Early-melting snowpatch plant communities are transitioning into novel ecosystems

doi:10.21203/rs.3.rs-2973415/v1

Download PDF

Article

Early-melting snowpatch plant communities are transitioning into novel ecosystems

https://doi.org/10.21203/rs.3.rs-2973415/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 02 Oct, 2023

Read the published version in Scientific Reports →

You are reading this latest preprint version

Snowpatch plant community distribution and composition is strongly tied to the duration of long-lasting snow cover in alpine areas; they are vulnerable to global climatic changes that result in warmer temperatures and longer growing seasons. We used a multi-decadal revisitation study to quantify early-melting snowpatch floristic and functional diversity change in southern Australia, and document shrub size-class distributions over time to detect evidence for their encroachment patterns into snowpatches, a key prediction with climatic change. Early-melting snowpatches have declined in areal extent, changed in species composition, shrub and tussock grass cover has increased, but the effects of these changes on functional trait diversity were less clear. Species gains accounted for most of the floristic change observed. Shrub recruitment into early-melting snowpatches is ongoing by most shrub species. Biotic differentiation is occurring; many early-melting snowpatches are transitioning to a novel community with changed composition and taller vegetation structure, but there is little evidence for species loss having occurred. Given enough time, however, the long-term loss of snowpatch species is likely (i.e. biotic homogenisation) if taller shrubs outcompete short-statured snowpatch species. We believe that the changes observed provide evidence that this alpine ecosystem is on the verge of collapse.

Snow cover is spatially and temporally heterogeneous in alpine landscapes, and snow accumulation and snowmelt patterns vary with respect to local topography^1,2. The timing of snowmelt is a critical determinant of the legacy effects of winter snow on the snow-free season^3,4, with differences in the duration of snow cover having profound effects on plant distributions, composition, phenology and productivity in high mountains^1,5–8. The distribution of snowpatch plant communities (also called snowbanks and snowbeds in the literature), in particular, is strongly tied to snow duration^2,9; they broadly exist in areas where snow is deep and lasts longest into the growing season. Late-lying snow cover provides insulation from spring frosts^10,11, a longer-lasting source of water in summer², and a shorter growing season⁹. This favours the dominance of herbaceous species over woody species^12–14. Because of their dependence on late-lying snowpacks, the extent of these communities are likely to decline with climatic change if less snow falls and earlier melt-out occurs^15–18.

We lack a comprehensive overview of how altered snow conditions will affect snowpatch communities^2,17 and cold climate ecosystems more broadly^19,20. In areas with a short growing season, climatic warming can be expected to favor species that take advantage of the longer growing season; the arrival and expansion of grassland and heathland species from adjoining plant communities into snowpatches is a likely outcome^5,21,22. This will have impacts on the structure, composition and ecological function of snowpatch communities⁴, but the mechanisms of change and outcomes for snowpatch communities remain poorly understood^23–25. A re-assembly of the original community may result because of immigration of new species and the persistence of snowpatch species (e.g. Kirkpatrick et al.²⁶), leading to a novel, structurally and compositionally different vegetation (i.e. biotic differentiation²⁷). If the immigration of new species into snowpatches translates to the loss of snowpatch specialists due to competitive exclusion²², it may lead to their biotic homogenisation, with compositional similarity increasing over time²⁸. Evidence for these types of change outcomes remain relatively scant (but see recent exceptions^22,23,29), as does the effect of environmental change on functional trait diversity, the measurable properties of plants (such as size, structure and biogeochemistry) that are intrinsically linked to the key functions of ecosystems²⁰. In all likelihood, biotic differentiation is the first outcome of climatic change for plant communities where immigrations outpace extinctions, with biotic homogenisation (and ecosystem collapse sensu Bergstrom et al.³⁰) a longer-term response to altered abiotic drivers and a milieu of changed biotic factors which may be mediated by climate.

In Australia, snowpatch communities have developed on south-east facing slopes at high elevations in subalpine and alpine zones ^31–33 (Fig S1). They are amongst the rarest and most restricted of Australian alpine plant communities, with subalpine snowpatches (the lowest elevation, early-melting examples of the community) likely to be the most vulnerable to the changes in snowcover timing and extent predicted (and already observed) by global change models ^15,18,32,34. Australian alpine regions have experienced ~1 °C warming (of both maximum and minimum temperatures) since the 1970s¹⁸ and there has been a substantial decline in the amount of snow and its persistence over this time³⁵. Continued environmental changes will likely cause snowpatches to re-organise due to the establishment of non-snowpatch species from adjacent plant communities^22,23; in the longer-term, warming will allow invasion of plants from lower elevations³⁶. In this study, we conduct a multi-decadal re-visitation survey of early-melting subalpine snowpatches to quantify their current floristic and functional composition relative to their historic state. Specifically, we ask: does the current floristic composition of snowpatches provide support for biotic differentiation or biotic homogenisation having occurred over time? Have such changes led to functional trait compositional change? We also examine the evidence for ongoing structural change by quantifying how woody species diversity and abundance has changed over time. We predict that incursions of new species (immigrations) with climatic changes will have outpaced local extinctions (snowpatch specialists will have long lag periods before their disappearance), meaning that biotic differentiation - the development of a novel ecosystem - is a likely outcome of multi-decadal change. We predict that shrub diversity, cover and density in snowpatches will have increased over time as the environmental filters that exclude woody plants from snowpatches (i.e. duration of snow cover, shortened growing seasons) are increasingly relaxed. From this evidence, we interpret whether snowpatches have undergone ecosystem collapse as defined by Bergstrom et al³⁰, i.e. potentially irreversible change to ecosystem structure, composition and function.

Areal extent of snowpatches. The area occupied by early-melting snowpatch plant communities has decreased over the last 40 yrs; 86% of snowpatches (Table S2) have declined in area, by an average of 41.6% (+/- 10.3%; 1SE).

Patterns of diversity. Early-melting snowpatches have gained plant species over time. Median quadrat alpha diversity was significantly higher in snowpatches in 2022 compared to 1982, in areas now with either low or high shrub cover (Fig. 1). Mean alpha diversity has increased by ~ 30%. Similarly, Hill-Shannon diversity (but not Hill-Simpson diversity) was significantly higher in snowpatches in 2022 compared to 1982, in areas now with either low or high shrub cover.

