1. Ecology

Lowland plant arrival in alpine ecosystems facilitates a decrease in soil carbon content under experimental climate warming

  1. Tom WN Walker  Is a corresponding author
  2. Konstantin Gavazov
  3. Thomas Guillaume
  4. Thibault Lambert
  5. Pierre Mariotte
  6. Devin Routh
  7. Constant Signarbieux
  8. Sebastián Block
  9. Tamara Münkemüller
  10. Hanna Nomoto
  11. Thomas W Crowther
  12. Andreas Richter
  13. Alexandre Buttler
  14. Jake M Alexander
  1. Institute of Integrative Biology, ETH Zürich, Switzerland
  2. Climate Impacts Research Centre, Department of Ecology and Environmental Sciences, Umeå Universitet, Sweden
  3. Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Switzerland
  4. École Polytechnique Fédérale de Lausanne EPFL, School of Architecture, Civil and Environmental Engineering ENAC, Laboratory of Ecological Systems ECOS and Plant Ecology Research Laboratory PERL, Switzerland
  5. Field-Crop Systems & Plant Nutrition, Switzerland
  6. Faculty of Geosciences & the Environment, Université de Lausanne, Switzerland
  7. Grazing Systems, Agroscrope, Switzerland
  8. Department of Ecology & Evolutionary Biology, Princeton University, United States
  9. Uni. Grenoble Alpes, CNRS, Uni. Savoie Mont Blanc, LECA, Laboratoire d’Ecologie Alpine, France
  10. Centre of Microbiology & Environmental Systems, Division of Terrestrial Ecosystem Research, University of Vienna, Austria
  11. International Institute for Applied Systems Analysis, Ecosystem Services and Management Program, Austria
https://doi.org/10.7554/eLife.78555
Share this article
Cite this article
  1. Tom WN Walker
  2. Konstantin Gavazov
  3. Thomas Guillaume
  4. Thibault Lambert
  5. Pierre Mariotte
  6. Devin Routh
  7. Constant Signarbieux
  8. Sebastián Block
  9. Tamara Münkemüller
  10. Hanna Nomoto
  11. Thomas W Crowther
  12. Andreas Richter
  13. Alexandre Buttler
  14. Jake M Alexander
(2022)
Lowland plant arrival in alpine ecosystems facilitates a decrease in soil carbon content under experimental climate warming
eLife 11:e78555.
https://doi.org/10.7554/eLife.78555
  2. Download BibTeX
  3. Download .RIS

Abstract

Climate warming is releasing carbon from soils around the world, constituting a positive climate feedback. Warming is also causing species to expand their ranges into new ecosystems. Yet, in most ecosystems, whether range expanding species will amplify or buffer expected soil carbon loss is unknown. Here, we used two whole-community transplant experiments and a follow-up glasshouse experiment to determine whether the establishment of herbaceous lowland plants in alpine ecosystems influences soil carbon content under warming. We found that warming (transplantation to low elevation) led to a negligible decrease in alpine soil carbon content, but its effects became significant and 52% ± 31% (mean ± 95% confidence intervals) larger after lowland plants were introduced at low density into the ecosystem. We present evidence that decreases in soil carbon content likely occurred via lowland plants increasing rates of root exudation, soil microbial respiration, and CO2 release under warming. Our findings suggest that warming-induced range expansions of herbaceous plants have the potential to alter climate feedbacks from this system, and that plant range expansions among herbaceous communities may be an overlooked mediator of warming effects on carbon dynamics.

Editor's evaluation

The authors transplanted alpine turfs from their cold environment to a lowland warm environment. They found that when lowland plants were inserted into these turfs under the thus simulated warming treatment they rapidly increased soil microbial decomposition of carbon stocks due to root exudates feeding the soil microbes. This finding is relevant because it suggests that global warming and shifts in plant species distributions may cause the release of soil-stored carbon into the atmosphere, thus further increasing warming.

https://doi.org/10.7554/eLife.78555.sa0

eLife digest

In a terrestrial ecosystem, the carbon cycle primarily represents the balance between plants consuming carbon dioxide from the atmosphere and soil microbes releasing carbon stored in the soil into the atmosphere (mostly as carbon dioxide). Given that carbon dioxide traps heat in the atmosphere, the balance of carbon inputs and outputs from an ecosystem can have important consequences for climate change.

Rising temperatures caused by climate warming have led plants from lowland ecosystems to migrate uphill and start growing in alpine ecosystems, where temperatures are lower and most carbon is stored in the soil. Soil microbes use carbon stored in the soil and exuded from plants to grow, and they release this carbon – in the form of carbon dioxide – into the atmosphere through respiration. Walker et al. wanted to know how the arrival of lowland plants in alpine ecosystems under climate warming would affect carbon stores in the soil.

To answer this question, Walker et al. simulated warmer temperatures by moving turfs (plants and soil) from alpine ecosystems to a warmer downhill site and planting lowland plants into the turfs. They compared the concentration of soil carbon in these turfs to that of soil in alpine turfs that had not been moved downhill and had no lowland plants. Their results showed that the warmed turfs containing lowland plants had a lower concentration of soil carbon. This suggests that climate warming will lead to more soil carbon being released into the atmosphere if lowland plants also migrate into alpine ecosystems.

Walker et al. also wanted to know the mechanism through which lowland plants were decreasing soil carbon concentration under warming. They find that lowland plants probably release more small molecules into the soil than alpine plants. Soil microbes use the carbon and nutrients in these molecules to break down more complex molecules in the soil, thereby releasing nutrients and carbon that can then be used in respiration. This finding suggests that soil microbes breakdown and respire native soil carbon faster in the presence of lowland plants, releasing more carbon dioxide into the atmosphere and reducing carbon stores in the soil.

Walker et al.’s results reveal a new mechanism through which uphill migration of lowland plants could increase the effects of climate change, in a feedback loop. Further research as to whether this mechanism occurs in different regions and ecosystems could help to quantify the magnitude of this feedback and allow scientists to make more accurate predictions about climate change.

Introduction

Climate warming is expected to accelerate the release of soil carbon to the atmosphere (Canadell et al., 2021). At the same time, warming is driving uphill and poleward range expansions of native plant species around the globe (Parmesan, 2006; Lenoir and Svenning, 2015; Pecl et al., 2017; Dolezal et al., 2016; Rumpf et al., 2018; Steinbauer et al., 2018). Yet while warming and plant range expansions are now co-occurring in many terrestrial ecosystems, the impact of range expanding plants on soil carbon storage under warming remains poorly understood (Wallingford et al., 2020). This is important because plant species adapted to different temperatures can have contrasting effects on carbon cycle processes (Walker et al., 2019), raising the potential for migrating species to have qualitatively different impacts on carbon dynamics from resident plants. For example, studies of non-native species invasions have revealed that the arrival of novel species in a community can alter carbon cycling in recipient ecosystems (Liao et al., 2008; Waller et al., 2020), while woody plant expansions into treeless Arctic ecosystems can reverse the direction of carbon cycle feedbacks to climate warming (Pearson et al., 2013; Cornelissen et al., 2007). However, the majority of warming-induced range expansions do not involve non-native species or novel growth forms, but occur among native communities of similar species (e.g. Steinbauer et al., 2018). In particular, we know remarkably little about how migrations of herbaceous plants into predominantly herbaceous ecosystems will affect carbon cycling, and so the magnitude and direction of resulting climate feedbacks are unknown.

Alpine ecosystems occur in the uppermost elevations of the planet as cool islands within warmer landscapes that are bounded by steep gradients in temperature (Körner, 2003). They are experiencing both rapid temperature change (0.4–0.6°C per decade) (Pepin, 2015; Vitasse et al., 2021) and rapid species immigration (Lenoir et al., 2008), the latter of which involves an influx of herbaceous warm-adapted species from lowlands into herbaceous cold-adapted communities at higher elevations (Dolezal et al., 2016; Rumpf et al., 2018; Steinbauer et al., 2018). Coupled to this, alpine ecosystems are globally relevant reservoirs of soil carbon and hold more than 90% of their total ecosystem carbon in the soil (Streit et al., 2014; Parker et al., 2015; Knowles et al., 2019; Hagedorn et al., 2019; Budge et al., 2011). As such, alpine ecosystems are not only themselves important for climate feedbacks, but also sentinels for climate warming impacts on other terrestrial ecosystems. Here, we used alpine ecosystems as a model system to determine how the establishment of range expanding native plant species could affect soil carbon storage in warming ecosystems. We first used two whole-community transplant experiments in different alpine regions (western/central Swiss Alps) to establish the effects of warming plus lowland plant arrival on alpine soil carbon content. Second, we used these experiments plus glasshouse and laboratory studies to explore potential mechanisms through which lowland plants influence the alpine soil carbon cycle, testing the hypothesis that they act via changes to root exudation (Hafner et al., 2012; Studer et al., 2014).

Results and discussion

Lowland plants facilitate warming-induced soil carbon loss

We exploited two independent field-based transplant experiments in which alpine plant communities plus rooting zone soil were transplanted from high to low elevation to simulate a climate warming scenario, or back into a high-elevation site as a negative control (Appendix 1—figure 1; Materials and methods). A portion of the alpine turfs at low elevation were planted with local lowland plant species to simulate the arrival of warm-adapted lowland plants in the warmed ecosystem, with the remaining portion being subjected to a planting disturbance control. Elevation-based transplant experiments are powerful tools for assessing climate warming effects on ecosystems because they expose plots to a real-world future temperature regime with natural diurnal and seasonal cycles while also warming both above- and belowground subsystems (Hannah, 2022; Shaver et al., 2000; Yang et al., 2018). This is especially true if they include rigorous disturbance controls (here, see Materials and methods) and are performed in multiple locations where the common change from high to low elevation is temperature (here, warming of 2.8°C in the central Alps and 5.3°C in the western Alps). While factors other than temperature can co-vary with elevation (Körner, 2003), such factors either do not vary consistently with elevation among experiments (e.g. precipitation, wind), are not expected to strongly influence plant performance (e.g. UV radiation) or in any case form part of a realistic climate warming scenario (e.g. growing-season length, snow cover) (Körner, 2003; Yang et al., 2018; Nomoto and Alexander, 2021). In short, the experiments used here examined how the arrival of warm-adapted lowland plants influences alpine ecosystems in a warmed climate matching lowland site conditions (i.e. turf transplantation to low elevation plus lowland plant addition) relative to warming-only (i.e. turf transplantation to low elevation) or control (i.e. turf transplantation within high elevation) scenarios.

We found that transplantation from high to low elevation (i.e. warming-only) caused a small decline in alpine soil carbon content, but in both regions this effect was stronger and only became statistically significant when lowland plants were also planted-in (Figure 1; Supplementary file 1; p < 0.001). Specifically, considering field data from both experiments together, the addition of lowland plants to transplanted alpine turfs increased warming-induced soil carbon loss from 4.1 ± 0.23 mg g−1 per °C to 6.3 mg g−1 per °C, which is a 52% ± 31% (mean ± 95% confidence intervals [CIs]) increase relative to warming alone (Figure 1—figure supplement 1; p = 0.001). We caution against equating changes to soil carbon content with changes to soil carbon stock in the absence of coupled measurements of soil bulk density (Materials and methods). Nevertheless, these findings show that once warm-adapted lowland plants establish in warming alpine communities, they facilitate warming effects on soil carbon loss on a per gram basis. We thus suggest that uphill migrations of lowland plants, which are already globally common (Rumpf et al., 2018; Lenoir et al., 2008), may have a hitherto unrecognized bearing over alpine soil carbon storage.

Figure 1 with 1 supplement see all
Download asset Open asset
Warming and lowland plant effects on alpine soil carbon content in the field experiment.

Mean ± standard error (SE) soil carbon content (mg C g−1 dry mass; i.e. mass-based per-mil) in alpine turfs transplanted to low elevation (warming, W; light grey), transplanted plus planted with lowland plants (warming plus lowland plant arrival, WL; dark grey), or replanted at high elevation (control, C; white). Data are displayed for two experiments in the western (left) and central (right) Alps, with letters indicating treatment differences (linear mixed effects models; N = 58).

Lowland and alpine plants have distinct effects on soil microbial respiration

Soil carbon loss occurs due to a shifting balance between soil carbon inputs via net primary production and soil carbon outputs via soil microbial respiration (Allison et al., 2010; Liang et al., 2017). We therefore hypothesized that lowland plants impacted alpine soil carbon loss by either decreasing net primary production or increasing soil microbial respiration relative to alpine plants. To test this, we performed a glasshouse experiment in which we grew lowland plants (Bromus erectus, Salvia pratensis, Medicago lupulina) or alpine plants (Phleum alpinum, Anthyllis vulneraria subsp. alpestris, Plantago alpina) in alpine soil collected from the western Alps region. The glasshouse experiment was not a direct simulation of field conditions, in that plants from lowland and alpine communities were grown separately at a constant temperature and humidity (Materials and methods). Instead, the glasshouse experiment allowed us to isolate how plants adapted to lowland versus alpine climates differentially affect the alpine soil system.

We first compared lowland and alpine plant traits to test whether lowland plants decreased carbon inputs into alpine soil. We found that lowland plant treatments produced more aboveground (Figure 2a; Supplementary file 2; p = 0.028) and belowground (Figure 2b; Supplementary file 2; p < 0.001) biomass than alpine plants, while also containing a higher proportion of root tissue (root-to-shoot ratio: Figure 2c; Supplementary file 2; p = 0.005). While biomass is an imperfect proxy for net primary production, these findings point away from the hypothesis that lowland plants cause alpine soil carbon loss by reducing carbon inputs into soil. On the contrary, the higher biomass and proportionately more root tissue present in lowland plant treatments suggest that lowland plants may in fact increase soil carbon inputs. This occurred despite lowland plant treatments possessing similar specific leaf area (SLA: Figure 2d; Supplementary file 2; p = 0.861) to alpine plant treatments, as well as lower maximal photosynthetic capacity (Amax: Figure 2e; Supplementary file 2; p = 0.026) and stomatal conductance (gs: Figure 2f; Supplementary file 2; p < 0.001). While the reason for decreased photosynthetic capacity and stomatal conductance in the lowland plant treatment is unclear, the fact that both variables were lower suggests that photosynthetic capacity was limited by stomata pore size rather than by photosynthetic machinery (Wong et al., 1979). Indeed, lowland plants accumulated more biomass despite having a lower maximal photosynthetic capacity, ruling it out as a limiting factor of lowland plant net primary production.

Figure 2
Download asset Open asset
Differences between lowland and alpine plant traits in the glasshouse experiment.

Mean ± standard error (SE) pot-level traits for alpine (A) and lowland (L) plant treatments in a glasshouse experiment. (a) Aboveground biomass (AGB; g; N = 19). (b) Belowground biomass (BGB; g; N = 10). (c) Root-to-shoot ratio (N = 19). Biomass-weighted (d) specific leaf area (SLA; cm2 g−1; N = 18), (e) maximum photosynthetic capacity (Amax; µmol m2 s−1; N = 19), and (f) maximum stomatal conductance (gs; µmol m2 s−1; N = 20). Treatment effects (p < 0.05) are indicated by asterisks (LMEs; see Supplementary file 2).

We next examined the soil system under lowland versus alpine plants to determine whether lowland plants alternatively increased carbon outputs from alpine soil. We found that lowland plants accelerated soil microbial respiration relative to alpine plants (Figure 3a; treatment × time: likelihood ratio [LR] = 147.99, p < 0.001). Accelerated microbial activity was also reflected in spectral analyses of soil pore water, which showed that soil under lowland plants contained more dissolved organic matter (DOM) (Figure 4a, b; Supplementary file 2; absorbance [a350]: p < 0.001, total fluorescence [Ftot]: p < 0.001) with a higher proportion of microbially derived products (Figure 4c; Supplementary file 2; fluorescence index [FI]; p = 0.002) than soil under alpine plants. Importantly, microbial respiration was enhanced despite no changes to microbial biomass between treatments (Figure 3d; Supplementary file 2; p = 0.693), showing that microbes respired intrinsically faster (i.e. per unit of biomass) in the presence of lowland plants. These findings support the hypothesis that lowland plants have the capacity to increase soil carbon outputs relative to alpine plants by stimulating soil microbial respiration and associated carbon dioxide release. While accelerated microbial respiration can alternatively be a signal of soil carbon accumulation via greater microbial growth (Liang et al., 2017), such a mechanism is unlikely to have been responsible here because it would have led to an increase in microbial biomass carbon under lowland plants, which we did not observe.

Figure 3
Download asset Open asset
Lowland and alpine plant effects on alpine soil microbes in the glasshouse experiment.

Effects of alpine (A) versus lowland (L) plant treatments on alpine soil microbes in a glasshouse experiment (mean ± standard error [SE]; grey points/ribbons: SEs for bare soil). (a) Total microbial respiration (R; µg C g−1 dry mass h−1) during follow-up incubations without plants (N = 260). (b) Size (µg C g−1 dry mass) and (c) decay rate (µg C g−1 dry mass d−1) of a carbon pool identified using two-pool models (N = 16). (d) Microbial biomass carbon (Cmic; mg C g−1 dry mass; N = 18). For (a), vertical line indicates time before which A–L comparison differed (LME, Tukey; see Main text). For (b–d), treatment effects (p < 0.05) are indicated by asterisks (LMEs; see Supplementary file 2).

Figure 4
Download asset Open asset
Lowland and alpine plants effects on alpine soil dissolved organic matter in the glasshouse experiment.

Effects of alpine (A) versus lowland (L) plant treatments on alpine soil dissolved organic matter extracted from soil pore water in a glasshouse experiment (mean ± standard error [SE]; grey ribbons: SEs for bare soil). (a) Absorbance (a350; m−1), (b) total fluorescence (Ftot; Raman Units) and (c) fluorescence index (FI) (N = 20). (d–i) Relative abundance (%) of six dissolved organic matter components (C1–C6), where components and associated compound families were derived from PARAFAC modelling (Materials and methods; Supplementary file 4). Treatment effects (p < 0.05) are indicated by asterisks (LMEs; see Supplementary file 2).

We speculate that lowland plants accelerated microbial activity in the glasshouse experiment by increasing the quantity of root exudates released into the soil, for four reasons. First, two-pool models of microbial respiration revealed that soils supporting any plant contained a soil carbon pool that was effectively absent from a bare soil control, and which was a factor of 1.5 ± 0.7 (mean ± standard error [SE]) times larger under lowland plants than alpine plants (Figure 3b; Supplementary file 2; p = 0.046). This carbon pool decayed at a similar rate between treatments (Figure 3c; Supplementary file 2; p = 0.631), indicating that it differed in size, but likely not composition, between lowland and alpine treatments. Second, the same carbon pool was consumed rapidly by soil microbes (half-life = 5.1 ± 0.4 days), and we also found that lowland plant effects on microbial respiration were lost after 9 days of incubation without plants (Figure 3a). Plants thus supplied compounds to microbes that were depleted within this timeframe. While not directly measured here, a 9-day decay period corresponds to the time expected for newly photosynthesized carbon dioxide to be released through root exudation and respired by soil microbes (Hafner et al., 2012; Studer et al., 2014), suggesting that this carbon pool was mostly root exudates. Third, lowland plant effects on the alpine soil system were rapid, occurring within a period of weeks in the glasshouse experiment and months in the field experiments. It is unlikely that lowland plants influenced microbial activity via litter inputs within this timeframe (De Deyn et al., 2008; Cornwell et al., 2008). Finally, DOM contained a higher relative abundance of protein-like substances than humic-like and fulvic acid-like substances (Figure 4d–i; Supplementary file 2; p < 0.01 in all cases) under lowland plants than alpine plants, also suggesting that lowland plants caused an enrichment in exudate-derived compounds (Jones et al., 2009). While further directed studies are required to resolve whether root exudates are truly involved, our findings collectively suggest that lowland plants have the capacity to increase total root exudation into alpine soil relative to resident alpine plants. In turn, root exudation is a well-established regulator of soil microbial decomposition and the associated release of soil carbon as carbon dioxide (Allison et al., 2010; Melillo et al., 2002).

