White, and Alvera 2002, DeBeer and Pomeroy 2010). At larger scales, south-facing ridges in the northern hemisphere experience faster snowmelt, especially on steeper slopes (Abudu et al. 2016). On northern slopes, snowbanks tend to persist longer, particularly in depressions that accumulate more snow (DeBeer and Pomeroy 2010). These terrain characteristics create consistent patterns of late snow disappearance that are decoupled from the typical seasonal timing of snowmelt, thus structuring local patterns in plant community composition (Erickson et al. 2005). Snow cover protects plants from exposed freezing temperatures in winter, regulating the length of the growing season, and serving as a primary source of soil moisture in summer (Körner 2003a, Inouye 2008, DeBeer and Pomeroy 2017).
Shrubs in the alpine grow best in areas with moderate to deep snow depths, where winter protection is provided and available summer moisture is extended (Kudo and Ito 1992). However, the relationship between snow and shrubs varies between different functional groups (Elmendorf et al. 2012). For instance, dwarf shrubs are more sensitive to growing-season length, having to wait until complete snow melt for phenological cues to start growth (Bell and Bliss 1977), whereas taller shrubs rely on more moisture provided by deep snow and slower melt rates (Hallinger et al. 2010, Conner et al. 2016). Shrub leaf habits also vary with solar exposure and snow bed characteristics (Saarinen et al. 2016). Light limitation is tolerated by evergreen shrubs, which can begin photosynthesis while still under snow (Starr and Oberbauer 2003, Christiansen et al. 2018, Dobbert et al. 2021). However, the ability of deciduous shrubs to tolerate light limitations depends more on their height (Dobbert et al. 2021). In alpine zones, shrub size is limited, in part due to a short growing season, but also temperature and moisture stressors (Boscutti et al. 2018). As such, instead of growing upwards, many alpine shrubs branch outward.
The relationship between shrubs and snow has influenced how shrubs respond to early snowmelt, thus changes in early snowmelt in recent decades have altered local patterns in shrub abundance (Martin et al. 2017, Scharnagl et al. 2019, Dobbert et al. 2021). Some shrub species have expanded their range, particularly into areas that were formerly snow-covered year-round or had snow cover until late in the growing season (Tape et al. 2012, Gentili et al. 2020). However, other shrub species are responding poorly to warming, with drought leading to early senescence, and in some cases, increased mortality (Gamm et al. 2018, Dobbert et al. 2021). Many shrubs are susceptible to frost damage in early spring (Wheeler et al. 2014). Some evergreen shrubs have expanded into the deepest snow beds due to warming temperatures and earlier snow loss, but are performing poorly reproductively (Christiansen et al. 2018). Overall, there has been a trend in shrub expansion within arctic-alpine communities, which, even if limited to a few species, can have substantial effects on the overall plant community (McManus et al. 2012, Myers-Smith and Hik 2018, Zong et al. 2022), including reducing herbaceous species (Ballantyne and Pickering 2015, Sholto-Douglas et al. 2017). However, the effect of shrub expansion on herbaceous composition in alpine areas remains uncertain, as some shrub species may enhance (facilitate) herbaceous diversity, while others might reduce it through competition (Ballantyne & Pickering 2015, Sholto-Douglas et al. 2017).
Not only are herbaceous species sensitive to competition with shrubs, but also to accelerated snowmelt. Early snowmelt, particularly in spring when frost events still occur, can cause frost damage to species that are active during that period (Inouye 2008, CaraDonna and Bain 2016, Pardee et al. 2018). Consequently, some species have shifted to later-melting snow beds to avoid frost and maintain adequate moisture (Verrall and Pickering 2020, Verrall et al. 2023). However, late snow beds generally have lower species richness and plant cover, demonstrating the presence of trade-offs (Alatalo et al. 2014, Liu et al. 2022, Cheng et al. 2023). Given recent changes in snow dynamics and shrub expansion, herbaceous species may be the most sensitive to change (Sholto-Douglas et al. 2017, Scharnagl et al. 2019). More work is needed to understand the interplay between snow and shrubs on herbaceous species, particularly the potential synergistic interactions (Hallinger et al. 2010, Moura et al. 2016, Boscutti et al. 2018). In this study, we examine how variation in terrain and snow cover influences the cover of different growth forms in shrubs, and, in turn, how this affects herbaceous plant cover and richness. We examine this at an alpine site along the eastern slopes of the Canadian Rocky Mountains using pathway diagrams with Structural Equation Models (SEMs) that link and quantify the direct and indirect relationships among terrain, snow, shrubs, and herbaceous plants.
Methods
Study Area
We conducted the study during the summers of 2022 and 2023 at Cardinal Divide (Latitude: 52.90797° N, Longitude: -117.2085° W), located 22 km south of Cadomin, Alberta, Canada, in the Nikannassin Range of the Canadian Rocky Mountains (Fig. 1). Elevations of study plots ranged from 1941 to 2189 meters. The site is adjacent to the Boreal Foothills ecoregion, where boreal and sub-alpine tree species are expanding upslope due to climate warming (Harsch et al. 2009, Myers-Smith et al. 2011).
Fig. 1
Aerial map (top) of Cardinal Divide, Alberta, illustrating the distribution of the 338 sampled 0.25 m2 (50 × 50 cm) quadrats (yellow points) outside of forested areas across three distinct regions of the divide that vary in slope, aspect, terrain curvature, and tree cover. (a) represents a southwest-facing slope that ascends to the upper ridge, with rocky, warm slopes that melt early with patches of krummholz forests. (b) represents the connecting northwest/southwest-facing ridge (divide) with distinct swales on the northwest side that accumulate snow and melt late, with (c) having the deepest snowbeds and latest melt dates. (d) represents the basin below the tallest peaks on the divide and generally slopes northeastward.
We sampled the open and semi-open/krummholz habitats across three distinct areas of the divide, each with unique slope and aspect characteristics. The "Upper Divide" (Fig. 1a) is a large southwest-facing dry, rocky slope near the top of the divide. Patches of krummholz tree cover occur up to its summit, intermixed with open alpine. The "Lower Divide," adjacent to the Upper Divide, features a distinct ridged slope (divide) having a northeast/southwest orientation (Fig. 1b). On the south side of this ridge, the open alpine of the ridgetop transitions rapidly to krummholz and then to dense forests. In contrast, the north side has a lower tree line and a rocky slope with a nearly systematic spacing of narrow dry ravine swales that are perpendicular to the ridge, reaching down to the tree line. These swales are characterized as having krummholz along the west side (east slope) of the swale, but are otherwise open, and capture substantial amounts of snow in winter and melt later in the summer. The third area of the divide is a northeast-facing open alpine basin located below Tripoli Mountain. This basin is distinguished by its deeper soils and lower concentration of bare rock, with swales and ravines running down its slope that collect winter snow (Fig. 1c). The combination of these areas spans the environmental gradients of the alpine community.
Study Design
Sampling Methods
We used random (2022) and stratified random (2023) sampling of non-forested alpine sites across the three parts of the divide. In 2022, we sampled 205 random sites, and in 2023, we used an initial remote sensing-based model of snow persistence to target the full range of snow persistence, sampling 133 additional sites (Fig. 1). This included 33 random plots around a large swale that retained the latest snow on the Cardinal Divide (Fig. 1c). At each selected site, we used a 0.25 m² quadrat (0.5 m × 0.5 m) to quantify alpine plant cover for all vascular plants during peak growing conditions between mid-July and early August. Plant cover was recorded within the quadrat up to a height of 50 cm, at 1% intervals for 0–10% cover and at 5% intervals for 10–100% cover. We then grouped plant species into herbaceous and shrub species, further defining shrub species by growth form based on leaf size and leaf habit. Locations of each quadrat were recorded in NAD83 UTM coordinates using a GNSS GPS (SX Blue II or Geode), achieving horizontal accuracy of less than 0.6 m.
Snow Persistence
We quantified snow persistence in each quadrat using a series of 0.6 m pan-sharpened SkySat satellite images obtained from Planet Labs, which were tasked with acquiring weekly images under clear-sky conditions (note that some weeks had no clear-sky conditions). We analyzed imagery from three ‘spring-to-summer’ periods: June 8 - July 18, 2022 (late-year snow melt); March 20 - June 12, 2023 (early-year snow melt); and April 1 - July 17, 2024 (late-year snow melt). Images were classified into binary rasters classified as snow (1) and no snow (0) cover for each acquisition date using the terra and sf packages in R (Hijmans et al. 2023, Pebesma et al. 2023). In total, we classified 30 image dates to develop a final gradient of snow disappearance for the divide (see Appendix A for a more detailed workflow and Fig. 5 for an illustration of the steps). Although we do not concentrate on specific dates of snow loss in any one year, our model emphasizes areas that are more consistent in holding late snow versus not, which tends to be consistent across years in alpine ecosystems (Erickson et al. 2005). Late snow beds in swales can persist until late July in some years (2022 & 2024), although in other years with poor snowfall and warm spring and early summer weather, snow may last until late June (2023). Because dense tree cover obscures snow cover detection with remote sensing, we limited our analyses to sites (plots) with less than 30% tree canopy cover (n = 338), as canopy cover below 30% did not affect snow persistence values.
Fig. 5
Illustration of the workflow used for classifying tasked satellite images into snow presence and persistence. First, images are classified into vector polygons to fix misclassifications, and then re-rasterized into snow binary images. The final step creates a single raster representing the proportion of snow persistence across all possible dates, thereby providing an index of early-to-late snowmelt during the spring-through-summer period. Note that 90% (0.9) here represents snow cover through late July-early August across 3 years (2022–2024).
The first step involved georeferencing satellite imagery. We used a series of pre-collected ground control points (GCPs) with a GNSS GPS unit, achieving an average accuracy of 0.6 m. In instances where GCPs were unavailable, we used reference images with distinct features and a cell size of at least 1 m2. All georeferencing was performed in QGIS. We accessed the georeferencing tool, loaded the raster image, and applied the GCPs. When GCPs were absent, we created them manually by zooming in to identify pixel patterns with high accuracy. We set the transformation type to 'Thin Plate Spline', the resampling method to 'Linear', and chose 'LZW' for compression.
