Aim: Mountain ecosystems are disproportionately exposed to global warming, which may put alpine plant species in particular at risk. Plant species’ migration and persistence depend on successful seed regeneration, with demographic traits such as germination being strongly sensitive to both present and ancestral climate. Germination may, thus, present a major bottleneck to alpine plant species under climate change, or they may cope through phenotypic plasticity. However, little is known about alpine plant germination under warming, its plasticity, and whether any of these are co-determined by the climatic origin of the seeds.

Location: Eastern Central Alps in Tyrol, Austria

Time Period: 2022-2023

Methods: We investigated intraspecific variation in germination proportion and timing of six co-occurring alpine plant species from different micro- and mesoclimatic origins as response to three temperature treatments using growth chambers. Elevation as well as topographic strata were used as proxies for mesoclimate and mean soil temperature, growing- as well as freezing-degree-days for microclimate.

Major taxa studied: Alpine plant species

Results: Our results show that species were plastic in germination proportion and timing in response to temperature, with all species germinating in the warmest treatment. Micro- and mesoclimatic origin affected germination traits but did not co-determine species responses to the temperature treatments. We also demonstrated maternal climatic effects, with seeds from warmer microclimates displaying higher phenotypic plasticity. Furthermore, seeds originating from colder mesoclimates, i.e., the upper elevation range, generally germinated less.

Main conclusions: Our results suggest that global warming will impact alpine plant regeneration from seeds through concurrent effects during germination, climatic origin as well as shifts in germination timing. Importantly, our results indicate that regeneration from seed may contribute to defining the upper elevational limit of alpine plant species through climatic origin and maternal microclimatic effects.

Study area and seed collection

Seeds originated from Schrankogel, located in the eastern Central Alps in Tyrol, Austria. The bedrock is composed of siliceous material, mostly amphibolite and gneiss. The regional climate is temperate sub-continental with an average annual temperature of 2.5°C and an average annual rainfall of 889 mm within the measurement period 1980–2021 (meteorological station Obergurgl, 1950 m a.s.l., 20 km south of the study area).

In July and August 2021 and 2022, 900 plots of 1 × 1 m² were established across the non-forest subalpine, alpine, and nival zones of Schrankogel as part of a broader research infrastructure (Chytrý et al. 2024; Helm et al. 2024). Each plot was equipped with a soil temperature logger (Geoprecision M-Log5W-SIMPLE) recording temperature at hourly intervals at approximately 10 cm depth. Plot positions were selected following a stratified random sampling design that considered 20 different topographic strata. These topographic strata were computed using K-means clustering of fine-scale maps of elevation, micro- and macro-slope, a fine-scale and a coarse-scale topographic position index (TPI), potential incoming solar radiation (May to September sum), and the maximum Normalized Difference Vegetation Index (NDVI; computed from Sentinel-2 satellite images of July and August 2020 at 10 m resolution). Micro- and macro-slopes were computed with a 3 m x 3 m and 99 m x 99 m neighborhood size, respectively. Fine-scale and coarse-scale TPI were computed as multi-scale versions (SAGA GIS (Conrad et al. 2015) implementation following Guisan et al. (1999)) with 1 m to 10 m neighborhood radius with three steps and with 1 m to 100 m neighborhood radius with four steps, respectively. In this study, we focused on a subset of four of these 20 topographic strata situated along the same macro-topographic hillslope in the alpine zone (Supporting information, Figure S1).

Seeds of six species (Rhododendron ferrugineum, Helictochloa versicolor, Silene acaulis subsp. exscapa, Phyteuma hemisphaericum, Primula minima, and Bartsia alpina) were collected between 10^th and 13^th August 2022. Species were chosen based on their occurrence and their seed phenology within the four focal strata during the collection. We aimed to sample seeds from nine mother plants (biological replicates) per species and stratum, ideally distributed across three plots with three mother plants per plot. The number of sampled plots and mother plants per plot varied depending on the availability of fruiting mothers and because not all species occurred in all strata (Table S1). Plots spanned a distance of approximately 895 m along the selected macro-topographic hillslope, with a minimum distance between plots of 16 m, covering an elevation range from 2205 m to 2775 m. Mother plants were selected in a regular-random manner within a 5 m radius of plots, but excluding areas belonging to a different stratum. To increase the number of mother plants, 16 mother plants across species were sampled at greater distance (max. 50 m) from the nearest plot. In total, we included seeds of 192 mother plants in this study (Table S1).

Germination experiment

In 2023, seeds were germinated under controlled conditions in petri dishes in plant growth chambers (Percival Scientific Model SE-59L1 SED: photoperiod 15 hours light/ 9 hours darkness; 60% air humidity) under three different temperature treatments: 17/7°C, 21/11°C, 25/15°C. The higher temperature was set during the light hours and the lower temperature during the darkness.

For each mother plant, well-developed seeds were equally divided among the three temperature treatments, enabling half- or full-sibling comparisons across treatments. Ten seeds per mother plant and treatment were placed together in a single petri dish (one dish per mother plant × treatment) on filter paper, without any growing medium other than water. For 11 mothers, less than 30 well-developed seeds were available, which were all used by equally distributing them among the three treatments. Helictochloa versicolor consistently produced less than 30 seeds, so that it was not included in the 21/11°C treatment. Instead, its seeds per mother plant were equally divided between the 17/7°C and the 25/15°C treatment.

Temperatures of our warmest treatment reflect daily field conditions in summer in the Alps of Lombardy (Northern Italy), used in a previous study on alpine plant germination (Tudela-Isanta et al. 2018a). We chose the other treatments at intervals of 4°C, which is close to the maximum difference in soil temperature among the strata (4.2°C) when averaging over the summer months (June-August) in 2022. The temperature differences mother plants may have experienced based on differences in elevation was maximum 3.4°C, assuming a lapse rate of ca 0.6°C per 100 m.

Phyteuma hemisphaericum and B. alpina received a cold stratification at 0°C in complete darkness for 10 and 11 weeks, respectively, following Tudela-Isanta et al. (2018a) who documented physiological dormancy in P. hemisphaericum and morphophysiological dormancy in B. alpina. The difference in stratification duration between these two species resulted from an unplanned delay in the start of the stratification treatment. For the remaining species, no (cold) stratification was applied because available information on dormancy type and stratification requirements is scarce or inconclusive. All seeds were stored at room temperature until the start of cold stratification or the start of the germination experiment, which was simultaneous for all species.

During the germination experiment, petri dishes were checked three times per week. Each seed was monitored individually, with the date of germination scored upon first emergence of the radicile. Petri dishes were watered as needed to remain constantly moist. The experiment lasted 42 days (six weeks), by which time germination had largely levelled off across species and treatments.

Seed quality was assessed at three stages: during field collection, during the setup of the germination experiment, and after the experiment using cut-tests. In the field and again when seeds were placed in petri dishes under an optical microscope, only well-developed, visually intact seeds without signs of deformation or infestation were selected, which may have resulted in a bias towards heavier seeds. After the germination experiment, all non-germinated seeds were cut and visually assessed under an optical microscope for the presence of a well-developed embryo and signs of infestation or tissue degradation. The non-germinated seeds were neither empty nor infested with insects or fungi, but all contained well-developed embryos, supporting the reliability of the first two quality assessments.