The characteristics of the frost regime (intensity, frequency, and timing) contribute to shaping tree species adaptations and distribution as well as ecosystem productivity and functions. However, climate change increases the variability in extreme events; therefore, the different characteristics of the frost regime may diverge under climate change. Using the BioSim 11 software, we simulated daily air temperature at 512 locations over northeastern North America between 1901–2021 to determine how the spatio-temporal trends in the frost regime varied over this complex landscape and if spatio-temporal trends in extreme climatic events such as frosts are stronger compared to changes in aggregated climate variables such as mean annual air temperature and growing degree-days. We also used an eco-physiological model to conduct a case study focussing on white spruce to determine if trees are currently more exposed to growing season frosts than they were in the past by modelling the timings of budburst using the thermal time model. Our results showed that, at 67% of locations (343 locations), the day of the year of the last frost in spring (minimum daily air temperature<0 °C) occurred, on average, earlier by seven days during 2001–2021 compared to 1901–1920, whereas it occurred, on average, later by four days at 33% of locations (169 locations). The average temporal trends in frost occurrence were similar in magnitude to the average trends in aggregated climate variables; however, their variances were larger compared to the aggregated climate metrics, showing that the frost regime does not change uniformly throughout our study area. Our case study also revealed that white spruce remains exposed to late frosts of low and intermediate intensities (minimum daily air temperature <0; <−2 °C) compared to the past but exposure to high-intensity frosts (minimum daily air temperature <−4 °C) is rare. Since extreme events such as late frosts diverge in their response to climate change compared to aggregated climate variables, the mean annual temperature is not sufficient to predict how climate change will impact ecosystems through frost regimes.

Summary

Intraspecific trait variability (ITV) provides the material for species' adaptation to environmental changes. To advance our understanding of how ITV can contribute to species' adaptation to a wide range of environmental conditions, we studied five widespread understorey forest species exposed to both continental-scale climate gradients, and local soil and disturbance gradients. We investigated the environmental drivers of between-site leaf and root trait variation, and tested whether higher between-site ITV was associated with increased trait sensitivity to environmental variation (i.e. environmental fit). We measured morphological (specific leaf area: SLA, specific root length: SRL) and chemical traits (Leaf and Root N, P, K, Mg, Ca) of five forest understorey vascular plant species at 78 sites across Canada. A total of 261 species-by-site combinations spanning ~4300 km were sampled, capturing important abiotic and biotic environmental gradients (neighbourhood composition, canopy structure, soil conditions, climate). We used multivariate and univariate linear mixed models to identify drivers of ITV and test the association of between-site ITV with environmental fit. Between-site ITV of leaf traits was primarily driven by canopy structure and climate. Comparatively, environmental drivers explained only a small proportion of variability in root traits: these relationships were trait specific and included soil conditions (Root P), canopy structure (Root N) and neighbourhood composition (SRL, Root K). Between-site ITV was associated with increased environmental fit only for a minority of traits, primarily in response to climate (SLA, Leaf N, SRL). Synthesis. By studying how ITV is structured along environmental gradients among species adapted to a wide range of conditions, we can begin to understand how individual species might respond to environmental change. Our results show that generalisable trait–environment relationships occur primarily aboveground, and only accounted for a small proportion of variability. For our group of species with broad ecological niches, variability in traits was only rarely associated with higher environmental fit, and primarily along climatic gradients. These results point to promising research avenues on the various ways in which trait variation can affect species' performance along different environmental gradients.