Next, we prepared our images for classification by updating a CSV file with the classification values for each image. Each image was duplicated, and the imagery was adjusted to single-band pseudo-color, selecting the red band (usually the first band). We set the classification interpolation to Discrete, the mode to Equal Interval, and the number of classes to 2, applying the Spectral color ramp. We adjusted the opacity to 50% and adjusted the minimum and maximum values to identify snow areas and to set appropriate thresholds. Large, misclassified areas were left to be handled by manually digitizing polygons. The CSV file was updated with location, date, and classification codes, and the map was saved for further adjustments. We then broke larger images into tiles for processing efficiency. We created multi-polygons with unique codes for each tile. We developed an R script to process the CSV classification, address misclassifications, and generate output files in shapefile format.
We manually corrected known misclassified areas by drawing new polygons in QGIS, using AutoHotkey to create a script. A new map was then developed containing the tiled polygons and the original raster image. We categorized the binary field into snow (1) and no snow (0) for each image. Misclassified polygons were organized in a separate folder, named according to area and date, and manually edited to rectify under- and over-classified areas. We ran a script to merge the classified vectors with the fixed (misclassified) vectors into a final binary raster for each date/tile, again reflecting snow (1) or no snow (0). We assessed the final accuracy of our binary snow classification model using GIS to sample 4000 random points across the field study and another script to randomly assign 100–200 points across all images. We engaged a second, independent reviewer to compare the imagery and manually classify the binary images, assessing whether snow was evident at the random locations in the original imagery. We sought an accuracy of > 95% between the original and classified fields and calculated the final snow classification model's accuracy at 96%.
The final step involved creating the snow persistence model by organizing snow binary images to reflect snow disappearance based on terrain features rather than on date. We compiled a CSV file listing the images in an order that mirrored their appearance based on terrain features. We then ran the final R script, which merged the binary images into a single composite image reflecting snow persistence. Then we divided that number by our sample size plus one to indicate that snow was never 100% persistent in the summer. This final image illustrates the spatial patterns in snow persistence over time.
Appendix B:
Measures of Local Terrain Features
Cardinal Divide has distinct directional slopes, tree lines, and swales that affect snow accumulation and persistence (Fig. 2). We quantified these features for each sample plot using a LiDAR-derived 1-m Digital Elevation Model (DEM) from Altalis (Government of Alberta) and subsequently using this model to derive surface curvature, slope, and aspect (Fig. 2a). Terrain curvature, modelled in ESRI ArcGIS Pro using the Curvature script in the Spatial Analyst Tool which expressed the relationship between concave (positive) and convex (negative) slopes. As such, the values of curvature are unitless and represent standard deviation from the mean of the elevation/slope metric. Comparisons between curvature and snow are shown in Fig. 2a and 2b, illustrating the strong relationship between them. Slope and aspect for each quadrat were calculated with a handheld compass at the center of each quadrat and used to calculate a Solar Severity Index (SSI) (Nielsen and Haney 1996). SSI shows the relationship between slope steepness and a ‘folded’ southwest aspect to emphasize the effects of cumulative daytime heating. The index ranges from − 1 (north-east 45˚ slope) to 1 (south-west 45˚ slope), with 0 representing either flat slopes (0˚) or north-west/south-east aspects (Fig. 2c).
Fig. 2
Spatial patterns in terrain and tree canopy at Cardinal Divide, Alberta, Canada. Terrain curvature (a) portrays surface unevenness, with lighter shades indicating swales and darker shades protruding surfaces. Solar severity (b) quantifies slope steepness and southwest exposure, ranging from steep southwest-facing slopes (yellow) to northeast-facing slopes (blue). Snow persistence (c) shows early snow melt (lighter shades) versus later melt (darker shades) based on the reclassified SkySat remote sensing imagery. It correlates well with lighter areas in map (a), highlighting swales that capture and retain snow. Shrub and tree presence (d) distinguishes among tall shrubs/short trees (light green), trees (dark green), and areas lacking visible vegetation, aligning closely with higher snow persistence.
Quadrat-Level Tree Canopy and Soil Depth Measures
Tree canopy cover was measured over each quadrat at waist height using a spherical densitometer (Lemon 1956). However, we excluded sites within forests (> 30% canopy cover) in our analysis, as our interest was in the alpine plant community and because estimating snow persistence from remote sensing is limited in areas of higher tree cover. Thus, only plots with densiometer canopy readings of 30% or less were retained. These types of sites were characterized as being krummholz-like patches of trees. Finally, we estimated the maximum soil depth (cm) within each quadrat by probing the ground surface with an 18-gauge metal prong and recording the deepest measure. Soil depth affects, among other things, the presence and cover of taller shrub species and thus is included here.
Statistical Analyses
We used Structural Equation Models (SEMs) to identify and quantify the causal relationships among our set of interrelated variables (Laughlin 2014, Moura et al. 2016). SEMs distinguish between direct and indirect effects, thereby creating a network of relationships among variables that are assumed to be correlated. Here, we used a Piecewise Structural Equation Model (Lefcheck 2016). We express our causal network as a set of linked hypotheses, as follows:
H1
Snow features can be explained by slope and aspect, which differentiate between cooler and warmer slopes, curvature that captures local terrain depressions accumulating more snow in winter, and canopy shade, which moderates solar exposure and spring melt. These factors influence the distribution and persistence of snow.
H2
The effect of snow on shrubs will depend on their growth traits (tall versus dwarf) and leaf habits (deciduous versus evergreen). This variation arises from differences in their tolerance to the length of the growing season, with some shrubs better adapted to shorter or longer snow cover durations.
H3
Deciduous shrubs, which benefit from earlier snowmelt and are often more competitive in areas of early snowmelt, are expected to have an indirect negative effect on herbaceous composition through their interaction with evergreen species. Evergreen shrubs, which tolerate, or even favour, prolonged snow conditions, may be indirectly influenced by the competitive presence of deciduous shrubs.
H4
Taller shrubs are predicted to negatively affect dwarf shrubs by overshadowing and outcompeting them for resources. Additionally, this competitive dynamic is expected to indirectly affect herbaceous plant composition by altering the competitive balance among different shrubs and herb species. Further, we expect deeper soils to support taller shrubs more strongly, as they have greater rooting depth requirements.
H5
Each shrub growth form (based on growth traits and leaf habits) will influence herbaceous composition differently. This is due to the varying cover strategies employed by each group, which affect the composition and abundance of herbaceous plants in their vicinity.
H6
Persistent snow is expected to reduce growing season length, and shade is predicted to limit light availability, both of which may negatively affect herbaceous composition. In contrast, warmer solar exposure is anticipated to promote herbaceous growth. These effects vary through the indirect pathways mediated by shrubs, as the interaction between shrubs and environmental factors (snow, shade, solar exposure) should influence resource availability and conditions for herbaceous growth.
To improve model fit, we tested for missing pathways and standardized coefficients for ease of variable comparisons. Specifically, four piecewise SEMs were fitted to compare potential network causalities between terrain features that affect snow persistence, shrub cover (among growth forms), and herbaceous cover/richness. We first fit separate pathway diagrams for different groups of variables, based on a priori knowledge of interactions among shrub growth forms (see Fig. 6 in Appendix B). We used the Akaike Information Criterion (AIC) to compare support among hypothesized models and assessed model fit using Shipley’s test of direct separation, which tests the probability that no pathways are missing from the hypothesized framework (Shipley 2000). Models were rejected if Fisher’s C statistic and χ2 tests were p < 0.05.
Fig. 6
Pathway diagrams of the four competing system hypotheses across different groups of shrub variables (a: shrub, b: size, c: leaf traits, d: mix of size and leaf habits), their relationship with terrain and snow, and how they interact with herbaceous species (abundance and richness). The arrows represent pathways between terrain, snow, shrubs, and herbaceous composition. Solar exposure was the only pathway predicted to affect abundance and not richness (grey line). Each hypothesis examines the weighted effects of different shrub growth forms on herb richness and on their interactions.
We then used linear models (R Core Team, 2024) to test: (1) the effects of terrain (solar severity index, curvature, canopy cover) on snow persistence, (2) the effect of snow and terrain on (a) shrub composition, (b) tall/dwarf shrub composition, (c) evergreen/deciduous shrub composition, and (d) combination of size and leaf habits of shrubs; and (3) the effects of terrain, snow, and shrubs growth forms (a-d) on herbaceous cover and richness. Although SEMs allow for multiple response variables, our primary final interest was herbaceous cover and richness. Herbaceous richness was normally distributed; thus, models were tested under Gaussian and Poisson distributions, with the Gaussian distribution being the most supported and with more homogenous residuals. All model variables are reported in standardized units. For a summary of responses and predictors, including mean and standard deviations, see Table 3 in Appendix B. For the most supported model, we estimated the effects of the indirect pathways on herbaceous abundance/richness by multiplying the standardized effect of the variable and the mediator, and of the mediator and herbaceous richness. We plot model predictions across the full gradient of snow and shrub cover between growth/phenology traits to better understand our results, highlighting the combined effects among variables on herbaceous abundance/richness.
