Structural shifts in plant functional diversity during biogeomorphic succession: Moving beyond taxonomic investigations in an alpine glacier foreland
Abstract
The complex interrelation between plants and geomorphic processes is described in the concept of biogeomorphic succession. While ecological research on succession and community assembly has transitioned towards functional approaches, studies on functional diversity in biogeomorphic settings, particularly in glacier forelands, remain limited. In this study, we investigated abundance of vascular plant species and functional traits in an alpine glacier foreland using data from 199 plots. Our objective was to unravel the development of functional diversity during biogeomorphic succession. Specifically, the study determined whether structural shifts in functional diversity are associated with stability thresholds related to plant cover, geomorphic influence, and examined trait spectra for stages of biogeomorphic succession. Our findings revealed a nonlinear trajectory of functional diversity along the plant cover gradient, marked by two distinct structural shifts at 30% and 74% cover, corresponding to established stability thresholds. Along the gradient of geomorphic influence, we observed an increase in functional diversity until 54% of the plot area was affected, beyond which functional diversity declined below the initial level. The analysis of community-weighted means of traits across four stages of biogeomorphic succession determined by plant cover and absence and presence of geomorphic influence revealed significant differences in trait values. In the transition to the biogeomorphic stage, associated with the identified initial structural shift, there is a shift from a prevalence of above-ground adaptation and reproductive traits, such as leaf longevity, structure, growth form and mixed reproductive strategies, to an increased dominance of competitor species and traits related to below-ground structures, including root type and structures, as well as vegetative reproduction. Our results contribute to understanding the relationship between vegetation succession and geomorphic influence by linking them to plant functional traits. This study advances beyond traditional taxonomic investigations by emphasizing functional approaches to biogeomorphic succession. Moreover, the functional trait data used in this study, easily downloadable from a public repository, can serve as a valuable template for future research in (bio)geomorphology, along with the employed methodologies.
1 INTRODUCTION
Biogeomorphology provides useful frameworks for studying landscape dynamics and establishes a link between biotic and abiotic processes (Coombes, 2016; Larsen et al., 2020; Stallins, 2006). The concept of biogeomorphic succession explores the temporal dynamics of the interaction between vegetation colonization and geomorphic processes (Corenblit et al., 2007; Corenblit et al., 2015; Corenblit & Steiger, 2009). These dynamics involve shifts in the frequency and magnitude of geomorphic processes, as well as alterations in plant composition (Eichel et al., 2016; Miller & Lane, 2018). The nonlinearity of biogeomorphic succession is associated with thresholds that might trigger structural shifts (Balke et al., 2011; Groffman et al., 2006). When plant cover exceeds specific thresholds, geomorphic processes decrease, leading to the stabilization of the terrain (Eichel, Draebing, et al., 2023; Hanusch et al., 2022; Haselberger et al., 2021; Ohler et al., 2023). Understanding changes in vegetation composition and their timing is essential to comprehend the implications for landscape development (Groffman et al., 2006).
Glacier forelands, marked by a gradient that extends from bare substrate and pioneer species to mature vegetation, offer the opportunity to track transformations in community composition, diversity and function over time (Chapin et al., 1994; Fastie, 1995; Jones & Del Moral, 2005; Junker et al., 2020; Matthews, 1992). In these areas, allogenic factors, especially the availability of resources like nutrients, water and substrate, restrict vegetation colonization (Erschbamer & Caccianiga, 2017; Wojcik et al., 2021). Additionally, disturbances from surface processes, including fluvial sediment transport, rockfalls and debris flows, contribute to the landscape's complexity (Carrivick & Heckmann, 2017; Wojcik et al., 2021). This leads to a patchwork of varying successional stages, which may not always correspond to a terrain age gradient (Ficetola et al., 2021; Rydgren et al., 2014; Wojcik et al., 2020). Plant species colonizing these areas possess functional traits that enable them to adapt to environmental constraints and/or modify habitat conditions by influencing geomorphic processes (Corenblit et al., 2015; Eichel et al., 2016; Losapio et al., 2021).
Ecosystems with substantial functional diversity exhibit resilience by maintaining functional redundancy and response diversity, serving as buffers against declines in ecosystem functions following geomorphic disturbance (Biswas & Mallik, 2010; Stallins & Corenblit, 2018). Response traits, such as dispersal, reproduction and root morphology, facilitate establishment in resource-limited and disturbed environments (Bernhardt-Römermann et al., 2011; Losapio et al., 2021; Wojcik et al., 2021). Effect traits directly modify the abiotic environment, including root stabilization, leaf protection against splash erosion and sediment binding through diverse vegetation dispersal modes (Burylo et al., 2012; Isselin-Nondedeu & Bédécarrats, 2007; Ohler et al., 2023).
Thus, plant communities' functional traits reveal ecological strategies and their environmental effects (Bruelheide et al., 2018; Diehl et al., 2017). Understanding these variations in plant traits among species and sites is essential for comprehending their influence on landforms and geomorphic processes (Violle et al., 2007; Westoby & Wright, 2006). Functional diversity metrics, which capture species distribution and abundances in functional space, provide more powerful indicators of ecosystem functioning and resilience than taxonomic metrics (Dı́az & Cabido, 2001; Mason et al., 2013; Tilman et al., 1997).
While geomorphic processes' influence on plant community traits has been studied in various environments (Corenblit et al., 2015; Diehl et al., 2017; Kemppinen et al., 2021; Schwarz et al., 2018), such investigations are scarce in glacier forelands (Eichel, Draebing, et al., 2023; He et al., 2023; Ohler et al., 2023).