A principal coordinates analysis of pairwise beta-diversity matrices shows greater dispersion for early-melting snowpatch sites in 2022 (both in low and high shrub cover plots) than in 1982; the variability in the species assemblages (i.e. beta-diversity) of snowpatches has increased over 40 years (Fig. 2). Mean distance to median centroid increased from 0.26 in 1982 to 0.32 (low shrub cover) and 0.33 (high shrub cover) in 2022 (Fig. 2). Pairwise comparison between these groups found this increased dispersion was significantly higher between 1982 plots and 2022 high shrub cover plots (p = 0.032; Fig. 2), and was marginally higher between 1982 and 2022 low shrub cover plots (p = 0.054; Fig. 2). There was no difference between dispersion for high or low plots in 2022 (p = 0.97; Fig. 2).

Floristic change (1982–2022). Frequency change highlights that species immigrations far outweigh local species losses in early-melting snowpatches (Table S3). Most species that were present in 1982 were also observed in 2022. ‘Increaser’ species included those entirely new to snowpatches, and species originally present in 1982 that increased in frequency and cover (Table S3). Taking a conservative approach to species gains (i.e. considering only species that were present at > 15% frequency in the 2022 survey), we observed 13 species that were not seen in 1982, comprising five grasses, one graminoid, one fern, four herbs and two shrubs. Shrubs and tussock grasses have increased in cover, while native herbs and some introduced species (e.g. Agrostis capillaris) have increased in frequency.

Overall, the floristic composition of early-melting snowpatches was more variable in 2022 than in 1982 (Fig. 3) and snowpatches share little overlap in composition between the two sampling periods (presence/absence data, Fig. 3a; cover data Fig. 3c). Vector fitting shows that dryland grass species cover and shrub cover help differentiate floristic patterns, particularly when using plant cover data (Fig. 3c). Historic snowpatches had both lower dryland species grass cover and shrub cover compared to current snowpatches.

Functional Traits. We found few significant changes to the Functional Diversity and Functional Evenness of early-melting snowpatches over time, regardless of shrub cover (low, high; Fig S2). Functional Richness, by contrast, was higher over time at high shrub cover (Fig S2). We found a significant increase in community trait-weighted mean height over time (1982 vs 2022) at both low and high shrub cover (Fig S3), but no significant differences in LDMC, SLA or seed mass across years or with differences in shrub cover (Fig S3).

Shrub dynamics in snowpatches. The total number of shrubs in early-melting snowpatches increased by 54% between 1990 and 2019. There was considerable site to site variation in shrub dynamics; 71% of the seven snowpatches studied gained shrubs (range + 33% to + 1014% increase in shrub numbers) while the remainder of snowpatches lost shrubs over time (range − 32% to -63% decrease in shrub numbers) (Table S3, Table S4). The number of shrub species in focal snowpatches increased by 33% (from 15 to 20 species); one tree species (Eucalyptus pauciflora) also established. Shrub numbers increased most for Acrothamnus montanus (+ 599% increase in individuals), Phebalium squamulosum (+ 344%), Olearia brevipedunculata (+ 301%) and Pimelea axiflora (+ 265%) (Table S5). Few shrub species declined in number over time; Olearia frostii (-60%) and Grevillea australis (-65%) were the chief examples.

The area of occupancy of shrubs (in 160 m² quadrats) increased over time (Table S6, table S7). In 1990, shrubs occupied a mean of 27.9 +/- 5.9 m² (i.e. 17.4% cover); this increased to 47.1 +/- 9.1 m² in 2019 (i.e. 29.4% cover), a mean increase of 69% in cover.

The size-class distribution of the most common shrubs observed (i.e. those species that made up more than 1% of individuals in 2019) shows that recruitment into early-melting snowpatches has been ongoing over time (Fig. 4). Type 1 Condit slopes (which are evidence of recent / accelerating recruitment) was the most frequent shrub regeneration dynamic observed; of the eight commonest shrubs in early-melting snowpatches, seven species exhibited this population dynamic, whereby small shrub size-classes are the most well represented (Fig. 5). Only Asterolasia trymalioides exhibited a Condit slope Type 3 whereby recruitment was severely limited i.e. there were few to no shrubs in the smaller size classes relative to larger shrubs. For four species, Type 1 slopes were substantially higher in 2019 compared to 1990, indicating that recruitment of new individuals is occurring at a higher level in the most recent survey (2019) than it was 30 yrs prior (1990).

Understanding how high mountain vegetation re-assembles with changes in climate and biotic filters is necessary to help interpret the new vegetation patterns that are emerging^5,13,23. Such changes occur due to floristic and functional changes resulting from both invasions by new species (immigration) and the extinction of resident species. This can result in biotic homogenisation (when compositional similarity increases over time²⁸) or biotic differentiation (when compositional similarity declines over time). Evidence of changes in high mountain vegetation with respect to these ecological change models is generally lacking. Our resurvey of early-melting snowpatches, comparing floristic, functional and structural changes over 40 yrs, revealed that contemporary snowpatches are the outcome of biotic differentiation. They have gained many non-snowpatch from surrounding plant communities, but have lost relatively few of the resident species that were initially recorded in them. Hence, a novel plant community has emerged over time. By contrast, Verrall et al.²³ found that biotic homogenisation was occurring in late-melting alpine snowpatches due to the increases in graminoids and declines in snowpatch specialists. They concluded that such changes were threatening the compositional distinctiveness of alpine snowpatches.