One explanation for increased soil carbon outputs under lowland plants is that lowland plants have more biomass than alpine plants and therefore stimulate microbial activity to a greater extent (Figure 2a–c). We investigated this possibility by summarizing pot-level traits from the glasshouse experiment using a principal components analysis (PCA) and testing for relationships between resulting axes of trait variation and alpine soil microbial respiration (Materials and methods). We found that descriptors of plant biomass represented the most important axis of trait variation (Appendix 1—figure 2) and were positively related to microbial respiration (Appendix 1—figure 2a; PC1: LR = 4.43, N = 20, p = 0.035). Thus, greater biomass of lowland plants may have contributed to amplified microbial activity in soils with lowland plants in the glasshouse experiment. However, it is unlikely that increased biomass was solely responsible for lowland plant effects in the field experiments because lowland plant cover was low (mean ± SE relative cover: 4.7% ± 0.7%). Furthermore, in the glasshouse experiment we found that trait variation represented by the second principal axis had opposing effects on microbial respiration between lowland and alpine treatments (Appendix 1—figure 2b; treatment × PC2: LR = 5.98, N = 20, p = 0.015). Lowland plants thus displayed relationships with the alpine soil system that were qualitatively distinct from resident alpine plants. While focussed work is now needed to pinpoint the physiological mechanism behind these differences, the existence of such a distinction raises the possibility that the arrival of warm-adapted lowland plants in warming alpine ecosystems initiates a change in plant community functioning via changes to traits.

Plants adapted to different temperatures can possess distinct traits that have opposing effects on carbon cycle processes (Walker et al., 2019). We therefore tested whether lowland plants in both the glasshouse and field experiments indeed differed from alpine plants in traits that could affect the alpine soil system. Publicly available trait data (Kattge et al., 2020) (plant height, leaf area, seed mass, SLA, leaf carbon, leaf nitrogen; Materials and methods) relating to all species in the field experiments revealed that lowland plants on average possessed distinct traits from alpine plants (Appendix 1—figure 3a; p = 0.025). This was also the case for measured traits of lowland versus alpine plants under standardized conditions in the glasshouse experiment (multidimensional analyses: Appendix 1—figure 3b; p = 0.001). Coupled to this, while lowland plants were present at low abundance in the field experiments (mean ± SE relative cover: 4.7% ± 0.7%), the magnitude of soil carbon loss caused by lowland plant presence was greater in plots where lowland plants were more abundant overall (Appendix 1—figure 3e; F1,14 = 5.79, p = 0.030). By contrast, we found no evidence that soil carbon loss in the field was alternatively related to individual lowland species identity (Appendix 1—figure 3c, d) or the composition of the recipient alpine plant community (Appendix 1—figure 3f, g). Based on these observations, we hypothesize that, at least in our experiments, the establishment of warm-adapted lowland plants in warmed alpine ecosystems introduces novel traits into the community that alter plant community functioning and thus carbon cycle processes, with effects that intensify as lowland plants become more abundant in the community (De Deyn et al., 2008; Cornwell et al., 2008). An alternative explanation is that, trait differences aside, the ability of alpine plants to facilitate soil carbon loss is suppressed when growing in warmer climates to which they are not adapted. While either explanation could ultimately accelerate soil carbon loss from warming alpine ecosystems following the arrival of lowland plants, future studies are needed to fully unravel the environmental dependence of plant trait effects on soil ecosystem processes.

Lowland plant-induced carbon loss is temporally dynamic

We have shown using two field experiments that lowland plants facilitate a decrease in soil carbon content from warming alpine ecosystems (Figure 1) and, with a glasshouse experiment, that they potentially act by accelerating soil carbon outputs via microbial respiration relative to alpine plants (Figures 3 and 4), as opposed to decreasing soil carbon inputs. We therefore hypothesized that lowland plant arrival in transplanted alpine plots in the field experiments would drive a net release of carbon dioxide from the ecosystem. As for the glasshouse experiment, we initially examined how lowland plants affected the balance between carbon inputs and outputs of the ecosystem using carbon dioxide flux measurements taken on multiple dates during the growing season (Materials and methods). We then used soil samples collected on one representative date to focus on intrinsic rates (i.e. per unit of biomass) of microbial respiration and growth as processes that promote soil carbon loss and gain, respectively (Allison et al., 2010; Liang et al., 2017), as well as microbial biomass as the scalar through which these processes affect soil carbon (Hartley et al., 2008; Walker et al., 2018).

We found that transplantation from high to low elevation (i.e. warming-only) affected alpine ecosystem net carbon dioxide exchange differently between regions (Figure 5a; Supplementary file 1; p = 0.040). In the western Alps, transplantation changed the ecosystem from a net sink to a net source of carbon dioxide, whereas in the central Alps it made the ecosystem a stronger net carbon dioxide sink. This effect occurred because although transplantation accelerated rates of both gross primary production (Figure 5b; Supplementary file 1; p < 0.001) and ecosystem respiration (Figure 5c; Supplementary file 1; p < 0.001), in the western Alps it had stronger effects on ecosystem respiration and in the central Alps stronger effects on gross primary production. Importantly, lowland plants had no significant bearing over net ecosystem exchange (Figure 5a), implying that although lowland plants were associated with soil carbon loss from warmed alpine plots (Figure 1), this must have occurred prior to carbon dioxide measurements being taken and was no longer actively occurring. Gross primary production in warmed alpine plots was also unaffected by lowland plant presence (Figure 5b), which is consistent with the low abundance of lowland plants in alpine communities (mean ± SE relative cover: 4.7% ± 0.7%). As such, lowland plants likely did not facilitate alpine soil carbon loss under warming by decreasing carbon inputs into the ecosystem. By contrast, ecosystem respiration in warmed alpine plots was depressed in the presence versus absence of lowland plants (Figure 5c). These findings generally support the hypothesis that lowland plants affect the alpine soil system by changing carbon outputs. However, they contrast with expectations that lowland plants perpetually increase carbon outputs from the ecosystem and thus raise questions about how soil carbon was lost from warmed plots containing lowland plants (Figure 1).

Figure 5
Download asset Open asset
Warming and lowland plant effects on alpine ecosystem carbon dioxide fluxes in the field experiment.

Effects of transplantation to low elevation (warming, W), transplantation plus lowland plants (warming plus lowlands, WL), or replantation at high elevation (control, C) on alpine ecosystem carbon dioxide fluxes (mean ± standard error [SE]) during one season. (a) Net ecosystem CO2 exchange (NEE; g CO2 m−2 hr−1). (b) Gross primary production (GPP; g CO2 m−2 hr−1). (c) Ecosystem respiration (ER; g CO2 m−2 hr−1). Measurements were taken in situ at both experiments on multiple dates (western Alps: N = 196, central Alps: N = 90), with negative values indicating carbon uptake into the ecosystem. Treatment effects (p < 0.05) are indicated by letters (LMEs, Tukey; see Supplementary file 1).

Carbon cycle processes are constrained by multiple feedbacks within the soil system, such as substrate availability (Hartley et al., 2008; Walker et al., 2018) and microbial acclimation (Crowther and Bradford, 2013; Bradford et al., 2019; Melillo et al., 2017), that over time can slow, or even arrest, soil carbon loss (Walker et al., 2018; Bradford et al., 2019; Melillo et al., 2017). We thus interrogated the state of the soil system in the field experiments in the western Alps experiment to explore whether such a feedback may be operating here, in particular to limit ecosystem respiration once soil carbon content had decreased in warmed alpine plots containing lowland plants. We found that biomass-specific rates of microbial respiration in the field were greater following transplantation to the warmer site in the presence, but not absence, of lowland plants (Figure 6a; Supplementary file 1; p = 0.038). By contrast, biomass-specific rates of microbial growth were slower following transplantation but remained unaffected by lowland plant presence (Figure 6b; Supplementary file 1; p = 0.016). Thus, despite lower rates of ecosystem respiration overall, alpine soil microbes still respired intrinsically faster in warmed plots containing lowland plants. Moreover, accelerated microbial respiration, but not growth, implies that alpine soils had a higher capacity to lose carbon under warming, but not to gain carbon via accumulation into microbial biomass, when lowland plants were present (Allison et al., 2010; Liang et al., 2017). These findings align with observations from the glasshouse experiment that lowland plants generally accelerated intrinsic rates of microbial respiration (Figure 3), although in field conditions this effect occurred in tandem with warming.

Figure 6
Download asset Open asset
Warming and lowland plant effects on the alpine soil system in the field experiment.

Effects of transplantation to low elevation (warming, W), transplantation plus lowland plants (warming plus lowlands, WL), or replantation at high elevation (control, C) on alpine soil pools and processes (means ± standard errors [SEs]) after one season. (a) Biomass-specific microbial respiration (Rmass; ng C µg−1 Cmic hr−1). (b) Biomass-specific microbial growth (Gmass; ng C µg−1 Cmic hr−1). (c) Cmic (µg C g−1). (d) Total microbial respiration (µg C g−1 h−1). (e) Total microbial growth (µg C g−1 h−1). (f) Microbial carbon use efficiency (CUE; %). Measurements were taken on soils sampled on one date, for (f) at both experiments (western Alps: N = 28; central Alps: N = 30) and for all other variables at the western Alps experiment (N = 28). Treatment effects (p < 0.05) are indicated by letters (LMEs, Tukey; see Supplementary file 1).

Faster microbial respiration can only yield a loss of soil carbon if it occurs on a per unit of soil basis (Walker et al., 2018; Melillo et al., 2017; Prommer et al., 2020), which was the case in the glasshouse experiment (Figure 3). However, in the field we found that lowland plants caused a reduction of microbial biomass carbon in alpine communities transplanted to the warmer site (Figure 6c; Supplementary file 1; p < 0.001). As such, although microbes generally respired intrinsically faster in the presence of lowland plants, in the field they had become less abundant and so we observed no change to microbial respiration when expressed per unit of soil (Figure 6d; Supplementary file 1; p = 0.550). The decline in microbial biomass likely took place because soil microbes are chiefly carbon limited (Hobbie and Hobbie, 2013), making their abundance tightly coupled to soil carbon content (here: r26 = 0.61, p = 0.001) (Walker et al., 2018). Moreover, since intrinsic rates of microbial growth were unaffected by lowland plant presence, a reduction in microbial biomass translated into a decline in microbial growth per unit of soil (Figure 6e; Supplementary file 1; Supplementary file 1; p = 0.009) – preventing any recovery of microbial biomass via growth (Liang et al., 2017). We thus suggest that as soil carbon was lost, microbial biomass also declined, and at the time of measurement this had eliminated the influence of accelerated biomass-specific microbial respiration over the soil system. In short, we propose based on these relationships that the decrease in soil carbon content observed in warmer transplanted plots containing lowland plants had both started and stopped within one season.

Taken together, one interpretation of our findings is that the establishment of lowland plants in warming alpine ecosystems accelerates intrinsic rates of microbial respiration (Figures 3 and 6a), leading to soil carbon release at baseline levels of microbial biomass (Figures 1 and 3c), a coupled decline in microbial biomass (Figure 6c) and a cessation of further carbon loss from the ecosystem (Figures 5a and 6d). Although such a mechanism has been reported in other ecosystems (Walker et al., 2018; Melillo et al., 2017), applying it here is speculative without additional timepoints because field soil measurements came from a single sampling event after soil carbon had already been lost from the ecosystem. For instance, an alternative mechanism could be that soil microbes acclimate to the presence of lowland plants and this decelerated microbial processes over time (Crowther and Bradford, 2013; Bradford et al., 2019; Melillo et al., 2017). Nevertheless, it remains true that biomass-specific microbial growth and respiration affect the soil system as a function of microbial biomass, and here changes to microbial biomass could explain the behaviour of the soil system in both glasshouse and field experiments. We also found no evidence that soil microbes instead changed their resource use strategy in the presence versus absence of lowland plants (i.e. carbon use efficiency; Figure 6f; Supplementary file 1; p < 0.001; Tukey: p = 0.996), pointing away from microbial acclimation as an alternative mechanism. Beyond the mechanism for lowland plant effects on alpine soil carbon loss, it is conceivable that soil carbon loss is not isolated to a single season, but will reoccur in the future even without further warming or lowland plant arrival. This is especially true in the western Alps experiment where warming yielded a net output of carbon dioxide from the ecosystem (Figure 5a). Moreover, in our field experiments we simulated a single event of lowland plant establishment and at relatively low abundance in the community (mean ± SE relative cover: 4.7% ± 0.7%), raising the possibility that increases in lowland plant cover or repeated establishment events in the future could facilitate further decreases in alpine soil carbon content under warming.

General conclusions

In summary, we show that uphill migrations of native lowland plants have the capacity to facilitate warming effects on soil carbon loss from alpine ecosystems. This is important because it demonstrates that plant range expansions between herbaceous plant communities have the potential to contribute to climate feedbacks. These results contrast with expectations from woody species expansions in the Arctic, where shrubs with recalcitrant litter can slow carbon turnover in herbaceous communities and yield a negative climate feedback (Pearson et al., 2013; Cornelissen et al., 2007). We thus suggest that the influence of range expanding species on ecosystem carbon dynamics depends on the relative trait composition of migrating species versus the recipient community (Bjorkman et al., 2018), as well as subsequent effects on environmental processes (Pearson et al., 2013). While our findings demonstrate that lowland plants affect the rate of soil carbon release in the short term, short-term experiments, such as ours, cannot resolve whether lowland plants will also affect the total amount of soil carbon lost in the long term. This includes whether processes such as genetic adaptation (in both alpine and lowland plants) (Walker et al., 2019) or community change (Walker et al., 2016; Bardgett et al., 2013) will moderate soil carbon responses to gradual or sustained warming. Nevertheless, here decreases in soil carbon content occurred rapidly despite a low abundance of lowland plants, and increased with total lowland plant cover. Given that both warming and associated plant migrations will intensify over time (Rumpf et al., 2018; Steinbauer et al., 2018), our estimates of lowland plant impacts may be conservative – particularly if increases in lowland plant cover or repeated establishment events cause recurrent pulses of soil carbon release. Moreover, while alpine ecosystems store only a small fraction of total global soil carbon (Streit et al., 2014; Parker et al., 2015), redistributions among herbaceous plant communities are occurring globally (Parmesan, 2006; Lenoir and Svenning, 2015; Pecl et al., 2017; Dolezal et al., 2016; Rumpf et al., 2018; Steinbauer et al., 2018), including in similarly cool ecosystems in the Arctic that contain vast soil carbon stocks (Tarnocai et al., 2009). This raises the potential for resulting climate feedbacks to also be widespread. Future work should focus on testing the conditions under which this feedback could occur in different mountain regions, as well as other ecosystems experiencing influxes of range expanding plant species, on quantifying how deeply it occurs in shallow alpine soils, and on estimating the magnitude of the climate feedback given both ongoing warming and variation in rates of species range shifts.

Materials and methods

Field site description, experimental design

Request a detailed protocol

We made use of two independent alpine grassland transplant experiments that were established in Switzerland in the western Alps (46°12′40″N, 7°03′13″E) (Nomoto and Alexander, 2021) and central Alps (46°52′20″N, 9°28′58″E) in autumn 2016. The experiments simulate effects of climate warming on alpine ecosystems via transplantation from high to low elevation, including the arrival of lowland plants due to warming-induced uphill plant migrations (Appendix 1—figure 1). In the western Alps experiment, the plant community is dominated by Sesleria caerulea, Carex sempervirens, and Vaccinium vitis-idea. In the central Alps experiment the dominant species are Nardus stricta, Alchemilla xanthochlora, Trifolium pratense, and A. vulneraria. Broadly, the experiments share the same design. Turfs (area = 1 m2, depth = 20 cm) of intact alpine vegetation plus rhizosphere soil were removed from a high-elevation control site (western Alps: 2200 m, MAT = 0.7°C, TAP = 1448 mm; central Alps: 2000 m, MAT = 1.6°C, TAP = 1151 mm; Karger et al., 2017) and replanted either at a low-elevation site to simulate climate warming (western Alps: 1400 m, MAT = 6.0°C, TAP = 1236 mm, N = 10; central Alps: 1400 m, MAT = 4.4°C, TAP = 1010 mm, N = 20; Karger et al., 2017) or back into a high-elevation site as a transplantation control (see below). At the low-elevation site, we created an additional treatment to simulate the arrival of lowland plants under warming by planting a low abundance of lowland plant species in a grid design (13 cm between individuals), creating a lowland plant density of approximately 56 individuals per m2 (4.7% ± 0.7% total cover) in a background alpine plant community. Lowland plants were positioned randomly in the grid, but with the same number of individuals per species per plot, and we controlled for planting disturbance in all remaining plots using alpine plants in place of lowland plants (see below). In the western Alps region, lowland plant species were Achillea millefolium, Bellis perennis, B. erectus, Dactylis glomerata, M. lupulina, Plantago media, and S. pratensis, and individuals were planted into one side of each plot in a split-plot design (n = 10). In the central Alps experiment, lowland plant species were Brachypodium pinnatum, Carex flacca, Carum carvi, D. glomerata, Hypericum perforatum, Plantago lanceolata, Primula veris, Ranunculus bulbosus, S. pratensis, Silene vulgaris, Trifolium montanum, and Viola hirta, and individuals were planted into half of the transplanted turfs (n = 10). While exact dispersal distances for selected lowland species are unknown, all species are widespread and are expected to migrate uphill under warming (Rumpf et al., 2018) and the horizontal distance between high and low sites in the field experiments was always less than 2 km. In the western Alps experiment, lowland individuals were pre-grown in a greenhouse in steam-sterilized field soil and planted into the plots in May 2017. In the central Alps experiment, lowland individuals were taken from the surrounding vegetation and planted into the plots in autumn 2016. In both regions, soil was partially removed from lowland individuals prior to planting, and we performed an additional experiment to eliminate the possibility that small contaminations of lowland soil biota during planting could have modified alpine soil microbial community composition (see Field experiment soil biota validation). Overall, the experimental design in both regions yielded three treatments (n = 10 for each) arranged in a block design: transplantation from high to low elevation (i.e. warming-only, W); transplantation plus low-elevation plants (i.e. warming plus lowland plant establishment, WL); and replantation at the high-elevation origin site (i.e. control, C). One turf at the low-elevation site in the western Alps experiment failed, leaving a total sample size of N = 28 and N = 30 in the western Alps and central Alps, respectively.