Incorporation of a trait-based approach has significant potential for advancing biogeomorphology (Larsen et al., 2020; Viles & Coombes, 2022). Previous studies in glacier forelands addressing the role of geomorphic processes in vegetation development have mainly focused on specific landforms and provided descriptive accounts of functional traits during stages of biogeomorphic succession (for a comprehensive overview, see Eichel, 2019; Eichel, Stoffel, & Wipf, 2023; Miller & Lane, 2018). With traits including clonal growth, burial tolerance, resprouting capacity and mat formation, Dryas octopetala L. has been recognized as an ecosystem engineer on lateral moraine slopes, influencing hillslope processes through sediment trapping, run-off reduction and moisture storage (Eichel et al., 2016). The initiation of turf-banked solifluction is linked to the related material accumulation (Eichel et al., 2017). Stabilization on lateral moraine slopes occurs stepwise in association with distinct species communities. These species communities, along with their traits and related on geomorphic processes, have been observed as common across ecosystems facing similar constraints (Eichel, Stoffel, & Wipf, 2023). Despite the identification of functional traits relevant for biogeomorphic succession, there is a notable research gap concerning the variability of trait diversity in vegetation communities throughout the biogeomorphic succession in glacier forelands.
We propose that stability thresholds identified at 30% to 40% and 70% to 80% plant cover by Eichel et al. (2016), Haselberger et al. (2021) and Eichel, Draebing, et al. (2023) during biogeomorphic succession are linked to structural shifts in the functional diversity and composition of functional traits within plant communities in glacier forelands. To test this hypothesis, we investigated successional changes in a small cirque-like glacier foreland in the Kaunertal Valley in Austria, where a broad spectrum of geomorphic processes and a mosaic of plant communities is present within a relatively compact, contiguous landscape. Consequently, the vegetation composition analysed in this study reflects a blend of primary and secondary succession, with secondary succession prominently observed in areas where geomorphic processes have reset primary succession. Our objectives included testing the presence of structural shifts in functional diversity along a plant cover gradient and a gradient of geomorphic influence. Additionally, we strived to analyse changes in the abundance of functional traits throughout different stages of biogeomorphic succession. Our approach aimed to promote trait-based analyses in (bio)geomorphology by gathering and discussing readily accessible effect traits and response traits. Additionally, we adopt a methodology from plant ecological research to analyse structural shifts and trait spectra throughout biogeomorphic succession.
2 STUDY AREA
The Kaunertal Valley, situated in Tyrol, Austria, is a north-to-south-oriented valley in the Eastern European Alps (Figure 1a). The study area is a hanging valley, located west of the Upper Kaunertal Valley (Figure 1b), which is covered by the Gepatschferner glacier, the second-largest glacier in Austria. The study catchment spans an elevation gradient from 2400 to 3200 m a.s.l. and covers an area of approximately 3 km2. The area is geologically homogeneous, mainly consisting of paragneiss and orthogneiss (Hammer, 1923) and belongs to the eastern Alps crystalline zone. During the Little Ice Age (LIA) maximum extent, one third of the catchment was ice covered. Since then, the glacier in the hanging valley has retreated by about 1500 m. More recent and detailed retreat rates of the glacier snout located in the main valley show a mean annual retreat rate of 50 m per year since 2000, with a maximum of 125 m in 2017 (Kellerer-Pirklbauer, 2019). Dusik et al. (2019) found that ground ice is likely to be present in the whole catchment on north-facing slopes, whereas south-facing slopes and flat terrain are less likely to contain ground ice. The study area's landforms are dominated by lateral and terminal moraine complexes, talus slopes, debris flow channels and fans, glaciofluvial incisions and floodplains (Baewert & Morche, 2014), which were formed during and after paraglacial adjustment. Peri-glacial landforms, such as thermokarst, solifluction lobes and an active rock glacier, are present, yet surface changes in the area are primarily characterized by water-mediated sediment transport, as indicated by the geomorphological mapping of Haselberger et al. (2023). The moraine and till material is poorly sorted, with grain sizes ranging from silt to boulders, and the proglacial area has weakly developed soils (Temme et al., 2016).
The study area has a central-alpine dry climate (Fliri, 1975). The meteorological station at Weisssee (2470 m a.s.l.), located 1.5 km west of the study area, recorded a long-term mean annual air temperature of −0.05°C (2007–2021) and mean annual precipitation rates ranging from 731 to 1118 mm (data source: Tiroler Wasserkraft AG). Precipitation maxima with more than 45 mm precipitation in 24 h occur during the summer months of July and August (Haselberger et al., 2021). The growing season for vegetation extends from mid-May to October, due to snow cover. The area is located above the local treeline and exhibits typical patterns of primary succession, with Cerastium uniflorum, Poa alpina and Veronica alpina being the dominant species in early successional stages. In mid-successional stages, Salix herbacea, P. alpina and Gnaphalium supinum dominate, while Nardus stricta, Cirsium spinosissimum and P. alpina dominate in late successional stages (outside LIA extent, Ohler et al., 2023). Local shepherds use the area beyond the LIA extent for grazing, supporting approximately 40 cows for a 2-week period each August season. The area used for analysis is confined by steep rock slopes, which limit the availability of manual and UAV-derived vegetation data (Figure 1a).
3 METHODS
3.1 Field observations and trait collection
In August 2021 and August 2022, a total of 199 plots ranging from 2440 to 2880 m a.s.l. were observed during the peak growing season (Figure 1a). To ensure an even distribution of sample locations representing various terrain ages post-glacier retreat within the catchment, a regular sample grid of 100 m × 100 m was predefined. Potential sampling locations were established at the central points of this grid (n = 368). Locations on water, bedrock or ice surfaces, as well as those on inaccessible steep scree and bedrock slopes with angles exceeding 50°, were omitted. To achieve a balanced distribution of both geomorphically active and inactive plots, we utilized a geomorphological map from Haselberger et al. (2023) and randomly selected 50% of locations as active and 50% as inactive prior to the sampling process. During field data collection, plots were placed near predetermined sampling locations to accurately reflect the surrounding conditions, encompassing vegetation cover and species composition, and to reflect predefined geomorphic activity. If any location exhibited signs of grazing, we excluded those specific plots from our analysis.
