Skip to main content
NCBI home page
Annals of Botany logoLink to Annals of Botany
. 2020 Jun 1;126(5):873–881. doi: 10.1093/aob/mcaa104

The role of perennation traits in plant community soil frost stress responses

Frederick Curtis Lubbe 1,2,✉, Hugh A L Henry 1
PMCID: PMC7539335  PMID: 32478386

Abstract

Background and Aims

Herbaceous plants can survive periods of prolonged freezing as below-ground structures or seed, which can be insulated from cold air by soil, litter or snow. Below-ground perennial structures vary in both form and their exposure to soil frost, and this structural variation thus may be important in determining the responses of plant communities to frost stress.

Methods

We conducted a suite of snow removal experiments in a northern temperate old field over 3 years to examine the relative freezing responses of different plant functional groups based on below-ground perennation traits. A litter removal treatment was added in the third year. Species-level percentage cover data were recorded in May, June and September then pooled by functional group.

Key Results

Snow removal decreased total plant cover, and this response was particularly strong and consistent among years for tap-rooted and rhizomatous species. The snow removal responses of cover for plants with root buds and new recruits from seed varied from positive to negative among years. The cover of rootstock plants consistently increased in response to snow removal. Rhizomatous species were generally the most vulnerable to litter removal.

Conclusions

This study is the first to explore the effects of variation in frost severity on the responses of different plant perennation trait functional groups. The responses of herbaceous species to frost may become increasingly important in northern temperate regions in the coming decades as a result of declining snow cover and increasing temperature variability. Our results reveal substantial variation in responses among perennation trait functional groups, which could drive changes in species abundance in response to variation in soil frost.

Keywords: Below-ground, community, herbaceous, frost, life form, perennation, snow removal, taproot

INTRODUCTION

Frost is an important stress that limits both the seasonal activity and global distribution of many plant species (Box, 1996). During periods of frequent or extended frost, perennial herbaceous plants often enter dormancy, at which point many of their above-ground tissues senesce (Raunkiær, 1934; Klimešová et al., 2015). Upon breaking dormancy, subsequent regeneration of tissue can depend highly on the survival and condition of perennial below-ground structures (Klimešová et al., 2015). Plants can circumvent freezing stress through physiological tolerance and avoidance of freezing at the cellular level (Pearce, 2001; Davik et al., 2013) or through spatial and temporal avoidance of freezing at the plant organ level (Raunkiær, 1934; Klimešová et al., 2015). Snow cover can be an important source of surface insulation from cold air temperatures during winter (Bertrand and Castonguay, 2003), and reductions in snow cover can therefore cause plant community shifts (Komac et al., 2015). Similarly, senesced stem and leaf tissue (plant litter) can decrease frost penetration into soil, thereby protecting vulnerable buds (Sharratt, 2002). The positioning of vulnerable tissue deep in the soil also can decrease its exposure to frost (Boydston et al., 2006).

Despite the potential benefits of plant structures being positioned deep in the soil or under thick litter for minimizing exposure to a range of biotic and abiotic stresses and disturbance (Vesk and Westoby, 2004; Boydston et al., 2006; Baseggio et al., 2015), these scenarios can result in delayed emergence and impaired above-ground growth in the spring, potentially reducing competitive ability (Hartnett and Keeler, 1995; Pan et al., 2009). Therefore, there may be a trade-off between protection from frost and competitive ability that is mediated by the thickness of soil and litter cover (Lubbe and Henry, 2019). Moreover, the balance of this trade-off can be dynamic; stress avoidance may be important for survival in one year, but in another year that features milder conditions a riskier, more competitive strategy may be advantageous (Grime, 1977).

Perennial structures are important for plant persistence, but also for clonal spread and recruitment (Klimešová et al., 2015, 2017). Rhizomes are an important organ of clonality, and they can achieve large amounts of growth and lateral spread via extension of internode length (Cornelissen et al., 2014). However, the abundance of horizontal stem tissue within a centimetre or two below the soil surface may increase vulnerability to stressors originating above-ground. Taproots are generally thickened, vertical roots can be branched to varying degrees (Chmelíková and Hejcman, 2012), and stem bud recruitment from taproots occurs at the soil surface. The shallow base is prone to damage and ageing, which can induce root-splitting, a form of clonal growth with minimal lateral spread (Chmelíková and Hejcman, 2012; Klimešová et al., 2017). Plants also can develop a thickened perennial stem base while forming fibrous roots. In eudicots, this structure can produce annual rings and can thicken like in woody plants (Schweingruber and Poschlod, 2005); the structure and terminology varies, but here we refer to this structure as a rootstock. Plants with monopodial branching and epigeogenous rhizomes with very short internodes are included, because of their similarity in structure. Some plants also can induce stem buds from lateral and adventitious roots (Bartušková et al., 2017). These root bud banks allow the recruitment of buds deep in the soil profile.

Life form classification categorizes plants based on the location of their vulnerable tissues during stressful seasons (Raunkiær, 1934), and in many temperate regions winter frost stress is used to determine this category. Hemicryptophytes have buds at or just below the soil surface, and this is the most common life form for perennial temperate herbs, especially in fields (Benson and Hartnett, 2006; Klimešová, 2018). Hemicryptophytes can be clonal or non-clonal, and they often have rhizomes (below-ground horizontal stems) (Raunkiær, 1934; Komac et al., 2015). Geophytes have buds positioned deep in the soil, and they often produce clonal structures, such as bulbs, stem tubers and rhizomes, with high storage capacity (Raunkiær, 1934). In temperate regions these species are most common in woodlands (Kamenetsky, 2013). Therophytes are annuals and do not overwinter as vegetative structures, but instead use seed (Raunkiær, 1934).