There were fewer species (alpha diversity) and less variability in the floristic composition of early-melting snowpatches (lower beta diversity) at the start of this study (1982) than that observed in 2022. A reduction of snowcover duration has been observed to increase alpha diversity and vegetation cover elsewhere, driven by the invasion of generalist species (e.g. ^3,29,38,39). In our study, species new to snowpatches encompassed species from all growth forms - herbs, grasses, shrubs - many of which are typically found in adjacent dryland heath and grassland communities^40–42. Shrubs are most likely responding to increasing growing season length in snowpatches as a result of the regional trend for declining snowfall^35,43 and increasing temperature¹⁸. Tussock grasses (e.g. Poa spp., Deyeuxia moniticola) may be responding to earlier snowmelt, but in a different way to shrubs. Earlier snowmelt may be reducing growing season soil moisture that typically comes from late-melting snow^12,44, allowing expansion of dryland tussock grasses into areas where they were previously absent because of site wetness. Increases in tussock-forming grass species has been previously identified as a signal of change in late-lying snowpatches^23,29, so their positive response in early-melting snowpatches over time are further confirmation of vegetation dynamics consistent with abiotic conditions now favouring species with different ecological strategies and/or traits to those in the past.

Several herbaceous dicot species entered and established in early-melting snowpatches over time (e.g. Brachyscome decipiens, Trachymene humils subsp. breviscapa). New species were typically not capable of short- to long-distance vegetative spread (sensu Grime⁴⁵). Indeed, five of the seven (71%) herbaceous dicot species that were new to early melting snowpatches (with > 10% frequency in 2022) were non-clonal. By contrast, resident herbaceous dicot species in snowpatches in 1982 were generally clonal (74% of species, Table S3). It is tempting to hypothesise that (a) seedling regeneration is the primary mechanism by which new herbaceous dicot species have entered early-melting snowpatches and (b) clonality allows snowpatch species to persist despite ongoing environmental change. While clonal growth provides advantages in cold environments⁴⁶, via the sharing of resources between ramets⁴⁷ or space occupation by the clumped distribution of ramets⁴⁸, clonality can lead to a trade-off in the reproductive investment in seeds^49,50 and their dispersal⁵¹. The changes we observed provide evidence that environmental filters are relaxing in snowpatches, and this is expressed as the immigration of mobile species that can now enter and establish in these habitats. Whether clonality delivers an extinction debt to snowpatch species is unclear. This gap in our understanding of vegetation dynamics seems crucial to resolve as it may affect long-term species coexistence patterns of herbs in snowpatches⁵².

More obvious is the change that is occurring to the vegetation structure in early-melting snowpatches; shrub density and cover are both increasing over time, and community trait-weighted mean heights were significantly higher in 2022 than 1982. The changes observed are in accordance with climate change predictions, particularly for woody plants in high mountains ²¹. We found that, in 2019, the most common shrub species in early-melting snowpatches typically exhibited a Type 1 Condit slope, and that the slope of these curves steepened for four species relative to those observed in 1990. This provides evidence of ongoing recruitment of shrubs into early-melting snowpatches, which is accelerating relative to earlier observations, and occurs across a number of woody species that were previously rare or absent from this community^31,40. Previously strong plant-microclimate interactions are likely eroding, enabling the potential niche of those species limited by the short growing season to expand. Shrub expansion is therefore likely to continue to be favoured by warming associated with global climate change⁵³.

There have been suggestions that recent floristic changes in snowpatch communities are evidence that these communities are collapsing (sensu Bergstrom et al. ³⁰), with taller, novel species outcompeting snowpatch taxa^13,15,29,33. Expansion of shrubs and graminoids is an increasingly common outcome in snowpatch habitats globally^3,21,39, typically resulting in a reduction of habitat for snowpatch specialists⁵⁴. Our data provide evidence that change to floristic composition of early-melting snowpatches is occurring, and that ongoing changes to vegetation structure (due to shrubs) are likely to be irreversible without management intervention. Compositional change in early-melting snowpatches involves the addition of new species but few local extinctions have yet occurred, leading to the formation of a novel community (i.e. biotic differentiation). Plant-plant asymmetric competition (at the scale of 20 m² quadrats) has yet to be demonstrated but such theory is likely to help explain both species gains (invasion of taller species into a short statured ecosystem where there is no size advantage of the resident community) and potential losses of rare, short-statured snowpatch species now growing under shrub canopies⁵⁵. There is the real prospect that ongoing invasion of the ecosystem by taller growing native shrubs and grasses will lead to a persistent structural state change and the eventual loss of snowpatch species. The implications of such changes for fauna remain completely unknown. It is hard not to conclude that this rare Australian ecosystem is on the verge of collapse.

Implications for early-melting snowpatches. Evidence is mounting that climate change is inducing changes in Australian snowpatch vegetation, causing biotic differentiation. Changing climate is also substantially altering vegetation structure, allowing encroachment of shrubs and tussock grasses into snowpatches. The expansion of taller growing native species, at the expense of shorter life forms, is a global phenomenon in temperature-limited systems at high altitude. Understanding the interplay between climate and other drivers is essential to better forecast the future of alpine plants. We conclude that ecosystem collapse seems inevitable. Collapse of early-melting snowpatches is a transformation of identity (structure and composition) and its replacement by a novel ecosystem. Williams et al.¹⁵ suggest that a transition from snowpatch herbfield to heathland occurs when shrub cover reaches 50%; our shrub dynamics plots had an average of 29% shrub cover in 2019, up by 63% from estimates in 1990. There is potentially time to implement management interventions to reduce the cover of taller growing native shrubs and tussock grasses (i.e. to ‘direct’ ecosystem change⁵⁶) if early-melting snowpatch communities are to be structurally (if not compositionally) maintained into the coming decades.

Study region. The study was conducted on the Bogong High Plains (36° 53’ S, 147° 19’ E), Alpine National Park, ∼320 km north-east of Melbourne, Victoria, Australia. The region is characterised by low mean monthly daily maximum temperatures (1.2–17.9°C), frequent frosts (> 100 per annum, that can occur at any time of the year) and high precipitation (> 1300 mm per annum), much of which falls as snow (Bureau of Meteorology recording station #083084, 1765 m asl; www.bom.gov.au). Soils are organic, highly acidic loams of variable depth⁵⁷.