The presence of geomorphic processes on each plot was assessed, resulting in 105 plots showing geomorphic influence and 94 plots showing no geomorphic influence. Geomorphic influence was determined based on geomorphological processes that frequently occur in proglacial areas and affect vegetation succession by moving sediment, namely rock fall, debris flows, fluvial process and slope wash. Given the limited occurrence of peri-glacial processes in the catchment, we adopted an a priori approach by excluding areas featuring peri-glacial landforms based on the geomorphological map by Haselberger et al. (2023). Additionally, we excluded locations displaying signs of peri-glacial activity during our fieldwork. For rockfall activity, the presence of freshly eroded rock fragments or fresh scars on rocky outcrops was documented. Debris flow activity was noted when unsorted sediment accumulations and the presence of deep channels and neighbouring levees were observed. Fluvial and slope wash activity was recorded when visible flow lines, deeper channels, the presence of flowing water and sorted sediment accumulations were detected. On older terrain, geomorphic influence is closely and inversely related to plant cover. Recognizing the impact of initial site conditions on plant colonization, especially on younger terrain, emphasizes the significance of geomorphic influence in understanding abiotic disturbances.
Finally, the extent to which each process affected the topsoil of a given plot was quantified as a proportion of the plot surface via visual inspection proposed by Virtanen et al. (2010) and Kemppinen et al. (2021). This cover of geomorphic influence ranges from 0% to 100% coverage (Table S1).
For each 1 × 1 m plot, we identified all vascular plant species and estimated their cover and the cover of functional groups based on their growth form (woody species, graminoids and forbs, Table S2). Visual estimation was employed, involving at least two observers contributing to an averaged cover estimate, a practice aimed at ensuring reproducibility in results. Sparsely distributed species with less than 1% cover were recorded with an interval of 0.5% cover, and rare species with few single individuals per plot were documented with an interval of 0.1%. The taxonomic nomenclature of vascular plants is based on Fischer et al. (2005). Further recorded parameters for the plots included maximum height of vegetation using a folding ruler, number of different species and ground covered with pebbles (>4 mm, Wentworth, 1922). To assess taxonomic diversity, we calculated the Shannon diversity index (Shannon, 2001), which quantifies the richness and evenness of species abundance across plots (R package vegan, Oksanen, 2012).
We used BIOLFLOR (Klotz et al., 2002), Flora indicativa (Landolt et al., 2010), LEDA (Kleyer et al., 2008), TRY (Kattge et al., 2011) databases and data from Ohler et al. (2023) and compiled a total of 15 traits for all 76 plant species (Table S3). The traits selected for this study were based on their significance in primary succession colonization and biogeomorphic interaction, encompassing those that exhibited responses to and impacts on geomorphic processes (Table 1). Categorical traits (n = 11) were dummified and ordinal traits (n = 4) and means of continuous traits were transformed by log10 prior to analysis (R package vegan, Oksanen, 2012). Given that all plots are located within the alpine growing habitat and the cover gradient is randomly distributed across the study area, we employed trait data without additional filtering.
| Group | Trait | Variable description | Attributes | Response to geomorphic processes | Selected literature | Effect on geomorphic processes | Selected literature |
|---|---|---|---|---|---|---|---|
| Life history | Growth forma | Categorical (three classes) | Forbs, graminoids and woody | Resource acquisition, adaptations to disturbance or seasonal changes in climate and life span | For example, Raunkiaer (1934), Lavorel et al. (1997), McIntyre et al. (1995), Polvi et al. (2014) | Potential to cover bare soil and anchor loose sediment | For example, Lavorel et al. (1997), McIntyre et al. (1995), Polvi et al. (2014), Stokes et al. (2009) |
| Life formb | Categorical (four classes) | Therophytes, hemicryptophytes, phanerophytes and chamaephytes | Survival during unfavourable seasons, resilience to disturbance or seasonal changes in climatic parameters and life span | For example, Raunkiær (1910, 1934), Noble and Slatyer (1980), McIntyre et al. (1995), Kervroëdan et al. (2018) | Facilitate plant establishment and subsequent stabilization, aiding in vegetation-driven resistance against sediment transport | For example, McIntyre et al. (1995), De Bello et al. (2023), Gilardelli et al. (2015), Kervroëdan et al. (2018) | |
| CSRc | Categorical (four classes) | Competitor, stress tolerator, ruderal and mixed CSR | Adaptation to environmental disturbance, resource stress and competition with other species | For example, Grime (1973), Caccianiga et al. (2006), Grime (2006), | Facilitate plant establishment and subsequent stabilization, aiding in vegetation-driven resistance against sediment transport | For example, Corenblit et al. (2015) | |
| Canopy and leaf morphology | Leaf persistenceb | Categorical (binary) | Deciduous and evergreen | Affects a plant's ability to respond to seasonal variations and disturbances | For example, Wright et al. (2004), Nelson et al. (2007), Catorci et al. (2017) | Persistent rainfall interception | For example, Nelson et al. (2007), Catorci et al. (2017) |
| Leaf anatomyb | Categorical (four classes) | Succulent, scleromorphic, mesomorphic and hygromorphic | Impact a plant's resistance to abrasion and physical damage from geomorphic processes | For example, Catorci et al. (2017), Wright et al. (2004) | Protection against splash erosion | For example, Foot and Morgan (2005), Burylo et al. (2012), Catorci et al. (2017) | |