Seed production can be considered another overwintering strategy employed by plants. Seed response to frost varies, in that while seeds can be damaged by severe freezing, moderate freezing can be required for germination (Chouard, 1960). Unlike clonal growth, which is limited by the speed of lateral spread, seed dispersal and seed banks can be particularly beneficial for the colonization of large soil patches denuded of live vegetation by disturbance. Although seeds are often a relatively tolerant life stage, newly germinated plants can be particularly vulnerable to frost, and seedlings often emerge in the spring, at a time when the risk of potentially damaging freeze–thaw cycles is high (Walck et al., 2011; Connolly and Orrock, 2015).

We performed snow removal experiments for 3 years in an intact, self-assembled, old field plant community to determine the effects of increased frost stress, both at the species level and with respect to groups based on categorical below-ground functional traits relevant to perennation. In the third year we also combined snow removal with a litter removal treatment, with the prediction that the latter would increase plant frost exposure. Plant responses were quantified using three cover surveys per growing season (a new set of plots was treated and assessed each year), and the cover data were then pooled based on categorical functional traits with respect to recruitment method, life form and organ of perennation. Overall, we predicted that increased frost stress would decrease total plant total cover and, more specifically, it would decrease disproportionately for plants with traits associated with shallow bud depth (i.e. within the top 2 cm of the soil surface) and vulnerable below-ground structures. Susceptible groups were expected to include hemicryptophytes, particularly taproot- and rhizome-bearing species. We also predicted that plants dependent on seeds would be the most resistant to frost stress.

MATERIALS AND METHODS

Site

Experiments were conducted in an old field at the Western University Environmental Sciences Western Field Station (ESW), near Ilderton, ON, Canada (43°04′37.6″ N, 81°20′16.1″ W). The soil at the site was characterized as Bryanston silt loam, which is a Brunisolic Gray Brown Luvisol (Hagerty and Kingston, 1992), and it had a mean pH of 7.5. The field, which was bordered by a woodland, cropland and maintained grass pathways and lawns, had been left fallow and permitted to naturalize since 2013. Dominant plant species were rhizomatous goldenrod (Solidago spp.), dandelion (Taraxacum officinale) and red clover (Trifolium pratense), the latter two species both growing from taproots. There was also a notable presence of asters (Symphyotrichum spp.) and thistles (Cirsium spp. and Sonchus spp.). Wild carrot (Daucus carota) and birdsfoot trefoil (Lotus corniculatus) also were abundant in patches.

Snow removal (year 1)

Six pairs of 1 m × 2 m plots (12 plots total) were laid out, with the pairings based on proximity and vegetation similarity. Over the winter of 2015/16, we administered either snow removal via shovelling or control (no snow removal) to one plot in each pair. Before treatment, all plots (snow removal and control) were overlain with white plastic netting with 1-cm openings (Protective Winter Wrap; Quest Plastics, Mississauga, ON, Canada) to denote the shovelling depth and to prevent the removal of litter and disturbance of the soil surface. Soil temperature probes (LogTag TRIX-8; MicroDAQ, NH, USA) were placed 2 cm deep in the centre of three plots from each treatment to record soil temperature hourly. Snow was removed opportunistically after snowfall events of >3 cm that were forecast to be followed by below-freezing temperatures, which resulted in nine snow removal events from 12 January 2016 until 17 February 2016. Snow removal was halted before the end of winter to minimize potential snow removal effects on soil moisture over the subsequent growing season. The plastic netting was removed on 18 April 2016 to avoid interference with above-ground plant growth. Plant cover was estimated based on the Domin–Krajina cover-abundance scale (Mueller‐Dombois and Ellenberg, 1974) and was surveyed to the nearest 5 % (but also including a 1 % cover category for species with ≤1 % cover within the plot) for each species present. Solidago altissima and Solidago canadensis were combined as Solidago spp., because of the inability to distinguish between them during the first two surveys of the year. We conducted cover surveys on 11–17 May, 8 June and 14–16 September 2016 to capture initial and peak growth of the common species and focal functional groups.

Snow removal (year 2)

The experiment was repeated a second time during the winter of 2016/17 for a new set of plots (12 pairs of 1 m × 1 m plots, 24 plots total). Snow was removed four times from 13 December 2016 until 14 March 2017. Winter wrap was removed on 24 April 2017. We conducted plant cover surveys on 23 May, 14 June and 12 September 2017.

Snow and winter litter removal (year 3)

The experiment was repeated during the winter of 2017/18 with another new set of plots, with the addition of a winter litter removal component (applied as a full factorial experiment in combination with the snow removal treatment; ten blocks of 1 m × 1 m plots, 40 plots total). Each treatment combination was applied randomly within each block. For winter litter removal, all above-ground biomass was cut on 27 November 2017 and placed adjacent to the plot to allow it to undergo decomposition under field conditions. For the remaining plots, litter was cut and removed, but placed immediately back onto the plot to produce the same disturbance effect as the litter removal plots. Litter primarily consisted of stem and leaf tissues; removal of seeds would have been minimal. Snow was removed seven times, from 26 December 2017 until 12 February 2018. The plastic netting was removed, and the litter was placed back on the winter removal plots on 11 April 2018. We conducted plant cover surveys on 14–17 May, 21–22 June and 11–12 September 2018.