Climate change context. The Australian Alps have been warming at about 0.2°C per decade over the past 50 years^33,58, a higher rate than many other areas of Australia¹⁸. The rate of temperature increase from 1970 to 2005 was approximately 2.6 times faster than the fastest rate of warming during the Medieval Climate Anomaly⁴³. There has been a significant decrease in both maximum snow depth, total snow accumulation³⁵ and snow persistence²³. Indeed, the Australian snowpack is now at a 2000 year low⁴³. Recent rapid decrease in snow cover over the past five decades is at least an order of magnitude greater than for similar periods over the past 2000 yrs⁴³. Spring thaw has been occurring, on average, two days earlier per decade, with very low snow years (e.g. 1999, 2006) represented by the earliest thaws on record³³. Predicted decreases in albedo caused by declines in snow depth and snow cover during the cold season⁵⁹ will mean that precipitation will increasingly fall as rain. Climate change predictions identify that the area sustaining snow in Australia for more than 60 days per annum may be reduced by up to 96% by 2050⁶⁰.

Community description. Snowpatches are one of 10 major structural assemblages of Australian alpine vegetation^41,61. Early-melting snowpatches are plant communities that occupy areas on moderate (10–15°), south-east-facing (85–170°) slopes with late lying snow and hence, reduced growing season length, enhanced water supply in summer, and protection from early growing-season frosts^15,40. Snowpatch vegetation in Australia is physiognomically similar to that of snowbed vegetation in alpine and subalpine regions elsewhere in the world (e.g. ^9,12,62,63).

In our study area, early-melting snowpatches were historically dominated by low statured (< 10 cm tall) graminoids (Agrostis venusta, Carex hebes, Poa hothamensis, Rytidosperma nudiflorum) and herbs (e.g. Celmisia spp.)^31,40. Tussock forming grasses and shrubs were virtually absent when surveyed by McDougall⁴⁰, although Wahren et al.³¹ note that while shrubs were relatively uncommon, some subalpine snowpatches were covered by open heath, with shrubs overlying herbfield. We focus on early-melting snowpatches (1670–1850 m), those at the lower elevation range of snowpatch distributions in Australia (Fig S1; ^23,31,32); i.e. the most marginal for snowpatches in Australia. Wahren ³¹ estimated that the average duration of snow days into the growing season (in the southern hemisphere, days post-October 15) for early-melting snowpatches in the study area was 23–53 days, while for alpine snowpatches, those that occur at higher elevation (> 1850 m), average number of snow days was 53–83.

Data on snowpatch fauna are sparse¹⁵. An undescribed spider from the genus Micropholcomma is thought to be an obligate snowpatch species, while one collembola - Australotomurus montanus - is a cold specialist (M.Nash, pers. comm.).

Extent of early-melting snowpatches. The areal extent of 14 early-melting snowpatches on the Bogong High Plains was compared between 1982 and 2022 to assess how the area of snowpatch vegetation is changing over time. The size of each snowpatch in 1982 was extracted from vegetation maps created by McDougall⁴⁰ who determined the types and extent of major vegetation types on the Bogong High Plains using aerial photography and ground-truthing. The area of each snowpatch in 1982 was calculated using shapefile versions of the vegetation maps⁴⁰ and calculating the area of snowpatch sites in QGIS 3.22.6. The area of each snowpatch in 2022 was assessed by on-ground assessment. At each site, the extent of snowpatch vegetation was determined using structure and composition to delineate the boundary between snowpatch vegetation (e.g., dominated by short-statured forbs and graminoids, such as Luzula acutifolia, Carex hebes, Celmisia spp., Poa fawcettiae, Poa hothamensis) and more widespread grassland or heathland vegetation (e.g., dominated by characteristic species such as Poa hiemata, Grevillea australis). The boundary of snowpatch and non-snowpatch vegetation was recorded on a Garmin GPS 64, with area calculated in QGIS 3.22.6.

Floristic changes over time. We surveyed 14 early-melting snowpatches on the Bogong High Plains in austral summer (January-February) 2022 for their floristic composition, and compared these surveys to those conducted by McDougall⁴⁰. McDougall⁴⁰ used a single 4 x 5 m quadrat to sample the composition and abundance (using Braun-Blanquet cover-abundance ranks) of ten subalpine snowpatches as part of a landscape-wide vegetation survey to describe and classify the plant communities of the Bogong High Plains based on the presence and cover of species. These quadrats were classified as ‘short turf snowpatch’. We extracted the original two-way table data from McDougall⁴⁰ to make comparisons to the current state. We sampled 14 subalpine snowpatches, selecting sites from vegetation maps generated by the McDougall⁴⁰ survey. The original snowpatch quadrats were not named or geo-referenced and hence, we cannot directly compare site-to-site snowpatch vegetation change. Instead, we sampled all larger early-melting snowpatches (1200–13500 m²) in the Bogong High Plains landscape, using the vegetation map as a guide to the historic extent of the snowpatch community, and assume McDougall⁴⁰ would have sampled most (probably all) of these snowpatches given their limited number in the landscape. Hence, we examine snowpatch floristic change on the Bogong High Plains at the community level rather than at the level of individual snowpatches.

Spatial heterogeneity in the vegetation can affect characterisation and quantification of change over time^64,65. We tried to minimise false estimates of change due to inherent spatial heterogeneity by sampling (depending on snowpatch size) from two to four (mean 3.2) 20 m² quadrats per site. Hence, the 2022 survey is more intensive than the original, allowing us (with some confidence) to detect losses of species from snowpatches. We assigned all species a Braun-Blanquet cover-abundance rank as was done by McDougall⁴⁰.

Taxonomic changes have occurred between the two surveys. We updated all species names to current nomenclature (https://vicflora.rbg.vic.gov.au), but where species have been split into several taxa since the original survey (e.g. Craspedia), we subsume these new species into their original synonyms. This, unavoidably, underestimates total species richness in snowpatches.