Field experiment controls and validation

Request a detailed protocol

We controlled for transplantation disturbance in both experiments by moving alpine control turfs back into a high-elevation control site. In the western Alps experiment, alpine turfs were planted into different positions at their origin high-elevation site. In the central Alps experiment, alpine turfs were planted into a new high-elevation site at the same elevation. As such, we eliminated the possibility that alpine turfs changed simply because they were disturbed or were moved to a different location (i.e. irrespective of elevation). The transplantation process had no effect on alpine soil carbon content, in that we found no difference in soil carbon content between replanted alpine control turfs at high-elevation sites and undisturbed alpine soil in the surrounding area (LR = 2.27, N = 16, p = 0.132). We also controlled for potential effects of planting disturbance in all turfs not planted with lowland plants (i.e. remaining turfs at low elevation and all control turfs at high elevation). In the western Alps experiment, we planted alpine species from the high-elevation control site into the other side of alpine turfs at the low-elevation site, as well as into all control turfs at the high-elevation site. This was done in autumn 2016 using the same approach and randomized hexagonal grid design. In the central Alps experiment, we mimicked planting disturbance by digging up and replanting existing individuals in alpine turfs at the low-elevation site not containing lowland plants, as well as all control turfs at the high-elevation site. This was done in spring 2017 using the same grid design. It is important to note that we did not perform a reverse transplantation (i.e. from low to high elevation), so we cannot entirely rule out the possibility that transplantation of any community to any new environment could yield a loss of soil carbon. Nevertheless, it is well established that climate warming promotes soil carbon loss (Melillo et al., 2002; Crowther et al., 2016; Davidson and Janssens, 2006), and we found that soil from low-elevation sites (i.e. the potential end-member for warming alpine ecosystems) contained less carbon than soil from the alpine sites (70.1 ± 3.05 mg C g−1 soil dry mass compared to 145.0 ± 5.29 mg C g−1 soil dry mass; LR = 37.2, N = 27, p < 0.001), validating our expectation that alpine soil would lose carbon under warming. Finally, while small variations in experimental design between regions, such in the order of planting in W and WL treatments, likely introduced some error when comparing experiments, we used both regions together so that we could identify treatment effects that were consistent among them and were thus not artefacts of experimental design.

Field experiment soil biota validation

Request a detailed protocol

We tested for the possibility that soil carbon loss was alternatively caused by low-elevation soil microbes hitchhiking on lowland plants during planting. Specifically, we performed an experiment at the central Alps site to determine whether inoculating alpine plots with low-elevation soil biota affected soil microbial community composition. We first generated a low-elevation soil biota inoculum from rhizosphere soil from the same area where lowland plants were sourced and gently mixing and sieving it with tap water (following refs van de Voorde et al., 2012; De Vries et al., 2015). We then added 3 L of inoculum to one half of each of 10 × 1 m2 alpine turfs equivalent to those in the main experiment, followed by 3 L of water to rinse the inoculum from the vegetation and facilitate its infiltration into the soil. The other half of the plot (separated by an impermeable rubber barrier buried 15 cm deep into the soil around the perimeter of each subplot) received 6 L of water as a control. Soil inoculation was performed in mid-May and repeated in mid-June 2017. In October 2017, we sampled soils by pooling three soil cores (ø = 1 cm, d = 5 cm) from each subplot, following which we determined soil microbial community composition by sequencing the 16S rRNA gene for bacteria/archaea and the ITS1 region for fungi (Herbold et al., 2015). DNA was extracted (NucleoSpin Soil kit, Macherey-Nagel, Düren, Germany) and amplified using quadruplicate PCRs, alongside extraction and PCR blank controls. PCR products were purified (MinEluteTM PCR purification kit, Qiagen, Hilden, Germany) prior to sequencing on a MiSeq sequencing platform (Illumina, San Diego, USA). Sequence data were processed using the OBITools software package (Boyer et al., 2016) and following Zinger et al., 2017. Paired-end reads were merged, assigned to their respective samples and dereplicated, and sequences with too short read lengths and with only one read were excluded. Remaining sequences were clustered using the SUMACLUST (Mercer et al., 2013) algorithm and summarized into operational taxonomic units (OTUs) using the Infomap Community Detection Algorithm (Rosvall and Bergstrom, 2008), following which we calculated per-sample relative OTU abundances for bacteria/archaea and fungi separately. We found that the alpine soil microbial community was resistant to the introduction of low-elevation soil biota, in that inoculation had no effect on soil bacterial (Appendix 1—figure 4a; PERMANOVA; F1,9 = 0.75, p = 0.573) or fungal (Appendix 1—figure 4b; F1,9 = 0.75, p = 0.774) community composition. The inoculation represented a strong soil biota treatment relative to contamination by hitchhiking microbes. As such, even if some contamination by low-elevation soil biota did occur, it was unlikely to have resulted in a change to the extant soil community, ruling out the possibility that contamination was alternatively responsible for the soil carbon loss observed when lowland plants were present in transplanted alpine plots.

Field plant community composition

Request a detailed protocol

We surveyed the composition of alpine communities in summer 2017. This was done by dividing each plot into 5 × 5 cm cells and visually estimating the canopy cover of all species rooted in every grid cell. Species cover was approximated with an ordinal scale in which the first three cover categories correspond to 1/16, 1/8, and 1/4 of the grid cell area (25 cm2), and subsequent cover categories were constant increases of 1/4 of the grid cell area. Plants were identified to species level whenever possible, although some taxa were identified only to genus or subfamily levels when species identity was unclear based on vegetative traits. Species cover (cm2) was then calculated as the species-wise sum of cover estimates across all grid cells in the quarter of the plot used for soil and ecosystem respiration measurements (see below).

Field soil carbon content, microbial physiology, microbial biomass

Request a detailed protocol

We collected soil samples from all treatments in both field experiments (N = 58) in August 2017. In each plot, we created a composite sample from three cores (ø = 1 cm, approx. d = 7 cm) no closer than 7 cm from a planted individual and from the same quarter of the plot used for ecosystem respiration measurements (see below; Appendix 1—figure 1). We focused on rhizosphere soil for two reasons. First, alpine soils are shallow and alpine plants are shallow-rooting, making the rhizosphere the most logical target for a first assessment of warming and lowland plant effects on alpine soil carbon cycling. Second, we wanted to avoid any potential contamination from the base of the turf (depth = 20 cm). Soils were kept cool, transported immediately to the laboratory, sieved (2 mm mesh size), adjusted to 60% of water holding capacity (WHC) and pre-incubated for 3 days at field temperature prior to measurements. We measured microbial biomass carbon (mg per g dry mass) via chloroform fumigation extraction followed by analysis on a TOC-VCPH/CPNTNM-1 analyzer (Shimadzu; 48 hr incubation period; 1 M KCl extraction for fumigated and non-fumigated samples; conversion factor 0.45) (Vance et al., 1987). While this approach also yielded measurements of dissolved organic carbon (DOC; mg C per g dry mass), treatment effects on this variable echoed those of microbial biomass carbon (Supplementary file 1) so we did not interpret them separately. Soil organic carbon content (%) was measured on dry soil (60°C for 48 hr) with a Carlo Erba 1,110 elemental analyzer (EA, CE Instruments) coupled to a Delta Plus IRMS via a Conflo III (Thermo Fisher). It was not possible to take widespread measurements of soil bulk density due to the destructive sampling required while other studies were underway (e.g. Nomoto and Alexander, 2021). Instead, we took additional soil cores (ø = 5 cm, d = 5 cm) from the central Alps experiment in 2021 once other studies were complete to indirectly explore whether lowland plant effects on soil carbon content in warmed alpine plots could have occurred due to changes in soil bulk density. We found that although transplantation to the warmer site increased alpine soil bulk density (LR = 7.18, p = 0.028, Tukey: p < 0.05), lowland plants had no effect (Tukey: p = 0.999). It is not possible to make direct inferences about the soil carbon stock using measurements made on different soil cores 4 years apart. Nevertheless, these results make it unlikely that lowland plant effects on soil carbon content in warmed alpine plots occurred simply due to a change in soil bulk density. We additionally quantified microbial growth, respiration and carbon use efficiency in soils from the western Alps experiment (N = 28) using the incorporation of 18O into DNA (Walker et al., 2018). Briefly, we incubated 500 mg of soil for 24 hr at field temperature following addition of either 18O-H2O to 20 atom% enrichment and 80% of WHC or the same volume of molecular grade non-labelled H2O as a natural abundance control. Gas samples were taken at the start and end of the 24 hr incubation and analyzed for CO2 concentration with a Trace GC Ultra (Thermo Fisher) to determine microbial respiration (µg C per g dry mass per hr). At the end of the incubation, soil was snap frozen in liquid nitrogen, and DNA was extracted (FastDNA SPIN Kit for Soil, MP Biomedical), quantified (Quant-iT PicoGreen dsDNA Assay Kit, Thermo Fisher) and analyzed for 18O abundance and total oxygen content using a thermochemical EA coupled to a Delta V Advantage IRMS via a Conflo III (Thermo Fisher). Following Walker et al., 2018, we calculated total DNA production during the 24-hr incubation (µg DNA per g dry mass) and coupled it with measures of microbial biomass carbon content to derive microbial growth (µg C per g dry mass per hr) and carbon use efficiency (%).

Field ecosystem respiration

Request a detailed protocol

We measured rates of ecosystem respiration (g CO2 per m2 per h) on multiple dates (western Alps: seven dates; central Alps: three dates) between July and August 2017. One month prior to the first measurement, we installed permanent gas sampling collars (h = 10 cm, ø = 30 cm) in each plot to a depth of 5 cm to create an airtight seal at the soil surface (Walker et al., 2019). Measurements were made by enclosing plants and soil within the collars with an opaque chamber (h = 50 cm, ø = 30 cm) and, following a 30-s equilibration period, monitoring the increase in CO2 concentration at 5-s intervals over a 60-s period (Vaisala GMP343 CO2 probe attached to an HM70 meter, Vaisala). Chambers contained a fan to mix the headspace during measurements and a small sealable hole on the top that was opened when placing the chamber on sampling collars to avoid pressurization inside the chamber. Soil temperature and moisture were recorded for each measurement and used alongside measurements of air temperature, air pressure, enclosure volume, and soil surface area to calculate the CO2 flux (Walker et al., 2019). Where necessary, we repeated measurements in the field to attain a visually consistent CO2 change over time, and during data processing we discarded fluxes with an r-squared of less than 0.9 for this relationship.

Glasshouse experiment design

Request a detailed protocol

We established a 6-week pot experiment to test for effects of lowland versus alpine plants on the alpine soil system. Specifically, we filled pots (ø = 10 cm, depth = 9 cm) with sieved soil (mesh size 4 mm) from the high-elevation site of the western Alps field experiment, and planted them with six lowland species (B. erectus, S. pratensis, M. lupulina) or six alpine species (P. alpinum, A. vulneraria ssp. alpestris, P. alpina) to create a lowland only treatment, an alpine only treatment and a bare soil control (N = 30). We selected a grass, herb, and legume species for each treatment, all of which were found naturally at the low (lowland treatment) or high (alpine treatment) elevation sites in the western Alps field experiment. Vegetated treatments contained two individuals of all lowland or alpine species planted in a circular pattern such they were never adjacent to conspecifics or plants of the same functional type (i.e. grass, herb, legume), making replicates within treatments identical in species composition and neighbourhoods. The original design also included a mixed community treatment containing one individual of each lowland and alpine species. However, the mixed treatment had double the species richness of other vegetated treatments and contained an equal cover of lowland and alpine plants (compared to <5% lowland plant cover in field plots), so we focussed our analyses on the comparison between alpine and lowland plant treatments to understand how lowland plants affect the soil differently from alpine plants. All plants were pre-grown from seed provided by commercial suppliers (https://www.hauenstein.ch, https://www.schutzfilisur.ch, https://www.saatbau.at) in steam-sterilized potting soil for 2–3 weeks to approximately the same biomass and root-washed immediately prior to planting. Pots were arranged in a randomized block design (n = 10) in a greenhouse under standardized growing conditions (22°C; 50% humidity; no artificial light), and were watered every 2–3 days as needed. After 6 weeks, we destructively harvested the plants and soil to characterize treatment effects on plant physiology and the soil system.

Glasshouse plant physiology

Request a detailed protocol

We measured maximum leaf-level photosynthesis (photosynthetic capacity; Amax; µmol CO2 per m2 per s) and stomatal conductance (gs; µmol H2O per m2 per s) under optimum conditions on one individual per species per pot using an open path infrared gas analyzer connected to a 2.5-cm chamber (CIRAS-2 and PLC-6, PP Systems) (Walker et al., 2019). Readings were taken once for each individual and corrected using measurements of leaf area (cm2) on the same leaf, which we also dried and weighed to determine SLA (cm2 per g). Following soil pore water sampling (below), we quantified total, aboveground and belowground biomass of all individuals in the experiment by destructively sampling, washing, drying (60°C for 48 hr), and weighing all plant tissue.

Glasshouse soil pore water analysis

Request a detailed protocol

We collected soil pore water from all pots prior to soil sampling (see below) to determine treatment effects on DOM quality. Soil water was sampled by leaching pots twice with 50 mL sterile tap water, combining and filtering leached water (filter pore size 0.2 µm) and measuring absorbance (200–900 nm with a 1-nm increment; 1-cm quartz cuvette; UV–vis spectrophotometer, Shimadzu) and fluorescence intensity (excitation: 240–450 nm with a 5-nm increment; emission: 290–550 nm with a 2.5-nm increment; Fluorolog-3 spectrometer, Horiba). We corrected absorbance spectra against a milli-Q water blank, calculated Napierian absorption coefficients and used these data to derive absorbance at 350 nm (a350; a proxy for DOM content). We expressed fluorescence data as total fluorescence (Ftot; Raman Units) and the FI, that is the ratio of emission intensity at 470 versus 520 nm at an excitation wavelength of 370 nm (a proxy for microbially derived DOM) (McKnight et al., 2001). We also derived excitation–emission matrices (EEMs) to quantify the relative abundance of different components of DOM through PARAFAC algorithms (Stedmon et al., 2003). Following preprocessing (i.e. Raman scattering removal and standardization to Raman Units) and normalization (Zepp et al., 2004), we used the drEEM toolbox (Murphy et al., 2013) in MATLAB (Mathworks) to obtain a six-component PARAFAC model. The model was validated with split-half analysis, random initialization and visualization of EEM residuals, after which we calculated the relative abundance of each component as the maximum fluorescence signal of the component against the total maximum fluorescence signal of all components. Finally, we classified components against classical EEM nomenclature (Supplementary file 4; Fellman et al., 2010; Coble et al., 2014). While PARAFAC modelling provides a means to discriminate between DOM sources (e.g. components C3 and C4; Supplementary file 4), we did not physically separate plant- and microbe-derived DOM when sampling soil pore water. As such, observed changes to DOM composition were likely a result of changes to both plant and microbial processes (see Main text).

Glasshouse soil carbon pools, microbial respiration

Request a detailed protocol

We took soil cores (ø = 1 cm, d = 7 cm) from the centre of each pot (i.e. equidistant from all individuals) to quantify soil carbon pools and microbial respiration. Soils were sieved (2-mm mesh size), adjusted to 60% of their WHC and pre-incubated at 22°C for 3 days prior to measurements. We quantified concentrations of microbial biomass carbon in all soils as for field soil (see above; mg C per g dry mass). We determined microbial respiration by incubating 8 g sieved soil in the dark at 22°C for a further 6-week period without plants. We measured microbial respiration daily for the first 4 days, then approximately biweekly until the end of the 6-week period. We measured CO2 concentrations (EGM-4, PP Systems) of gas samples taken approximately 24 hr after flushing enclosed vials (v = 126 ml) with CO2-free air, and calculated microbial respiration (µg C per g dry mass per hr) using accompanying measurements of enclosure volume, air pressure, and soil mass (Walker et al., 2019). After each measurement, soils were weighed and, when necessary, re-adjusted to 60% of their WHC.

Two-pool mixed models of soil carbon mineralization

Request a detailed protocol

We described the kinetics of soil carbon mineralization during the follow-up incubation of glasshouse soil by employing a two-pool mixed model to microbial respiration data (Bonde and Rosswall, 1987). Specifically, we constructed a model that considered a first pool that followed first-order decomposition kinetics (i.e. a small soil carbon pool with a fast turnover) and a second pool that followed zero-order kinetics (i.e. decay rate was constant over the incubation period, representing a large soil carbon pool with a slow turnover):

Cmin=Cl(1−e−tkl)+B⋅t

where Cmin is the cumulative CO2 respired (µg C g−1) over the incubation time t (d), Cl and kl are the size (µg C g−1), and decomposition constant (d−1) of the first-order kinetics carbon pool, respectively, and B is the basal respiration rate (µg C g−1 d−1) of the zero-order kinetics carbon pool. Given that Cmin and t were known, we ran the two-pool mixed model with the nlsList function (maxiter = 100, tol = 1 × 10−5, minFactor = 1/1024) in R Development Core Team, 2014 using the package ‘nlme’ (Pinheiro et al., 2019) for each pot separately to estimate values for Cl, kl, and B. Models failed to converge for one sample, and we excluded four additional samples for which models made negative pool estimates (N = 25).

Statistical analysis

Request a detailed protocol

The majority of statistical test outputs appear as p values in the Main text, with full outputs being provided in Supplementary files 1–3. See text below for exceptions. Data analyses and visualization were undertaken in R Development Core Team, 2014 using the packages ‘cowplot’ (Wilke, 2019), ‘drake’ (Michael Landau, 2018), ‘emmeans’ (Lenth, 2020), ‘nlme’ (Pinheiro et al., 2019), ‘tidyverse’ (Wickham et al., 2019), and ‘vegan’ (Oksanen, 2019). Where necessary, region was accounted for in statistical models as a fixed effect, since two levels was too few to be included as a random intercept term. As such, region effects are presented (Supplementary files 1 and 3), but not interpreted. Statistical methods are described as they appear in the Main text.

Warming and warming plus lowland plant effects on soil carbon content

We used field data to test for treatment (C, W, WL) effects on soil carbon content using an LME model including region as a fixed effect and block as a random intercept term (see Supplementary file 1 for test output). We also calculated how much more soil carbon loss occurred in the presence versus absence of lowland plants by fitting a single LME testing for effects of soil temperature (plot means of all sampling dates), the presence or absence of lowland plants and their interaction on soil carbon content (see Appendix 1—figure 3 for test outputs). In order to compare slopes between C–W and C–WL treatment combinations, we duplicated C treatment values, assigned one set to each C–W and C–WL pair and accounted for this using a random intercept term. We also included region as a random slope term. We then tested how much more soil carbon loss would occur per °C of warming when lowland plants are present by calculating the relative difference (% ± 95% CIs) between the C–WL and C–W slopes.