1. Considering intraspecific trait variability (ITV) in ecological studies has improved our understanding of species persistence and coexistence. These advances are based on the growing number of leaf ITV studies over local gradients, but logistical constraints have prevented a solid examination of ITV in root traits or at scales reflecting species’ geographic ranges.
2. We compared the magnitude of ITV in above? and below?ground plant organs across three spatial scales (biophysical region, locality and plot). We focused on six understorey species (four herbs and two shrubs) that occur both in disturbed and undisturbed habitats across boreal and temperate Canadian forests. We aimed to document ITV structure over broad ecological and geographical scales by asking: (a) What is the breadth of ITV across species range?scale? (b) What proportion of ITV is captured at different spatial scales, particularly when local scale disturbances are considered? and (c) Is the variance structure consistent between analogous leaf and root traits, and between morphological and chemical traits?
3. Following standardized methods, we sampled 818 populations across 79 forest plots simultaneously, including disturbed and undisturbed stands, spanning four biophysical regions (~5,200 km). Traits measured included specific leaf area (SLA), specific root length (SRL) and leaf and root nutrient concentrations (N, P, K, Mg, Ca). We used variance decomposition techniques to characterize ITV structure across scales.
4. Our results show that an important proportion of ITV occurred at the local scale when sampling included contrasting environmental conditions resulting from local disturbance. A certain proportion of the variability in both leaf and root traits remained unaccounted for by the three sampling scales included in the design (36% on average), with the largest amount for SRL (54%). Substantial differences in magnitude of ITV were found among the six species, and between analogous traits, suggesting that trait distribution was influenced by species strategy and reflects the extent of understorey environment heterogeneity.
5. Even for species with broad geographical distributions, a large proportion of within?species trait variability can be captured by sampling locally across ecological gradients. This has practical implications for sampling design and trait selection for both local studies and continental?scale modelling.

Aims

Vegetation?plot records provide information on the presence and cover or abundance of plants co?occurring in the same community. Vegetation?plot data are spread across research groups, environmental agencies and biodiversity research centers and, thus, are rarely accessible at continental or global scales. Here we present the sPlot database, which collates vegetation plots worldwide to allow for the exploration of global patterns in taxonomic, functional and phylogenetic diversity at the plant community level.

Results

sPlot version 2.1 contains records from 1,121,244 vegetation plots, which comprise 23,586,216 records of plant species and their relative cover or abundance in plots collected worldwide between 1885 and 2015. We complemented the information for each plot by retrieving climate and soil conditions and the biogeographic context (e.g., biomes) from external sources, and by calculating community?weighted means and variances of traits using gap?filled data from the global plant trait database TRY. Moreover, we created a phylogenetic tree for 50,167 out of the 54,519 species identified in the plots. We present the first maps of global patterns of community richness and community?weighted means of key traits.

Conclusions

The availability of vegetation plot data in sPlot offers new avenues for vegetation analysis at the global scale.

Because of their scarcity, protected areas alone cannot maintain biodiversity. Therefore, it is necessary to create conditions appropriate for plants and wildlife in managed landscapes. We compared the effects of different intensities of forest management on functional responses of vascular understory plants using the fourth-corner method. We analysed functional community composition along a management gradient that spanned semi-natural forests to extensively managed forests (naturally regenerated cuts) to intensively managed forests (planted forests) in Canada. Results showed trait filtering along the gradient of forest management intensity. In natural and extensively managed forests, where forest retention was high in time and space, persistence traits (e.g. perennial geophytes or chamaephytes, non-leafy stem foliage structure) were maintained. At the opposite end of the gradient, in intensively managed plantations where forest retention elements (e.g. amount of dead wood) were reduced, trait filtering led to species associated with colonization, such as tall species with limited lateral extension. These results suggest that intensive forestry conducted over a large extent may change the functional composition of understory plants.

This study was conducted in six different forest types in Abitibi, Que., (i) to identify the factors that most influence understory light transmission in the southern boreal forest and (ii) to develop light extinction coefficients (k), which could be used to simulate light transmission in the understory. Light availability and understory vegetation (cover, composition, vertical distribution, and leaf area index) were characterized within three strata (0.05-5 m) in a total of 180 quadrats. Calculated k values were based on measured light availability and leaf area index. These values varied among forest types, strata, understory vegetation types, and cover in the upper stratum. The highest k values were generally associated with a dense stratum of Acer spicatum Lam. We developed five sets of k values based on the factors that most affected light transmission. Measured transmission (T_m) was compared with transmission predicted (Tp) from each set of k values. Light transmission predicted using a single k value (mean k = 0.54) underestimated T_m. More accurate predictions were obtained when we used the other four sets of k values. Our results indicate that, in the southern boreal forest, the understory vegetation can be quite heterogeneous and patterns of light transmission cannot be accurately simulated using a unique k value. However, the various sets of k values developed in this study could be used in prediction models of forest dynamics to obtain relatively good predictions of understory light extinction in forest types similar to the ones studied here. ©2000 NRC Canada