| Canopy structure/rosettesb | Categorical (binary) | Erosulate and rosettes/hemirosettes | Affect a plant's ability to dissipate energy from water flow, influencing its response to geomorphic forces | For example, Polvi et al. (2014), Burylo et al. (2012) | Protection against splash erosion and sediment trapping | For example, Bochet et al. (2000), Isselin-Nondedeu and Bédécarrats (2007), Burylo et al. (2012), Polvi et al. (2014) | |
| Heightd,e | Continuous | m | Influence exposure to water flow and sediment transport, affecting vulnerability and response to geomorphic processes | For example, Westoby and Wright (2006), Schenk and Jackson (2002b), Vesk (2006), Kervroëdan et al. (2018), Nardin and Edmonds (2014) | Submergence condition, velocity profile, depositional patterns, protection against splash erosion and higher resilience against disturbance | For example, Nepf and Vivoni (2000), Diehl et al. (2017), Schenk and Jackson (2002b), Kervroëdan et al. (2018), Nardin and Edmonds (2014) (Diehl et al., 2017) | |
| Leaf dry matter content (LDMC) | Continuous | g/g | Reflects the structural and resource allocation strategy of leaves, influencing their resistance to damage from geomorphic processes | For example, Ohler et al. (2023), Burylo et al. (2012), Ciccarelli (2015), Garnier et al. (2004) | High LDMC is related to lower decomposition and higher litter accumulation, hence decreasing kinetic energy of raindrops | For example, Cornelissen et al. (2003), Ohler et al. (2023), Burylo et al. (2012), Quétier et al. (2007) | |
| Specific leaf area (SLA) | Continuous | cm2/mg | Influence a plant's capacity to capture resources and respond to disturbances | For example, Ohler et al. (2023), Ciccarelli (2015), Kyle and Leishman (2009), Wright et al. (2004), Garnier et al. (2004) | SLA is associated with faster growth rates, facilitating rapid progression towards a closed canopy. Additionally, SLA is associated with enhanced sediment retention, through increased hydraulic roughness | For example, Chapin et al. (1993), Erktan et al. (2013), Kervroëdan et al. (2018), Garnier et al. (2004), Quétier et al. (2007) | |
| Vegetative morphology | Shoot meta-morphosesb | Categorical (three classes) | Runner, tuft and rhizome | Adaptations to disturbance or seasonal changes in climatic parameters | For example, Bouma et al. (2013), Gurnell (2014), Corenblit et al. (2015), Catorci et al. (2017) | Lower pore water pressure, increase soil aggregate stability and provide additional soil cohesion through root reinforcement, fix topsoil superficially and laterally with horizontally spreading roots | For example, Gyssels et al. (2005), Burylo et al. (2012), Gurnell (2014) |
| Vegetative spreadc | Categorical (five classes) | Basal shoots, runners, tufts/cushions and no veg. dispersal | Affects potential to colonize disturbed areas | For example, Haugland (2006), Bouma et al. (2013), Gurnell (2014), Eichel et al. (2016), Kervroëdan et al. (2018) | Sediment anchoring and trapping and protection against splash erosion | For example, Haugland (2006), Bouma et al. (2013), Gurnell (2014), Kervroëdan et al. (2018), Schwarz et al. (2018) | |
| Root morphology | Root depthc | Ordinal | 1–4 | Enhancing ability to withstand erosive forces and access to water and nutrients, enhancing resistance to geomorphic processes | For example, Wright and Westoby (1999); Gurnell (2014), Holloway et al. (2017), Schenk and Jackson (2002a), Stokes et al. (2009) | Mechanical stability | For example, Diehl et al. (2017), Stokes et al. (2009), Gyssels et al. (2005), Baets et al. (2007), Vannoppen et al. (2017) |
| Root architecture | Categorical (three classes) | Fibrous, tap and mixed roots | |||||
| Reproduction | establishmentb | Categorical (three classes) | Seed, vegetative and mixed | Increase chances of successful establishment and tempo of recolonization after disturbances | For example, Klimešová et al. (2017), Caccianiga et al. (2006), Erschbamer and Mayer (2011), Corenblit et al. (2015) | Facilitate plant establishment and subsequent stabilization, aiding in vegetation-driven resistance against sediment transport | For example, Caccianiga et al. (2006), Schwienbacher et al. (2011), Erschbamer and Mayer (2011), Schwarz et al. (2018) |
| Seed productionb | Categorical (binary) | High seed capacity and low seed capacity |
3.2 Analyses of community-weighted means (CWMs) and functional diversity
To assess the average trait values for plant communities, we calculated CWMs using the R package FD (Laliberté et al., 2014). We used gower distances for trait dissimilarities of mixed data based on the gawdis function (R package gawdis, De Bello et al., 2023). Gawdis considers optimal weights and trait correlations via groups. Groups of traits were predefined and reflect a species' life history, canopy and leaf structure, vegetative characteristics, root structure and traits related to establishment and reproduction (Table 1). CWM represents the average trait value within the community, weighted by species abundances. We used functional dispersion (FDIS) as index for functional diversity, which was computed using the R package FD (Laliberté et al., 2014). Only sites with at least two species were included into the calculation. FDIS assesses how different the functional traits are among species, indicating the range of functional strategies present (Mason et al., 2013).
3.3 Analysis of functional diversity across gradients of plant cover and geomorphic influence
- m1 assumes a constant linear increase of the independent variable without a change point;
- m2 considers one shift in the independent variable with an initial increase followed by an immediate decrease;
- m3 incorporates one shift with an initial increase of the independent variable and a shift to a plateau, following a broken-stick model;
- m4 Incorporates two shifts, representing an initial increase of the independent variable, a shift to a plateau and a subsequent late decrease, forming a three-segment broken-stick model.