Plant mechanisms for soil carbon loss

Request a detailed protocol

We expressed soil carbon content as the block-wise percent changes between WL and W treatments to derive a metric for lowland plant-induced soil carbon loss (i.e. % soil carbon remaining). We tested for lowland species identity effects on soil carbon remaining using separate LMEs for each region (due to different lowland species pools) that tested for individual species cover (cm2) effects on soil carbon remaining in WL plots, including species as a random intercept term (see Appendix 1—figure 3 for test outputs). We also performed separate linear models including species as fixed effects, none of which were statistically significant (Supplementary file 3). We tested for effects of total lowland species cover (cm2) on soil carbon remaining in WL plots using a linear model also including region as a fixed effect. We summarized alpine community composition in W and WL treatments using an NMDS (Hellinger distance matrix, two axes), following which we expressed NMDS scores as the block-wise percent changes between W and WL treatments to derive metrics for lowland plant-induced alpine plant community change, and used a linear model that tested for effects of changes to NMDS scores on soil carbon remaining in WL plots including region as a fixed effect (see Appendix 1—figure 3 for test outputs). For reference, raw NMDS scores did not significantly differ between W and WL treatments (LME: NMDS #1: LR = 1.57, p = 0.210; NMDS #2: LR = 2.67, p = 0.102). Finally, we tested for differences between alpine and lowland species traits in the field experiment using a PERMANOVA on a PCA of species mean trait values (log10 plant height, log10 leaf area, log10 seed mass, SLA, leaf carbon, leaf nitrogen) from the TRY database (Kattge et al., 2020) for all species present (where necessary, missing values were gap-filled following Schrodt et al., 2015), and in the glasshouse experiment using a PERMANOVA on a PCA of physiological measurements (total biomass, leaf biomass, root biomass, SLA, Amax, gs) for all plants (see Appendix 1—figure 3 for test outputs). Traits were selected to explore broad life history differences between alpine and lowland plant species (Diaz et al., 2016), but were not used to identify the specific trait(s) responsible for lowland plant effects on alpine soil carbon loss. We used TRY traits for field species due to the large number of species present in the communities (N = 118), but used measured traits in the glasshouse experiment to provide direct evidence of physiological differences between alpine and lowland species.

Lowland versus alpine plant effects on the soil system

Request a detailed protocol

We used the glasshouse experiment to test for lowland versus alpine plant treatment effects on soil microbial biomass carbon, the size (Cl) and decay rate (kl) of the first modelled soil carbon pool and all variables relating to soil DOM quality using LMEs including block as a random intercept term. We tested for effects of treatment, time, and their interaction on soil microbial respiration using the same LME structure but including the bare soil treatment. In all cases, including/excluding the bare soil treatment had no effect over model outcomes, but mean ± SE values for this treatment are included in figures for reference. Finally, we explored how plant physiology in lowland and alpine treatments at the pot level affected initial rates of soil microbial respiration. We defined initial microbial respiration as the pot-wise mean respiration rates of the first 9 days of soil incubation (i.e. while treatment effects persisted; see Figure 3a), and expressed pot-level plant physiology using scores from the first two axes of a PCA performed on sums of total, root and shoot biomass and biomass-weighted means of SLA, Amax, and gs. We then tested for effects of treatment, PC1 and PC2 scores and their interaction on initial microbial respiration using an LME including block as a random intercept term. Statistical test outputs for LMEs are provided in Supplementary file 2, except for treatment and plant physiology effects on microbial respiration, which appear in the Main text.

Sequence of warming and lowland plant effects on alpine soil carbon loss

Request a detailed protocol

We used field data to test for treatment (C, W, WL) effects on soil microbial biomass carbon and soil DOC using LME models including region as a fixed effect and block as a random intercept term. We tested for treatment effects on ecosystem respiration using an LME that included region as a fixed effect and sampling date nested in block as a random intercept term. We tested for treatment effects on soil microbial growth and respiration (per unit of soil and per unit of microbial biomass) and carbon use efficiency in the western Alps experiment using LMEs including block as a random intercept term. Statistical test outputs for these variables are provided in Supplementary file 1. We also used Pearson Product Moment correlations to test for associations between soil microbial biomass and soil carbon content, microbial respiration and growth in the western Alps experiment, irrespective of experimental treatment.

Data availability

Request a detailed protocol

Data supporting the findings of this study are stored on the Open Science Framework website (https://www.osf.io; DOI: https://doi.org/10.17605/OSF.IO/S54CH). The R code used to process and analyse the data are available in a Github repository with the URL: https://github.com/tom-n-walker/uphill-plants-soil-carbon (copy archived at swh:1:rev:a9e7a872d22a45dbd91bb00751d522812138eefd; Walker, 2022).

Appendix 1

Appendix 1—figure 1
Download asset Open asset
Experimental design.

(a) In both field experiments, intact alpine turfs containing alpine plant communities plus rhizosphere soil were removed from their high-elevation home sites and either replanted at the home site (negative control; C) or transplanted to a low-elevation recipient site to simulate climate warming (W, WL). At low elevation, half of the turfs were planted with a low abundance of lowland plant species to additionally simulate the arrival of migrating plant species in the ecosystem (WL). Both other treatments (i.e. C, W) were planted with the same abundance of alpine plant species as a disturbance control. (b) The layout of W and WL treatments at the low-elevation site differed between western and central alps experiments. In the central alps, W and WL treatments were established in distinct turfs planted adjacently in a block design (N = 10). In the western alps, W and WL treatments were established in different halves of the same turf in a split-plot design (N = 10).

Appendix 1—figure 2
Download asset Open asset
Alpine and lowland plant effects on alpine soil microbial respiration.

Relationships between plant physiology, expressed as (a) PC1 and (b) PC2 scores from PCAs of pot-level measurements, and rates of soil microbial respiration during the first nine days of incubation (initial R; µg C g−1 h−1) in alpine (green) versus lowland (yellow) treatments of the glasshouse experiment. Fit lines describe relationships as determined by a single linear mixed-effects model including treatment, PC1, PC2, and treatment × PC# interactions (N = 20; Main Text), with different colours in (b) indicating a significant treatment × PC2 interaction. Loadings of plant physiological measurements on (c) PC1 and (d) PC2 axes. Total, belowground, and aboveground biomass are sums of all plants per pot and specific leaf area (SLA), maximum photosynthetic capacity (Amax), and stomatal conductance (gs) are pot-wise community-weighted mean (CWM) values.

Appendix 1—figure 3
Download asset Open asset
Alpine and lowland plant effects on soil carbon content.

Trait separation of lowland (yellow) and alpine (green) plant species, displayed as PC1 and PC2 scores from PCAs containing (a) TRY data for species in field experiments (N = 242; F1,240 = 3.98, p = 0.003) or (b) measured data from a glasshouse experiment on selected species (N = 109; F1,107 = 17.69, p = 0.001). Arrows show loadings and p values refer to alpine–lowland comparison (PERMANOVAs). Relationships between soil carbon loss (% initial soil carbon content remaining) and cover (cm2) of each lowland plant species in the (c) western Alps (N = 36; LME: LR = 0.62, p = 0.4302) and (d) central Alps (N = 80; LME: LR = 0.33, p = 0.5666) experiments, (e) total lowland plant cover (cm2; N = 19; see Main text) and (f, g) alpine plant community composition (% change in NMDS axes; N = 18; linear models: NMDS #1: F1,14 = 1.19, p = 0.2934, NMDS #2: F1,14 = 3.17, p = 0.0966; region: F1,14 = 1.29, p = 0.2755). For (c, d), non-significant species-wise fit lines (yellow) are also displayed.

Appendix 1—figure 4
Download asset Open asset
Soil inoculation effects on alpine soil community composition.

Effects of low-elevation soil biota inoculation on (a) bacterial and (b) fungal community composition in alpine plots (N = 10 in both cases). Axes show the first two components of principal coordinates analyses (PCoA; Bray–Curtis distance) performed on the relative abundances of bacterial and fungal operational taxonomic units, respectively. p values are from PERMANOVAs testing for inoculation effects on the same distance matrices.

Data availability

All data contributing to the findings of this study have been deposited in the OSF under the DOI: https://doi.org/10.17605/OSF.IO/S54CH. The R code used to process and analyse the data are available in a Github repository with the URL: https://github.com/tom-n-walker/uphill-plants-soil-carbon (copy archived at swh:1:rev:a9e7a872d22a45dbd91bb00751d522812138eefd).

The following data sets were generated
    1. Walker TWN
    (2021) Open Science Framework
    Data: uphill plant migrations soil carbon loss.
    https://doi.org/10.17605/OSF.IO/S54CH