|
Variable
|
Minimum
|
Maximum
|
Mean
|
SD
|
|---|---|---|---|---|
|
Surface Terrain
|
-19.98
|
19.97
|
-0.93
|
3.71
|
|
Solar Severity
|
-0.92
|
0.57
|
-0.05
|
0.26
|
|
Shade
|
0.00
|
0.27
|
0.02
|
0.05
|
|
Snow
|
0.16
|
0.97
|
0.48
|
0.21
|
|
Depth
|
4.00
|
57.00
|
21.57
|
12.30
|
|
Depth.t
|
1.39
|
4.04
|
2.90
|
0.61
|
|
Shrub
|
0.00
|
1.26
|
0.45
|
0.27
|
|
Shrub Tall
|
0.00
|
1.00
|
0.17
|
0.23
|
|
Shrub Dwarf
|
0.00
|
1.00
|
0.29
|
0.26
|
|
Shrub Deciduous
|
0.00
|
1.00
|
0.13
|
0.17
|
|
Shrub Evergreen
|
0.00
|
1.00
|
0.28
|
0.25
|
|
Tall Deciduous
|
0.00
|
1.00
|
0.12
|
0.19
|
|
Tall Evergreen
|
0.00
|
1.00
|
0.11
|
0.23
|
|
Dwarf Deciduous
|
0.00
|
0.95
|
0.08
|
0.17
|
|
Dwarf Evergreen
|
0.00
|
1.00
|
0.21
|
0.24
|
|
Abundance
|
0.00
|
1.00
|
0.17
|
0.15
|
|
Richness
|
0.00
|
14.00
|
5.71
|
2.35
|
Results
General Patterns in Plant Composition
We identified 101 vascular plant species across 338 quadrats, of which 74 were herbs, 16 shrubs, eight graminoids, and three tree species (Table 2 in Appendix B). Herbs were present in 95% of quadrats, averaging 30 ± 1% cover when present, with an average richness of 6 (± 2) species. Shrubs were present in 78% of the quadrats, averaging 57 ± 2% cover when present, and an average richness of 3 ± 1 species (Table 4 in Appendix B). There were seven tall/deciduous, two tall/evergreen, three dwarf deciduous, and three dwarf/evergreen shrubs (Table 5 in Appendix B). Tall evergreen species had the largest average cover when present (46 ± 3%), but the lowest prevalence (31%), while dwarf evergreen species had the second highest average cover (28 ± 1%) and the highest prevalence of shrubs (75%). Tall deciduous shrubs had a higher prevalence than dwarf shrubs (73% to 57%), but similar average cover (Table 6 in Appendix B).
A
|
Species
|
Lifeform
|
Prevalence
|
Mean
|
SD
|
SE
|
|---|---|---|---|---|---|
|
Abies bisfolia
|
Conifer
|
0.12
|
0.10
|
0.20
|
0.03
|
|
Achillea borealis
|
Herb
|
0.03
|
0.05
|
0.04
|
0.01
|
|
Aconitum delphiniifolium
|
Herb
|
0.03
|
0.08
|
0.10
|
0.03
|
|
Agoseris glauca
|
Herb
|
0.01
|
0.06
|
0.03
|
0.02
|
|
Androsace chamaejasme
|
Herb
|
0.44
|
0.02
|
0.01
|
0.00
|
|
Anemone multifida
|
Herb
|
0.01
|
0.01
|
0.00
|
0.00
|
|
Anemone parviflora
|
Herb
|
0.32
|
0.03
|
0.03
|
0.00
|
|
Antennaria alpina
|
Herb
|
0.01
|
0.06
|
0.06
|
0.05
|
|
Antennaria lanata
|
Herb
|
0.05
|
0.03
|
0.02
|
0.01
|
|
Antennaria umbrinella
|
Herb
|
0.06
|
0.01
|
0.01
|
0.00
|
|
Aquilegia flavescens
|
Herb
|
0.03
|
0.05
|
0.03
|
0.01
|
|
Arctous rubra
|
Shrub
|
0.13
|
0.35
|
0.27
|
0.04
|
|
Arctostaphylos uva-ursi
|
Shrub
|
0.05
|
0.42
|
0.36
|
0.09
|
|
Arnica angustifolia
|
Herb
|
0.08
|
0.03
|
0.02
|
0.00
|
|
Arnica cordifolia
|
Herb
|
0.02
|
0.17
|
0.17
|
0.07
|
|
Arnica latifolia
|
Herb
|
0.02
|
0.19
|
0.19
|
0.07
|
|
Arnica ovata
|
Herb
|
0.03
|
0.07
|
0.05
|
0.02
|
|
Artemisia norvegica
|
Herb
|
0.16
|
0.16
|
0.17
|
0.02
|
|
Astragalus alpinum
|
Herb
|
0.01
|
0.03
|
0.02
|
0.01
|
|
Astragalus australis
|
Herb
|
0.01
|
0.05
|
0.06
|
0.03
|
|
Astragalus vexilliflexus
|
Herb
|
0.03
|
0.05
|
0.06
|
0.02
|
|
Betula glandulosa
|
Shrub
|
0.28
|
0.16
|
0.19
|
0.02
|
|
Bisorta vivipara
|
Herb
|
0.81
|
0.03
|
0.02
|
0.00
|
|
Botrychium lunaria
|
Herb
|
0.01
|
0.01
|
0.01
|
0.00
|
|
Campanula lasiocarpa
|
Herb
|
0.01
|
0.09
|
0.14
|
0.08
|
|
Carex microptera
|
Graminoid
|
0.25
|
0.07
|
0.11
|
0.01
|
|
Carex petricosa
|
Graminoid
|
0.40
|
0.06
|
0.07
|
0.01
|
|
Care phaeocephala
|
Graminoid
|
0.24
|
0.07
|
0.11
|
0.01
|
|
Castilleja occidentalis
|
Herb
|
0.05
|
0.04
|
0.03
|
0.01
|
|
Castilleja rhexifolia
|
Herb
|
0.01
|
0.06
|
0.05
|
0.04
|
|
Cassiope tetragona
|
Shrub
|
0.15
|
0.31
|
0.26
|
0.04
|
|
Chamaenerion angustifolium
|
Herb
|
0.01
|
0.04
|
0.03
|
0.02
|
|
Coeloglossum viride
|
Herb
|
0.02
|
0.02
|
0.01
|
0.00
|
|
Dasiphora fruticosa
|
Shrub
|
0.22
|
0.06
|
0.07
|
0.01
|
|
Delphinium glaucum
|
Herb
|
0.01
|
0.09
|
0.08
|
0.04
|
|
Dryas hookeriana
|
Shrub
|
0.14
|
0.09
|
0.11
|
0.02
|
|
Dryas integrifolia
|
Shrub
|
0.56
|
0.31
|
0.22
|
0.02
|
|
Equisetum arvense
|
Herb
|
0.01
|
0.17
|
0.27
|
0.12
|
|
Equisetum scirpoides
|
Herb
|
0.11
|
0.07
|
0.11
|
0.02
|
|
Erigeron glacialis
|
Herb
|
0.01
|
0.15
|
0.07
|
0.05
|
|
Erigeron grandiflorus
|
Herb
|
0.01
|
0.01
|
0.01
|
0.00
|
|
Erigeron lanatus
|
Herb
|
0.11
|
0.02
|
0.02
|
0.00
|
|
Erigeron peregrinus
|
Herb
|
0.18
|
0.07
|
0.08
|
0.01
|
|
Euphrasia subarctica
|
Herb
|
0.01
|
0.01
|
0.01
|
0.00
|
|
Fragaria virginiana
|
Herb
|
0.02
|
0.06
|
0.05
|
0.02
|
|
Gentianella propinqua
|
Herb
|
0.09
|
0.02
|
0.02
|
0.00
|
|
Gentianella prostrata
|
Herb
|
0.01
|
0.01
|
0.00
|
0.00
|
|
Hedysarum alpinum
|
Herb
|
0.54
|
0.09
|
0.08
|
0.01
|
|
Hedysarum boreale
|
Herb
|
0.21
|
0.07
|
0.04
|
0.01
|
|
Heracleum maximum
|
Herb
|
0.01
|
0.17
|
0.18
|
0.12
|
|
Heiracium triste
|
Herb
|
0.02
|
0.02
|
0.02
|
0.01
|
|
Juncus biglumis
|
Graminoid
|
0.03
|
0.06
|
0.06
|
0.02
|
|
Juniperus communis
|
Conifer
|
0.02
|
0.36
|
0.37
|
0.14
|
|
Juncus drummondii
|
Graminoid
|
0.04
|
0.05
|
0.04
|
0.01
|
|
Leymus innovatus
|
Graminoid
|
0.30
|
0.09
|
0.12
|
0.01
|
|
Mertensia paniculata
|
Herb
|
0.02
|
0.11
|
0.15
|
0.06
|
|
Moneses uniflora
|
Herb
|
0.03
|
0.02
|
0.01
|
0.00
|
|
Oxytropis campetris var. davisii
|
Herb
|
0.09
|
0.03
|
0.02
|
0.00
|
|
Oxyria digyna
|
Herb
|
0.01
|
0.04
|
0.04
|
0.02
|
|
Oxytropis podocarpa
|
Herb
|
0.07
|
0.03
|
0.02
|
0.00
|
|
Parnassia fimbriata
|
Herb
|
0.06
|
0.07
|
0.05
|
0.01
|
|
Pedicularis bracteosa
|
Herb
|
0.13
|
0.05
|
0.05
|
0.01
|
|
Peicularis flammea
|
Herb
|
0.03
|
0.06
|
0.10
|
0.03
|
|
Pedicularis lanata
|
Herb
|
0.09
|
0.03
|
0.02
|
0.00
|
|
Petasites frigidus
|
Herb
|
0.02
|
0.09
|
0.08
|
0.03
|
|
Phleum alpinum
|
Graminoid
|
0.01
|
0.06
|
0.06
|
0.03
|
|
Phylodoce glandulifora
|
Shrub
|
0.16
|
0.61
|
0.29
|
0.04
|
|
Picea engelmannii
|
Conifer
|
0.14
|
0.05
|
0.10
|
0.01
|
|
Poa alpina
|
Graminoid
|
0.21
|
0.05
|