For each model, we performed three separate runs of Markov chain Monte Carlo estimators with a uniform prior, totalling n = 10 000 generations. We discarded the initial n = 2000 generations as burn-in. To assess model convergence, we visually examined the trace plots and ensured that all model parameters had reached a state of stationarity. To evaluate the predictive performance of the four models, we employed leave-one-out cross-validation (LOOCV; Vehtari et al., 2017). Additionally, we employed Bayes Factors to precisely determine the location of structural shifts and relate them to known stability thresholds identified at 30% to 40% and 70% to 80% plant cover (Eichel, Draebing, et al., 2023). We applied a uniform prior, assigning equal probability density to all potential values across the entire range of the plant cover gradient, treating each value as equally likely before any data observation.
3.4 Analysis of functional diversity across stages of biogeomorphic succession
To test trait spectra differences during biogeomorphic succession, we classified vegetation plots into four groups representing specific stages during biogeomorphic succession, which were conceptualized by Haselberger et al. (2023). The groups were determined by two factors: a plant cover gradient (low: <35% plant cover; high: >35% plant cover) and absence and presence of geomorphic influence. 35% plant cover threshold was selected as limit, as it signifies the initial stability threshold at which geomorphic activity is reduced (Haselberger et al., 2021). Given that rockfalls, debris flows, fluvial dynamics and slope wash operate on much larger scales than our 1 × 1 m vegetation plot (Ballantyne, 2002), we believe that any presence would markedly affect the vegetation both within the plot and its surrounding area. Therefore, in the event of any geomorphic processes, we classified the entire plot as geomorphologically influenced. The focus on the interplay between geomorphic influence and vegetation cover reflects the prevalent conceptual notion that the dominance between abiotic and biotic processes shifts during primary succession (Matthews, 1992; Miller & Lane, 2018; Raab et al., 2013; Wojcik et al., 2021). In accordance with the concept of biogeomorphic succession and Haselberger et al. (2023) classification, we assume that a scenario before the first stability threshold, coupled with present geomorphic influence, represents the geomorphic or pioneer stage. Contrarily, a scenario preceding the first stability threshold without geomorphic influence does not fall within the scope of biogeomorphic succession; instead, represents a separate successional pathway driven by initial site conditions Beyond the first stability threshold, present geomorphic influence represents the biogeomorphic stage, whereas no geomorphic activity beyond this threshold indicates the ecologic stage.
Differences in the multivariate trait spectrum of the four groups were tested using canonical analysis of principal coordinates, which aims to find linear combinations of principal coordinate axes that maximize the differences between predefined groups in the original data space (CAP, Anderson & Willis, 2003). Statistical significance was determined using a permutation test (n = 999 permutations) with a Bray–Curtis distance to account for non-parametric data and group structure. Group membership was included as a fixed factor. CAP loadings were examined to identify the specific plant traits driving the differences. Differences of single CWM traits between groups were tested using multilevel pairwise comparison (permanova). Analysis was conducted using the ‘vegan’ package in R (Oksanen, 2012) with a significance level of alpha = 0.05.
4 RESULTS
4.1 Functional diversity across a plant cover gradient
Functional dispersion exhibits a nonlinear trajectory along the plant cover gradient, characterized by two distinct structural shifts. The first change point was identified at a relative plant cover of 30% (with lower and upper bounds of 19.5% and 46% cover, respectively), while a second, less pronounced change point was observed at a relative plant cover of 74% (with lower and upper bounds of 29.5% and 88.7% cover, respectively; Figure 2 and Table S4). Resulting piecewise linear models show a significant linear increase in functional dispersion during low levels of plant cover (first segment: p < 0.001, r2 = 0.27), a stationary functional dispersion for mid-levels of plant cover (non-significant relationship with increasing cover; second segment: p = 0.63, r2 = 0.004) and a significant decrease in functional dispersion for the high levels of plant cover (third segment: p < 0.001, r2 = 0.20; Figure 2 and Table S4). Taxonomic diversity is associated with functional diversity (r = 0.72, p < 0.001) and shows a similar initial increase till 34% cover followed by a plateau and second shift at 50% cover beyond which taxonomic diversity declines (Table S2 and Figure S4), while exhibiting a weaker correlation with plant cover (r = 0.26, p < 0.001). CWM of traits exhibits nonlinearity along the plant cover gradient, with distinct model relationships indicated by LOOCV (Table S4). Significant changes occur at stability thresholds identified by change point analysis and Bayes factors (Figure S1). The following traits show structural shifts near the stability thresholds: rhizome-like shoots (36%), root depth (68%), chamaephytes (71%), therophytes (78%), woody species (76%) and mixed mode of reproduction (68%).
4.2 Functional diversity in relation to geomorphic influence
Functional dispersion along the gradient of geomorphic influence is characterized by a nonlinear trajectory with a distinct structural shift at 54% (with lower and upper bounds of 39.5% and 69% cover, respectively) as demonstrated by LOOCV (Figure 3). Resulting piecewise linear models show a slight and not significant linear increase in functional dispersion during low levels of geomorphic influence and a significant decrease in functional dispersion for the high levels of geomorphic influence (first segment: p = 0.54, r2 = 0.003 and second segment: p < 0.001, r2 = 0.17; Table S4). Taxonomic diversity correlates with geomorphic influence (r = 0.22, p < 0.001).
Comparison of plots where geomorphic influence is recorded and where geomorphic influence is not recorded show no significant differences in functional dispersion according to Wilcoxon rank-sum test (Table S4).