References

    1. Bjorkman AD
    2. Myers-Smith IH
    3. Elmendorf SC
    4. Normand S
    5. Rüger N
    6. Beck PSA
    7. Blach-Overgaard A
    8. Blok D
    9. Cornelissen JHC
    10. Forbes BC
    11. Georges D
    12. Goetz SJ
    13. Guay KC
    14. Henry GHR
    15. HilleRisLambers J
    16. Hollister RD
    17. Karger DN
    18. Kattge J
    19. Manning P
    20. Prevéy JS
    21. Rixen C
    22. Schaepman-Strub G
    23. Thomas HJD
    24. Vellend M
    25. Wilmking M
    26. Wipf S
    27. Carbognani M
    28. Hermanutz L
    29. Lévesque E
    30. Molau U
    31. Petraglia A
    32. Soudzilovskaia NA
    33. Spasojevic MJ
    34. Tomaselli M
    35. Vowles T
    36. Alatalo JM
    37. Alexander HD
    38. Anadon-Rosell A
    39. Angers-Blondin S
    40. Beest MT
    41. Berner L
    42. Björk RG
    43. Buchwal A
    44. Buras A
    45. Christie K
    46. Cooper EJ
    47. Dullinger S
    48. Elberling B
    49. Eskelinen A
    50. Frei ER
    51. Grau O
    52. Grogan P
    53. Hallinger M
    54. Harper KA
    55. Heijmans M
    56. Hudson J
    57. Hülber K
    58. Iturrate-Garcia M
    59. Iversen CM
    60. Jaroszynska F
    61. Johnstone JF
    62. Jørgensen RH
    63. Kaarlejärvi E
    64. Klady R
    65. Kuleza S
    66. Kulonen A
    67. Lamarque LJ
    68. Lantz T
    69. Little CJ
    70. Speed JDM
    71. Michelsen A
    72. Milbau A
    73. Nabe-Nielsen J
    74. Nielsen SS
    75. Ninot JM
    76. Oberbauer SF
    77. Olofsson J
    78. Onipchenko VG
    79. Rumpf SB
    80. Semenchuk P
    81. Shetti R
    82. Collier LS
    83. Street LE
    84. Suding KN
    85. Tape KD
    86. Trant A
    87. Treier UA
    88. Tremblay JP
    89. Tremblay M
    90. Venn S
    91. Weijers S
    92. Zamin T
    93. Boulanger-Lapointe N
    94. Gould WA
    95. Hik DS
    96. Hofgaard A
    97. Jónsdóttir IS
    98. Jorgenson J
    99. Klein J
    100. Magnusson B
    101. Tweedie C
    102. Wookey PA
    103. Bahn M
    104. Blonder B
    105. van Bodegom PM
    106. Bond-Lamberty B
    107. Campetella G
    108. Cerabolini BEL
    109. Chapin FS
    110. Cornwell WK
    111. Craine J
    112. Dainese M
    113. de Vries FT
    114. Díaz S
    115. Enquist BJ
    116. Green W
    117. Milla R
    118. Niinemets Ü
    119. Onoda Y
    120. Ordoñez JC
    121. Ozinga WA
    122. Penuelas J
    123. Poorter H
    124. Poschlod P
    125. Reich PB
    126. Sandel B
    127. Schamp B
    128. Sheremetev S
    129. Weiher E
    (2018) Plant functional trait change across a warming tundra biome
    Nature 562:57–62.
    https://doi.org/10.1038/s41586-018-0563-7
  8. Book
    1. Canadell JG
    2. Monteiro PMS
    3. Costa MH
    4. Cotrim da Cunha L
    5. Cox PM
    6. Eliseev S
    7. Henson S
    8. Ishii M
    9. Jaccard S
    10. Koven A
    11. Lohila A
    12. Patra PK
    13. Piao S
    14. Rogelj J
    15. Syampungani S
    16. Zaehle S
    17. Zickfeld K
    (2021) Global Carbon and other Biogeochemical Cycles and Feedbacks
    In: Masson-Delmotte V, Zhai P, Pirani A, Connors SL, Péan C, Berger S, Caud N, Chen Y, Goldfarb L, Gomis MI, Huang M, Leitzell K, Lonnoy E, Matthews JBR, Maycock TK, Waterfield T, Yelekçi O, Yu R, Zhou B, editors. Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press. pp. 673–816.
    https://doi.org/10.1017/9781009157896.007
    1. Kattge J
    2. Bönisch G
    3. Díaz S
    4. Lavorel S
    5. Prentice IC
    6. Leadley P
    7. Tautenhahn S
    8. Werner GDA
    9. Aakala T
    10. Abedi M
    11. Acosta ATR
    12. Adamidis GC
    13. Adamson K
    14. Aiba M
    15. Albert CH
    16. Alcántara JM
    17. Alcázar C C
    18. Aleixo I
    19. Ali H
    20. Amiaud B
    21. Ammer C
    22. Amoroso MM
    23. Anand M
    24. Anderson C
    25. Anten N
    26. Antos J
    27. Apgaua DMG
    28. Ashman T-L
    29. Asmara DH
    30. Asner GP
    31. Aspinwall M
    32. Atkin O
    33. Aubin I
    34. Baastrup-Spohr L
    35. Bahalkeh K
    36. Bahn M
    37. Baker T
    38. Baker WJ
    39. Bakker JP
    40. Baldocchi D
    41. Baltzer J
    42. Banerjee A
    43. Baranger A
    44. Barlow J
    45. Barneche DR
    46. Baruch Z
    47. Bastianelli D
    48. Battles J
    49. Bauerle W
    50. Bauters M
    51. Bazzato E
    52. Beckmann M
    53. Beeckman H
    54. Beierkuhnlein C
    55. Bekker R
    56. Belfry G
    57. Belluau M
    58. Beloiu M
    59. Benavides R
    60. Benomar L
    61. Berdugo-Lattke ML
    62. Berenguer E
    63. Bergamin R
    64. Bergmann J
    65. Bergmann Carlucci M
    66. Berner L
    67. Bernhardt-Römermann M
    68. Bigler C
    69. Bjorkman AD
    70. Blackman C
    71. Blanco C
    72. Blonder B
    73. Blumenthal D
    74. Bocanegra-González KT
    75. Boeckx P
    76. Bohlman S
    77. Böhning-Gaese K
    78. Boisvert-Marsh L
    79. Bond W
    80. Bond-Lamberty B
    81. Boom A
    82. Boonman CCF
    83. Bordin K
    84. Boughton EH
    85. Boukili V
    86. Bowman DMJS
    87. Bravo S
    88. Brendel MR
    89. Broadley MR
    90. Brown KA
    91. Bruelheide H
    92. Brumnich F
    93. Bruun HH
    94. Bruy D
    95. Buchanan SW
    96. Bucher SF
    97. Buchmann N
    98. Buitenwerf R
    99. Bunker DE
    100. Bürger J
    101. Burrascano S
    102. Burslem DFRP
    103. Butterfield BJ
    104. Byun C
    105. Marques M
    106. Scalon MC
    107. Caccianiga M
    108. Cadotte M
    109. Cailleret M
    110. Camac J
    111. Camarero JJ
    112. Campany C
    113. Campetella G
    114. Campos JA
    115. Cano-Arboleda L
    116. Canullo R
    117. Carbognani M
    118. Carvalho F
    119. Casanoves F
    120. Castagneyrol B
    121. Catford JA
    122. Cavender-Bares J
    123. Cerabolini BEL
    124. Cervellini M
    125. Chacón-Madrigal E
    126. Chapin K
    127. Chapin FS
    128. Chelli S
    129. Chen S-C
    130. Chen A
    131. Cherubini P
    132. Chianucci F
    133. Choat B
    134. Chung K-S
    135. Chytrý M
    136. Ciccarelli D
    137. Coll L
    138. Collins CG
    139. Conti L
    140. Coomes D
    141. Cornelissen JHC
    142. Cornwell WK
    143. Corona P
    144. Coyea M
    145. Craine J
    146. Craven D
    147. Cromsigt JPGM
    148. Csecserits A
    149. Cufar K
    150. Cuntz M
    151. da Silva AC
    152. Dahlin KM
    153. Dainese M
    154. Dalke I
    155. Dalle Fratte M
    156. Dang-Le AT
    157. Danihelka J
    158. Dannoura M
    159. Dawson S
    160. de Beer AJ
    161. De Frutos A
    162. De Long JR
    163. Dechant B
    164. Delagrange S
    165. Delpierre N
    166. Derroire G
    167. Dias AS
    168. Diaz-Toribio MH
    169. Dimitrakopoulos PG
    170. Dobrowolski M
    171. Doktor D
    172. Dřevojan P
    173. Dong N
    174. Dransfield J
    175. Dressler S
    176. Duarte L
    177. Ducouret E
    178. Dullinger S
    179. Durka W
    180. Duursma R
    181. Dymova O
    182. E-Vojtkó A
    183. Eckstein RL
    184. Ejtehadi H
    185. Elser J
    186. Emilio T
    187. Engemann K
    188. Erfanian MB
    189. Erfmeier A
    190. Esquivel-Muelbert A
    191. Esser G
    192. Estiarte M
    193. Domingues TF
    194. Fagan WF
    195. Fagúndez J
    196. Falster DS
    197. Fan Y
    198. Fang J
    199. Farris E
    200. Fazlioglu F
    201. Feng Y
    202. Fernandez-Mendez F
    203. Ferrara C
    204. Ferreira J
    205. Fidelis A
    206. Finegan B
    207. Firn J
    208. Flowers TJ
    209. Flynn DFB
    210. Fontana V
    211. Forey E
    212. Forgiarini C
    213. François L
    214. Frangipani M
    215. Frank D
    216. Frenette-Dussault C
    217. Freschet GT
    218. Fry EL
    219. Fyllas NM
    220. Mazzochini GG
    221. Gachet S
    222. Gallagher R
    223. Ganade G
    224. Ganga F
    225. García-Palacios P
    226. Gargaglione V
    227. Garnier E
    228. Garrido JL
    229. de Gasper AL
    230. Gea-Izquierdo G
    231. Gibson D
    232. Gillison AN
    233. Giroldo A
    234. Glasenhardt M-C
    235. Gleason S
    236. Gliesch M
    237. Goldberg E
    238. Göldel B
    239. Gonzalez-Akre E
    240. Gonzalez-Andujar JL
    241. González-Melo A
    242. González-Robles A
    243. Graae BJ
    244. Granda E
    245. Graves S
    246. Green WA
    247. Gregor T
    248. Gross N
    249. Guerin GR
    250. Günther A
    251. Gutiérrez AG
    252. Haddock L
    253. Haines A
    254. Hall J
    255. Hambuckers A
    256. Han W
    257. Harrison SP
    258. Hattingh W
    259. Hawes JE
    260. He T
    261. He P
    262. Heberling JM
    263. Helm A
    264. Hempel S
    265. Hentschel J
    266. Hérault B
    267. Hereş A-M
    268. Herz K
    269. Heuertz M
    270. Hickler T
    271. Hietz P
    272. Higuchi P
    273. Hipp AL
    274. Hirons A
    275. Hock M
    276. Hogan JA
    277. Holl K
    278. Honnay O
    279. Hornstein D
    280. Hou E
    281. Hough-Snee N
    282. Hovstad KA
    283. Ichie T
    284. Igić B
    285. Illa E
    286. Isaac M
    287. Ishihara M
    288. Ivanov L
    289. Ivanova L
    290. Iversen CM
    291. Izquierdo J
    292. Jackson RB
    293. Jackson B
    294. Jactel H
    295. Jagodzinski AM
    296. Jandt U
    297. Jansen S
    298. Jenkins T
    299. Jentsch A
    300. Jespersen JRP
    301. Jiang G-F
    302. Johansen JL
    303. Johnson D
    304. Jokela EJ
    305. Joly CA
    306. Jordan GJ
    307. Joseph GS
    308. Junaedi D
    309. Junker RR
    310. Justes E
    311. Kabzems R
    312. Kane J
    313. Kaplan Z
    314. Kattenborn T
    315. Kavelenova L
    316. Kearsley E
    317. Kempel A
    318. Kenzo T
    319. Kerkhoff A
    320. Khalil MI
    321. Kinlock NL
    322. Kissling WD
    323. Kitajima K
    324. Kitzberger T
    325. Kjøller R
    326. Klein T
    327. Kleyer M
    328. Klimešová J
    329. Klipel J
    330. Kloeppel B
    331. Klotz S
    332. Knops JMH
    333. Kohyama T
    334. Koike F
    335. Kollmann J
    336. Komac B
    337. Komatsu K
    338. König C
    339. Kraft NJB
    340. Kramer K
    341. Kreft H
    342. Kühn I
    343. Kumarathunge D
    344. Kuppler J
    345. Kurokawa H
    346. Kurosawa Y
    347. Kuyah S
    348. Laclau J-P
    349. Lafleur B
    350. Lallai E
    351. Lamb E
    352. Lamprecht A
    353. Larkin DJ
    354. Laughlin D
    355. Le Bagousse-Pinguet Y
    356. le Maire G
    357. le Roux PC
    358. le Roux E
    359. Lee T
    360. Lens F
    361. Lewis SL
    362. Lhotsky B
    363. Li Y
    364. Li X
    365. Lichstein JW
    366. Liebergesell M
    367. Lim JY
    368. Lin Y-S
    369. Linares JC
    370. Liu C
    371. Liu D
    372. Liu U
    373. Livingstone S
    374. Llusià J
    375. Lohbeck M
    376. López-García Á
    377. Lopez-Gonzalez G
    378. Lososová Z
    379. Louault F
    380. Lukács BA
    381. Lukeš P
    382. Luo Y
    383. Lussu M
    384. Ma S
    385. Maciel Rabelo Pereira C
    386. Mack M
    387. Maire V
    388. Mäkelä A
    389. Mäkinen H
    390. Malhado ACM
    391. Mallik A
    392. Manning P
    393. Manzoni S
    394. Marchetti Z
    395. Marchino L
    396. Marcilio-Silva V
    397. Marcon E
    398. Marignani M
    399. Markesteijn L
    400. Martin A
    401. Martínez-Garza C
    402. Martínez-Vilalta J
    403. Mašková T
    404. Mason K
    405. Mason N
    406. Massad TJ
    407. Masse J
    408. Mayrose I
    409. McCarthy J
    410. McCormack ML
    411. McCulloh K
    412. McFadden IR
    413. McGill BJ
    414. McPartland MY
    415. Medeiros JS
    416. Medlyn B
    417. Meerts P
    418. Mehrabi Z
    419. Meir P
    420. Melo FPL
    421. Mencuccini M
    422. Meredieu C
    423. Messier J
    424. Mészáros I
    425. Metsaranta J
    426. Michaletz ST
    427. Michelaki C
    428. Migalina S
    429. Milla R
    430. Miller JED
    431. Minden V
    432. Ming R
    433. Mokany K
    434. Moles AT
    435. Molnár A
    436. Molofsky J
    437. Molz M
    438. Montgomery RA
    439. Monty A
    440. Moravcová L
    441. Moreno-Martínez A
    442. Moretti M
    443. Mori AS
    444. Mori S
    445. Morris D
    446. Morrison J
    447. Mucina L
    448. Mueller S
    449. Muir CD
    450. Müller SC
    451. Munoz F
    452. Myers-Smith IH
    453. Myster RW
    454. Nagano M
    455. Naidu S
    456. Narayanan A
    457. Natesan B
    458. Negoita L
    459. Nelson AS
    460. Neuschulz EL
    461. Ni J
    462. Niedrist G
    463. Nieto J
    464. Niinemets Ü
    465. Nolan R
    466. Nottebrock H
    467. Nouvellon Y
    468. Novakovskiy A
    469. Nystuen KO
    470. O’Grady A
    471. O’Hara K
    472. O’Reilly-Nugent A
    473. Oakley S
    474. Oberhuber W
    475. Ohtsuka T
    476. Oliveira R
    477. Öllerer K
    478. Olson ME
    479. Onipchenko V
    480. Onoda Y
    481. Onstein RE
    482. Ordonez JC
    483. Osada N
    484. Ostonen I
    485. Ottaviani G
    486. Otto S
    487. Overbeck GE
    488. Ozinga WA
    489. Pahl AT
    490. Paine CET
    491. Pakeman RJ
    492. Papageorgiou AC
    493. Parfionova E
    494. Pärtel M
    495. Patacca M
    496. Paula S
    497. Paule J
    498. Pauli H
    499. Pausas JG
    500. Peco B
    501. Penuelas J
    502. Perea A
    503. Peri PL
    504. Petisco-Souza AC
    505. Petraglia A
    506. Petritan AM
    507. Phillips OL
    508. Pierce S
    509. Pillar VD
    510. Pisek J
    511. Pomogaybin A
    512. Poorter H
    513. Portsmuth A
    514. Poschlod P
    515. Potvin C
    516. Pounds D
    517. Powell AS
    518. Power SA
    519. Prinzing A
    520. Puglielli G
    521. Pyšek P
    522. Raevel V
    523. Rammig A
    524. Ransijn J
    525. Ray CA
    526. Reich PB
    527. Reichstein M
    528. Reid DEB
    529. Réjou-Méchain M
    530. de Dios VR
    531. Ribeiro S
    532. Richardson S
    533. Riibak K
    534. Rillig MC
    535. Riviera F
    536. Robert EMR
    537. Roberts S
    538. Robroek B
    539. Roddy A
    540. Rodrigues AV
    541. Rogers A
    542. Rollinson E
    543. Rolo V
    544. Römermann C
    545. Ronzhina D
    546. Roscher C
    547. Rosell JA
    548. Rosenfield MF
    549. Rossi C
    550. Roy DB
    551. Royer-Tardif S
    552. Rüger N
    553. Ruiz-Peinado R
    554. Rumpf SB
    555. Rusch GM
    556. Ryo M
    557. Sack L
    558. Saldaña A
    559. Salgado-Negret B
    560. Salguero-Gomez R
    561. Santa-Regina I
    562. Santacruz-García AC
    563. Santos J
    564. Sardans J
    565. Schamp B
    566. Scherer-Lorenzen M
    567. Schleuning M
    568. Schmid B
    569. Schmidt M
    570. Schmitt S
    571. Schneider JV
    572. Schowanek SD
    573. Schrader J
    574. Schrodt F
    575. Schuldt B
    576. Schurr F
    577. Selaya Garvizu G
    578. Semchenko M
    579. Seymour C
    580. Sfair JC
    581. Sharpe JM
    582. Sheppard CS
    583. Sheremetiev S
    584. Shiodera S
    585. Shipley B
    586. Shovon TA
    587. Siebenkäs A
    588. Sierra C
    589. Silva V
    590. Sitzia T
    591. Sjöman H
    592. Slot M
    593. Smith NG
    594. Sodhi D
    595. Soltis P
    596. Soltis D
    597. Somers B
    598. Sonnier G
    599. Sørensen MV
    600. Sosinski EE
    601. Soudzilovskaia NA
    602. Souza AF
    603. Spasojevic M
    604. Sperandii MG
    605. Stan AB
    606. Stegen J
    607. Steinbauer K
    608. Stephan JG
    609. Sterck F
    610. Stojanovic DB
    611. Strydom T
    612. Suarez ML
    613. Svenning J-C
    614. Svitková I
    615. Svitok M
    616. Svoboda M
    617. Swaine E
    618. Swenson N
    619. Tabarelli M
    620. Takagi K
    621. Tappeiner U
    622. Tarifa R
    623. Tauugourdeau S
    624. Tavsanoglu C
    625. Te Beest M
    626. Tedersoo L
    627. Thiffault N
    628. Thom D
    629. Thomas E
    630. Thompson K
    631. Thornton PE
    632. Thuiller W
    633. Tichý L
    634. Tissue D
    635. Tjoelker MG
    636. Tng DYP
    637. Tobias J
    638. Török P
    639. Tarin T
    640. Torres-Ruiz JM
    641. Tóthmérész B
    642. Treurnicht M
    643. Trivellone V
    644. Trolliet F
    645. Trotsiuk V
    646. Tsakalos JL
    647. Tsiripidis I
    648. Tysklind N
    649. Umehara T
    650. Usoltsev V
    651. Vadeboncoeur M
    652. Vaezi J
    653. Valladares F
    654. Vamosi J
    655. van Bodegom PM
    656. van Breugel M
    657. Van Cleemput E
    658. van de Weg M
    659. van der Merwe S
    660. van der Plas F
    661. van der Sande MT
    662. van Kleunen M
    663. Van Meerbeek K
    664. Vanderwel M
    665. Vanselow KA
    666. Vårhammar A
    667. Varone L
    668. Vasquez Valderrama MY
    669. Vassilev K
    670. Vellend M
    671. Veneklaas EJ
    672. Verbeeck H
    673. Verheyen K
    674. Vibrans A
    675. Vieira I
    676. Villacís J
    677. Violle C
    678. Vivek P
    679. Wagner K
    680. Waldram M
    681. Waldron A
    682. Walker AP
    683. Waller M
    684. Walther G
    685. Wang H
    686. Wang F
    687. Wang W
    688. Watkins H
    689. Watkins J
    690. Weber U
    691. Weedon JT
    692. Wei L
    693. Weigelt P
    694. Weiher E
    695. Wells AW
    696. Wellstein C
    697. Wenk E
    698. Westoby M
    699. Westwood A
    700. White PJ
    701. Whitten M
    702. Williams M
    703. Winkler DE
    704. Winter K
    705. Womack C
    706. Wright IJ
    707. Wright SJ
    708. Wright J
    709. Pinho BX
    710. Ximenes F
    711. Yamada T
    712. Yamaji K
    713. Yanai R
    714. Yankov N
    715. Yguel B
    716. Zanini KJ
    717. Zanne AE
    718. Zelený D
    719. Zhao Y-P
    720. Zheng J
    721. Zheng J
    722. Ziemińska K
    723. Zirbel CR
    724. Zizka G
    725. Zo-Bi IC
    726. Zotz G
    727. Wirth C
    728. Nutrient Network
    (2020) TRY plant trait database - enhanced coverage and open access
    Global Change Biology 26:119–188.
    https://doi.org/10.1111/gcb.14904
  33. Software
    1. Lenth R
    (2020)
    Estimated Marginal Means, aka Least-Squares Means
    R Project.
  39. Book
    1. Mercer C
    2. Boyer F
    3. Bonin A
    4. Coissac E
    (2013)
    SUMATRA and SUMACLUST: Fast and Exact Comparison and Clustering of Sequences
    LECA.
  43. Software
    1. Oksanen J
    (2019)
    vegan: Community Ecology Package
    R Project.
  49. Software
    1. Pinheiro J
    2. Bates D
    3. DebRoy S
    4. Sarkar D
    (2019)
    nlme: Linear and Nonlinear Mixed Effects Models
    R Project.
  51. Software
    1. R Development Core Team
    (2014) R: A language and environment for statistical computing
    R Foundation for Statistical Computing, Vienna, Austria.
    http://www.r-project.org
  71. Software
    1. Wilke C
    (2019)
    cowplot: Streamlined Plot Theme and Plot Annotations for “ggplot2"
    R Project.

Decision letter

  1. Bernhard Schmid
    Reviewing Editor; University of Zurich, Switzerland
  2. Meredith C Schuman
    Senior Editor; University of Zurich, Switzerland
  3. Bernhard Schmid
    Reviewer; University of Zurich, Switzerland
  4. Xin Jing
    Reviewer
  5. Biao Zhu
    Reviewer; Peking University, China

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting the paper "Lowland plant migrations into alpine ecosystems amplify soil carbon loss under climate warming" for consideration by eLife. Your article has been reviewed by 4 peer reviewers, including Bernhard Schmid as the Reviewing Editor and Reviewer #1, and the evaluation has been overseen by a Senior Editor. The following individual involved in review of your submission has agreed to reveal their identity: Xin Jing (Reviewer #4).

Comments to the Authors:

You will see that the reviewers found your study highly innovative and well conducted. However, they also raise substantial concerns that lead us to decline the manuscript in its current form. If you feel you could incorporate the suggestions of the reviewers in a new manuscript, we would be happy to consider a new submission.

Reviewer #1:

The authors of this study carried out two carefully designed field and a glasshouse experiment simulating effects of rapid warming on soil carbon loss. They did this by transplanting alpine turfs from their cold environment to lowland warm environment. They found that when lowland plants were inserted into alpine turfs under these lowland climatic conditions (referred to as warming treatment combined with warm-adapted plant introduction) they rapidly increased soil microbial decomposition of carbon stocks due to root exudates feeding the microbes.

The question is how well this experimental setup mimics what would happen if lowland plants would be inserted into alpine turfs in situ (which have already experienced considerable warming over the past decades), perhaps with an additional warming treatment there. A further question is if alpine plants inserted in turfs at alpine climatic conditions would have a similar effect as lowland plants inserted in turfs at lowland climatic conditions.

I suggest that the authors consider these questions when they draw conclusions about the results from their experiments. It would also be interesting to discuss the relevance of sudden strong warming effects relative to slower warming, potentially allowing ecosystems to adjust via changes in genetic composition of species (i.e. evolution) or species composition of communities (i.e. community assembly).

Overall, I was very impressed by this study regarding experimental designs and analysis. However, as I read the three reviews of colleagues and the interpretations of the results I found that the authors in my view draw general conclusions that are not justified with this confidence.

Thus, I find some of the phrasing misleading, in particular already the shortcut to use warming for transplanting alpine turfs to lowland conditions. Thus, one of the other reviewers first thought you had transplanted lowland plants to alpine turfs (which also have warmed already over the past decades). You end paragraphs very nicely with suggestions or implications, but I think more parsimonious interpretations would often relate to the inability of alpine plants to have the same effects as lowland plants under the lowland climatic conditions because of lack of adaptation. For example, on lines 111-113 you could write "alpine plants transplanted to lowland climatic conditions reduce growth, root exudation and thus microbial activity and soil carbon loss". Or, on lines 133-135: "resident alpine plants cannot induce soil carbon loss from soil transferred to environmental conditions of lowland plants." Or, lines 151-152: "that the home temperature environment is more important than the home soil environment."

Regarding the statistical analysis I have difficulties to believe the extremely high significances of effects for lowland vs. alpine plants in the glasshouse experiment given that there were only n=3 replicate species for the two groups. I do believe that significances would remain if species identity is added as random-effect to the LME error model, but currently I did not find out if this was the case.

Reviewer #2:

The authors were trying to test whether the migration of lowland plants into alpine ecosystems affects the warming impact on soil carbon. To achieve this goal, the authors first did two field experiments (moving intact turf from high-elevation to low-elevation to simulate warming) in the Alps, and then did a greenhouse pot study to explore the potential mechanisms for the results observed in the field experiments.

The main strengths of this work are the combination of a field experiment (conducted at two sites) and a greenhouse pot experiment (to explore the detailed mechanisms). Moreover, a number of techniques were used to measure plant traits, soil DOM and microbial properties (e.g. CUE, growth) which help to find the potential mechanisms.

The main weaknesses of this work are below:

1) The two field experiments are very short-term (<1 year), but the results were that warming and/or warming+lowland plants led to very high amount of soil C loss (up to ~40%, Figure 1). I was shocked to see these results as many field warming studies have shown undetectable change in SOC even after years or decades. The authors did not provide a good explanation for this rapid and large change in SOC.

2) The greenhouse experiment was used to explore the potential reasons for the amplified loss of soil C in the field experiment. However, a key result was based on incubation of disturbed soils (8 g) and a two-pool modeling of the respiration data from the short-term incubation. This may not provide a good estimate of the true turnover rate of SOC under different plant species (even in the greenhouse condition). If rhizosphere priming was the proposed mechanism (as hinted by the authors), a better approach (such as 13C labeling) is needed to measure microbial respiration from intact soils (with plant/root presence).

3) Some details of the sampling or measurement are very crucial and affect the results/interpretations. For example, in the field experiment, the soil core was only 1-cm diameter. Considering the spatial heterogeneity of soil carbon in field plots, this small volume may not well represent the true soil condition. Moreover, in the field plots, did soil bulk density change after planting of lowland plants or warming? This will affect the measured SOC concentration (mg/g) even the SOC stock (g/m2) did not change.

The authors provide a lot of data from two well-replicated experiments, which is applaudable. However, I have concerns on some of the key results and conclusions as explained above. I would like the authors to further elaborate these concerns.

L32: what does this 52% mean? If the baseline is low (e.g. 2% loss in SOC relative to control), a 52% increase in the warming effect (i.e. 3% loss in SOC relative to control) is not a big deal. I think the percent change relative to the unwarmed control really matters here.

Figure 1: I was shocked to see these SOC data. Other data of C pools (e.g. AGB, BGB, MBC) may help interpret this.

Reviewer #3:

The authors investigated the effect of warming and herbaceous plant migration on soil carbon (C) content using an ecosystem monolith transplant experiment along an elevation gradient in the Swiss Alp mountains. They observed, approximately 1 year after the transplant, that warming alone had little effect on soil carbon content (monoliths transplanted to a lower elevation with higher temperature remained unchanged in C content) but that the presence of lowland (warm-adapted) herbaceous plants in combination with warming had a negative effect on soil C content. The authors then conducted a glasshouse experiment and used a series of field and laboratory measurements to explore potential mechanisms explaining the observed changes in soil C content in the field. They concluded that soil C losses under lowland plant migration were likely mediated via increased microbial activity and CO2 release from soil C decomposition.

The research questions are extremely relevant to our understanding of the feedback between soil C dynamics and climate warming and remain an unexplored part of this debate. Moreover, both field and laboratory experimental designs are robust, with all the relevant and necessary validation checks needed for transplant experiments; the laboratory techniques employed to measure the range of microbial and plant variables potentially explaining soil C dynamics are adequate and modern; and the statistical analyses are appropriate. These elements make the present data set very relevant and valuable. The manuscript is also very well and clearly written.