Snow cover and air and soil temperatures: years 1–3

The snow removal experiments were conducted in three contrasting winters. The first was relatively mild (average daily temperature 0 °C) with regular snow cover late in the season (3.5 cm average snow cover from December to March); the second was colder (average daily temperature −1 °C) but with relatively low snowfall later in the season (4.0 cm average snow cover from December to March); and the third was cold (average daily temperature −4 °C) but with relatively high snow cover (6.9 cm average snow cover from December to March) (Environment Canada, 2019). Snow removal decreased minimum soil temperature and increased the number of soil freeze–thaw cycles, although the effect sizes varied by year (Table 1). Litter removal also decreased minimum temperature and increased the number of freeze–thaw cycles, both combined with and separate from snow removal (Table 1).

Table 1.

Minimum and average temperatures and number of freeze–thaw cycles at 2 cm soil depth under the different litter and snow removal treatments (standard error displayed; n = 3). Freeze–thaw cycles were defined as any drop below 0 °C followed by an increase to above 0 °C

Control Removal
Minimum (°C) Average (°C) Cycles Minimum (°C) Average (°C) Cycles
Year 1 −1.2 ± 0.1 1.9 ± 0.1 10 ± 2 −2.0 ± 0.2 1.9 ± 0.1 11 ± 3
Year 2 −3.5 ± 0.3 1.8 ± 0.1 21 ± 1 −6.0 ± 0.7 2.5 ± 0.1 66 ± 13
Year 3 with litter −1.6 ± 0.8 1.2 ± 0.1 8 ± 4 −5.3 ± 0.1 0.5 ± 0.1 16 ± 1
Year 3 without litter −3.1 ± 0.4 0.8 ± 0.1 21 ± 4 −6.7 ± 1 0.4 ± 0.1 22 ± 4
Open in a new tab

Plant functional group categorization

The plant functional type of each species present was assessed with respect to three different categories: recruitment method, life form, and organ of perennation (Supplementary Data Table S1). Percentage cover for the species exhibiting each trait was totalled. Species trait data were acquired from databases (Fitter and Peat, 1994; Klimešová et al., 2017; Native Plant Trust, 2019), the literature (Bhowmik and Bandeen, 1976; Turkington and Cavers, 1979; Werner et al., 1980; Lemna and Messersmith, 1990; Lemieux et al., 1993; Chmielewski and Semple, 2001a, b, 2003; Stewart-Wade et al., 2002; Klimešová, 2018), and personal observation. Species identification was not possible for all plants present, and thus phylogenetic corrections were not applied. Obligate biennials were assessed by emergence and size. Recruitment methods included (1) germination (genets that overwintered as seed), and (2) resprouting (ramets that overwintered as vegetative structures). Life form was based on the location of sensitive tissue (buds) during the harshest season (winter in this region) (Raunkiær, 1934), and included (1) hemicryptophytes (buds at or just below the soil surface), (2) geophytes (buds deep in the soil) and (3) therophytes (overwintering as seeds). Organs of perennation included only structures that persisted through the winter. The structures present were (1) rhizomes (below-ground horizontal stems of the hypogeogenous type that have long internodes and high potential for lateral spread), (2) rootstocks (concentrated below-ground stem bases with greater likelihood of woodiness, including epigeogenous stems that have very short internodes), (3) taproots, (4) root buds (stem bud-bearing roots), and (5) none (overwintering as seeds). Rhizome- and rootstock-bearing plants both exhibited fibrous roots. Symphyotrichum lanceolatum has hypogeogenous rhizomes, but these have low conductivity and short lifespans (Chmielewski and Semple, 2001a); thus, this species functions more similarly to a rootstock species and is classified as one in this system. There was some overlap between categories. Specifically, the resprouting species category was composed of both hemicryptophyte and geophyte species. The hemicryptophyte category contained rhizome-, taproot- and rootstock-bearing species. The only geophytes in this study were root bud-bearing species. The seed germination, no organ of perennation (overwintering as seeds) and therophyte categories all represented species that depend on seeds for overwintering, and thus contained the same species.

Data analyses

Total cover and species cover, both individual and summed for each functional trait group, were the response variables for each treatment. Total, species and functional trait cover for surveys in years 1 and 2 was analysed with paired one-tailed t-tests (for the species-level analyses, only the dominant species, which were present in the majority of plots, were examined). Species and functional trait cover surveys for year 3 were analysed with two-way ANOVAs with block as a random factor and further analysed with paired one-tailed t-tests for each treatment variable (snow removal or litter removal) when no interaction was present. Normality was verified visually by analysis of normal quantile plots. Analyses were conducted using JMP version 13 (SAS Institute, 1989–2019).