4.3 Functional diversity across stages of biogeomorphic succession
The CAP analysis reveals that the four stages explain 12.12% of trait dissimilarity, with a total inertia of 7.75 (Table S4). CAP1 explains 92.81% of the variation, and the following traits show the highest positive correlation with CAP1: evergreen leaves (0.787205), chamaephytes (0.606419), mixed reproduction (0.585488), LDMC (0.444879) and forbs (0.372044). Traits with the highest negative correlation with the first axis include root depth (−0.903137), basal shoots (−0.463666), rhizome (−0.424026), competitors (−0.376355) and vegetative reproduction (−0.364508). When comparing the four stages of biogeomorphic succession, significant differences were found for all groups, except those with low plant cover and no geomorphic influence present. However, pairwise t tests reveal differences in certain traits between groups with high plant cover and either high or low geomorphic influence (Figure 4 and Table S4). Specifically, hygromorphic leaves, vegetative reproduction, competitor species and plant height show significant differences.
5 DISCUSSION
5.1 Two structural shifts in functional diversity are associated with stability thresholds during biogeomorphic succession
Our results deliver an empirical test of functional shifts in biogeomorphic succession as proposed by Corenblit et al. (2007), Corenblit and Steiger (2009) and Corenblit et al. (2015). They showed that threshold dynamics play an important role during biogeomorphic succession in glacier foreland environments, which can be detected with functional diversity indices (Figure 2).
Here, we identify a structural shift once succession reaches around 30% plant cover, at which point functional diversity becomes temporarily stable. Another, less distinct structural shift arises around 74% plant cover, beyond which functional diversity decreases again. While we cannot make a direct correlation between quantified geomorphic process rates and levels of functional diversity, we identify changes in functional diversity in locations where Haselberger et al. (2021) reported shifts in geomorphic process rates.
Based on our results the development of functional diversity along the plant cover gradient can be subdivided into three distinct phases: an initial increase (I), a sustained plateau (II) and a final reduction (III). These phases can be linked to conceptual models on biogeomorphic succession (Corenblit et al., 2007; Corenblit et al., 2009).
5.1.1 Phase I (geomorphic/pioneer)
Initial development of functional diversity is linked to stochasticity, where both allogenic factors (i.e., initial site conditions and geomorphic disturbances) and random factors (e.g., seed dispersal) govern colonization (Del Moral, 2009; Dini-Andreote et al., 2015; Hanusch et al., 2022; Marteinsdóttir et al., 2010; Mong & Vetaas, 2006; Wojcik et al., 2021). Initially, plant colonization is inhibited by high geomorphic activity during a geomorphic stage (Corenblit et al., 2007; Corenblit et al., 2009). We argue that the lowest functional diversity in this stage is attributed to a confined pool of pioneer species with adaptive traits that allow them to adapt to harsh environmental conditions (Chase & Myers, 2011; Vellend, 2010).
The subsequent rise in functional diversity suggests a diversification of strategies, including biogenic microhabitat amelioration and plant–plant facilitation, which have been shown to enhance functional diversity. The adaptive response during this pioneer stage (Corenblit et al., 2007; Corenblit et al., 2009) is likely employed to cope with conditions influenced by allogenic factors, encompassing initial environmental conditions and geomorphological disturbances (Matthews, 1992; Raab et al., 2013). Species increasingly colonize in safe sites that offer more favourable environmental conditions (Wojcik et al., 2021).
Forbs and graminoids predominate in the geomorphic and pioneer stages. However, during the initial structural shift, forb species exhibit an ascending trend, reaching a peak, while graminoids concurrently experience a decline (Figure S1). This aligns with the initial dominance of species characterized by fibrous root systems, pivotal for initial soil binding, as well as nutrient and water uptake (Schenk & Jackson, 2002b; Wright & Westoby, 1999). Simultaneously, tap roots, associated with deeper soil penetration, exhibit a gradual increase towards the onset of the first structural shift, indicating a transition towards a more stable soil environment (Jonasson & Callaghan, 1992).
5.1.2 Phase II (biogeomorphic)
Upon surpassing the initial structural shift in functional diversity, this study observes a subsequent plateau-like phase that we associate with the biogeomorphic stage (Corenblit et al., 2007, Corenblit et al., 2009). This structural shift has been observed to be consistent across diverse organismal groups, including plant, microbe and invertebrate taxa, along a terrain age gradient, occurring 60 years post-glacier retreat (Hanusch et al., 2022). The structural shift at 30% plant cover observed in our study aligns with the stability threshold that was qualitatively tested by Eichel et al. (2016) and Eichel, Draebing, et al. (2023) and quantitatively confirmed by Haselberger et al. (2021). Here, the biogeomorphic feedback window starts and functional diversity reaches an optimum. This high trait diversity offers beneficial effects for slope stabilization (Martin et al., 2010; Ohler et al., 2023; Zhu et al., 2015). At this point, the rates of geomorphic processes already decreased, resulting in reduced allogenic pressure (geomorphic disturbance, Eichel, Stoffel, & Wipf, 2023; Haselberger et al., 2021). This may be a result of autogenic factors, specifically ecosystem engineering (Osterkamp et al., 2012), species facilitation that allows the colonization of new species (Brooker et al., 2008; Corenblit et al., 2015; Wheeler et al., 2015) and an accelerated expansion in plant cover, which can be attributed to a more effective utilization of resources via niche differentiation (Díaz et al., 2007; Ravenek et al., 2014). In the biogeomorphic phase, we note an increase in woody species, particularly dwarf shrubs such as Salix spp., is observed. These species are thought to possess engineering traits for sustainable terrain stabilization. Notably, they feature deeply rooted woody roots and a mat-like growth structure characterized by low-lying stems, branches and numerous leaves (Beerling, 1998; Eichel, 2019).