0.06
|
0.01
|
|
Polemonium acutiflorum
|
Herb
|
0.01
|
0.16
|
0.21
|
0.14
|
|
Potentilla glaucophylla
|
Herb
|
0.22
|
0.03
|
0.03
|
0.00
|
|
Potentilla nivea
|
Herb
|
0.07
|
0.02
|
0.01
|
0.00
|
|
Potentilla pensylvanica
|
Herb
|
0.03
|
0.04
|
0.04
|
0.01
|
|
Rannunculus acris
|
Herb
|
0.01
|
0.05
|
0.03
|
0.01
|
|
Ranunculus eschsholtzii
|
Herb
|
0.04
|
0.09
|
0.14
|
0.04
|
|
Ranunculus pygmaeus
|
Herb
|
0.01
|
0.03
|
0.02
|
0.01
|
|
Rhinatnthus minor
|
Herb
|
0.03
|
0.01
|
0.01
|
0.00
|
|
Ribes montigenum
|
Shrub
|
0.01
|
0.20
|
0.14
|
0.10
|
|
Sabulina rubella
|
Herb
|
0.01
|
0.01
|
0.00
|
0.00
|
|
Salix alaxensis
|
Shrub
|
0.01
|
0.16
|
0.19
|
0.13
|
|
Salix arctica
|
Shrub
|
0.07
|
0.13
|
0.11
|
0.02
|
|
Salix barrattiana
|
Shrub
|
0.11
|
0.28
|
0.18
|
0.03
|
|
Salix drummondiana
|
Shrub
|
0.03
|
0.23
|
0.18
|
0.06
|
|
Salix glauca
|
Shrub
|
0.07
|
0.24
|
0.2
|
0.04
|
|
Salix nivalis
|
Shrub
|
0.36
|
0.07
|
0.08
|
0.01
|
|
Salix vestita
|
Shrub
|
0.01
|
0.22
|
0.2
|
0.09
|
|
Saxifraga oppositifolia
|
Herb
|
0.02
|
0.01
|
0.00
|
0.00
|
|
Saxifraga tricuspidata
|
Herb
|
0.03
|
0.04
|
0.04
|
0.01
|
|
Selaginella densa
|
Herb
|
0.01
|
0.04
|
0.03
|
0.02
|
|
Senecio lugens
|
Herb
|
0.14
|
0.04
|
0.03
|
0.00
|
|
Senecio triangularis
|
Herb
|
0.02
|
0.26
|
0.21
|
0.09
|
|
Sibbaldia procumbens
|
Herb
|
0.11
|
0.05
|
0.06
|
0.01
|
|
Silene acaulis
|
Herb
|
0.07
|
0.06
|
0.05
|
0.01
|
|
Solidago multiradiata
|
Herb
|
0.30
|
0.04
|
0.04
|
0.00
|
|
Solidago simplex
|
Herb
|
0.01
|
0.04
|
0.02
|
0.01
|
|
Stellaria longipes
|
Herb
|
0.01
|
0.02
|
0.02
|
0.01
|
|
Tofieldia pusilla
|
Herb
|
0.04
|
0.03
|
0.02
|
0.01
|
|
Trollius albiflorus
|
Herb
|
0.05
|
0.14
|
0.21
|
0.05
|
|
Veronica alpina
|
Herb
|
0.02
|
0.02
|
0.02
|
0.01
|
|
Viola renifolia
|
Herb
|
0.01
|
0.03
|
0.01
|
0.00
|
|
Zigadenus elegans
|
Herb
|
0.25
|
0.06
|
0.05
|
0.01
|
|
Lifeform
|
Prevalence
|
Richness
|
Mean Richness
|
Mean Cover
|
SD Cover
|
SE Cover
|
|---|---|---|---|---|---|---|
|
Conifer
|
0.28
|
3
|
1
|
0.10
|
0.19
|
0.02
|
|
Graminoid
|
0.74
|
8
|
2
|
0.12
|
0.10
|
0.01
|
|
Herb
|
0.95
|
74
|
6
|
0.30
|
0.07
|
0.01
|
|
Shrub
|
0.78
|
16
|
3
|
0.57
|
0.24
|
0.02
|
|
Shrub Trait
|
Species N
|
Prevalence
|
Mean
|
SD
|
SE
|
|---|---|---|---|---|---|
|
Dwarf & Deciduous
|
3
|
0.57
|
0.14
|
0.19
|
0.01
|
|
Dwarf & Evergreen
|
3
|
0.75
|
0.28
|
0.23
|
0.01
|
|
Tall & Deciduous
|
7
|
0.73
|
0.16
|
0.18
|
0.01
|
|
Tall & Evergreen
|
2
|
0.31
|
0.46
|
0.31
|
0.03
|
|
Species
|
Growth
|
Leaf Habit
|
Prevalence
|
Mean
|
SD
|
SE
|
|---|---|---|---|---|---|---|
|
Arctous rubra
|
Dwarf
|
Deciduous
|
0.13
|
0.35
|
0.27
|
0.04
|
|
Arctostaphylos uva-ursi
|
Dwarf
|
Evergreen
|
0.05
|
0.42
|
0.36
|
0.09
|
|
Betula glandulosa
|
Tall
|
Deciduous
|
0.28
|
0.16
|
0.19
|
0.02
|
|
Cassiope tetragona
|
Tall
|
Evergreen
|
0.15
|
0.31
|
0.26
|
0.04
|
|
Dasiphora fruticosa
|
Tall
|
Deciduous
|
0.22
|
0.06
|
0.07
|
0.01
|
|
Dryas hookeriana
|
Dwarf
|
Evergreen
|
0.14
|
0.09
|
0.11
|
0.02
|
|
Dryas integrifolia
|
Dwarf
|
Evergreen
|
0.56
|
0.31
|
0.22
|
0.02
|
|
Phylodoce glandulifora
|
Tall
|
Evergreen
|
0.16
|
0.61
|
0.29
|
0.04
|
|
Ribes montigenum
|
Tall
|
Deciduous
|
0.01
|
0.20
|
0.14
|
0.10
|
|
Salix alaxensis
|
Tall
|
Deciduous
|
0.01
|
0.16
|
0.19
|
0.14
|
|
Salix arctica
|
Dwarf
|
Deciduous
|
0.07
|
0.13
|
0.11
|
0.02
|
|
Salix barrattiana
|
Tall
|
Deciduous
|
0.11
|
0.28
|
0.18
|
0.03
|
|
Salix drummondiana
|
Tall
|
Deciduous
|
0.03
|
0.23
|
0.18
|
0.06
|
|
Salix glauca
|
Tall
|
Deciduous
|
0.07
|
0.24
|
0.20
|
0.04
|
|
Salix nivalis
|
Dwarf
|
Deciduous
|
0.36
|
0.07
|
0.08
|
0.01
|
|
Salix vestita
|
Tall
|
Deciduous
|
0.01
|
0.22
|
0.20
|
0.09
|
Structural Equation Model (SEM) of the Alpine System
Model Selection of SEM and Model Fit
For both herbaceous cover and richness, our most supported model was one with both size and leaf habit growth forms of shrubs (Fig. 6d and Tables 7 & 8 in Appendix B). The goodness of fit for the herbaceous cover SEM was Fischer’s C = 32.18 (p = 0.187), and for the herbaceous richness SEM was Fischer’s C = 35.71 (p = 0.218). Pathways between terrain, snow, and shrubs remained the same for both models (Fig. 3). Models explained 29% of the variation in snow persistence, 8% in tall deciduous cover, 30% in tall evergreen shrub cover, and 35% in dwarf evergreen shrub cover. None of our variables were significantly related to dwarf deciduous shrub cover (Tables 9 & 10 in Appendix B). The final SEM models explained 33% variation of herbaceous cover (Fig. 3a) and 17% variation of herbaceous richness (Fig. 3b).
|
SEM Hypothesis (herb cover)
|
AIC
|
K
|
ΔAIC
|
|---|---|---|---|
|
Shrub (grouped)
|
-557
|
15
|
820
|
|
Shrub Growth Form
|
-832
|
22
|
545
|
|
Shrub Leaf Habit
|
-908
|
21
|
469
|
|
Shrub Growth Form & Leaf Habit
|
-1377
|
32
|
0
|
Fig. 3
Final Structural Equation Model (SEM) from the four a priori competing hypotheses (Hypothesis 4, Fig. 6d in Appendix B). The diagram shows the pathways and standardized coefficients between terrain (grey boxes), snow (blue boxes), shrub cover (green boxes), and herbaceous cover (yellow boxes) (a). Pathways between terrain, snow and shrub were consistent across both herbaceous cover (a, pink box) and richness (b, pink box) models and thus only highlighted in (a), while (b) only shows variables affecting herbaceous richness. Solid black lines represent the hypothesized pathways that were statistically significant (p < 0.05), with line thickness scaled to the strength of relationships. Dashed lines indicate hypothesized pathways that were non-significant. Two pathways were added (grey solid line) when a test of directed separation confirmed that surface curvature affected both herbaceous cover and richness, and dwarf shrub cover (a). Indirect effects are not illustrated due to their complexity, but are described in Table 1. All model coefficients and standard errors in the final models are presented in Tables 9 & 10.