Functional diversity. Data on four continuous traits (leaf dry matter content, specific leaf area, height, seed mass) for all species observed across the two surveys (n = 80; Table S3) were collected. Leaf area, the one-sided projected surface area of a single leaf, is related to light harvesting strategy and the amount of mass allocated to that purpose⁶⁶. It has important consequences for leaf energy budgets and whole-plant water balance⁶⁷. Leaf area was scanned for each species using University of Sheffield Leaf Area Measurement program (version 1.3). We measured 10–20 mature, fully-expanded leaves from 5–10 individuals of each species to obtain average leaf area. Plant height at maturity is an indirect measure of the species’ overall competitive ability, mostly for light⁶⁸. We quantified maximum height of canopy from measurements provided in local floras, combined with expert opinion. Seed mass, the mean dried mass of a seed⁶⁷, has important implications for seed dispersal in space and time⁶⁶. Often, small seeds can disperse further and tend to be buried deeper in the soil profile, aiding to their longevity in the soil seed bank⁶⁷, but stored resources in larger seeds can help early seedling survival and establishment despite environmental hazards⁶⁶. We estimated seed mass from published⁶⁹ and unpublished datasets that we have assembled for the regional species pool. We standardized all traits using Z-scores so that each trait had the same weight in the functional diversity estimation and the units used to measure traits had no influence⁷⁰.

Shrub demography in early-melting snowpatches (1990 to 2019). McDougall⁴⁰ estimated that mean shrub cover in early-melting snowpatches was 5%. With climatic change, increases in shrubs (number, diversity, cover) are expected in snowpatches as growing season length increases; this may result in a structural change in the community with consequent effects on short-statured species. Due to the lack of long-term studies in snowpatch vegetation, few temporal patterns have been discerned. An effective means of documenting trends is to determine the population structural changes occurring at snowpatches. Population structure of shrubs (i.e. the frequency distribution of size or age in a population) can be a guide to regeneration status⁷¹. Changes in shrubs over time were therefore quantified using revisitation studies at seven early-melting snowpatches. The size-structure of shrub populations was surveyed in summer 1990 by Wahren⁷², and these same snowpatches were re-surveyed in the summer of 2019. The original plots were not permanently marked but Wahren⁷² provides sufficient detail to minimise relocation uncertainty.

In the centre of each snowpatch, a 10 x 16 m plot was established, and divided into forty 2 x 2 m quadrats. Within each quadrat, all shrubs were identified, and two widths and one height measurement recorded (to the nearest cm) for each shrub. We define a shrub as woody plant that typically has multiple stems or branches arising from near the base of the plant. The data were used to produce size-class frequency distributions for all shrubs, as well as individual species. Only maximum canopy width size-class frequency distributions are shown. Pimelea alpina was removed from the dataset due to inconsistent recording of this species.

Data analyses

All data analyses were carried out using R version 4.2.3⁷³.

Patterns of diversity. We quantified diversity indices between 1982 and 2022 using the floristic dataset. To investigate whether shrub encroachment is driving changes in diversity, we split the 2022 dataset (which included multiple quadrats per site) into two for further analysis. In each 2022 dataset, only one quadrat was taken per site to standardise sampling effort to the 1982 dataset. Only the quadrats at each site with the highest shrub cover (‘2022 high’ dataset) and lowest shrub cover (‘2022 low’ dataset) were selected for further analysis. Local scale diversity indices (alpha diversity, Hill-Shannon, Hill-Simpson) were calculated for each site and compared between 1982, 2022 high shrub cover and 2022 low shrub cover. Alpha diversity is the number of species per plot. Hill-Shannon and Hill-Simpson indices are modified versions of the traditional Shannon and Simpsons indices which give each index value in units of ‘effective species number’. Hill-Shannon is calculated as the exponential of Shannon’s entropy index^74,75. Hill-Simpson is equivalent to the inverse of the traditional Simpson index^74,75. These three indices were calculated using the ‘renyi’ function in the vegan package⁷⁶. Differences were compared pairwise between the three groups for each index using an ANOVA (‘aov’ function) and Tukey’s Honest Significance Difference test (‘TukeyHSD’ in vegan).

To assess heterogeneity of local communities, we assessed beta-diversity of the three datasets (1982, 2022 high shrubs, 2022 low shrubs) using ‘beta.pair’ from the betapart package to create a distance matrix accounting for total dissimilarity⁷⁷. We then implemented Anderson’s procedure for the analysis of homogeneity of group dispersions using ‘betadisper’ from the vegan package. Beta diversity can be compared between communities by comparing average dissimilarity from plots and their group community centroid⁷⁸. To test differences in beta-diversity, we performed pairwise comparisons of group mean dispersions using Tukey’s Honest Significance Difference between groups using ‘TukeyHSD’ in vegan⁷⁶.

Floristic changes (1982 to 2022). The variation in floristic composition among sites across sample periods (1982, 2022) was explored using ordination (Non-metric Multidimensional Scaling⁷⁹). Dissimilarities between all pairs of samples were based on the Bray-Curtis dissimilarity index for ordinal abundance data and Jaccard for presence/absence data^79,80. For each ordination, the relationships between the biotic factors (total shrub cover, total dryland grass cover) and the floristic patterns were tested using vector fitting, a procedure which determines the direction and correlation of the variables with the configuration of the ordination ^31,81,82.

Functional diversity change. We assessed functional diversity of early-melting snowpatches (1982 to 2022), using multiple continuous traits, with the indices proposed by Mason et al.⁸³ and modified by Villeger et al.⁷⁰ - functional richness, functional evenness and functional divergence. These functional diversity indices are complementary and describe the distribution of species and their abundances within the functional space and have the potential to reveal the processes that structure biological communities⁸⁴. Functional richness represents the amount of functional space occupied by the community, functional evenness corresponds to how regularly species abundances are distributed in the functional space, and functional divergence defines how far high species abundances are from the centre of the functional space (see ⁷⁰ for a full description of these indices). Analyses were performed in R using the functional diversity code provided by Villeger et al.⁷⁰

Community trait-weighted means for four continuous plant traits (plant height, LDMC, SLA, seed mass) were calculated for the 1982 dataset and the 2022 dataset using 1) the lowest shrub cover quadrat at each site (low) and 2) the highest shrub cover quadrat at each site (high). Community trait-weighted means were computed using the ‘cwm’ function from the BAT package⁸⁵. Trait data used in this analysis were standardised for each trait using the ‘scale’ function in R. This standardisation procedure scales the data by subtracting the mean from the original data and dividing by the standard deviation, shifting the centre of the distribution to 0 and scaling the standard deviation to 1.