However, I have two major concerns, casting doubt respectively on the main field results and on the proposed explanatory mechanisms.

First, at no point is bulk density mentioned and it does not appear to have been measured. This is critical because changes in soil C concentration (which was measured and reported here, in mg C g-1 soil) does not necessarily indicate an actual change in the quantity of C present in the soil (C stock, in unit mass C per unit soil volume, or per unit surface area to a constant depth) if this is accompanied by a change in bulk density: if less C per unit mass of soil (lower C concentration) is concurrent with more mass of soil in a constant volume (higher bulk density), this could mean that no change in C stocks actually occurs (or that even an increase occurs). In the present study, it is possible that the presence of lowland plants increased bulk density as compared to only alpine plants, compensating the lower C concentration and resulting in no change in C stocks. This is perhaps not likely, but it is too critical an issue not to be quantified (or at the very least discussed).

Second, even assuming that no changes in bulk density occurred and that indeed soil C stocks decreased under warming combined with lowland plant migration, the interpretation of the results are, in my view, at least incomplete. Certainly, the results do not support the claim that soil C losses were mediated via increased microbial decomposition of soil C with the certainty suggested by the authors. Generally speaking, I see three issues with the interpretation:

– Very schematically, increased microbial respiration and soil C losses from decomposition is only one of two equally likely pathways potentially explaining soil C losses (the other being decreased C inputs to the soil from the plant community). The possibility that decreased soil C content was simply mediated by decreased inputs of C to the soil is hardly explored at all in the study (there is a quick mention of it (L155)), but differences in plant biomass are interpreted only for their correlations with microbial activity (L160-166), not as a component of the C balance. Plant traits are measured and analysed but not in a way that can be used to test the hypothesis of changing C inputs. The presence of "more productive traits" (L141) for the lowland plants does not directly relate to differences in the quantity of C inputs to the soil, nor is it interpreted in relation to inputs. Even the interpretation of changes in ecosystem respiration seem to omit the possibility of changes in plant respiration (L208): "depressed microbial respiration per unit of soil was also evident at the ecosystem scale in that warming accelerated total ecosystem respiration but its effect was dampened in plots containing lowland plants". This statement was made despite no significant differences in microbial respiration per unit soil in the field data, and disregards the possibility that the dampened effect in plots with lowland plants could be due to lower plant respiration.

– For the glasshouse experiment, I agree that the results indicate that (L115); "lowland plants accelerated microbial activity by increasing the quantity of root exudates", but not that (L112): "these findings together imply that lowland plants accelerate alpine soil C loss" because stimulating microbial activity is not per se an indicator of soil C loss. It is now well-known that the activity of microbes is not only a motor for soil C losses, but also a key mechanism leading to transformation of C inputs from plants that leads to the subsequent stabilisation of C in the soil. This is actually clearly stated further down in the manuscript when interpreting the field microbial data (L190). Furthermore, there is no direct evidence that the pots with lowland plants were losing more C than those without. Therefore, results from the glasshouse experiment could be interpreted differently: a larger fast cycling pool of soil C constituted of recently photosynthetically fixed exudates associated with higher microbial activity could well be interpreted as an early indicator of more C stabilisation, particularly since the absorbance index seems to indicate more microbially derived product in the DOC. It would have been great to measure microbial biomass C over time (as well as CUE, and mass specific growth and respiration), to see if higher respiratory activity was associated with higher biomass. The lack of differences in microbial biomass between the plant community treatments at the end of the 6 weeks does not show that the quantity of microbial biomass produced over the whole incubation period remained constant. In a word, more respiration of a larger fast cycling pool is not an indicator of future soil C loss (in the presence of plants).

– The interpretation of the microbial variables measured in the field line up better with current conceptualisations of the role of microbes in C cycling (but overall interpretation still lacks consideration for plant C inputs). However, interpreting those data measured once 1 year after the transplant to explain the changes that happened gradually over this whole year is a risky and difficult exercise. How do we know that CUE, Rmass, Gmass etc… measured then represent what they were a day, a week, a month before? There is an attempt to deal with this timing issue by comparison with the glasshouse experiment, but only Cmic and Rmass can really be compared and it only very partially fills in the gap in time. Besides, the interpretation of this comparison can be questioned: in the glasshouse, Rmass was higher for the lowland plant pots (as compared to alpine plant at constant temperature) but actually remained constant between the comparable treatments W and WL in the field (Figure 2m). The results from the field, therefore, do not "support observations from the glasshouse experiment" in this context (L197) and neither do they "confirm (…) that this persists for at least one season" (L199). Finally, the thinking around the pulsed nature of C losses seems misplaced because there are no evidence that soil C losses had stopped after a year in the field (no measurements of soil C content are presented after that year).

In my view, it is critical that the authors address the bulk density issue. Ideally, bulk density data have been collected and the soil C content results (Figure 1) can be converted to C stocks. If bulk density data have not been collected, the effect of lowland plant migration at similar densities than in the transplant experiment on bulk density should be tested and confirmed to be neutral.

That being done, dC/dt = inputs – decomposition.

Nearly every aspect of the dataset should be re-interpreted in terms of a shifting balance between inputs and decomposition, rather than simply in terms of more or less microbial respiration. It could mean having to conduct additional analyses on the plant traits. To put it bluntly, it feels like the microbial data were interpreted as if collected from a soil incubation experiment with no plants (this is not totally fair of course because plant traits were measured, but not really interpreted in a way that is helpful to conclude on the input/decomposition balance).

In my view, the manuscript needs nearly a complete re-write, but I reiterate that this is a novel, valuable and properly (but perhaps not completely) analysed (statistically) data set.

Reviewer #4:

This manuscript took alpine grasslands as a model system and investigated whether lowland herbaceous plants contributed to the short-term dynamics of soil carbon under the context of climate warming. The authors find that warming individually does not render significant changes in alpine soil carbon, but corporately causes ~52% of carbon loss with lowland herbaceous plants in two short periods of field experiments. They further show that alpine soil carbon loss is likely mediated by lowland herbaceous plants through root exudation, soil microbial respiration, and CO2 release. This work adds in an interesting way to the ongoing debate on whether a positive climate feedback will be mediated by plant uphill range expansion in alpine grasslands, where climate warming may lead to a rapid loss of soil carbon.

The claims of this manuscript are well supported, but some aspects of background information in the studied alpine systems and field experiment design need to be clarified.

1) There is an extremely high level of carbon stored in the alpine soils (Figure 1). Climate warming will certainly lead to a great loss of soil carbon in the study systems that could contribute to the positive climate feedback. However, it is unclear for me how the effects of climate warming on soil carbon are relevant to the ongoing climate change in the studied alpine grasslands. It is therefore reasonable to provide more background information about ongoing climate change, and whether the simulated climate warming (i.e., 2.8 oC in central alps and 5.3 oC in western alps, Line 328-329) is realized as real-world climate change in the local systems. In addition, it seems that the manuscript aims to address a question that is of global concern, but my concern is about how the findings could be generalized to other regions.

2) I understand that the manuscript considers elevation as a natural gradient of climate change, which makes it possible to compare soil carbon dynamics in lowlands with alpine grasslands under climate warming. I also understand that the authors have done everything they can to control for the disturbances caused by transplanting that has been well justified by the supplementary data (e.g., Figure S6). However, it is unclear how the authors controlled for the influences of other factors given there are huge differences between lowlands and alpine grasslands, such as differences in wind, solar radiation, humidity, and the length of growing season.

3) It is generally known that different species respond to climate warming differently. Some species may be sensitive to climate warming and have traits aiding to dispersion that could expand their living ranges to some degree, while others may adjust themselves to adapt to climate warming and may not migrate to alpine systems. It is therefore cautious to assume that all the lowland species have the same dispersal ability. In other words, it is unclear how lowland plant species are selected for the field transplanting experiment (Line 284-290). Do all the lowland plant species selected have the potential to migrate to alpine systems?

4) The authors acknowledge that "we did not perform a reverse transplantation (that is, from low to high elevation), so we cannot entirely rule out the possibility that transplantation of any community to any new environment could yield a loss of soil carbon" (Line 318-320). When I read the title "lowland plant migrations into alpine grasslands …", I thought lowland plant species that were transplanted from low to high elevation. In fact, it is just the opposite to my thoughts. Without performing a reverse transplantation experiment, I am not sure the conclusion will stand that "lowland plant migrations into alpine grasslands amplify soil carbon loss under climate warming". In addition, it is unclear whether lowland plant effects stand alone or depend on climate warming based on the results in Figure 1 that lowland plant treatment is missing, and it is impossible to test the interactions between lowland plant and climate warming.

https://doi.org/10.7554/eLife.78555.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Reviewer #1:

The authors of this study carried out two carefully designed field and a glasshouse experiment simulating effects of rapid warming on soil carbon loss. They did this by transplanting alpine turfs from their cold environment to lowland warm environment. They found that when lowland plants were inserted into alpine turfs under these lowland climatic conditions (referred to as warming treatment combined with warm-adapted plant introduction) they rapidly increased soil microbial decomposition of carbon stocks due to root exudates feeding the microbes.

The question is how well this experimental setup mimics what would happen if lowland plants would be inserted into alpine turfs in situ (which have already experienced considerable warming over the past decades), perhaps with an additional warming treatment there.

The Reviewer alludes to two pertinent points here. The Reviewer’s first point considers whether lowland plants would function similarly (and, by extension, have the same effect on the soil system) if moved from the warmer lowland site to the cooler alpine site. This is a fascinating question in its own right, in that it raises questions about how migrations of non-adapted genotypes far beyond range edges (e.g. via human activity) impact recipient ecosystems. However, although we agree that alpine ecosystems have warmed considerably in recent decades, we cannot be confident that the high elevation sites in our study are already within the climate niche of the lowland focal species. As such, to address our research questions in situ at the high sites would have required additional warming treatments, which come with their own set of disadvantages (see our second point to this comment, below). We also refer the Reviewer to specific questions about adaptation below (see R6), although we see that we were not careful enough about the rationale for our design in the previous version of the manuscript. We have therefore added a clarifying sentence to the Main Text as follows:

L101: “In short, the experiments used here examined how the arrival of warm-adapted lowland plants influences alpine ecosystems in a warmed climate matching lowland site conditions (i.e. turf transplantation to low elevation plus lowland plant addition) relative to warming-only (i.e. turf transplantation to low elevation) or control (i.e. turf transplantation within high elevation) scenarios.”

Second, the Reviewer implicitly raises a point about whether our chosen approach of simulating warming plus lowland plant arrival (i.e. transplantation plus addition of lowland plants) is the most appropriate, specifically by suggesting an alternative option of adding lowland plants to (possibly experimentally-warmed) alpine turfs at the high elevation origin site. Here, it was essential to create a climate scenario in which lowland plants would survive and operate within their climatic niche (i.e. relative to their home conditions) once planted into alpine turfs, rather than perform suboptimally (e.g. be in a potentially inferior competitive position) or be unable to persist at all. The most parsimonious and reliable way to ensure this was to transplant alpine turfs to a site with a lowland temperature regime, with transplantations also being shown to outperform other methods when novel species interactions are involved (Yang et al. 2018). Most importantly, it was crucial to select a method that warmed the entire plant-soil system rather than only the air (e.g. open-top chambers, IR lamps; Marion et al. 1997; Aronson et al. 2009) or soil (e.g. heating cables; Hanson et al. 2017), and did so realistically throughout the year regardless of the weather (e.g. open-top chambers only work on sunny days in the summer; Marion et al. 1997) or a power supply (e.g. IR lamps, heating cables). Transplantation remains the only way to achieve this (Hannah 2022; Shaver et al. 2000). We now clarify our logic in the manuscript as follows:

L91: “Elevation-based transplant experiments are powerful tools for assessing climate warming effects on ecosystems because they expose plots to a real-world future temperature regime with natural diurnal and seasonal cycles while also warming both aboveground and belowground subsystems25–27. This is especially true if they include rigorous disturbance controls (here, see Methods) and are performed in multiple locations where the common change from high to low elevation is temperature (here, warming of 2.8 ºC in the central Alps and 5.3 ºC in the western Alps). While factors other than temperature can co-vary with elevation17, such factors either do not vary consistently with elevation among experiments (e.g. precipitation, wind), are not expected to strongly influence plant performance (e.g. UV radiation) or in any case form part of a realistic climate warming scenario (e.g. growing-season length, snow cover)17,27,28.”

A further question is if alpine plants inserted in turfs at alpine climatic conditions would have a similar effect as lowland plants inserted in turfs at lowland climatic conditions.

We interpret “turfs” to mean “lowland turfs” here, since we did insert lowland plants into alpine turfs under lowland climatic conditions (i.e. the WL treatment). We found that adding alpine plants to alpine turfs in alpine climatic conditions (i.e. planting disturbance control, see Methods) had no effect on alpine soil carbon content (L458). By extension, we would expect that adding lowland plants to lowland turfs in lowland climatic conditions would have no effect on lowland soil carbon content. While not explicitly tested, including this treatment would not change our finding that adding lowland plants to alpine turfs causes a reduction in soil carbon content relative to adding alpine plants to alpine turfs. Given this, we have left the text as is, but are happy to revisit this issue based on further discussion with the Reviewer/Editor.

I suggest that the authors consider these questions when they draw conclusions about the results from their experiments. It would also be interesting to discuss the relevance of sudden strong warming effects relative to slower warming, potentially allowing ecosystems to adjust via changes in genetic composition of species (i.e. evolution) or species composition of communities (i.e. community assembly).

Thank you for this excellent suggestion. We absolutely agree that anything short of a decadal experiment is unable to detect the role of longer-term evolutionary or community processes on soil carbon dynamics. While this doesn’t eliminate the need for experiments that consider shorter timescales, it is important to explicitly state this limitation. As suggested, we have added a sentence discussing this possibility in the concluding paragraph:

L387: “While our findings demonstrate that lowland plants affect the rate of soil carbon release in the short term, short-term experiments, such as ours, cannot resolve whether lowland plants will also affect the total amount of soil carbon lost in the long term. This includes whether processes such as genetic adaptation (in both alpine and lowland plants)12 or community change44,45 will moderate soil carbon responses to gradual or sustained warming.”

We also agree that it is extremely challenging to undertake warming experiments that do not initially “shock” the system through a sudden change in temperature. Having said this, alpine ecosystems are adapted to rapid within- and between-season temperature changes, making such shocks less relevant here.

Overall, I was very impressed by this study regarding experimental designs and analysis. However, as I read the three reviews of colleagues and the interpretations of the results I found that the authors in my view draw general conclusions that are not justified with this confidence.

Thus, I find some of the phrasing misleading, in particular already the shortcut to use warming for transplanting alpine turfs to lowland conditions. Thus, one of the other reviewers first thought you had transplanted lowland plants to alpine turfs (which also have warmed already over the past decades).

We agree with the Reviewer that there was a leap in logic in the original submission between transplantation and warming. We have resolved this in two ways. First, we now explicitly state the rationale behind using transplant experiments in the first paragraph of the Results:

L91: “Elevation-based transplant experiments are powerful tools for assessing climate warming effects on ecosystems because they expose plots to a real-world future temperature regime with natural diurnal and seasonal cycles while also warming both aboveground and belowground subsystems25–27. This is especially true if they include rigorous disturbance controls (here, see Methods) and are performed in multiple locations where the common change from high to low elevation is temperature (here, warming of 2.8 ºC in the central Alps and 5.3 ºC in the western Alps). While factors other than temperature can co-vary with elevation17, such factors either do not vary consistently with elevation among experiments (e.g. precipitation, wind), are not expected to strongly influence plant performance (e.g. UV radiation) or in any case form part of a realistic climate warming scenario (e.g. growing-season length, snow cover)17,27,28. In short, the experiments used here examined how the arrival of warm-adapted lowland plants influences alpine ecosystems in a warmed climate matching lowland site conditions (i.e. turf transplantation to low elevation plus lowland plant addition) relative to warming-only (i.e. turf transplantation to low elevation) or control (i.e. turf transplantation within high elevation) scenarios.”

Second, we have altered the language used to describe results throughout the manuscript to qualify that warming was simulated through transplantation. Specifically, we no longer use transplantation as a shortcut for warming, but describe the results in terms of the experimental manipulations rather than in terms of the scenarios they are intended to simulate. Correspondingly, our conclusions are now interpreted as implications of the experimental treatments. We now see the potential for our previous formulations to be misleading, and have therefore ensured that all conclusions are phrased in a way that is fully justified by the results. For instance:

L77: “Figure 1 | Warming and lowland plant effects on alpine soil carbon content in the field experiment. Mean ± SE soil carbon content (mg C g-1 dry mass; i.e. mass-based per-mil) in alpine turfs transplanted to low elevation (warming, W; light grey), transplanted plus planted with lowland plants (warming plus lowland plant arrival, WL; dark grey) or replanted at high elevation (control, C; white). Data are displayed for two experiments in the western (left) and central (right) Alps, with letters indicating treatment differences (LMEs; N = 58).”

L107: “We found that transplantation from high to low elevation (i.e. warming-only) caused a small decline in alpine soil carbon content, but in both regions this effect was stronger and only became statistically significant when lowland plants were also planted-in (Figure 1; Table S1; P < 0.001). Specifically, considering field data from both experiments together, the addition of lowland plants to transplanted alpine turfs increased warming-induced soil carbon loss from 4.1 ± 0.23 mg g-1 per ºC to 6.3 mg g-1 per ºC, which is a 52% ± 31% (mean ± 95% confidence intervals) increase relative to warming alone (Supplementary Figure S3; P = 0.001).”

L254: “Based on these observations, we hypothesise that, at least in our experiments, the establishment of warm-adapted lowland plants in warmed alpine ecosystems introduces novel traits into the community that alter plant community functioning and thus carbon cycle processes, with effects that intensify as lowland plants become more abundant in the community31,32. An alternative explanation is that, trait differences aside, the ability of alpine plants to facilitate soil carbon loss is suppressed when growing in warmer climates to which they are not adapted.”

L303: “… Effects of transplantation to low elevation (warming, W), transplantation plus lowland plants (warming plus lowlands, WL) or replantation at high elevation (control, C) on alpine ecosystem carbon dioxide fluxes (mean ± SE) during one season.”

L311: “… Effects of transplantation to low elevation (warming, W), transplantation plus lowland plants (warming plus lowlands, WL) or replantation at high elevation (control, C) on alpine soil pools and processes (means ± SEs) after one season.”

L325: “We found that biomass-specific rates of microbial respiration in the field were greater following transplantation to the warmer site in the presence, but not absence, of lowland plants (Figure 6a; Table S1; P = 0.038). By contrast, biomass-specific rates of microbial growth were slower following transplantation but remained unaffected by lowland plant presence (Figure 6b; Table S1; P = 0.016).”

L338: “However, in the field we found that lowland plants caused a reduction of microbial biomass carbon in alpine communities transplanted to the warmer site (Figure 6c; Table S1; P < 0.001).”