|
SEM Hypothesis (herb richness)
|
AIC
|
K
|
ΔAIC
|
|---|---|---|---|
|
Shrub (grouped)
|
1375
|
15
|
807
|
|
Shrub Growth Form
|
1087
|
22
|
520
|
|
Shrub Leaf Habit
|
1017
|
21
|
449
|
|
Shrub Growth Form & Leaf Habit
|
567
|
30
|
0
|
|
Response
|
Predictor
|
β
|
SE
|
df
|
Z
|
Std β
|
p
|
|---|---|---|---|---|---|---|---|
|
Snow
|
(Intercept)
|
0.44
|
0.01
|
335
|
43.74
|
0.00
|
< 0.001
|
|
Snow
|
Surface Terrain
|
-0.02
|
0.00
|
335
|
-8.06
|
-0.37
|
< 0.001
|
|
Snow
|
Solar Severity
|
-0.30
|
0.04
|
335
|
-8.20
|
-0.38
|
< 0.001
|
|
Tall Deciduous
|
(Intercept)
|
-0.14
|
0.05
|
336
|
-2.78
|
0.00
|
0.006
|
|
Tall Deciduous
|
Depth.t
|
0.09
|
0.02
|
336
|
5.26
|
0.28
|
< 0.001
|
|
Tall Evergreen
|
(Intercept)
|
-0.40
|
0.05
|
334
|
-7.57
|
0.00
|
< 0.001
|
|
Tall Evergreen
|
Snow
|
0.45
|
0.05
|
334
|
8.32
|
0.40
|
< 0.001
|
|
Tall Evergreen
|
Depth.t
|
0.11
|
0.02
|
334
|
5.75
|
0.29
|
< 0.001
|
|
Tall Evergreen
|
Tall Deciduous
|
-0.20
|
0.06
|
334
|
-3.40
|
-0.16
|
0.001
|
|
Dwarf Deciduous
|
(Intercept)
|
0.05
|
0.02
|
336
|
2.42
|
0.00
|
0.016
|
|
Dwarf Deciduous
|
Snow
|
0.06
|
0.04
|
336
|
1.36
|
0.07
|
0.174
|
|
Dwarf Evergreen
|
(Intercept)
|
0.47
|
0.03
|
332
|
16.70
|
0.00
|
< 0.001
|
|
Dwarf Evergreen
|
Snow
|
-0.42
|
0.06
|
332
|
-6.81
|
-0.37
|
< 0.001
|
|
Dwarf Evergreen
|
Tall Evergreen
|
-0.17
|
0.05
|
332
|
-3.31
|
-0.17
|
0.001
|
|
Dwarf Evergreen
|
Tall Deciduous
|
-0.20
|
0.06
|
332
|
-3.51
|
-0.16
|
< 0.001
|
|
Dwarf Evergreen
|
Dwarf Deciduous
|
-0.22
|
0.06
|
332
|
-3.49
|
-0.16
|
< 0.001
|
|
Dwarf Evergreen
|
Surface Terrain
|
0.01
|
0.00
|
332
|
2.17
|
0.10
|
0.031
|
|
Abundance
|
(Intercept)
|
0.03
|
0.04
|
329
|
0.70
|
0.00
|
0.485
|
|
Abundance
|
Snow
|
0.12
|
0.04
|
329
|
2.85
|
0.18
|
0.005
|
|
Abundance
|
Dwarf Evergreen
|
-0.16
|
0.03
|
329
|
-4.73
|
-0.27
|
< 0.001
|
|
Abundance
|
Tall Evergreen
|
-0.28
|
0.03
|
329
|
-8.06
|
-0.45
|
< 0.001
|
|
Abundance
|
Tall Deciduous
|
-0.08
|
0.04
|
329
|
-2.24
|
-0.11
|
0.025
|
|
Abundance
|
Dwarf Deciduous
|
-0.26
|
0.04
|
329
|
-6.33
|
-0.29
|
< 0.001
|
|
Abundance
|
Surface Terrain
|
-0.01
|
0.00
|
329
|
-3.25
|
-0.16
|
0.001
|
|
Abundance
|
Solar Severity
|
-0.06
|
0.03
|
329
|
-2.19
|
-0.11
|
0.029
|
|
Abundance
|
Depth.t
|
0.06
|
0.01
|
329
|
4.55
|
0.24
|
< 0.001
|
|
Response
|
Predictor
|
β
|
SE
|
df
|
Z
|
Std β
|
p
|
|---|---|---|---|---|---|---|---|
|
Snow
|
(Intercept)
|
0.44
|
0.01
|
335
|
43.74
|
0.00
|
< 0.001
|
|
Snow
|
Surface Terrain
|
-0.02
|
0.00
|
335
|
-8.06
|
-0.37
|
< 0.001
|
|
Snow
|
Solar Severity
|
-0.30
|
0.04
|
335
|
-8.20
|
-0.38
|
< 0.001
|
|
Tall Deciduous
|
(Intercept)
|
-0.14
|
0.05
|
336
|
-2.78
|
0.00
|
0.006
|
|
Tall Deciduous
|
Depth.t
|
0.09
|
0.02
|
336
|
5.26
|
0.28
|
< 0.001
|
|
Tall Evergreen
|
(Intercept)
|
-0.40
|
0.05
|
334
|
-7.57
|
0.00
|
< 0.001
|
|
Tall Evergreen
|
Snow
|
0.45
|
0.05
|
334
|
8.32
|
0.40
|
< 0.001
|
|
Tall Evergreen
|
Depth.t
|
0.11
|
0.02
|
334
|
5.75
|
0.29
|
< 0.001
|
|
Tall Evergreen
|
Tall Deciduous
|
-0.20
|
0.06
|
334
|
-3.40
|
-0.16
|
0.001
|
|
Dwarf Deciduous
|
(Intercept)
|
0.05
|
0.02
|
336
|
2.42
|
0.00
|
0.016
|
|
Dwarf Deciduous
|
Snow
|
0.06
|
0.04
|
336
|
1.36
|
0.07
|
0.174
|
|
Dwarf Evergreen
|
(Intercept)
|
0.47
|
0.03
|
332
|
16.70
|
0.00
|
< 0.001
|
|
Dwarf Evergreen
|
Snow
|
-0.42
|
0.06
|
332
|
-6.81
|
-0.37
|
< 0.001
|
|
Dwarf Evergreen
|
Tall Evergreen
|
-0.17
|
0.05
|
332
|
-3.31
|
-0.17
|
0.001
|
|
Dwarf Evergreen
|
Tall Deciduous
|
-0.20
|
0.06
|
332
|
-3.51
|
-0.16
|
< 0.001
|
|
Dwarf Evergreen
|
Dwarf Deciduous
|
-0.22
|
0.06
|
332
|
-3.49
|
-0.16
|
< 0.001
|
|
Dwarf Evergreen
|
Surface Terrain
|
0.01
|
0.00
|
332
|
2.17
|
0.10
|
0.031
|
|
Richness
|
(Intercept)
|
2.88
|
0.65
|
331
|
4.42
|
0.00
|
< 0.001
|
|
Richness
|
Snow
|
3.74
|
0.69
|
331
|
5.46
|
0.33
|
< 0.001
|
|
Richness
|
Tall Evergreen
|
-3.48
|
0.61
|
331
|
-5.74
|
-0.35
|
< 0.001
|
|
Richness
|
Dwarf Deciduous
|
-2.15
|
0.72
|
331
|
-2.99
|
-0.15
|
0.003
|
|
Richness
|
Surface Terrain
|
-0.08
|
0.03
|
331
|
-2.21
|
-0.12
|
0.028
|
|
Richness
|
Depth.t
|
0.58
|
0.22
|
331
|
2.60
|
0.15
|
0.010
|
|
Richness
|
Tall Deciduous
|
-1.43
|
0.65
|
331
|
-2.19
|
-0.12
|
0.029
|
Effects of Terrain Features on Snow Persistence
The average snow persistence across all sites was 48 ± 1% from mid-March to mid-July, with the model values ranging from 10 to 90% snow cover. Surface curvature and solar exposure significantly affected snow persistence patterns across sites (quadrats). Concave curvatures increased snow persistence by 37% per standardized unit increase in curvature, while cooler and steeper slopes increased snow persistence by 38% per standardized unit.
Effects of Terrain and Snow on Shrub Abundance by Growth Forms
Shrub growth forms responded differently to snow and terrain features (Fig. 3a; Tables 9 & 10 in Appendix B). Tall deciduous shrubs were influenced by soil depth, with cover increasing by 28% per standardized unit increase in soil depth. Similarly, tall evergreen shrubs increased by 29% per standardized unit increase in soil depth. However, the cover of tall evergreen shrubs also increased by 40% per standardized unit increase in snow persistence. In contrast, tall deciduous shrub cover was negatively affected by snow cover, decreasing by 16% per standardized unit increase in snow cover. Dwarf evergreen shrubs were negatively affected by snow (37% per standardized unit increase), as was the cover of tall evergreen shrubs (17% per standardized unit increase), tall deciduous shrubs (16% per standardized unit increase), and dwarf deciduous shrubs (16% per standardized unit increase). Convex slopes were the only factor positively influencing dwarf evergreen shrubs, increasing their cover by 10% per standardized unit increase in convexity. Finally, dwarf deciduous shrub cover was not affected by any of the variables measured.
Direct Pathways of Terrain, Snow, and Shrubs on Herbaceous Cover and Richness
Snow and soil depth positively influenced herbaceous cover, with increases of 18% per standardized unit for snow and 24% per standardized unit for soil depth. Tall evergreen shrub cover had the most adverse effect on herbaceous cover, reducing herbaceous cover by 45% per standardized unit change in tall evergreen shrub cover, followed by dwarf deciduous and evergreen shrub cover, which decreased cover of the herbaceous community by 29% and 27% per standardized unit of tall evergreen shrub cover, respectively. Tall deciduous shrub cover had the lowest adverse effect on herbaceous cover, with an 11% decrease per standardized unit. Terrain features affected herbaceous cover with solar severity and surface curvature, reducing herbaceous cover by 11% per standardized unit on warm slopes (with increased cover on cool slopes) and by 16% per standardized unit on convex slopes (with increased cover on concave slopes). Detailed pathways and standardized coefficients are shown in Fig. 3, while the contrasting effects of snow persistence (positive) and shrub cover (negative) are illustrated in Fig. 4a-d.