Size-class distributions. While many studies use histograms to make inferences about recruitment dynamics, there is emerging evidence that histograms can be misinterpreted^86,87. To overcome this, we regressed the number of individuals (pooled across all sites) for shrub species that contributed at least 1% of individuals in the 2019 survey against the size class mid-point (x-axis) and used the slope of the regression line as a measure of recruitment success³⁷. We used eight shrub diameter size classes (modified from Condit et al.³⁷): 1–4, 5–9, 10–14, 15–19, 20–29, 30–39, 40–49, 50–100 cm. The mid-point value of each size class (di) was calculated. The number of shrubs within each class was divided by the width of the size class to obtain an adjusted total number of shrubs (ni) for each class (see Condit et al.³⁷). We chose to retain the 0 values where a size-class had ni = 0, as the removal of 0 values can influence regression slopes creating an impression that a population is experiencing more or less recruitment than it is⁸⁸. To retain values, 1 was added to all adjusted total shrub counts (ni + 1). Both ni + 1 and di were then log10 transformed. We then calculated a regression log10 (di) as the independent variable and log10 (ni + 1) as the dependent variable. The slopes of these regressions were then used to define recruitment type; all slopes are henceforth referred to as ‘Condit slopes’ which can be broadly defined as:

Type 1: Recruitment not limited: large numbers of small shrubs relative to larger shrubs suggesting ongoing recruitment over time - this might be evidence of climate release for shrubs in snowpatches; slopes are steeper with large negative values (− 1.0).

Type 2: Recruitment moderately limited: fewer smaller shrubs relative to larger shrubs; slopes are flatter, with values falling between 0 and − 1.0.

Type 3: Recruitment severely limited: few to no observable shrubs in the smaller size classes relative to larger shrubs, suggesting that recruitment has occurred in the past but is currently limited - this might be a signature of post-grazing disturbance recovery (bare ground gap regeneration) but with fewer recent opportunities as bare ground availability declines; the slopes are flat with positive values (> 0).

Data availability statement

All data and code used in this study are available at: https://github.com/zacwalkr/snowpatch_loss/