L350: “In short, we propose based on these relationships that the decrease in soil carbon content observed in warmer transplanted plots containing lowland plants had both started and stopped within one season.”

L408: “The experiments simulate effects of climate warming on alpine ecosystems via transplantation from high to low elevation, including the arrival of lowland plants due to warming induced uphill plant migrations (Supplementary Figure S1).”

L442: “Overall, the experimental design in both regions yielded three treatments (n = 10 for each) arranged in a block design: transplantation from high to low elevation (i.e. warming-only, W); transplantation plus low elevation plants (i.e. warming plus lowland plant establishment, WL); and replantation at the high elevation origin site (i.e. control, C).”

You end paragraphs very nicely with suggestions or implications, but I think more parsimonious interpretations would often relate to the inability of alpine plants to have the same effects as lowland plants under the lowland climatic conditions because of lack of adaptation. For example, on lines 111-113 you could write "alpine plants transplanted to lowland climatic conditions reduce growth, root exudation and thus microbial activity and soil carbon loss". Or, on lines 133-135: "resident alpine plants cannot induce soil carbon loss from soil transferred to environmental conditions of lowland plants." Or, lines 151-152: "that the home temperature environment is more important than the home soil environment."

We thank the Reviewer for this important and thought-provoking comment. On reflection, we agree that the experimental design precludes us from identifying whether lowland plants affect warmed alpine soils because of their traits, per se, or because their performance is conditional on the warmer low site conditions. While this distinction does not influence the main findings of the study (i.e. that the arrival of warm-adapted lowland plants in warming alpine ecosystems facilitates soil carbon loss), it is a subtle but still important qualifier that was missing from the original submission. We have rectified this in the revised manuscript in four ways.

First, we have clarified that the experimental design explicitly tests the scenario whereby lowland plants arrive in alpine ecosystems experiencing a warmer climate akin to their home conditions:

L101: “In short, the experiments used here examined how the arrival of warm-adapted lowland plants influences alpine ecosystems in a warmed climate matching lowland site conditions (i.e. turf transplantation to low elevation plus lowland plant addition) relative to warming-only (i.e. turf transplantation to low elevation) or control (i.e. turf transplantation within high elevation) scenarios.”

Second, we agree with the Reviewer that there are two directions with which to interpret the findings. On the one hand, lowland plants facilitate soil carbon loss under the warmed conditions to which they are adapted. On the other hand, alpine plants cannot induce soil carbon loss to the same extent as lowland plants under warmer conditions due to a lack of adaptation. Having said this, in our opinion alpine plants must remain the baseline reference point for interpretation, since the focus is on how alpine ecosystems change once warming occurs and lowland plants arrive. So, while we absolutely agree that adaptation of lowland plants to the lowland site may play an important role in the response observed, lowland plants still represent the “altered state”. We have thus decided against reversing the direction of comparison (i.e. that alpine plants are suppressed relative to lowland plants at lowland conditions), although we now present this issue as an alternative perspective for consideration. Specifically:

L254: “Based on these observations, we hypothesise that, at least in our experiments, the establishment of warm-adapted lowland plants in warmed alpine ecosystems introduces novel traits into the community that alter plant community functioning and thus carbon cycle processes, with effects that intensify as lowland plants become more abundant in the community31,32. An alternative explanation is that, trait differences aside, the ability of alpine plants to facilitate soil carbon loss is suppressed when growing in warmer climates to which they are not adapted. While either explanation could ultimately accelerate soil carbon loss from warming alpine ecosystems following the arrival of lowland plants, future studies are needed to fully unravel the environmental dependence of plant trait effects on soil ecosystem processes.”

Third, we have nuanced our language throughout the manuscript to make it clear that the effect of lowland plants on the alpine turfs is conditional on the warmer lowland site conditions. That is to say, we avoid implying that alpine and lowland plants have intrinsically different effects on the soil system (even though the functional trait literature would support such an interpretation), and rather emphasise that the differences we observe are conditional on the environmental conditions at the site (which is what our experimental design allows us to conclude). For instance:

L67: “We first used two whole-community transplant experiments in different alpine regions (western/central Swiss Alps) to establish the effects of warming plus lowland plant arrival on alpine soil carbon content.”

L89: “A portion of the alpine turfs at low elevation were planted with local lowland plant species to simulate the arrival of warm-adapted lowland plants in the warmed ecosystem, with the remaining portion being subjected to a planting disturbance control.”

L115: “Nevertheless, these findings show that once warm-adapted lowland plants establish in warming alpine communities, they facilitate warming effects on soil carbon loss on a per gram basis.”

Fourth, with reference to L111-L113 in the original manuscript, the findings in question relate to the glasshouse experiment, which was undertaken in standardised conditions without a warming treatment per se. We see that this was not clear in the Main Text of the original submission, so we have added a sentence to clarify this as follows:

L127: “The glasshouse experiment was not a direct simulation of field conditions, in that plants from lowland and alpine communities were grown separately at a constant temperature and humidity (Methods). Instead, the glasshouse experiment allowed us to isolate how plants adapted to lowland versus alpine climates differentially affect the alpine soil system.”

We have also nuanced the concluding sentence of this paragraph as follows:

L181: “These findings support the hypothesis that lowland plants have the capacity to increase soil carbon outputs relative to alpine plants by stimulating soil microbial respiration and associated CO2 release.”

With reference to the remaining examples given by the Reviewer, we have retained the original direction of comparison (see above) but added nuance to allow for adaptation to be responsible. Specifically:

L215: “While further directed studies are required to resolve whether root exudates are truly involved, our findings collectively suggest that lowland plants have the capacity to increase total root exudation into alpine soil relative to resident alpine plants.”

L256: “Based on these observations, we hypothesise that, at least in our experiments, the establishment of warm-adapted lowland plants in warmed alpine ecosystems introduces novel traits into the community that alter plant community functioning and thus carbon cycle processes, with effects that intensify as lowland plants become more abundant in the community31,32.”

Regarding the statistical analysis I have difficulties to believe the extremely high significances of effects for lowland vs. alpine plants in the glasshouse experiment given that there were only n=3 replicate species for the two groups. I do believe that significances would remain if species identity is added as random-effect to the LME error model, but currently I did not find out if this was the case.

We apologise for the confusion here. The glasshouse experimental design involved two planted treatments (not including the bare soil treatment) – one planted with lowland species and another planted with alpine species. We used six species from each alpine and lowland group, but every species was planted into every pot within its respective group. As such, there was no within treatment variation in species composition, with species composition only varying between alpine and lowland plant treatments. In other words, each of the 10 replicates for the lowland or alpine plant treatments were identical in terms of species composition – and varied only in their position in the greenhouse. We accounted for such a blocked design using linear mixed effects models. We have now clarified this in the revised manuscript:

L587: “Vegetated treatments contained two individuals of all lowland or alpine species planted in a circular pattern such they were never adjacent to conspecifics or plants of the same functional type (i.e. grass, herb, legume), making replicates within treatments identical in species composition and neighbourhoods.”

L718: “… we used the glasshouse experiment to test for lowland versus alpine plant treatment effects on soil microbial biomass carbon, the size (Cl) and decay rate (kl) of the first modelled soil carbon pool and all variables relating to soil DOM quality using LMEs including block as a random intercept term.”

L725: “Finally, we explored how plant physiology in lowland and alpine treatments at the pot level affected initial rates of soil microbial respiration.”

Reviewer #2:

The authors were trying to test whether the migration of lowland plants into alpine ecosystems affects the warming impact on soil carbon. To achieve this goal, the authors first did two field experiments (moving intact turf from high-elevation to low-elevation to simulate warming) in the Alps, and then did a greenhouse pot study to explore the potential mechanisms for the results observed in the field experiments.

The main strengths of this work are the combination of a field experiment (conducted at two sites) and a greenhouse pot experiment (to explore the detailed mechanisms). Moreover, a number of techniques were used to measure plant traits, soil DOM and microbial properties (e.g. CUE, growth) which help to find the potential mechanisms.

We thank the Reviewer for this positive comment.

The main weaknesses of this work are below:

1) The two field experiments are very short-term (<1 year), but the results were that warming and/or warming+lowland plants led to very high amount of soil C loss (up to ~40%, Figure 1). I was shocked to see these results as many field warming studies have shown undetectable change in SOC even after years or decades. The authors did not provide a good explanation for this rapid and large change in SOC.

We apologise for the confusion. We’re unsure where “up to ~40%” comes from here, so we have taken the Reviewer’s later suggestion of changing the annotation on Figure 1 to contrast C versus WL treatments (Western Alps = 25.6 ± 7.2 mg g-1; Central Alps = 25.3 ± 8.6 mg g-1) rather than W versus WL treatments.

With regards to the magnitude of soil carbon loss observed, we express soil carbon content in mg g-1 (i.e. mass-based per-mil), not cg g-1 (i.e. mass-based percent). This is so that we could use percent changes in the text to highlight the numeric magnitude of differences between treatments without confusing them with mass-based percent soil carbon – although we appreciate that this also caused confusion (see R15). To clarify, converting the above C versus WL treatment contrasts from mg g-1 to mass-based percent yields 2.56% ± 0.72% for the Western Alps experiment and 2.53% ± 0.86% for the Central Alps experiment. While it is striking that the WL treatments lost ~2.5% (~25 mg g-1) soil carbon in one year, such a loss is not extraordinary. To avoid future confusion, we have clarified the units in the Figure 1 caption as follows:

L77: “Mean ± SE soil carbon content (mg C g-1 dry mass; i.e. mass-based per-mil) in alpine turfs transplanted to low elevation (warming, W; light grey), transplanted plus planted with lowland plants (warming plus lowland plant arrival, WL; dark grey) or replanted at high elevation (control, C; white). Data are displayed for two experiments in the western (left) and central (right) Alps, with letters indicating treatment differences (LMEs; N = 58).”

2) The greenhouse experiment was used to explore the potential reasons for the amplified loss of soil C in the field experiment. However, a key result was based on incubation of disturbed soils (8 g) and a two-pool modeling of the respiration data from the short-term incubation. This may not provide a good estimate of the true turnover rate of SOC under different plant species (even in the greenhouse condition). If rhizosphere priming was the proposed mechanism (as hinted by the authors), a better approach (such as 13C labeling) is needed to measure microbial respiration from intact soils (with plant/root presence).

We agree with the Reviewer that using an approach such as 13C-labelling would have provided more direct evidence that lowland plants cause a rhizosphere priming effect. However, although some of our evidence comes from disturbed soils (i.e. microbial respiration), some (i.e. soil pore water) also comes from intact pots prior to harvest and we now also include another line of evidence from plant root biomass (see R18). In short, we draw on multiple lines of evidence suggesting that root exudates were involved, and note that Reviewer #3 thought our approach and interpretation on this aspect of the study was robust.

Having said this, we acknowledge that we were too confident in our interpretation here, so we have added caveats to the text as follows:

L207: “While not directly measured here, a nine-day decay period corresponds to the time expected for newly photosynthesised CO2 to be released through root exudation and respired by soil microbes23,24, suggesting that this carbon pool was mostly root exudates.”

L215: “While further directed studies are required to resolve whether root exudates are truly involved, our findings collectively suggest that lowland plants have the capacity to increase total root exudation into alpine soil relative to resident alpine plants.”

3) Some details of the sampling or measurement are very crucial and affect the results/interpretations. For example, in the field experiment, the soil core was only 1-cm diameter. Considering the spatial heterogeneity of soil carbon in field plots, this small volume may not well represent the true soil condition. Moreover, in the field plots, did soil bulk density change after planting of lowland plants or warming? This will affect the measured SOC concentration (mg/g) even the SOC stock (g/m2) did not change.

We agree with the Reviewer that taking a single soil core of 1 cm diameter in each plot would not have been robust. We did not do this. While we used 1 cm diameter cores to minimise disturbance, we took three cores per plot to account for within-plot heterogeneity and combined them into a composite sample. This is stated in the Methods as follows:

L523: “In each plot, we created a composite sample from three cores (ø = 1 cm, approx. d = 7 cm) no closer than 7 cm from a planted individual and from the same quarter of the plot used for ecosystem respiration measurements (see below; Supplementary Figure S1).”

We also agree that bulk density measurements were an important omission in the initial submission. We note that this point was fleshed out by Reviewer #3, below, so we refer the Reviewer to our response to that comment for further details.

The authors provide a lot of data from two well-replicated experiments, which is applaudable. However, I have concerns on some of the key results and conclusions as explained above. I would like the authors to further elaborate these concerns.

We hope that our responses to the Reviewer’s detailed comments address these concerns.

L32: what does this 52% mean? If the baseline is low (e.g. 2% loss in SOC relative to control), a 52% increase in the warming effect (i.e. 3% loss in SOC relative to control) is not a big deal. I think the percent change relative to the unwarmed control really matters here.

We refer the Reviewer to our response to their previous comment which we hope has resolved this issue.

Figure 1: I was shocked to see these SOC data. Other data of C pools (e.g. AGB, BGB, MBC) may help interpret this.

We apologise for this. We believe that this comment is an extension to the Reviewer’s previous comments about the magnitude of soil carbon lost from the system. Either way, in response to comments from Reviewer #3 we now present new data on ecosystem carbon dioxide fluxes (Figure 5) and draw further attention to field soil pools and processes (Figure 6), which we hope resolves this issue.

Reviewer #3:

The authors investigated the effect of warming and herbaceous plant migration on soil carbon (C) content using an ecosystem monolith transplant experiment along an elevation gradient in the Swiss Alp mountains. They observed, approximately 1 year after the transplant, that warming alone had little effect on soil carbon content (monoliths transplanted to a lower elevation with higher temperature remained unchanged in C content) but that the presence of lowland (warm-adapted) herbaceous plants in combination with warming had a negative effect on soil C content. The authors then conducted a glasshouse experiment and used a series of field and laboratory measurements to explore potential mechanisms explaining the observed changes in soil C content in the field. They concluded that soil C losses under lowland plant migration were likely mediated via increased microbial activity and CO2 release from soil C decomposition.

The research questions are extremely relevant to our understanding of the feedback between soil C dynamics and climate warming and remain an unexplored part of this debate. Moreover, both field and laboratory experimental designs are robust, with all the relevant and necessary validation checks needed for transplant experiments; the laboratory techniques employed to measure the range of microbial and plant variables potentially explaining soil C dynamics are adequate and modern; and the statistical analyses are appropriate. These elements make the present data set very relevant and valuable. The manuscript is also very well and clearly written.

We thank the Reviewer and are delighted that they think the study is extremely relevant, novel, experimentally robust, cutting-edge and valuable.

However, I have two major concerns, casting doubt respectively on the main field results and on the proposed explanatory mechanisms.

First, at no point is bulk density mentioned and it does not appear to have been measured. This is critical because changes in soil C concentration (which was measured and reported here, in mg C g-1 soil) does not necessarily indicate an actual change in the quantity of C present in the soil (C stock, in unit mass C per unit soil volume, or per unit surface area to a constant depth) if this is accompanied by a change in bulk density: if less C per unit mass of soil (lower C concentration) is concurrent with more mass of soil in a constant volume (higher bulk density), this could mean that no change in C stocks actually occurs (or that even an increase occurs). In the present study, it is possible that the presence of lowland plants increased bulk density as compared to only alpine plants, compensating the lower C concentration and resulting in no change in C stocks. This is perhaps not likely, but it is too critical an issue not to be quantified (or at the very least discussed).

This is an excellent point, and one also raised by Reviewer #2. To clarify, we initially decided against measuring bulk density because it is destructive and the experiments were still being used for other studies. Having said this, we agree with the Reviewer that more consideration of soil bulk density was needed, so we have rectified this in three ways. First, although the western Alps experiment has now been taken down, to address this comment we took new soil cores to measure bulk density in the central Alps experiment in 2021 to indirectly confirm that no changes occurred in the presence versus absence of lowland plants. They did not, and we now include these data in the Methods as follows:

L539: “It was not possible to take widespread measurements of soil bulk density due to the destructive sampling required while other studies were underway (e.g. ref 28). Instead, we took additional soil cores (ø = 5 cm, d = 5 cm) from the central Alps experiment in 2021 once other studies were complete to indirectly explore whether lowland plant effects on soil carbon content in warmed alpine plots could have occurred due to changes in soil bulk density. We found that although transplantation to the warmer site increased alpine soil bulk density (LR = 7.18, P = 0.028, Tukey: P < 0.05), lowland plants had no effect (Tukey: P = 0.999). It is not possible to make direct inferences about the soil carbon stock using measurements made on different soil cores four years apart. Nevertheless, these results make it unlikely that lowland plant effects on soil carbon content in warmed alpine plots occurred simply due to a change in soil bulk density.”

Second, in the Main Text we now caution readers against translating soil carbon content changes to soil carbon stock in absence of coupled measurements of soil bulk density as follows:

L113: “We caution against equating changes to soil carbon content with changes to soil carbon stock in the absence of coupled measurements of soil bulk density (Methods). Nevertheless, these findings show that once warm-adapted lowland plants establish in warming alpine communities, they facilitate warming effects on soil carbon loss on a per gram basis.”

Finally, we have altered the language throughout the manuscript (including the title) to make it clearer that we focussed on soil carbon content/concentration – not stock.

Second, even assuming that no changes in bulk density occurred and that indeed soil C stocks decreased under warming combined with lowland plant migration, the interpretation of the results are, in my view, at least incomplete. Certainly, the results do not support the claim that soil C losses were mediated via increased microbial decomposition of soil C with the certainty suggested by the authors. Generally speaking, I see three issues with the interpretation:

– Very schematically, increased microbial respiration and soil C losses from decomposition is only one of two equally likely pathways potentially explaining soil C losses (the other being decreased C inputs to the soil from the plant community). The possibility that decreased soil C content was simply mediated by decreased inputs of C to the soil is hardly explored at all in the study (there is a quick mention of it (L155)), but differences in plant biomass are interpreted only for their correlations with microbial activity (L160-166), not as a component of the C balance. Plant traits are measured and analysed but not in a way that can be used to test the hypothesis of changing C inputs. The presence of "more productive traits" (L141) for the lowland plants does not directly relate to differences in the quantity of C inputs to the soil, nor is it interpreted in relation to inputs. Even the interpretation of changes in ecosystem respiration seem to omit the possibility of changes in plant respiration (L208): "depressed microbial respiration per unit of soil was also evident at the ecosystem scale in that warming accelerated total ecosystem respiration but its effect was dampened in plots containing lowland plants". This statement was made despite no significant differences in microbial respiration per unit soil in the field data, and disregards the possibility that the dampened effect in plots with lowland plants could be due to lower plant respiration.

This is an excellent point. We have performed new analyses of the plant trait/biomass data from the field experiment, included additional measurements/analyses of NEE and GPP from the field experiments (originally omitted due to space, which was a mistake!) and have rewritten all relevant sections in the manuscript to change the focus to a shifting balance between soil carbon inputs and outputs. We refer the Reviewer particularly to paragraphs starting L121, L143, L171, L266, L279 and L320, but note that other sections of the manuscript have also been modified to account for these changes. Importantly, our original interpretation remains robust – i.e. that lowland plants most likely operate by accelerating soil carbon outputs, not decelerating soil carbon inputs – but we are careful to present our conclusions with an appropriate level of caution.