A
Snow persistence and soil depth both had positive effects on herbaceous richness, increasing it by 0.33 species per standardized unit for snow persistence and 0.15 species per standardized unit of soil depth. Cover of tall evergreen shrub had the most substantial effect on herbaceous richness, reducing herbaceous richness by 0.35 species per standardized unit of tall evergreen shrub cover. Tall deciduous and dwarf evergreen shrub cover had more minor negative effects, reducing herbaceous richness by 0.12 and 0.15 species per standardized unit, respectively. Dwarf evergreen shrub cover did not significantly affect herbaceous richness. Surface terrain also played a role, with concave slopes decreasing herbaceous richness by 0.12 species per standardized unit of concavity, whereas convex slopes increased it. Pathways and standardized coefficients for herbaceous richness are shown in Fig. 3b, and the relationships between types of shrubs and snow on herbaceous richness are detailed in Fig. 4e-g. Tall deciduous shrub cover experienced the largest decrease in total effect size, with indirect pathways through tall and dwarf evergreen shrubs nullifying its direct effect and rendering it neutral with respect to herbaceous cover. Total direct and indirect effects for each variable on herbaceous cover and richness are presented in Table 1 and visualized in Fig. 8 in Appendix B.Fig. 4
Predicted cover (left side plots) and richness (right side plots) of herbaceous alpine plants (color gradient outlines) across different levels of snow persistence (x-axis) and cover (percent) for four growth forms of shrubs (y-axis): (a & e) dwarf deciduous shrub cover, (b & f) tall deciduous shrub cover, (c & g) tall evergreen shrub cover, and (d) dwarf evergreen shrub cover. Note that the lower limits of predicted herbaceous richness and abundance are under conditions of a site having no snow cover and high shrub cover (extrapolation since not observed in plots).
|
Predictor
|
Pathway to herbaceous cover
|
Effect
|
|---|---|---|
|
Surface Curvature
|
Direct
|
-0.16
|
|
Indirect: Snow
|
-0.07
|
|
|
Indirect: Dwarf/Evergreen Shrubs
|
-0.03
|
|
|
Total
|
-0.26
|
|
|
Solar Severity
|
Direct
|
-0.11
|
|
Indirect: Snow
|
-0.07
|
|
|
Total
|
-0.18
|
|
|
Soil Depth
|
Direct
|
0.24
|
|
Indirect: Tall/Evergreen Shrubs
|
-0.13
|
|
|
Indirect: Tall/Deciduous Shrubs
|
-0.03
|
|
|
Total
|
0.08
|
|
|
Snow Persistence
|
Direct
|
0.17
|
|
Indirect: Dwarf/Evergreen Shrubs
|
0.10
|
|
|
Indirect: Tall/Evergreen Shrubs
|
-0.18
|
|
|
Total
|
0.09
|
|
|
Tall/Deciduous Shrubs
|
Direct:
|
-0.11
|
|
Indirect: Tall/Evergreen Shrubs
|
0.07
|
|
|
Indirect: Dwarf/Evergreen Shrubs
|
0.04
|
|
|
Total
|
0
|
|
|
Tall/Evergreen Shrubs
|
Direct
|
-0.45
|
|
Indirect: Dwarf/Evergreen Shrubs
|
0.05
|
|
|
Total
|
-0.40
|
|
|
Dwarf/Deciduous Shrubs
|
Direct
|
-0.29
|
|
Indirect: Dwarf/Evergreen Shrubs
|
0.04
|
|
|
Total
|
-0.25
|
|
|
Dwarf/Evergreen Shrubs
|
Direct
|
-0.27
|
|
Predictor
|
Pathway to herbaceous richness
|
Effect
|
|
Surface Curvature
|
Direct
|
-0.12
|
|
Indirect: Snow
|
-0.13
|
|
|
Total
|
-0.25
|
|
|
Solar Severity
|
Indirect: Snow
|
-0.13
|
|
Soil Depth
|
Direct
|
0.15
|
|
Indirect: Tall/Deciduous Shrub
|
-0.03
|
|
|
Indirect: Tall/Evergreen Shrubs
|
-0.10
|
|
|
Total
|
0.02
|
|
|
Snow
|
Direct
|
0.34
|
|
Indirect: Tall/Evergreen Shrubs
|
-0.14
|
|
|
Total
|
0.20
|
|
|
Tall/Deciduous Shrubs
|
Direct
|
-0.12
|
|
Indirect: Tall/Evergreen Shrubs
|
0.06
|
|
|
Total
|
-0.06
|
|
|
Tall/Evergreen Shrubs
|
Direct
|
-0.35
|
|
Dwarf/Deciduous Shrubs
|
Direct
|
-0.15
|
Discussion
General Findings
We found that herbaceous alpine communities at Cardinal Divide in Alberta, Canada, were negatively affected by shrub cover, and positively influenced by snow and terrain features that supported higher snow persistence and deeper soils. On average, herbaceous cover was lower than shrub cover, although herbaceous richness was higher. Given the relatively low average richness of six species per 0.25 m2, the effects of our predictors on herbaceous richness, though seemingly small, are still considerable.
Effect of Shrubs on Alpine Herbaceous Communities
As hypothesized, deciduous shrubs negatively affected their evergreen shrub counterparts, supporting previous observations (Dobbert et al. 2021). We also found that tall deciduous and evergreen shrubs negatively affected dwarf evergreen shrubs (Tremblay et al. 2012), but not deciduous shrubs. Further, the dwarf-deciduous shrubs were not significantly affected by any of our predictors. This was surprising given the moderate occurrence and relatively high coverage of dwarf-deciduous shrubs in the study area, which led us to assume that snow and tall shrubs negatively affect dwarf-deciduous shrubs (Press et al. 1998, Saarinen et al. 2016, Boscutti et al. 2018). As anticipated, tall deciduous shrub cover was positively influenced by soil depth. Unexpectedly, no relationship was observed between snow and tall deciduous shrubs, in contrast to studies indicating mutual effects in which snow provides protection and moisture, and shrubs offer shade that reduces snowmelt time (Hallinger et al. 2010, Wheeler et al. 2016). Both deciduous shrub forms had similar direct effects on herbaceous cover and richness. However, because of indirect pathways through both forms of evergreen shrubs, tall deciduous shrubs did not affect herbaceous cover and had a reduced effect on herbaceous richness compared to dwarf deciduous shrubs. Tall evergreen shrubs, although less prevalent among sites, had the highest cover when present and exerted the most substantial adverse effects on both herbaceous cover and richness. Tall evergreen shrubs were positively associated with areas of higher snow persistence and deeper soils, as found in previous studies (Mallik et al. 2011, Christiansen et al. 2018), and were negatively associated with tall deciduous shrubs, consistent with our hypothesis and other work (Dobbert et al. 2021). This resulted in tall evergreen shrubs reducing the positive effects of snow and soil depth on herbaceous composition, thereby mediating indirect pathways. The direct effect of tall evergreen shrubs on herbaceous cover and richness was attenuated by indirect pathways involving dwarf evergreen shrubs. Dwarf evergreen shrubs, which were the most frequently encountered shrub and the second most abundant, were negatively affected by all shrub types, snow, and surface curvature, establishing them as primary mediators of indirect effects. Despite their significant mediating role, dwarf evergreen shrubs did not affect herbaceous richness, which in turn limited the number of indirect pathways that influenced it. Nevertheless, dwarf evergreen shrubs had the second-largest adverse effect on herbaceous cover, underscoring their coexistence with numerous small, growing herbaceous species (Wookey et al. 1995).
Influence of Snow Persistence on the Alpine System
As hypothesized, snow was positively related to swale-like features on cooler, north-facing slopes, supporting the observation that snow accumulates in these sites (Erickson et al. 2005, Björk and Molau 2007, DeBeer and Pomeroy 2017). Snow was positively associated with tall evergreen shrubs and negatively related to dwarf evergreen shrubs, reinforcing the idea of the relationships between snow and shrub heights and leaf habits (Wheeler et al. 2014, Saarinen et al. 2016). In contrast to other studies, snow did not affect the local abundance of dwarf deciduous shrubs (Saarinen et al. 2016, Gamm et al. 2018, Dobbert et al. 2021). This may be due to the high prevalence of evergreen species in our system or to differences in snow disappearance rates relative to snow persistence. In our study, areas of short snow persistence often emerged just one week after complete snow cover, presumably due to high wind exposure (Dadic et al. 2010). Rapid changes in snow cover, including winter snow loss, should affect the limitations on deciduous shrubs.
Snow was positively associated with both herbaceous cover and richness. This contradicts some prior studies (Alatalo et al. 2014, Good et al. 2019, Cheng et al. 2023), but aligns with recent research indicating shifts in community composition along snow gradients (Verrall et al. 2023, Morgan and Walker 2023). While we cannot confirm whether our system has undergone compositional shifts over time, the persistence of snow until late July in some years underscores changes in melt timing observed at other sites (Inouye 2020). Earlier melt times appear to alter the trade-off between growing season length, frost protection, and moisture reserves provided by the presence and/or persistence of snow (Kudo and Ito 1992, Wheeler et al. 2014, Conner et al. 2016). Despite the positive effect of snow on herbaceous cover and richness, this relationship was diminished by the indirect effects of tall evergreen shrubs. For instance, the negative indirect pathway through tall evergreen shrubs slightly outweighed the positive direct effect of snow on herbaceous cover. Without the positive indirect pathway through dwarf evergreen shrubs, the overall positive effect of snow on herbaceous composition could be negated, potentially becoming negative. Our findings suggest that areas with both high snow persistence and tall evergreen shrubs can have the lowest herbaceous cover and richness. These insights would not have been intuitive without the indirect pathways provided by structural equation models (Wang et al. 2019). Traditional approaches may have led us to expect areas of high snow persistence and tall evergreen species would be associated with moderate herb cover and richness. In contrast, the lowest herbaceous cover and richness would occur in areas with very low snow persistence and very high shrub cover (as shown in Fig. 3a). However, the causal pathways revealed that (1) high cover of tall evergreen shrubs is related to high snow cover, making the scenario of lowest herb richness unlikely, and (2) the positive association between snow and tall evergreen shrubs creates an adverse indirect effect on herbaceous cover that outweighs the positive direct effect of snow. Consequently, areas with high snow persistence and tall evergreen shrub cover are more likely to exhibit the lowest herbaceous cover and potentially richness.