RESULTS

Total cover and dominant species responses

Total cover decreased with snow removal significantly in May in year 1 (t10 = −3.25, P = 0.004) and there was a trend of a decrease in September (t10 = −1.73, P = 0.06) (Fig. 1). Trifolium pratense cover decreased in response to snow removal for all three cover surveys (t10 = −4.46, P = 0.0006, t10 = −4.20, P = 0.001 and t10 = −4.42, P = 0.0006, respectively) and there was a marginally significant decrease for Taraxacum officinale (t10 = −1.68, P = 0.06) for the September survey (Table 2). There were no significant snow removal effects on total cover in year 2 (Fig. 1), although Solidago spp. cover decreased significantly with snow removal for the May survey (t22 = −2.05, P = 0.03) and there was a marginally significant decrease for this species (t22 = −1.40, P = 0.09) for the September survey (Table 2). There also were marginally significant decreases in Trifolium pratense cover in May and June in year 2 (t22 = −1.66, P = 0.06 and t22 = −1.68, P = 0.05, respectively) (Table 2). In year 3, snow removal decreased total cover for all three surveys (t38 = −3.01, P = 0.002, t38 = −2.73, P = 0.005 and t38 = −2.35, P = 0.01) (Fig. 1), with a significant decrease in Trifolium pratense cover in May (P = 0.004) and decreases in Taraxacum officinale cover in May (P = 0.07) and June (P = 0.04) (Table 2). There was a marginally significant decrease in total cover in response to litter removal in May of year 3 (t38 = −1.51, P = 0.07).

Fig. 1.

Fig. 1.

Open in a new tab

Mean and standard error for total plant percentage cover for years 1, 2 and 3. For years 1 and 2: +P < 0.1, *P < 0.05. Subscripts for P values in year 3 are snow removal (S), litter removal (L) and snow removal × litter removal interaction (I).

Table 2.

Mean and standard error for percentage cover for the dominant species in response to snow and litter removal. All dominant species are resprouting hemicryptophytes. Taraxacum officinale and Trifolium pratense have taproots while Solidago spp. are rhizomatous

May June September
Species Control Removal Control Removal Control Removal
Solidago spp. 35 ± 4 31 ± 2 47 ± 6 53 ± 7 50 ± 3 53 ± 7
Year 1 Taraxacum officinale 18 ± 5 13 ± 4 19 ± 5 17 ± 4 19 ± 6+ 13 ± 3
Trifolium pratense 10 ± 3*** 2 ± 1 10 ± 3*** 3 ± 1 20 ± 5*** 4 ± 2
Solidago spp. 57 ± 3* 50 ± 4 38 ± 5 37 ± 5 43 ± 5+ 38 ± 5
Year 2 Taraxacum officinale 13 ± 2 13 ± 2 12 ± 2 13 ± 2 2 ± 1 2 ± 1
Trifolium pratense 9 ± 3+ 5 ± 2 20 ± 6+ 12 ± 3 25 ± 5 28 ± 5
Year 3 with litter Solidago spp. 21 ± 2+ 20 ± 2+ 30 ± 6 31 ± 5 30 ± 7 27 ± 3
Taraxacum officinale 14 ± 2+ 11 ± 2 21 ± 3* 16 ± 2 14 ± 3 15 ± 3
Trifolium pratense 27 ± 8*** 21 ± 6 34 ± 9 31 ± 8 43 ± 7 38 ± 5
Year 3 without litter Solidago spp. 17 ± 2 19 ± 2 31 ± 6 27 ± 5 26 ± 4 30 ± 6
Taraxacum officinale 12 ± 2+ 12 ± 3 17 ± 3 17 ± 2 12 ± 3 14 ± 3
Trifolium pratense 27 ± 7*** 19 ± 6 36 ± 9 28 ± 8 43 ± 6 40 ± 6
Open in a new tab

+ P < 0.1; *P < 0.05; ***P < 0.001.

Recruitment method

The cover of germinated plants exhibited a trend of decline in response to snow removal in June of year 1 (t10 = −1.49, P = 0.08), but it increased in May and June in year 2 (t22 = 2.78, P = 0.005 and t22 = 1.74, P = 0.05, respectively) (Fig. 2A). Snow and litter removal had interactive effects on the cover of germinated plants in May of year 3 (P = 0.021) and they decreased in response to snow removal in June and September (t38 −2.59, P = 0.007 and t38 = −3.06, P = 0.002, respectively) and increased in response to litter removal in June (t38 = 1.72, P = 0.05). The cover of resprouted plants decreased in response to snow removal in May of each year (t10 = −3.57, P = 0.002, t22 = −3.94, P = 0.0003 and t38 = −3.25, P = 0.001, respectively), and there was a trend of a decrease in response to litter removal in May (t38 = −1.31, P = 0.1) (Fig. 2B).

Fig. 2.

Fig. 2.

Open in a new tab

Mean and standard error for trait groups based on recruitment method for years 1, 2 and 3. (A) Resprouting, (B) germination. For years 1 and 2: +P < 0.1, *P < 0.05. Subscripts for P values in year 3 are snow removal (S), litter removal (L) and snow removal × litter removal interaction (I).