5.1.3 Phase III (ecologic)
The noted decrease in functional diversity following the biogeomorphic stage can be linked to a reduction in geomorphic activity, which has been identified as a second stability threshold by Haselberger et al. (2021) and aligned with Wojcik et al. (2021) concept, that an autogenic/biotic phase is initiated and biotic factors, such as species competition, gain dominance while the relative significance of allogenic factors declines. This also signifies the closure of the ‘biogeomorphic feedback window’, identified as a competition threshold by Eichel et al. (2016) initiating the ecological stage wherein vegetation takes control of the geomorphological environment (Corenblit et al., 2007, Corenblit et al., 2009). This might also be represented in the decline of functional diversity at high levels of plant cover, which can be tied to the competitive dynamics of late successional communities where a few dominant species with similar functional traits begin to predominate (Caccianiga et al., 2006; Corenblit et al., 2015; Losapio et al., 2021; Matthews, 1992; Raab et al., 2013).
The close relationship between functional diversity and taxonomic diversity, both showing a hump-backed curve along the cover gradient, has been previously demonstrated (Echeverría-Londoño et al., 2018; Swenson et al., 2012), supporting the shift from environmental filtering during early stages to interspecific competition in late successional stages (Erdős et al., 2023; Weiher & Keddy, 1995). However, it is demonstrated that shifts in functional diversity are closer associated with stability thresholds and are more pronounced. In contrast to classical succession studies that focus on terrain age as guiding environmental gradient (Greinwald et al., 2021; Jones & Del Moral, 2005; Losapio et al., 2021), our study focuses on a plant cover gradient which takes into account the disturbing nature of geomorphological processes in glacier foreland (Rydgren et al., 2014; Wojcik et al., 2021). This way we consider sites, which possess older terrain ages but show lower levels of succession, based on present plant cover, where succession is limited by available resources and/or has been reset by geomorphic disturbance (Gurnell et al., 2000; Osterkamp et al., 2012; Temme & Lange, 2014). Therefore, the chosen plant cover gradient in this study shows a mixed signal of primary and secondary succession. While plant cover does not strictly correspond with terrain ages in our plot selection, we observe an initial increase in functional diversity, followed by a plateau where functional diversity stabilizes at a consistent level. This mirrors findings from a strict terrain age gradient in another glacier foreland (Hanusch et al., 2022). We argue that this indicates autogenic processes have a dominant role in shaping the rate of successional development. Additionally, the period since the last significant disturbance, whether from glacier retreat or geomorphic events, offers a more precise indication of successional timing than terrain age gradients based only on glacier retreat.
5.2 Changing influence of geomorphology on functional diversity
An examination of geomorphologically active plot surfaces reveals a slight increase in functional diversity at lower levels of geomorphic influence, but a significant decrease of functional diversity when approximately 50% or more of the plot surface is active (Figure 3). This indicates that low levels of geomorphic influence, such as the accumulation of sediment through slope wash and fluvial deposition on a small scale, can be beneficial for increasing functional diversity; providing resources such as nutrients, water and substrate (Gurnell et al., 2000; Kemppinen et al., 2021; Stawska, 2017; Virtanen et al., 2010); or necessitating diverse adaptation strategies in response to geomorphic forcing (Corenblit et al., 2007; Eichel et al., 2013; Gurnell, 1997). The structuring effect of geomorphic disturbance on the functional composition of vegetation has been consistently observed across various environments (Bernhardt-Römermann et al., 2011; Kemppinen et al., 2021). Acknowledging the potential for circularity, caution is warranted due to the data inherent potential for inverse proportionality between plant cover and geomorphic influence. Despite the lack of information on initial site conditions and the prevalence of limited plant cover in many plots due to unfavourable substrate and potential resource constraints, we note an increasing functional diversity at low levels of geomorphic influence, indicating a positive geomorphic influence up to a threshold. Future research should prioritize untangling the effects of initial site conditions and abiotic disturbance to strengthen our interpretations.
The observed decrease in functional diversity beyond the identified structural shift can be attributed to allogenic forcing, which either destroys pre-existing plants or permits only a limited set of species to establish or persist (Losapio et al., 2021; Wojcik et al., 2021). The extent of the geomorphic disturbance determines whether it is possible for neighbouring plants to re-colonize through means such as seed dispersal or vegetative spread (Polzin & Rood, 2006; Stöcklin & Bäumler, 1996). The results also demonstrated lower functional diversity in areas with rockfall deposits. The coarse and blocky material in these deposits can be associated with stochastic assemblage mechanisms, which are characterized by sparse vegetation comprising a limited number of species that randomly migrate and temporarily survive on such extremely unstable substrates (Bricca et al., 2021; Valachovič et al., 1997; Zanzottera et al., 2020).
While there is an understanding that geomorphic processes in glacier forelands, including rock fall, debris flows, fluvial processes and slope wash, influence plant colonization by modifying microhabitat conditions (Gentili et al., 2013; Le Roux et al., 2013), the complex spatiotemporal characteristics of these geomorphic processes and the potential confounding effects of terrain age, elevation and plant richness make it challenging to establish a consistent relationship with plant colonization (Giaccone et al., 2019).
Unlike fluvial processes (Corenblit et al., 2007) and salt marsh dynamics (Balke et al., 2011), landforms shaped by geomorphic activities in glacier forelands present a more complex template of microtopography (Gentili et al., 2010; Lane et al., 2016). These conditions can either be favourable or non-favourable to plant colonization (Marler & Del Moral, 2018). Given plants' dependence on substrates and resources (Körner, 2003), we propose a need for more detailed analyses of the spatial heterogeneity in site conditions geomorphic processes create for plant growth. As evidenced in biogeomorphic investigations within fluvial systems (Bätz et al., 2016; Corenblit et al., 2007), an in-depth examination of the intricate allogenic template produced by specific geomorphic processes is necessary. Sampling along longitudinal and cross profiles has proven to be an effective way to encompass this complex template, which varies in aspects like water availability, presence of fine sediment, frequency and magnitude of inundation, among other factors (Corenblit et al., 2015; Gentili et al., 2010; Lane et al., 2016).