Effects of Terrain on Snow Persistence and the Structuring of Alpine Plant Communities
Our study revealed that concave terrain surfaces, such as swales on the Lower Divide, and cool-facing slopes positively influenced herbaceous composition. The persistence of snow may protect against frost (Körner 2003a). Swales collect runoff from snowmelt, particularly at their base, thus providing greater soil moisture (DeBeer and Pomeroy 2010, Abudu et al. 2016). The cool, moist conditions of these swales should benefit herbaceous composition, as some herbs are sensitive to intense heat and drought (Notarnicola et al. 2021). North-facing slopes intuitively support higher herbaceous cover due to reduced solar exposure. While our proxy for solar exposure did not directly affect herbaceous richness, it positively influenced richness through an indirect pathway involving snow. Unexpectedly, our proxy for solar exposure did not affect any of the shrub groups. Surface curvature and soil depth outperformed solar exposure in the models, and these are indirectly related to solar exposure. Instead, dwarf evergreen shrubs were positively associated with convex curvatures, which correlated with reduced snow persistence. Our hypothesis that soil depth positively affects tall-shrub and herbaceous composition was supported. Still, soil depth did not influence dwarf-species cover, which typically have shorter and/or lateral root systems, whereas taller shrubs and many herbs are deep-rooted and thus prefer deeper soils. However, our results indicated that the negative indirect pathways through tall deciduous and evergreen shrubs nearly offset the positive direct effect of soil depth on herbaceous cover and richness.
Implications for Climate Change
Ongoing changes in snow dynamics and shrub expansion have significant implications for alpine ecosystems (Wipf et al. 2009, Ballantyne and Pickering 2015). As temperatures rise and snow patterns shift, plant community composition is likely to change further. For instance, tall evergreen shrubs, such as heather species in krummholz habitat with late snowbeds, are unlikely to expand their habitat due to their association (protection) with later snow melt (Christiansen et al. 2018, Dobbert et al. 2021). Conversely, some dwarf shrub species appear to be threatened by warming temperatures in some places, yet are exhibiting growth and expansion in others (Press et al. 1998, Saarinen et al. 2016). In alpine systems, which are high-stress environments with relatively few species, even minor shifts in plant composition can have major consequences for the plant community (Myers-Smith et al. 2011).
Herbaceous communities appear to face the greatest threat from climate change. With earlier snowmelt, some herbaceous species are shifting their ranges to deeper snow patches, which offer trade-offs among a longer growing season, frost protection, and moisture reserves during prolonged droughts (Verrall et al. 2023, Morgan and Walker 2023). Although shrubs often outcompete herbs, some shrubs can promote herbaceous cover and richness (Ballantyne and Pickering 2015, Wang et al. 2019). Tall deciduous shrubs, with their erect stems and canopy shade, may provide refugia-like conditions that protect herbs from excessive heat, keep soils cool, reduce evapotranspiration, and slow snowmelt, thereby maintaining higher moisture content (Weinig 2000, Myers-Smith and Hik 2013, Sholto-Douglas et al. 2017). Increased shrub cover, especially deciduous shrubs, has been linked to changes in surface albedo, soil composition, and reduced herbaceous diversity (Sturm et al. 2005, Myers-Smith and Hik 2013, Sholto-Douglas et al. 2017). Long-term monitoring will be essential for understanding and potentially mitigating the effects of climate change on alpine biodiversity (Scharnagl et al. 2019, Zong et al. 2022).
While our study provides valuable insights, it has limitations. Deciduous shrubs do not appear to be affected by terrain features or snow in our site, suggesting that other factors, such as soil characteristics or surface temperatures (MacDonald et al. 2025), might be more influential (Francon et al. 2021). Variability in annual snowmelt patterns that we didn’t measure may be important (Abudu et al. 2016, Verrall et al. 2023). Future research should focus on long-term studies to capture temporal changes in snow dynamics, soil properties, and temperature, and examine their effects on plant communities, particularly deciduous shrubs (Sholto-Douglas et al., 2017). Experimental approaches could clarify the causal relationships between snow, shrubs, and herbs, enhancing our understanding of these complex interactions (Bell and Bliss 1977, Pardee et al. 2018). Despite these limitations, our findings illustrate critical linkages (pathways) among terrain, snow, shrubs, and herbs, thereby providing a better understanding of the complex ecological processes at play in alpine ecosystems (Körner 2003b).
Conclusion
Our study underscores the complex, synergistic relationships between terrain features, snow dynamics, shrub growth forms, and herbaceous communities. By employing Structural Equation Models (SEM), we uncovered a set of indirect pathways that influence the direct effects of these factors on herbaceous alpine communities, revealing the complexity of abiotic-to-biotic relationships. SEMs provide a valuable approach for elucidating complex direct and indirect ecological relationships. Our findings highlight the importance of alpine shrubs in shaping alpine community composition. While evergreen shrubs negatively affect herbaceous cover and richness, our results suggest that deciduous shrub cover and snow moderate their influence. Finally, despite the perceived threat posed by deciduous shrubs to alpine plant diversity (Sholto-Douglas et al. 2017), deciduous shrubs in our site had minimal effects on herbaceous cover and richness when present alongside evergreen shrubs. Recognizing these nuances is crucial for understanding community dynamics and prioritizing future research.
Disclosure Statement
The authors have no competing interests to declare that are relevant to the content of this article.
A
Funding
This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC Discovery Grant #RGPIN-2019-06040 to Scott E. Nielsen) and by the Alberta Conservation Association (Grant #ACABIO 1265-1/1202).
A
Data Availability
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.
A
Author Contribution
Z.J.M. and S.E.N. conceptualized and designed the study. Z.J.M. led the analysis and first manuscript draft with support from S.E.N.S.E.N. led in project administration, supervision, and funding.
Reference:
Abudu S, Sheng Z, Cui C et al (2016) Integration of aspect and slope in snowmelt runoff modeling in a mountain watershed. Water Sci Eng 9:265–273. https://doi.org/10.1016/j.wse.2016.07.002
Alatalo JM, Little CJ, Jägerbrand AK, Molau U (2014) Dominance hierarchies, diversity and species richness of vascular plants in an alpine meadow: contrasting short and medium term responses to simulated global change. PeerJ 2:e406. https://doi.org/10.7717/peerj.406
Anderton SP, White SM, Alvera B (2002) Micro-scale spatial variability and the timing of snow melt runoff in a high mountain catchment. J Hydrol 268:158–176. https://doi.org/10.1016/S0022-1694(02)00179-8
Ballantyne M, Pickering CM (2015) Shrub facilitation is an important driver of alpine plant community diversity and functional composition. Biodivers Conserv 24:1859–1875. https://doi.org/10.1007/s10531-015-0910-z
Bell KL, Bliss LC (1977) Overwinter Phenology of Plants in a Polar Semi-Desert. Arctic 30:118–121
Björk RG, Molau U (2007) Ecology of Alpine Snowbeds and the Impact of Global Change. Arct Antarct Alp Res 39:34–43. https://doi.org/10.1657/1523-0430(2007)39%255B34:EOASAT%255D2.0.CO;2
Bjorkman AD, García Criado M, Myers-Smith IH et al (2020) Status and trends in Arctic vegetation: Evidence from experimental warming and long-term monitoring. Ambio 49:678–692. https://doi.org/10.1007/s13280-019-01161-6
Boscutti F, Casolo V, Beraldo P et al (2018) Shrub growth and plant diversity along an elevation gradient: Evidence of indirect effects of climate on alpine ecosystems. PLoS ONE 13:e0196653. https://doi.org/10.1371/journal.pone.0196653
Cheng C, He N, Li M et al (2023) Plant species richness on the Tibetan Plateau: patterns and determinants. Ecography 2023:e06265. https://doi.org/10.1111/ecog.06265
Christiansen CT, Lafreniére MJ, Henry GHR, Grogan P (2018) Long-term deepened snow promotes tundra evergreen shrub growth and summertime ecosystem net CO2 gain but reduces soil carbon and nutrient pools. Glob Change Biol 24:3508–3525. https://doi.org/10.1111/gcb.14084
Conner LG, Gill RA, Belnap J (2016) Soil moisture response to experimentally altered snowmelt timing is mediated by soil, vegetation, and regional climate patterns. Ecohydrology 9:1006–1016. https://doi.org/10.1002/eco.1697
Dadic R, Mott R, Lehning M, Burlando P (2010) Wind influence on snow depth distribution and accumulation over glaciers. J Geophys Res Earth Surf 115. https://doi.org/10.1029/2009JF001261
DeBeer CM, Pomeroy JW (2010) Simulation of the snowmelt runoff contributing area in a small alpine basin. Hydrol Earth Syst Sci 14:1205–1219. https://doi.org/10.5194/hess-14-1205-2010
DeBeer CM, Pomeroy JW (2017) Influence of snowpack and melt energy heterogeneity on snow cover depletion and snowmelt runoff simulation in a cold mountain environment. J Hydrol 553:199–213. https://doi.org/10.1016/j.jhydrol.2017.07.051
Dobbert S, Pape R, Löffler J (2021) Contrasting growth response of evergreen and deciduous arctic-alpine shrub species to climate variability. Ecosphere 12:e03688. https://doi.org/10.1002/ecs2.3688