Life form

Hemicryptophyte cover decreased with snow removal in every May survey (t10 = −3.72, P = 0.002, t22 = −4.13, P = 0.0002 and t38 = −2.70, P = 0.005, respectively), in the June survey in year 3 (t38 = −2.36, P = 0.01) and in the September survey in year 1 (t10 = −1.89, P = 0.04) (Fig. 3A). Litter removal decreased the cover of these species in May of year 3 (t38 = −1.49, P = 0.07). Geophyte cover increased in response to snow removal in year 1 in all surveys (t10 = 2.24, P = 0.02, t10 = 3.64, P = 0.002 and t10 = 3.00, P = 0.006) and in year 2 there was a trend of an increase in May in response to snow removal (t22 = 1.41, P = 0.09) (Fig. 3B). Geophyte cover decreased in response to snow removal in May in year 3 (t38 = −2.80, P = 0.004). Totals and results for therophyte cover (Fig. 3C) were the same as for the cover of germinated plants described above (i.e. these groups contained the same set of species).

Fig. 3.

Fig. 3.

Open in a new tab

Mean and standard error for life forms for years 1, 2 and 3. (A) Hemicryptophytes, (B) geophytes, (C) therophytes. For years 1 and 2: +P < 0.1, *P < 0.05. Subscripts for P values in year 3 are snow removal (S), litter removal (L) and snow removal × litter removal interaction (I).

Organ of perennation

The cover of plants that produce rhizomes decreased with snow removal in May of year 1 (t10 = −1.63, P = 0.07) and there was a significant decrease in the cover of these species in May of year 2 (t22 = −2.22, P = 0.02). Rhizomatous species cover decreased in May of year 3 with litter removal (t38 = −2.16, P = 0.02) (Fig. 4A). The cover of plants that produce rootstocks increased in response to snow removal in June of year 1 (t10 = 2.24, P = 0.02) and May of year 2 (t22 = 1.96, P = 0.03) and there was a trend of an increase in May of year 3 (t38 = 1.50, P = 0.07) (Fig. 4B). The cover of plants that produce taproots generally decreased with snow removal (Fig. 4C). The cover of these species decreased in response to snow removal in all three May surveys (t10 = −2.87, P = −0.008, t22 = −2.02, P = 0.03 and t38 = −3.69, P = 0.0004 respectively), with trends of decreases in June in years 1 and 2 (t10 = −1.70, P = 0.06 and t22 = −1.38, P = 0.09, respectively). Cover of these species also decreased significantly in response to snow removal in June of year 3 (t38 = −2.48, P = 0.009) and September of year 1 (t10 = −5.81, P = 0.0001). The results for species with root buds (Fig. 4D) were the same as those for geophytes, described above, and the results for species with no organ of perennation were the same as for therophytes (Fig. 4E) (i.e. in both cases they represented the same species).

Fig. 4.

Fig. 4.

Open in a new tab

Mean and standard error for trait groups based on organ of perennation for years 1, 2 and 3. (A) rhizomes, (B) rootstock, (C) taproot, (D) root buds, (E) seed (no organ). For years 1 and 2: +P < 0.1, *P < 0.05. Subscripts for P values in year 3 are snow removal (S), litter removal (L) and snow removal × litter removal interaction (I).

DISCUSSION

Consistent with our prediction and previous results in the literature (Malyshev and Henry, 2012; Henry et al., 2018; Reinmann et al., 2019), snow removal generally had a negative effect on total plant cover, although both the magnitude and significance of this effect varied based on year and survey date. In particular, in year 2, which featured both cold temperatures and low snow cover, there was no snow removal effect on total plant cover, and plant cover did not increase appreciably across survey dates, unlike the other two years. Our interpretation here is that although very cold temperatures were experienced in the snow removal plots in year 2, the ambient plots also experienced similarly cold temperatures (i.e. colder than the snow removal plots in year 1) as a result of the low ambient snow cover. This could have resulted in the threshold temperature for plant damage (which often lies between 0 and −5 °C for the species at the study site; Malyshev and Henry, 2012; Kong and Henry, 2019a) being exceeded in all plots. While soil and plant tissue do not typically freeze at precisely 0 °C (McKersie and Lesham, 1994; Barnes, 2010), this threshold was informative for assessing the relative effects of the snow removal treatment on changes in soil freezing dynamics. Decreased total cover later in the growing season as a result of drought stress also could have masked any snow removal effect [this year experienced abnormally high rainfall in May (133 mm as opposed to 31 and 54 mm in years 1 and 3, respectively) followed by a drought for the next 4 months (192 mm as opposed to 395 and 335 mm in years 1 and 3, respectively); Environment Canada, 2019]. Litter cover also affected plant cover, but it was most influential for species overwintering as either seed or rhizomes.

With respect to recruitment method, resprouting decreased consistently in response to snow removal in each year, possibly from delayed growth caused by frost damage (e.g. damage to shallow buds resulting in sprouting of deeper buds). Although such an effect also was observed for germination in year 3 and there was no significant snow removal effect on germination in year 1, in year 2 germinated plants increased in response to snow removal. The latter response could have been driven by competitive release as a result of the decrease in resprouted plants, and thus an indirect response to snow removal (e.g. Kong and Henry, 2019b). In support of this hypothesis, we observed a general delay of emergence via germination compared with resprouting from vegetative structures, which could have provided a competitive advantage to resprouting plants in the absence of frost damage. In year 3, the recruitment of seeds in general was very low compared with the other years, whereas functional groups with vegetative structures for recruitment were much more successful. In May, germination recruitment was greatest when both snow and litter were present, but the effects of litter became more complicated in June when germination recruitment increased in both the control plots and the litter removal plots.