5.3 Distinct trait composition during biogeomorphic succession
This study also examines changes in trait composition across four stages of biogeomorphic succession, characterized by a sequence of evolving abiotic and biotic dominance. We were able to distinguish trait compositions among the four stages of biogeomorphic succession (Figure 4). Geomorphic and pioneer stages exhibit no significant differences in trait spectra. We argue that this reflects the dominance of either the prevailing site conditions or autogenic processes as influencing factor (Wojcik et al., 2021) and the fact that our set-up does make a clear differentiation between geomorphic and pioneer stages difficult.
Geomorphic and pioneer stages during biogeomorphic succession characterized by low plant cover show a predominance of above-ground adaptation and reproductive traits, including leaf longevity and structure, growth form and mixed reproductive strategies. In contrast, biogeomorphic and ecologic stages emphasize competitor species and traits related to below-ground structures, such as root type and structures and vegetative reproduction. The transition from the biogeomorphic stage with ruderal species emphasizing dispersal and rapid establishment to the ecologic stage with stress-tolerant species prioritizing resource acquisition and competition aligns with established plant ecological frameworks (Caccianiga et al., 2006; Greinwald et al., 2021; Grime, 2006; Westoby & Wright, 2006). This indicates a shift from disturbance-resistant species, typically associated with smaller leaf sizes, to more resilient species, characterized by traits like increased specific leaf area and decreased leaf dry matter content (Bernhardt-Römermann et al., 2011; Losapio et al., 2021).
Between pioneer and biogeomorphic stage significant differences can be observed, including changes in plant height, the occurrence of hygromorphic leaves, the capacity for vegetative dispersal and the presence of competitor species. Plant height tends to be lower in the biogeomorphic stage, a phenomenon often linked to increased stress (Körner, 2003) or indicative of environments where rapid plant establishment is necessary due to recurring disturbances (Anthelme et al., 2021; Franzén et al., 2019). The rise in the presence of competitor species during the biogeomorphic stage could be attributed to the fact that gaps created by disturbances are primarily colonized by the dominant species already present in the vicinity (Cichini et al., 2011). Similarly, the presence of species utilizing mixed ways of reproduction declines during the biogeomorphic stage, hinting at a potential change towards more specialized seed-dispersing or vegetatively dispersing species in response to disturbances (Stöcklin & Bäumler, 1996).
6 CONCLUSION
This research provides an important step towards integrating concepts from functional ecology into biogeomorphic research, especially in glacier forelands.
Our study indicates functional changes at known stability and competition thresholds during biogeomorphic succession by utilizing openly accessible trait data. Unlike earlier studies that focused predominantly on taxonomic analyses of biogeomorphic succession, our research underscores significant shifts in the dominance of certain traits during this process. We urge geomorphologists to redirect their focus from taxonomic investigations to examining species functionalities, as traits are instrumental in determining a species' response to and impact on geomorphic processes (Violle et al., 2007).
The functional approach can facilitate generalizable interpretations and ease the identification of favourable species traits for targeted environmental management within the context of nature-based solutions—a recognized critical aspect (Bernhardt-Römermann et al., 2011; Larsen et al., 2020; Viles & Coombes, 2022).
This study represents one of the first steps towards incorporating a functional perspective into biogeomorphic succession research. Based on the empirical evidence supporting existing conceptual models, we propose several promising directions for future research: Comparative studies across various glacier forelands or even different environments may confirm whether functional shifts in biogeomorphic succession are indeed consistent and generalizable (I). The quantification of geomorphic process rates will aid in specifically linking physical forcing to changes in plant functionalities (II). Conversely, testing the influence of certain plant traits on geomorphic process rates in a controlled environment will help pinpoint the physical mechanisms behind biogeomorphic interactions (III). From a catchment perspective, a subsequent step would involve considering the geomorphic legacy—that is, the history of frequency and magnitude of geomorphic processes—on the distribution of functional traits (IV). We need to progress beyond directly linking geomorphic processes with vegetation distribution and plant traits and more thoroughly consider the complex site conditions created by a single geomorphic process (e.g., lateral and longitudinal cross profiles of debris flow landforms; V). It is essential to disentangle the allogenic influence (i.e., initial site conditions and geomorphic disturbance) through the quantification of geomorphic forcing and the collective quantification of critical site conditions (moisture, nutrients, substrate and temperature; VI). Phenotypic variability and plasticity offer direct assessment of how geomorphic processes influence intra-species trait variations during biogeomorphic succession in glacier forelands. By employing phenotypic plasticity, it becomes feasible to associate specific geomorphic situations with shifts in established response and effect traits (VII).
ACKNOWLEDGEMENTS
We are particularly grateful to Matthias Marbach for his technical support, Jonas Machold for assistance in the field, all students and volunteers who supported fieldwork, the ENGAGE working group at the University of Vienna for the use of their facilities and local stakeholders, namely, the municipality of Feichten and the operator of the Kaunertal Valley glacier road (Kaunertaler Gletscherbahnen GmbH) for their cooperation. The present study is part of the PHUSICOS (Solutions to reduce risk in mountain landscapes) project, funded by the European taxpayer via the European Union's Horizon 2020 research and innovation programme under grant agreement No. 776681.
Open Research
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available in the Supporting Information of this article or from the corresponding author upon request.