Acknowledgements

This work was conducted on the traditional lands of the Jaithmathang people and we acknowledge their elders past and present, as well as their ongoing support for our research. We thank Keith McDougall and Henrik Wahren whose original data allowed this revisitation study to be undertaken. Michael Nash provided information about snow country fauna. We thank Karen Ford, Alex Blackburn-Smith, Nina Kerr, Suzie Moss, Courtney Taylor and Iris Hickman for fieldwork support. This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Körner, C. Alpine plant life. (Springer Verlag, 1999).
Rixen, C. et al. Intraspecific trait variation in alpine plants relates to their elevational distribution. Journal of Ecology 110, 860–875 (2022).
Wipf, S., Stoeckli, V. & Bebi, P. Winter climate change in alpine tundra: plant responses to changes in snow depth and snowmelt timing. Climatic Change 94, 105–121 (2009).
Wilcox, E. J. et al. Tundra shrub expansion may amplify permafrost thaw by advancing snowmelt timing. Arctic Science 5, 202–217 (2019).
Walker, D. A., Halfpenny, J. C., Walker, M. D. & Wessman, C. A. Long-Term Studies of Snow-Vegetation Interactions. BioScience 43, 287–301 (1993).
Grippa, M. et al. The impact of snow depth and snowmelt on the vegetation variability over central Siberia. Geophysical Research Letters 32, (2005).
Kankaanpää, T. et al. Spatiotemporal snowmelt patterns within a high Arctic landscape, with implications for flora and fauna. Arctic, Antarctic, and Alpine Research 50, e1415624 (2018).
Assmann, J. J. et al. Local snow melt and temperature—but not regional sea ice—explain variation in spring phenology in coastal Arctic tundra. Global Change Biology 25, 2258–2274 (2019).
Bliss, L. C. Alpine Plant Communities of the Presidential Range, New Hampshire. Ecology 44, 678–697 (1963).
Larcher, W. & Siegwolf, R. Development of Acute Frost Drought in Rhododendron ferrugineum at the Alpine Timberline. Oecologia 67, 298–300 (1985).
Bokhorst, S., Bjerke, J. W., Street, L. E., Callaghan, T. V. & Phoenix, G. K. Impacts of multiple extreme winter warming events on sub-Arctic heathland: phenology, reproduction, growth, and CO2 flux responses. Global Change Biology 17, 2817–2830 (2011).
Billings, W. D. & Bliss, L. C. An Alpine Snowbank Environment and Its Effects on Vegetation, Plant Development, and Productivity. Ecology 40, 388–397 (1959).
Venn, S. E., Green, K., Pickering, C. M. & Morgan, J. W. Using plant functional traits to explain community composition across a strong environmental filter in Australian alpine snowpatches. Plant Ecol 212, 1491–1499 (2011).
Good, M., Morgan, J. W., Venn, S. & Green, P. Timing of snowmelt affects species composition via plant strategy filtering. Basic and Applied Ecology 35, 54–62 (2019).
Williams, R. J. et al. An International Union for the Conservation of Nature Red List ecosystems risk assessment for alpine snow patch herbfields, South-Eastern Australia. Austral Ecology 40, 433–443 (2015).
Camac, J. S. et al. Predicting species and community responses to global change using structured expert judgement: An Australian mountain ecosystems case study. Global Change Biology 27, 4420–4434 (2021).
Körner, C. & Hiltbrunner, E. Why Is the Alpine Flora Comparatively Robust against Climatic Warming? Diversity 13, 383 (2021).
Ji, F. et al. Rapid Warming in the Australian Alps from Observation and NARCliM Simulations. Atmosphere 13, 1686 (2022).
Slatyer, R. A., Umbers, K. D. L. & Arnold, P. A. Ecological responses to variation in seasonal snow cover. Conservation Biology 36, e13727 (2022).
Niittynen, P., Heikkinen, R. K. & Luoto, M. Decreasing snow cover alters functional composition and diversity of Arctic tundra. Proceedings of the National Academy of Sciences 117, 21480–21487 (2020).
Myers-Smith, I. H. et al. Shrub expansion in tundra ecosystems: dynamics, impacts and research priorities. Environ. Res. Lett. 6, 045509 (2011).
Matteodo, M., Ammann, K., Verrecchia, E. P. & Vittoz, P. Snowbeds are more affected than other subalpine–alpine plant communities by climate change in the Swiss Alps. Ecology and Evolution 6, 6969–6982 (2016).
Verrall, B., Green, K. & Pickering, C. M. Temporal dynamics in alpine snowpatch plants along a snowmelt gradient explained by functional traits and strategies. Oecologia 201, 155–171 (2023).
Galen, C. & Stanton, M. L. Responses of Snowbed Plant Species to Changes in Growing-Season Length. Ecology 76, 1546–1557 (1995).
Björk, R. G. & Molau, U. Ecology of Alpine Snowbeds and the Impact of Global Change. Arctic, Antarctic, and Alpine Research 39, 34–43 (2007).
Kirkpatrick, J. B. et al. Causes and consequences of variation in snow incidence on the high mountains of Tasmania, 1983–2013. Aust. J. Bot. 65, 214–224 (2017).
Olden, J. D. & Poff, N. L. Toward a Mechanistic Understanding and Prediction of Biotic Homogenization. The American Naturalist 162, 442–460 (2003).
Olden, J. D. Biotic homogenization: a new research agenda for conservation biogeography. Journal of Biogeography 33, 2027–2039 (2006).
Pickering, C., Green, K., Barros, A. A. & Venn, S. A resurvey of late-lying snowpatches reveals changes in both species and functional composition across snowmelt zones. Alp Botany 124, 93–103 (2014).
Bergstrom, D. M. et al. Combating ecosystem collapse from the tropics to the Antarctic. Global Change Biology 27, 1692–1703 (2021).
Wahren, C.-H., Williams, R. J. & Papst, W. A. Alpine and Subalpine Snow Patch Vegetation on the Bogong High Plains, SE Australia. Journal of Vegetation Science 12, 779–790 (2001).
Edmonds, T., Lunt, I. D., Roshier, D. A. & Louis, J. Annual variation in the distribution of summer snowdrifts in the Kosciuszko alpine area, Australia, and its effect on the composition and structure of alpine vegetation. Austral Ecology 31, 837–848 (2006).
Green, K. & Pickering, C. Vegetation, microclimate and soils associated with the latest-lying snowpatches in Australia. Plant Ecology & Diversity 2, 289–300 (2009).
Bormann, K. J., Brown, R. D., Derksen, C. & Painter, T. H. Estimating snow-cover trends from space. Nature Clim Change 8, 924–928 (2018).
Fiddes, S. L., Pezza, A. B. & Barras, V. A new perspective on Australian snow. Atmospheric Science Letters 16, 246–252 (2015).
Jump, A. S. & Peñuelas, J. Running to stand still: adaptation and the response of plants to rapid climate change. Ecology Letters 8, 1010–1020 (2005).
Condit, R., Sukumar, R., Hubbell, S. P. & Foster, R. B. Predicting Population Trends from Size Distributions: A Direct Test in a Tropical Tree Community. The American Naturalist 152, 495–509 (1998).
Schöb, C., Kammer, P. M., Choler, P. & Veit, H. Small-scale plant species distribution in snowbeds and its sensitivity to climate change. Plant Ecol 200, 91–104 (2009).
Formica, A., Farrer, E. C., Ashton, I. W. & Suding, K. N. Shrub expansion over the past 62 years in Rocky Mountain alpine tundra: possible causes and consequences. Arctic, Antarctic, and Alpine Research 46, 616–631 (2014).
McDougall, K. L. The alpine vegetation of the Bogong High Plains. Environmental Studies Publication No. 357. (1982).
McDougall, K. L. & Walsh, N. G. Treeless vegetation of the Australian Alps. Cunninghamia 10, 1–57 (2007).
Venn, S. E., Kirkpatrick, J. B., McDougall, K. L., Whinam, J. & Williams, R. J. Alpine, sub-alpine and sub-Antarctic vegetation of Australia. in Australian Vegetation 461–490 (Cambridge University Press, 2017).