Snow removal effects also differed among life form categories and organ of perennation categories, with the cover of species with bud-bearing organs at or near the soil surface (hemicryptophytes) being most sensitive to snow removal, which was consistent with the relatively high intensity of frost exposure at the soil surface. Hemicryptophytes represent the majority of herbaceous species in many temperate habitats (Komac et al., 2015; Hameed et al., 2016; Klimešová, 2018), and they feature substantial structural diversity. Secondary thickening, composition and structure persistence all can vary among and within species, especially with age (Klimešová and Klimeš, 1996; Klimeš et al., 1999). Species with rootstocks generally had an early growth advantage in snow removal plots, as opposed to the other hemicryptophytes. This result likely occurred because of organ structure, with increased woodiness possibly offering greater protection from the cold (Wisniewski et al., 2003). Rhizomatous species responded roughly the same as hemicryptophytes as a whole during years 1 and 2, likely because of the presence of buds near the surface (within the top 2 cm), as well as their horizontal stems. In year 3, rhizomatous cover decreased with litter removal, suggesting that litter cover is important for insulation from cold air in hypogeogenous rhizomatous species. Soil surface temperatures may be particularly relevant for rhizomatous hemicryptophyte species, because of their high concentration of sensitive structures near the surface, and branching occurs parallel to the surface, as opposed to downward. We also observed particularly high sensitivity to freezing in some of these species (e.g. Trifolium species), whereas Solidago was less sensitive to snow removal. Taproots have a concentration of stem buds near the soil surface, and these structures are often quite vulnerable to damage (Chmelíková and Hejcman, 2012; Klimešová et al., 2017), including frost (Perfect et al., 1987). These structures are vertically long and cylindrical, and this root structure also may contribute to increased vulnerability to frost heave, which could push these structures even higher and increase exposure to freezing air temperatures, as compared with fibrous root systems (Perfect et al., 1987). It is also possible that these structures are damaged directly as a result of frost heave. Recruitment by vegetative structures in general was vulnerable to reduced growth in response to snow removal in the emergence and early growth stage, but it recovered later in the growing season. Species dependent upon recruitment from buds closer to the surface varied in their responses among years, but they remained negatively affected by snow removal in June in year 3 and even in September of year 1. Aside from cover, the effects on reproductive structures and investment in storage organs (not investigated here) also could reveal the potential for long-term legacy frost effects on plant community structure.

Species bearing buds from roots (the only geophytes present) were most successful in year 1 under snow removal, which may have occurred because these species form buds along the soil depth gradient (Moore, 1975; Lemna and Messersmith, 1990), allowing them to regenerate from deeper in the soil profile than their competitors, and thus avoid frost stress to a greater extent. The decrease in the cover of root-budding species in the first season of year 3 may have been driven by increased frost exposure of their shallow structures; successful root budding would then occur deeper in the soil, and this increased resprouting depth could come at the cost of later emergence. Although we did not assess decreases in soil freezing intensity with increasing soil depth, in a snow removal study conducted at the same field site minimum soil temperatures in the snow removal plots were −6.5, −3.9 and −1.0 °C for soil depths of 2, 5 and 15 cm, respectively, and the corresponding numbers of freeze–thaw cycles were 38, 10 and 4 (Lubbe and Henry, 2019), which demonstrates the extent to which greater soil depth could have provided escape from frost severity. For comparison, regarding the geophyte/root-budding species from this study, Sonchus arvensis can form buds along its stem base and root system as deep as 50 cm (Lemna and Messersmith, 1990), while the root system of Cirsium arvense can extend to at least to 38 cm depth (Moore, 1975); therefore, both species could have formed buds as low as 15 cm depth, where soil freezing stress was relatively mild in our study. In addition to the direct effects of temperature, close to the soil surface, soil heave also may have fragmented roots and severed connectivity to the parent plant, thus slowing growth.

Conclusions

Studies of plant below-ground traits generally have been under-represented in plant trait studies, and although interest has increased in recent years, much of this has been specifically for root traits. In contrast, there is still relatively little known about plant storage organs and perennation traits, and to our knowledge our study is the first to examine the response of these traits to frost stress. Knowledge of these responses at the functional trait level with respect to regeneration and perennation will provide a better understanding of how plant communities may respond to changes in soil frost in the coming decades, which are predicted based on decreased snow cover and increased temperature variability. Rare taproot-bearing species may be at greatest risk and thus possibly an important focus for conservation efforts.

SUPPLEMENTARY DATA

Supplementary data are available online at https://academic.oup.com/aob and consist of the following. Table S1: list of species and their functional groups in three categories: regeneration mode, organ of perennation and life form.

mcaa104_suppl_Supplementary_Data
Click here for additional data file. (10.2KB, xlsx)

FUNDING

This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC).

