1. Observed climate change in northern high latitudes is strongest in winter, but still relatively little is known about the effects of winter climate change on tundra ecosystems. Ongoing changes in winter climate and snow cover will change the intensity, duration, and frequency of frost events. Bryophytes form a major component of northern ecosystems but their responses to winter climate changes are largely unknown. 2. Here, we studied how changes in overall winter climate and snow regime affect frost damage in three common bryophyte taxa that differ in desiccation tolerance in a subarctic tundra ecosystem. We used a snow manipulation experiment where bryophyte cores were transplanted from just above the treeline to similar elevation (i.e., current cold climate) and lower elevation (i.e., near-future warmer climate scenario) in Abisko, Sweden. Here we measured frost damage in shoots of Ptilidium ciliare, Hylocomium splendens and Sphagnum fuscum with the relative electrolyte leakage (REL) method, during late winter and spring in two consecutive years. We hypothesized that frost damage would be lower in a milder climate (low site), higher under reduced snow cover, and that taxa from moister habitats with assumed low desiccation tolerance would be more sensitive to lower temperature and thinner snow cover than those from drier and more exposed habitats. 3. Contrary to our expectations, frost damage was highest at low elevation, while the effect of snow treatment differed across sites and taxa. At the high site, frost damage was reduced under snow addition in the taxon with the assumed lowest desiccation tolerance, S. fuscum. Surprisingly, frost damage increased with mean temperature in the bryophyte core of the preceding 14 days leading up to REL measurements and decreased with higher frost degree sums, i.e., was highest in the milder climate at the low site. 4. Synthesis. Our results imply that climate warming in late winter and spring increases frost damage in bryophytes. Given the high abundance of bryophytes in tundra ecosystems, higher frost damage could alter the appearance and functioning of the tundra landscape, although the short and long-term effects on bryophyte fitness remain to be studied.

We used a snow manipulation experiment where bryophyte cores were transplanted from just above the treeline to similar elevation (i.e., current cold climate) and lower elevation (i.e., near-future warmer climate scenario) in Abisko, Sweden. For full details on the experimental design, see the corresponding publication in Journal of Ecology (Van Zuijlen et al. 2023).

Sampling and frost damage assessment

To capture periods when bryophytes may be more vulnerable to freezing, we sampled multiple times after snowmelt (i.e., when all plots had melted out at least partially), and across two consecutive years to increase the generality of our findings. In 2015 sampling took place on May 10 and 27 at the low site and on May 11 and 29 at the high site. In 2016, we sampled on May 2, 11 and 16 and June 1 at the low site and on May 3, 10 and 17 and June 2 at the high site. At each sampling occasion, we collected bryophyte shoots from each species in each plot. Five living (i.e., healthy looking) individual shoots of approximately 2.5 cm length were selected randomly from each core, at a minimum of 5 cm distance from the core edge. After sampling, bryophyte shoots were stored at 4 °C overnight until the REL assay was started the following day.

For REL estimation, each sample, consisting of 5 shoots, was soaked in 16 ml of 0.1% (v/v) Triton X-100 (VWR chemicals) for 24 hours at 20 °C to allow electrolyte leakage to equilibrate, after which we measured initial conductivity (SevenCompact S230, Mettler-Toledo AG, Switzerland). We then placed the samples in a water bath at 97 °C for one hour, with the purpose to lyse all cells. Final conductivity was measured after another 24 hours at 20 °C. Along with each batch of 45 samples (one sampling moment of one site), three blank samples, consisting of 16 ml of 0.1% (v/v) Triton X-100, were taken of which the mean initial and mean final conductivity was calculated. REL was then calculated as (initial conductivity – initial blank conductivity) / (final conductivity – final blank conductivity). 

Bryophyte performance

To assess the impact of snow and elevation on growth as a measure of actual fitness of the bryophytes, we tied cotton threads around two H. splendens and three P. ciliare shoots. Shoot growth was not measured on S. fuscum as shoots were not strong enough to withstand this handling. Distance from thread to tip was measured at the beginning and end of the growing season of 2014 and the difference was calculated to represent shoot length growth. For each plot, the shoot with the highest growth was used to assess growth response. This was done to avoid counting growth of shoots that were potentially affected by the measurement, i.e., tying a thread around the moss shoot may disturb or even kill the shoot. In this way, we were more likely to be assessing actual treatment responses. 

Microclimate in the bryophyte cores

In spring 2015 temperature (°C) and moisture (volumetric water content (VWC), in m³m^-3) in the bryophyte cores were measured hourly (5TM sensors and ECH2O Em50 loggers, Decagon Devices Inc., WA, USA). Sensors were placed vertically in bryophyte cores so that measurements were taken at 0–5 cm depth (integrated) from the bryophyte surface and down. Sensors were placed in four out of five blocks in all bryophyte species and snow treatments at both sites. In spring 2016, bryophyte core temperature was measured as in 2015 (but using Tinytags, Plus2 TGP‐4017 external sensors, Gemini Co., UK; and ECH2O Em50 loggers, Decagon Devices Inc., WA, USA) and in H. splendens cores only, and only in three out of five blocks, from mid-April until the beginning of June. All data were checked visually for discrepancies; untenable moisture values (i.e. VWC<0 and VWC>1) and bryophyte core moisture values measured when the soil was frozen (i.e. concurrent temperature ≤0°C) were removed. We further removed unreliable data from one control plot (showing exceptionally high temperatures right at the start of the spring snow treatments).

After data cleaning, we calculated the following microclimate variables: average daily mean temperature; absolute minimum temperature; no. of diurnal freeze-thaw cycles (if temperature reaches below and above 0°C in one calendar day); number of frost days (daily average temperature < 0°C); frost degree sums (the absolute value of the sum of daily minimum temperatures below 0°C; i.e. higher values correspond to more frost); and finally mean moisture content. We calculated these variables over the entire experimental period in 2015, that is April 16 to and including May 26, i.e., right after the start of the spring snow treatments and just before the final REL sampling in 2015. In addition, for each REL sampling occasion in 2015 and 2016, we calculated all the same microclimate variables over the 14 days leading up to sampling, as well as an additional variable for each measurement date that is the number of days since the last diurnal freeze-thaw cycle.

Frost damage in unmanipulated bryophyte shoots

To assess the REL of healthy, non-transplanted bryophytes in spring, shoots of H. splendens and P. ciliare were collected on the 28th of April 2016 just before snowmelt, from a half-open birch forest close to the low site. For each moss species, shoots were sampled from one site with thin snow cover (7-10 cm) and one site with thick snow cover (53-40 cm). For each snow regime and species, 6 samples containing 5 shoots each of about 5 cm length were collected. The samples were consistently subjected to +5 °C for 144 h with alternating tubular lighting every 12 h in environmental test chambers (MLR-350, Sanyo Electric Biomedical Co., Ltd., Japan) before REL measurements were taken following the same protocol as described above for bryophytes exposed to experimental treatments in the field common gardens.