McGowan, H. et al. Global warming in the context of 2000 years of Australian alpine temperature and snow cover. Sci Rep 8, 4394 (2018).
Amagai, Y., Kudo, G. & Sato, K. Changes in alpine plant communities under climate change: Dynamics of snow-meadow vegetation in northern Japan over the last 40 years. Applied Vegetation Science 21, 561–571 (2018).
Grime, J. P. Plant strategies and vegetation processes. (John Wiley & Son, 1979).
Zhang, H., Chen, S.-C., Bonser, S. P., Hitchcock, T. & Moles, A. T. Factors that shape large-scale gradients in clonality. Journal of Biogeography 50, 827–837 (2023).
Goldberg, D. E., Batzer, E., Elgersma, K., Martina, J. & Klimešová, J. Allocation to clonal growth: Critical questions and protocols to answer them. Perspectives in Plant Ecology, Evolution and Systematics 43, 125511 (2020).
Klimešová, J., Martínková, J. & Herben, T. Horizontal growth: An overlooked dimension in plant trait space. Perspectives in Plant Ecology, Evolution and Systematics 32, 18–21 (2018).
Aarssen, L. W. Death without sex—the ‘problem of the small’ and selection for reproductive economy in flowering plants. Evol Ecol 22, 279–298 (2008).
Herben, T., Šerá, B. & Klimešová, J. Clonal growth and sexual reproduction: tradeoffs and environmental constraints. Oikos 124, 469–476 (2015).
Silvertown, J. The Evolutionary Maintenance of Sexual Reproduction: Evidence from the Ecological Distribution of Asexual Reproduction in Clonal Plants. International Journal of Plant Sciences 169, 157–168 (2008).
Kuussaari, M. et al. Extinction debt: a challenge for biodiversity conservation. Trends in Ecology & Evolution 24, 564–571 (2009).
Camac, J. S. et al. Modeling rates of life form cover change in burned and unburned alpine heathland subject to experimental warming. Oecologia 178, 615–628 (2015).
Niittynen, P., Heikkinen, R. K. & Luoto, M. Snow cover is a neglected driver of Arctic biodiversity loss. Nature Clim Change 8, 997–1001 (2018).
Brown, C. & Cahill Jr, J. F. Competitive size asymmetry, not intensity, is linked to species loss and gain in a native grassland community. Ecology 103, e3675 (2022).
Schuurman, G. W. et al. Navigating Ecological Transformation: Resist–Accept–Direct as a Path to a New Resource Management Paradigm. BioScience 72, 16–29 (2022).
Williams, R. J. & Costin, A. B. Alpine and subalpine vegetation. in Australian Vegetation 465–500 (Cambridge University Press., 1994).
CSIRO. State of the Climate Technical Report. (2020).
Luca, A. D., Evans, J. P. & Ji, F. Australian snowpack in the NARCliM ensemble: evaluation, bias correction and future projections. Clim Dyn 51, 639–666 (2018).
Hennessy, K. et al. The impact of climate change on snow conditions in mainland Australia. (CSIRO Publishing, 2003).
Costin, A. B., Gray, M., Totterdell, C. J. & Wimbush, D. J. Kosciuszko alpine flora. (CSIRO Publishing, 2000).
Billings, W. D. Vegetational pattern near alpine timberline as affected by fire-snowdrift interactions. Vegetatio Acta Geobot 19, 192–207 (1969).
Mark, A. F. & Bliss, L. C. The high-alpine vegetation of Central Otago, New Zealand. New Zealand Journal of Botany 8, 381–451 (1970).
Ross, L. C., Woodin, S. J., Hester, A., Thompson, D. B. A. & Birks, H. J. B. How important is plot relocation accuracy when interpreting re-visitation studies of vegetation change? Plant Ecology & Diversity 3, 1–8 (2010).
Kapfer, J. et al. Resurveying historical vegetation data – opportunities and challenges. Applied Vegetation Science 20, 164–171 (2017).
Westoby, M., Falster, D. S., Moles, A. T., Vesk, P. A. & Wright, I. J. Plant ecological strategies: Some leading dimensions of variation between species. Annual Review of Ecology and Systematics 33, 125–159 (2002).
Cornelissen, J. H. C. et al. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust. J. Bot. 51, 335–380 (2003).
Weiher, E. et al. Challenging Theophrastus: A common core list of plant traits for functional ecology. Journal of Vegetation Science 10, 609–620 (1999).
Morgan, J. W. & Venn, S. E. Alpine plant species have limited capacity for long-distance seed dispersal. Plant Ecology 218, 813–819 (2017).
Villéger, S., Mason, N. W. H. & Mouillot, D. New Multidimensional Functional Diversity Indices for a Multifaceted Framework in Functional Ecology. Ecology 89, 2290–2301 (2008).
Harper, J. L. The population biology of plants. (Academic Press, 1977).
Wahren, C.-H. Vegetation dynamics on the Bogong High Plains. (Monash University, 1997).
R Core Team. R: A language and environment for statistical computing. (2023).
Jost, L. Entropy and diversity. Oikos 113, 363–375 (2006).
Roswell, M., Dushoff, J. & Winfree, R. A conceptual guide to measuring species diversity. Oikos 130, 321–338 (2021).
Oksanen, J. et al. vegan: Community Ecology Package. (2022).
Baselga, A. et al. betapart: Partitioning Beta Diversity into Turnover and Nestedness Components. (2023).
Anderson, M. J., Ellingsen, K. E. & McArdle, B. H. Multivariate dispersion as a measure of beta diversity. Ecology Letters 9, 683–693 (2006).
Minchin, P. R. An evaluation of the relative robustness of techniques for ecological ordination. Vegetatio 69, 89–107 (1987).
Quinn, G. P., Lake, P. S. & Schreiber, E. S. G. Littoral benthos of a Victorian lake and its outlet stream: Spatial and temporal variation. Australian Journal of Ecology 21, 292–301 (1996).
Dargie, T. C. D. On the integrated interpretation of indirect site ordinations: a case study using semi-arid vegetation in southeastern Spain. Vegetatio 55, 37–55 (1984).
Faith, D. P., Minchin, P. R. & Belbin, L. Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69, 57–68 (1987).
Mason, N. W. H., Mouillot, D., Lee, W. G. & Wilson, J. B. Functional richness, functional evenness and functional divergence: the primary components of functional diversity. Oikos 111, 112–118 (2005).
Mouchet, M. A., Villéger, S., Mason, N. W. H. & Mouillot, D. Functional diversity measures: an overview of their redundancy and their ability to discriminate community assembly rules. Functional Ecology 24, 867–876 (2010).
Cardoso, P., Mammola, S., Rigal, F. & Carvalho, J. BAT: Biodiversity Assessment Tools. (2022).
Boels, L., Bakker, A., Van Dooren, W. & Drijvers, P. Conceptual difficulties when interpreting histograms: A review. Educational Research Review 28, 100291 (2019).
Lem, S., Onghena, P., Verschaffel, L. & Van Dooren, W. On the misinterpretation of histograms and box plots. Educational Psychology 33, 155–174 (2013).
Lykke, A. M. Assessment of species composition change in savanna vegetation by means of woody plants’ size class distributions and local information. Biodiversity and Conservation 7, 1261–1275 (1998).

No competing interests reported.

Snowpatchlosssupplementarymaterial.docx

Download PDF

Journal Publication

published 02 Oct, 2023

Read the published version in Scientific Reports →

Editorial decision: Major revision
03 Jul, 2023
Reviews received at journal
03 Jul, 2023
Reviews received at journal
18 Jun, 2023
Reviews received at journal
15 Jun, 2023
Reviewers agreed at journal
07 Jun, 2023
Reviewers agreed at journal
04 Jun, 2023
Reviewers invited by journal
03 Jun, 2023
Editor assigned by journal
03 Jun, 2023
Editor invited by journal
31 May, 2023
Submission checks completed at journal
31 May, 2023
First submitted to journal
23 May, 2023

You are reading this latest preprint version

Early-melting snowpatch plant communities are transitioning into novel ecosystems

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Data analyses

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1