LITERATURE CITED

  1. Barnes G. 2010. Soil mechanics: principles and practice, 3rd edn.  New York: Palgrave Macmillan. [Google Scholar]
  2. Bartušková A, Malíková L, Klimešová J. 2017. Checklist of root-sprouters in the Czech flora: mapping the gaps in our knowledge. Folia Geobotanica  52: 337–343. [Google Scholar]
  3. Baseggio M, Newman Y, Sollenberger LE, Fraisse C, Obreza T. 2015. Planting rate and depth effects on Tifton 85 bermudagrass establishment using rhizomes. Crop Science  55: 1338–1345. [Google Scholar]
  4. Benson EJ, Hartnett DC. 2006. The role of seed and vegetative reproduction in plant recruitment and demography in tallgrass prairie. Plant Ecology  187: 163–178. [Google Scholar]
  5. Bertrand A, Castonguay Y. 2003. Plant adaptations to overwintering stresses and implications of climate change. Canadian Journal of Botany  81: 1145–1152. [Google Scholar]
  6. Bhowmik PC, Bandeen JD. 1976. The biology of Canadian weeds: 19. Asclepias syriaca L. Canadian Journal of Plant Science  56: 579–589. [Google Scholar]
  7. Box EO. 1996. Plant functional types and climate at the global scale. Journal of Vegetation Science  7: 309–320. [Google Scholar]
  8. Boydston RA, Seymour MD, Brown CR, Alva AK. 2006. Freezing behavior of potato (Solanum tuberosum) tubers in soil. American Journal of Potato Research  83: 305–315. [Google Scholar]
  9. Chmelíková L, Hejcman M. 2012. Root system variability in common legumes in Central Europe. Biologia  67: 116–125. [Google Scholar]
  10. Chmielewski JG, Semple JC. 2001a The biology of Canadian weeds. 113. Symphyotrichum lanceolatum (Willd.) Nesom [Aster lanceolatus Willd.] and S. lateriflorum (L.) Löve & Löve [Aster lateriflorus (L.) Britt.]. Canadian Journal of Plant Science  81: 829–849. [Google Scholar]
  11. Chmielewski JG, Semple JC. 2001b The biology of Canadian weeds. 114. Symphyotrichum pilosum (Willd.) Nesom (Aster pilosus Willd.). Canadian Journal of Plant Science  81: 851–865. [Google Scholar]
  12. Chmielewski JG, Semple JC. 2003. The biology of Canadian weeds. 125. Symphyotrichum ericoides (L.) Nesom (Aster ericoides L.) and S. novae-angliae (L.) Nesom (A. novae-angliae L.). Canadian Journal of Plant Science  83: 1017–1037. [Google Scholar]
  13. Chouard P. 1960. Vernalization and its relations to dormancy. Annual Review of Plant Physiology  11: 191–238. [Google Scholar]
  14. Connolly BM, Orrock JL. 2015. Climatic variation and seed persistence: freeze–thaw cycles lower survival via the joint action of abiotic stress and fungal pathogens. Oecologia  179: 609–616. [DOI] [PubMed] [Google Scholar]
  15. Cornelissen JHC, Song YB, Yu FH, Dong M. 2014. Plant traits and ecosystem effects of clonality: a new research agenda. Annals of Botany  114: 369–376. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Davik J, Koehler G, From B, et al.  2013. Dehydrin, alcohol dehydrogenase, and central metabolite levels are associated with cold tolerance in diploid strawberry (Fragaria spp.). Planta  237: 265–277. [DOI] [PubMed] [Google Scholar]
  17. Environment Canada 2019. Canadian climate normals 1981–2010. National climate data and information archive.  http://climate.weather.gc.ca/climate_normals/index_e.html  (21 May 2019).
  18. Fitter AH, Peat HJ. 1994. The ecological flora database. Journal of Ecology  82: 415–425. [Google Scholar]
  19. Grime JP. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. American Naturalist  111: 1169–1194. [Google Scholar]
  20. Hagerty T, Kingston M. 1992. The soils of Middlesex County.  Ottawa: Ministry of Agriculture and Food. [Google Scholar]
  21. Hameed MA, Uzun A, Saeed JF. 2016. Vascular plant species of Hujran Basin-Erbil/Iraq. Kahramanmaraş Sütçü İmam Üniversitesi Doğa Bilimleri Dergisi  19: 279–295. [Google Scholar]
  22. Hartnett DC, Keeler KH. 1995. Population processes. In: Joern A, Keeler KH, eds. The changing prairie: North American grasslands. Oxford: Oxford University Press, 82–99. [Google Scholar]
  23. Henry HAL, Abedi M, Alados CL, et al.  2018. Increased soil frost versus summer drought as drivers of plant biomass responses to reduced precipitation: results from a globally coordinated field experiment. Ecosystems  21: 1432–1444. [Google Scholar]
  24. Kamenetsky R. 2013. Biodiversity of geophytes: phytogeography, morphology, and survival strategies. In: Kamenetsky R, Okubo H, eds. Ornamental geophytes: from basic science to sustainable production. Boca Raton: Taylor & Francis, 77: –121. [Google Scholar]
  25. Klimeš L, Klimešová J, Čížková H. 1999. Carbohydrate storage in rhizomes of Phragmites australis: the effects of altitude and rhizome age. Aquatic Botany  64: 105–110. [Google Scholar]
  26. Klimešová J. 2018. Temperate herbs: an architectural analysis.  Praha: Academia. [Google Scholar]
  27. Klimešová J, Klimeš L. 1996. Effects of rhizome age and nutrient availability on carbohydrate reserves in Rumex alpinus rhizomes. Biologia  51: 457–461. [Google Scholar]