Elmendorf SC, Henry GHR, Hollister RD et al (2012) Plot-scale evidence of tundra vegetation change and links to recent summer warming. Nat Clim Change 2:453–457. https://doi.org/10.1038/nclimate1465
Erickson TA, Williams MW, Winstral A (2005) Persistence of topographic controls on the spatial distribution of snow in rugged mountain terrain, Colorado, United States. Water Resour Res 41. https://doi.org/10.1029/2003WR002973
Francon L, Corona C, Till-Bottraud I et al (2021) Shrub growth in the Alps diverges from air temperature since the 1990s. Environ Res Lett 16:074026. https://doi.org/10.1088/1748-9326/ac0b67
Gamm CM, Sullivan PF, Buchwal A et al (2018) Declining growth of deciduous shrubs in the warming climate of continental western Greenland. J Ecol 106:640–654. https://doi.org/10.1111/1365-2745.12882
Gentili R, Baroni C, Panigada C et al (2020) Glacier shrinkage and slope processes create habitat at high elevation and microrefugia across treeline for alpine plants during warm stages. CATENA 193:104626. https://doi.org/10.1016/j.catena.2020.104626
Good M, Morgan JW, Venn S, Green P (2019) Timing of snowmelt affects species composition via plant strategy filtering. Basic Appl Ecol 35:54–62. https://doi.org/10.1016/j.baae.2019.01.004
Hallinger M, Manthey M, Wilmking M (2010) Establishing a missing link: warm summers and winter snow cover promote shrub expansion into alpine tundra in Scandinavia. New Phytol 186:890–899. https://doi.org/10.1111/j.1469-8137.2010.03223.x
Inouye DW (2008) Effects of Climate Change on Phenology, Frost Damage, and Floral Abundance of Montane Wildflowers. Ecology 89:353–362. https://doi.org/10.1890/06-2128.1
Inouye DW (2020) Effects of climate change on alpine plants and their pollinators. Ann N Y Acad Sci 1469:26–37. https://doi.org/10.1111/nyas.14104
Körner C (2003a) Life under snow: Protection and limitation. In: Körner C (ed) Alpine Plant Life: Functional Plant Ecology of High Mountain Ecosystems. Springer, Berlin, Heidelberg, pp 47–62
Körner C (2003b) The climate plants experience. In: Körner C (ed) Alpine Plant Life: Functional Plant Ecology of High Mountain Ecosystems. Springer, Berlin, Heidelberg, pp 31–46
Kudo G, Ito K (1992) Plant distribution in relation to the length of the growing season in a snow-bed in the Taisetsu Mountains, northern Japan. Vegetatio 98:165–174. https://doi.org/10.1007/BF00045554
Lefcheck JS (2016) piecewiseSEM: Piecewise structural equation modelling in r for ecology, evolution, and systematics. Methods Ecol Evol 7:573–579. https://doi.org/10.1111/2041-210X.12512
Liu L, Xu X, Zhang L et al (2022) Global patterns of species richness of the holarctic alpine herb Saxifraga: the role of temperature and habitat heterogeneity. J Plant Ecol 15:237–252. https://doi.org/10.1093/jpe/rtab085
MacDonald AAC, Stralberg D, Nielsen SE (2025) Alpine Summer Surface Temperature Amplification Is Spatially Heterogeneous and Intensified by Wind and Sun. Ecol Evol 15:e72542. https://doi.org/10.1002/ece3.72542
Mallik AU, Wdowiak JV, Cooper EJ (2011) Growth and Reproductive Responses of Cassiope tetragona, a Circumpolar Evergreen Shrub, to Experimentally Delayed Snowmelt. Arct Antarct Alp Res 43:404–409. https://doi.org/10.1657/1938-4246-43.3.404
Martin AC, Jeffers ES, Petrokofsky G et al (2017) Shrub growth and expansion in the Arctic tundra: an assessment of controlling factors using an evidence-based approach. Environ Res Lett 12:085007. https://doi.org/10.1088/1748-9326/aa7989
McManus K, Morton D, Masek J et al (2012) Satellite-based evidence for shrub and graminoid tundra expansion in northern Quebec from 1986 to 2010. Glob Change Biol 18:2313–2323. https://doi.org/10.1111/j.1365-2486.2012.02708.x
Morgan J, Walker Z (2023) Early-melting snowpatch plant communities are transitioning into novel states. Sci Rep 13:16520. https://doi.org/10.1038/s41598-023-42808-5
Moura MR, Villalobos F, Costa GC, Garcia PCA (2016) Disentangling the Role of Climate, Topography and Vegetation in Species Richness Gradients. PLoS ONE 11:e0152468. https://doi.org/10.1371/journal.pone.0152468
Myers-Smith IH, Forbes BC, Wilmking M et al (2011) Shrub expansion in tundra ecosystems: dynamics, impacts and research priorities. Environ Res Lett 6:045509. https://doi.org/10.1088/1748-9326/6/4/045509
Myers-Smith IH, Grabowski MM, Thomas HJD et al (2019) Eighteen years of ecological monitoring reveals multiple lines of evidence for tundra vegetation change. Ecol Monogr 89:e01351. https://doi.org/10.1002/ecm.1351
Myers-Smith IH, Hik DS (2018) Climate warming as a driver of tundra shrubline advance. J Ecol 106:547–560. https://doi.org/10.1111/1365-2745.12817
Myers-Smith IH, Hik DS (2013) Shrub canopies influence soil temperatures but not nutrient dynamics: An experimental test of tundra snow–shrub interactions. Ecol Evol 3:3683–3700. https://doi.org/10.1002/ece3.710
Notarnicola RF, Nicotra AB, Kruuk LEB, Arnold PA (2021) Tolerance of Warmer Temperatures Does Not Confer Resilience to Heatwaves in an Alpine Herb. Front Ecol Evol 9
Pardee GL, Inouye DW, Irwin RE (2018) Direct and indirect effects of episodic frost on plant growth and reproduction in subalpine wildflowers. Glob Change Biol 24:848–857. https://doi.org/10.1111/gcb.13865
Post E, Forchhammer MC, Bret-Harte MS et al (2009) Ecological Dynamics Across the Arctic Associated with Recent Climate Change. Science 325:1355–1358. https://doi.org/10.1126/science.1173113
Press Mc, Potter Ja, Burke M et al (1998) j. w., Responses of a subarctic dwarf shrub heath community to simulated environmental change. J Ecol 86:315–327. https://doi.org/10.1046/j.1365-2745.1998.00261.x
R Core Team (2024) R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.r-project.org/
Saarinen T, Rasmus S, Lundell R et al (2016) Photosynthetic and phenological responses of dwarf shrubs to the depth and properties of snow. Oikos 125. https://doi.org/10.1111/oik.02233
Scharnagl K, Johnson D, Ebert-May D (2019) Shrub expansion and alpine plant community change: 40-year record from Niwot Ridge, Colorado. Plant Ecol Divers 12:407–416. https://doi.org/10.1080/17550874.2019.1641757
Sholto-Douglas C, Shackleton CM, Ruwanza S, Dold T (2017) The Effects of Expansive Shrubs on Plant Species Richness and Soils in Semi-arid Communal Lands, South Africa. Land Degrad Dev 28:2191–2206. https://doi.org/10.1002/ldr.2745
Starr G, Oberbauer SF (2003) Photosynthesis of Arctic Evergreens Under Snow: Implications for Tundra Ecosystem Carbon Balance. Ecology 84:1415–1420. https://doi.org/10.1890/02-3154
Sturm M, Douglas T, Racine C, Liston GE (2005) Changing snow and shrub conditions affect albedo with global implications. J Geophys Res Biogeosciences 110. https://doi.org/10.1029/2005JG000013
Tape KD, Hallinger M, Welker JM, Ruess RW (2012) Landscape Heterogeneity of Shrub Expansion in Arctic Alaska. Ecosystems 15:711–724. https://doi.org/10.1007/s10021-012-9540-4
Tremblay B, Lévesque E, Boudreau S (2012) Recent expansion of erect shrubs in the Low Arctic: evidence from Eastern Nunavik. Environ Res Lett 7:035501. https://doi.org/10.1088/1748-9326/7/3/035501
Verrall B, Green K, Pickering CM (2023) Temporal dynamics in alpine snowpatch plants along a snowmelt gradient explained by functional traits and strategies. Oecologia 201:155–171. https://doi.org/10.1007/s00442-022-05297-3
Verrall B, Pickering CM (2020) Alpine vegetation in the context of climate change: A global review of past research and future directions. Sci Total Environ 748:141344. https://doi.org/10.1016/j.scitotenv.2020.141344
Wang X, Michalet R, Liu Z et al (2019) Stature of dependent forbs is more related to the direct and indirect above- and below-ground effects of a subalpine shrub than are foliage traits. J Veg Sci 30:403–412. https://doi.org/10.1111/jvs.12739
Weinig C (2000) Differing Selection in Alternative Competitive Environments: Shade-Avoidance Responses and Germination Timing. Evolution 54:124–136. https://doi.org/10.1111/j.0014-3820.2000.tb00013.x
Wheeler JA, Cortés AJ, Sedlacek J et al (2016) The snow and the willows: earlier spring snowmelt reduces performance in the low-lying alpine shrub Salix herbacea. J Ecol 104:1041–1050. https://doi.org/10.1111/1365-2745.12579
Wheeler JA, Hoch G, Cortés AJ et al (2014) Increased spring freezing vulnerability for alpine shrubs under early snowmelt. Oecologia 175:219–229. https://doi.org/10.1007/s00442-013-2872-8
Wipf S, Stoeckli V, Bebi P (2009) Winter climate change in alpine tundra: plant responses to changes in snow depth and snowmelt timing. Clim Change 94:105–121. https://doi.org/10.1007/s10584-009-9546-x
Wookey PA, Robinson CH, Parsons AN et al (1995) Environmental constraints on the growth, photosynthesis and reproductive development of Dryas octopetala at a high Arctic polar semi-desert. Svalbard Oecologia 102:478–489. https://doi.org/10.1007/BF00341360
Zong S, Lembrechts JJ, Du H et al (2022) Upward range shift of a dominant alpine shrub related to 50 years of snow cover change. Remote Sens Environ 268:112773. https://doi.org/10.1016/j.rse.2021.112773
Appendix A
Supplemental Methods of Classifying Snow Cover (Persistence) from Remote Sensing Images
We developed a method to generate binary raster images that distinguish snow from non-snow pixels across a series of remote sensing images (dates) using geographic information systems (GIS) and automation in R. Figure 5 below provides a visualization of the workflow.
Supplemental Figures & Tables
A
A
Fig. 7
Direct effects (dots) between variables and herbaceous cover (a) and herbaceous richness (b). Colored lines indicate indirect effect pathways that either add to or subtract from the direct effect (whisker lines and bars).