  28. Klimešová J, Tackenberg O, Herben T. 2015. Herbs are different: clonal and bud bank traits can matter more than leaf–height–seed traits. New Phytologist  210: 13–17. [DOI] [PubMed] [Google Scholar]
  29. Klimešová J, Danihelka J, Chrtek J, de Bello F, Herben T. 2017. CLO-PLA: a database of clonal and bud bank traits of Central European flora. Ecology  98: 1179. doi: 10.1002/ecy.1745. [DOI] [PubMed] [Google Scholar]
  30. Komac B, Pladevall C, Peñuelas J, Conesa JV, Domènech M. 2015. Variations in functional diversity in snowbed plant communities determining snowbed continuity. Plant Ecology  216: 1257–74. [Google Scholar]
  31. Kong RS, Henry HAL. 2019a Interactions of plant growth responses to spring freezing and summer drought: a multispecies comparison. American Journal of Botany  106: 531–539. [DOI] [PubMed] [Google Scholar]
  32. Kong RS, Henry HAL. 2019b Cross acclimation effects of spring freezing and summer drought on plant functional groups and ecosystem properties. Environmental and Experimental Botany  164: 52–57. [Google Scholar]
  33. Lemieux C, Cloutier DC, Leroux GD. 1993. Distribution and survival of quackgrass (Elytrigia repens) rhizome buds. Weed Science  41: 600–606. [Google Scholar]
  34. Lemna WK, Messersmith CG. 1990. The biology of Canadian weeds. 94. Sonchus arvensis L. Canadian Journal of Plant Science  70: 509–532. [Google Scholar]
  35. Lubbe FC, Henry HAL. 2019. The cost of depth: frost avoidance trade-offs in herbaceous plants. Plant and Soil  444: 213–224. [Google Scholar]
  36. Malyshev AV, Henry HAL. 2012. N uptake and growth responses to sub-lethal freezing in the grass Poa pratensis L. Plant and Soil  360: 175–85. [Google Scholar]
  37. McKersie BD, Lesham YY. 1994. Stress and stress coping in cultivated plants.  Dordrecht: Kluwer Academic Publishers. [Google Scholar]
  38. Moore RJ. 1975. The biology of Canadian weeds. 13. Cirsium arvense (L.) Scop. Canadian Journal of Plant Science  55: 1033–1048. [Google Scholar]
  39. Mueller‐Dombois D, Ellenberg H, 1974. Aims and methods of vegetation ecology.  New York: John Wiley & Sons. [Google Scholar]
  40. Native Plant Trust 2019. Go botany (3.1.1) https://gobotany.nativeplanttrust.org/  (21 May 2019).
  41. Pan Y, Geng YP, Li B, Chen JK. 2009. Effect of root fragment length and planting depth on clonal establishment of alligatorweed. Journal of Aquatic Plant Management  47: 96–100. [Google Scholar]
  42. Pearce R. 2001. Plant freezing and damage. Annals of Botany  87: 417–424. [Google Scholar]
  43. Perfect E, Miller RD, Burton B. 1987. Root morphology and vigor effects on winter heaving of established alfalfa. Agronomy Journal  79: 1061–1067. [Google Scholar]
  44. Raunkiær C. 1934. The life forms of plants and statistical plant geography.  Oxford: Oxford University Press. [Google Scholar]
  45. Reinmann AB, Susser JR, Demaria EMC, Templer PH. 2019. Declines in northern forest tree growth following snowpack decline and soil freezing. Global Change Biology  25: 420–430. [DOI] [PubMed] [Google Scholar]
  46. SAS Institute, 1989–2019. JMP®, Version 13. Cary, NC: SAS Institute. [Google Scholar]
  47. Schweingruber FH, Poschlod P. 2005. Growth rings in herbs and shrubs: life span, age determination and stem anatomy. Forest Snow and Landscape Research  79: 195–415. [Google Scholar]
  48. Sharratt BS. 2002. Corn stubble height and residue placement in the northern US corn belt: Part I. Soil physical environment during winter. Soil and Tillage Research  64: 243–252. [Google Scholar]
  49. Stewart-Wade SM, Neumann S, Collins LL, Boland GJ. 2002. The biology of Canadian weeds. 117. Taraxacum officinale GH Weber ex Wiggers. Canadian Journal of Plant Science  82: 825–853. [Google Scholar]
  50. Turkington R, Cavers PB. 1979. The biology of Canadian weeds: 33. Medicago lupulina L. Canadian Journal of Plant Science  59: 99–110. [Google Scholar]
  51. Vesk PA, Westoby M. 2004. Sprouting by plants: the effects of modular organization. Functional Ecology  18: 939–945. [Google Scholar]
  52. Walck JL, Hidayati SN, Dixon KW, Thompson K, Poschlod P. 2011. Climate change and plant regeneration from seed: climate change and plant regeneration. Global Change Biology  17: 2145–2161. [Google Scholar]
  53. Werner PA, Gross RS, Bradbury IK. 1980. The biology of Canadian weeds: 45. Solidago canadensis L. Canadian Journal of Plant Science  60: 1393–1409. [Google Scholar]
  54. Wisniewski M, Bassett C, Gusta LV. 2003. An overview of cold hardiness in woody plants: seeing the forest through the trees. Hortscience  38: 952–959. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

mcaa104_suppl_Supplementary_Data
Click here for additional data file. (10.2KB, xlsx)

Articles from Annals of Botany are provided here courtesy of Oxford University Press

RESOURCES