– For the glasshouse experiment, I agree that the results indicate that (L115); "lowland plants accelerated microbial activity by increasing the quantity of root exudates", but not that (L112): "these findings together imply that lowland plants accelerate alpine soil C loss" because stimulating microbial activity is not per se an indicator of soil C loss. It is now well-known that the activity of microbes is not only a motor for soil C losses, but also a key mechanism leading to transformation of C inputs from plants that leads to the subsequent stabilisation of C in the soil. This is actually clearly stated further down in the manuscript when interpreting the field microbial data (L190). Furthermore, there is no direct evidence that the pots with lowland plants were losing more C than those without. Therefore, results from the glasshouse experiment could be interpreted differently: a larger fast cycling pool of soil C constituted of recently photosynthetically fixed exudates associated with higher microbial activity could well be interpreted as an early indicator of more C stabilisation, particularly since the absorbance index seems to indicate more microbially derived product in the DOC. It would have been great to measure microbial biomass C over time (as well as CUE, and mass specific growth and respiration), to see if higher respiratory activity was associated with higher biomass. The lack of differences in microbial biomass between the plant community treatments at the end of the 6 weeks does not show that the quantity of microbial biomass produced over the whole incubation period remained constant. In a word, more respiration of a larger fast cycling pool is not an indicator of future soil C loss (in the presence of plants).

We thank the Reviewer for raising this important point. On reflection, we agree that the previous version of the manuscript did not give sufficient consideration to the possibility for increased microbial activity (and, indeed, respiration) in the glasshouse experiment to signal soil carbon accumulation via increased microbial growth. Having said this, all pots began with the same soil and microbial biomass remained unchanged between alpine and lowland plant treatments at the end of the six-week experiment. By extension, no net microbial growth occurred during this timeframe, making it unlikely that the accelerated respiration observed under lowland plants was indicative of soil carbon accumulation. Sadly, while we can deduce that intrinsic rates of respiration were higher, we can only speculate that growth remained unchanged (no new measurements can be done since growth measurements require fresh soil). We have rewritten the respective section in the manuscript in light of this and the Reviewer’s other comments, which includes the following caveat:

L181: “These findings support the hypothesis that lowland plants have the capacity to increase soil carbon outputs relative to alpine plants by stimulating soil microbial respiration and associated CO2 release. While accelerated microbial respiration can alternatively be a signal of soil carbon accumulation via greater microbial growth30, such a mechanism is unlikely to have been responsible here because it would have led to an increase in microbial biomass carbon under lowland plants, which we did not observe.”

– The interpretation of the microbial variables measured in the field line up better with current conceptualisations of the role of microbes in C cycling (but overall interpretation still lacks consideration for plant C inputs). However, interpreting those data measured once 1 year after the transplant to explain the changes that happened gradually over this whole year is a risky and difficult exercise. How do we know that CUE, Rmass, Gmass etc… measured then represent what they were a day, a week, a month before? There is an attempt to deal with this timing issue by comparison with the glasshouse experiment, but only Cmic and Rmass can really be compared and it only very partially fills in the gap in time. Besides, the interpretation of this comparison can be questioned: in the glasshouse, Rmass was higher for the lowland plant pots (as compared to alpine plant at constant temperature) but actually remained constant between the comparable treatments W and WL in the field (Figure 2m). The results from the field, therefore, do not "support observations from the glasshouse experiment" in this context (L197) and neither do they "confirm (…) that this persists for at least one season" (L199). Finally, the thinking around the pulsed nature of C losses seems misplaced because there are no evidence that soil C losses had stopped after a year in the field (no measurements of soil C content are presented after that year).

With regards to plant carbon inputs, we refer the Reviewer to their previous comment for corresponding revisions. With regards to specific comparisons between the glasshouse and field experiments, we have now deleted the sentences in question and have interpreted our results as follows:

L329: “Thus, despite lower rates of ecosystem respiration overall, alpine soil microbes still respired intrinsically faster in warmed plots containing lowland plants. Moreover, accelerated microbial respiration, but not growth, implies that alpine soils had a higher capacity to lose carbon under warming, but not to gain carbon via accumulation into microbial biomass, when lowland plants were present29,30. These findings align with observations from the glasshouse experiment that lowland plants generally accelerated intrinsic rates of microbial respiration (Figure 3), although in field conditions this effect occurred in tandem with warming.”

With regards to soil carbon loss being pulsed, while there is support for such a mechanism (e.g. see manuscript references 37-40), we agree that this is one of several hypotheses and with only two timepoints we were too confident about it in the original submission. We have now reshaped this section of the manuscript entirely to be more cautious about the temporal dynamics involved (L265 onwards). For instance, the section title now reads “Lowland plant-induced soil carbon loss is temporally dynamic”. Some other notable changes are:

L286: “Importantly, lowland plants had no significant bearing over net ecosystem exchange (Figure 5a), implying that although lowland plants were associated with soil carbon loss from warmed alpine plots (Figure 1), this must have occurred prior to carbon dioxide measurements being taken and was no longer actively occurring.”

L293: “By contrast, ecosystem respiration in warmed alpine plots was depressed in the presence versus absence of lowland plants (Figure 5c). These findings generally support the hypothesis that lowland plants affect the alpine soil system by changing carbon outputs. However, they contrast with expectations that lowland plants perpetually increase carbon outputs from the ecosystem and thus raise questions about how soil carbon was lost from warmed plots containing lowland plants (Figure 1).”

L320: “Carbon cycle processes are constrained by multiple feedbacks within the soil system, such as substrate availability36,37 and microbial acclimation38–40, that over time can slow, or even arrest, soil carbon loss37,39,40. We thus interrogated the state of the soil system in the field experiments in the western Alps experiment to explore whether such a feedback may be operating here, in particular to limit ecosystem respiration once soil carbon content had decreased in warmed alpine plots containing lowland plants.”

L354: “Taken together, one interpretation of our findings is that the establishment of lowland plants in warming alpine ecosystems accelerates intrinsic rates of microbial respiration (Figure 3, Figure 6a), leading to soil carbon release at baseline levels of microbial biomass (Figure 1, Figure 3c), a coupled decline in microbial biomass (Figure 6c) and a cessation of further carbon loss from the ecosystem (Figure 5a, Figure 6d).”

L358: “Although such a mechanism has been reported in other ecosystems37,40, applying it here is speculative without additional timepoints because field soil measurements came from a single sampling event after soil carbon had already been lost from the ecosystem. For instance, an alternative mechanism could be that soil microbes acclimate to the presence of lowland plants and this decelerated microbial processes over time38–40.”

L368: “Beyond the mechanism for lowland plant effects on alpine soil carbon loss, it is conceivable that soil carbon loss is not isolated to a single season, but will reoccur in the future even without further warming or lowland plant arrival. This is especially true in the western Alps experiment where warming yielded a net output of carbon dioxide from the ecosystem (Figure 5a). Moreover, in our field experiments we simulated a single event of lowland plant establishment and at relatively low abundance in the community (mean ± SE relative cover: 4.7% ± 0.7%), raising the possibility that increases in lowland plant cover or repeated establishment events in the future could facilitate further decreases in alpine soil carbon content under warming.”

In my view, it is critical that the authors address the bulk density issue. Ideally, bulk density data have been collected and the soil C content results (Figure 1) can be converted to C stocks. If bulk density data have not been collected, the effect of lowland plant migration at similar densities than in the transplant experiment on bulk density should be tested and confirmed to be neutral.

That being done, dC/dt = inputs – decomposition.

We refer the Reviewer to our response to their previous comment, which we hope has resolved this issue.

Nearly every aspect of the dataset should be re-interpreted in terms of a shifting balance between inputs and decomposition, rather than simply in terms of more or less microbial respiration. It could mean having to conduct additional analyses on the plant traits. To put it bluntly, it feels like the microbial data were interpreted as if collected from a soil incubation experiment with no plants (this is not totally fair of course because plant traits were measured, but not really interpreted in a way that is helpful to conclude on the input/decomposition balance).

We refer the Reviewer to our response to their previous comments, which we hope has resolved this issue.

In my view, the manuscript needs nearly a complete re-write, but I reiterate that this is a novel, valuable and properly (but perhaps not completely) analysed (statistically) data set.

We thank the Reviewer again for this positive comment. We have completely overhauled and re-written the manuscript in response to these and the other reviewer comments and are confident that all issues have been appropriately addressed.

Reviewer #4:

This manuscript took alpine grasslands as a model system and investigated whether lowland herbaceous plants contributed to the short-term dynamics of soil carbon under the context of climate warming. The authors find that warming individually does not render significant changes in alpine soil carbon, but corporately causes ~52% of carbon loss with lowland herbaceous plants in two short periods of field experiments. They further show that alpine soil carbon loss is likely mediated by lowland herbaceous plants through root exudation, soil microbial respiration, and CO2 release. This work adds in an interesting way to the ongoing debate on whether a positive climate feedback will be mediated by plant uphill range expansion in alpine grasslands, where climate warming may lead to a rapid loss of soil carbon.

The claims of this manuscript are well supported, but some aspects of background information in the studied alpine systems and field experiment design need to be clarified.

1) There is an extremely high level of carbon stored in the alpine soils (Figure 1). Climate warming will certainly lead to a great loss of soil carbon in the study systems that could contribute to the positive climate feedback. However, it is unclear for me how the effects of climate warming on soil carbon are relevant to the ongoing climate change in the studied alpine grasslands. It is therefore reasonable to provide more background information about ongoing climate change, and whether the simulated climate warming (i.e., 2.8 oC in central alps and 5.3 oC in western alps, Line 328-329) is realized as real-world climate change in the local systems. In addition, it seems that the manuscript aims to address a question that is of global concern, but my concern is about how the findings could be generalized to other regions.

We thank the Reviewer for pointing this out. With regards to the amount of soil carbon stored in the alpine soils, we refer the Reviewer to comments from Reviewer #2. With regards to the magnitude of warming expected in mountain regions, we agree with the Reviewer that the original submission lacked context. We have therefore added specific values as suggested:

L59: “They are experiencing both rapid temperature change (0.4 to 0.6 ºC per decade)18,19 and rapid species immigration20…”

With regards to how findings could be generalised to other regions or ecosystems, this is an important point that requires further research – and which we raise in the concluding paragraph. However, we see that we could have been more explicit about validating our findings in other mountain regions, so we have amended the sentence in question as follows:

L400: “Future work should focus on testing the conditions under which this feedback could occur in different mountain regions, as well as other ecosystems, experiencing influxes of range expanding plant species, on quantifying how deeply it occurs in shallow alpine soils, and on estimating the magnitude of the climate feedback given both ongoing warming and variation in rates of species range shifts.”

2) I understand that the manuscript considers elevation as a natural gradient of climate change, which makes it possible to compare soil carbon dynamics in lowlands with alpine grasslands under climate warming. I also understand that the authors have done everything they can to control for the disturbances caused by transplanting that has been well justified by the supplementary data (e.g., Figure S6). However, it is unclear how the authors controlled for the influences of other factors given there are huge differences between lowlands and alpine grasslands, such as differences in wind, solar radiation, humidity, and the length of growing season.

This is an excellent point. We note that Reviewer #1 also raised this point, so we refer the Reviewer to our response to that comment for further details.

3) It is generally known that different species respond to climate warming differently. Some species may be sensitive to climate warming and have traits aiding to dispersion that could expand their living ranges to some degree, while others may adjust themselves to adapt to climate warming and may not migrate to alpine systems. It is therefore cautious to assume that all the lowland species have the same dispersal ability. In other words, it is unclear how lowland plant species are selected for the field transplanting experiment (Line 284-290). Do all the lowland plant species selected have the potential to migrate to alpine systems?

This is an excellent question. In short, the specific dispersal abilities of lowland species used are currently unknown and will certainly vary. However, all are widespread and we assume have the capacity to migrate to higher elevations, given that horizontal distances between high and low elevation sites were in both cases less than 2 km. We now clarify this in the manuscript as follows:

L433: “While exact dispersal distances for selected lowland species are unknown, all species are widespread and are expected to migrate uphill under warming8 and the horizontal distance between high and low sites in the field experiments was always less than 2 km.”

4) The authors acknowledge that "we did not perform a reverse transplantation (that is, from low to high elevation), so we cannot entirely rule out the possibility that transplantation of any community to any new environment could yield a loss of soil carbon" (Line 318-320). When I read the title "lowland plant migrations into alpine grasslands …", I thought lowland plant species that were transplanted from low to high elevation. In fact, it is just the opposite to my thoughts. Without performing a reverse transplantation experiment, I am not sure the conclusion will stand that "lowland plant migrations into alpine grasslands amplify soil carbon loss under climate warming". In addition, it is unclear whether lowland plant effects stand alone or depend on climate warming based on the results in Figure 1 that lowland plant treatment is missing, and it is impossible to test the interactions between lowland plant and climate warming.

We apologise for the confusion. This comment echoes other comments from Reviewer #1 asking us to be more explicit about the treatments used when interpreting findings, to caveat the step in logic from transplantation to warming and to acknowledge throughout the manuscript that lowland plant effects were dependent on transplantation in the field experiment.

We therefore refer the Reviewer to our responses to those comments for details on how we resolved this. We have also modified the title and abstract to more accurately represent the experimental design, as follows:

Title: “Lowland plant arrival in alpine ecosystems facilitates soil carbon loss under experimental climate warming”

L30: “Here we used two whole-community transplant experiments and a follow-up glasshouse experiment to determine whether the establishment of herbaceous lowland plants in alpine ecosystems influences soil carbon content under warming. We found that warming (transplantation to low elevation) led to a negligible decrease in alpine soil carbon content, but its effects became significant and 52% ± 31% (mean ± 95% CIs) larger after lowland plants were introduced at low density into the ecosystem.”

With regards to testing the interaction between warming and lowland plants, while we acknowledge that not performing a fully-factorial design limited our ability to explicitly separate lowland plant versus warming effects on alpine soil, both are occurring simultaneously due to climate warming and we thus focussed effort on simulating such a scenario with greater experimental replication and at multiple locations. We note that Reviewers #1, #2 and #3 thought that this approach was robust. Importantly, the statistical analyses performed are valid for such an experimental design, and we have clarified and nuanced our interpretation throughout to avoid reaching beyond it.

https://doi.org/10.7554/eLife.78555.sa2

Article and author information

Author details

  1. Tom WN Walker

    Institute of Integrative Biology, ETH Zürich, Zürich, Switzerland
    Contribution
    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Validation, Visualization, Writing – original draft, Writing – review and editing
    For correspondence
    thomas.walker@unine.ch
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8095-6026

  2. Konstantin Gavazov

    1. Climate Impacts Research Centre, Department of Ecology and Environmental Sciences, Umeå Universitet, Abisko, Sweden
    2. Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Lausanne, Switzerland
    3. École Polytechnique Fédérale de Lausanne EPFL, School of Architecture, Civil and Environmental Engineering ENAC, Laboratory of Ecological Systems ECOS and Plant Ecology Research Laboratory PERL, Lausanne, Switzerland
    Contribution
    Data curation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-4479-7202

  3. Thomas Guillaume

    1. Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Lausanne, Switzerland
    2. École Polytechnique Fédérale de Lausanne EPFL, School of Architecture, Civil and Environmental Engineering ENAC, Laboratory of Ecological Systems ECOS and Plant Ecology Research Laboratory PERL, Lausanne, Switzerland
    3. Field-Crop Systems & Plant Nutrition, Nyon, Switzerland
    Contribution
    Data curation, Formal analysis, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  4. Thibault Lambert

    Faculty of Geosciences & the Environment, Université de Lausanne, Lausanne, Switzerland
    Contribution
    Data curation, Formal analysis, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  5. Pierre Mariotte

    1. Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Lausanne, Switzerland
    2. École Polytechnique Fédérale de Lausanne EPFL, School of Architecture, Civil and Environmental Engineering ENAC, Laboratory of Ecological Systems ECOS and Plant Ecology Research Laboratory PERL, Lausanne, Switzerland
    3. Grazing Systems, Agroscrope, Posieux, Switzerland
    Contribution
    Data curation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-8570-8742

  6. Devin Routh

    Institute of Integrative Biology, ETH Zürich, Zürich, Switzerland
    Contribution
    Data curation, Formal analysis
    Competing interests
    No competing interests declared

  7. Constant Signarbieux

    1. Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Lausanne, Switzerland
    2. École Polytechnique Fédérale de Lausanne EPFL, School of Architecture, Civil and Environmental Engineering ENAC, Laboratory of Ecological Systems ECOS and Plant Ecology Research Laboratory PERL, Lausanne, Switzerland
    Contribution
    Data curation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  8. Sebastián Block

    1. Institute of Integrative Biology, ETH Zürich, Zürich, Switzerland
    2. Department of Ecology & Evolutionary Biology, Princeton University, Princeton, United States
    Contribution
    Conceptualization, Data curation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  9. Tamara Münkemüller

    Uni. Grenoble Alpes, CNRS, Uni. Savoie Mont Blanc, LECA, Laboratoire d’Ecologie Alpine, Grenoble, France
    Contribution
    Conceptualization, Data curation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  10. Hanna Nomoto

    Institute of Integrative Biology, ETH Zürich, Zürich, Switzerland
    Contribution
    Data curation, Investigation, Writing – review and editing
    Competing interests
    No competing interests declared

  11. Thomas W Crowther

    Institute of Integrative Biology, ETH Zürich, Zürich, Switzerland
    Contribution
    Data curation, Formal analysis, Writing – review and editing
    Competing interests
    No competing interests declared

  12. Andreas Richter

    1. Centre of Microbiology & Environmental Systems, Division of Terrestrial Ecosystem Research, University of Vienna, Vienna, Austria
    2. International Institute for Applied Systems Analysis, Ecosystem Services and Management Program, Laxenburg, Austria
    Contribution
    Methodology, Project administration, Resources, Writing – review and editing
    Competing interests
    No competing interests declared

  13. Alexandre Buttler

    1. Swiss Federal Institute for Forest, Snow and Landscape Research (WSL), Lausanne, Switzerland
    2. École Polytechnique Fédérale de Lausanne EPFL, School of Architecture, Civil and Environmental Engineering ENAC, Laboratory of Ecological Systems ECOS and Plant Ecology Research Laboratory PERL, Lausanne, Switzerland
    Contribution
    Conceptualization, Project administration, Resources, Writing – review and editing
    Competing interests
    No competing interests declared

  14. Jake M Alexander

    Institute of Integrative Biology, ETH Zürich, Zürich, Switzerland
    Contribution
    Conceptualization, Funding acquisition, Investigation, Project administration, Resources, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared

Funding

European Union Horizon 2020 (678841)

  • Jake M Alexander

Swiss National Science Foundation (31003A-176044)

  • Jake M Alexander
  • Tom WN Walker

Swiss National Science Foundation (PZ00P2-174047)

  • Konstantin Gavazov

Swiss National Science Foundation (31003A-173210)

  • Sebastián Block

French National Research Agency (ANR-20-CE02-0021)

  • Tamara Münkemüller

The funders had no role in study design, data collection, and interpretation, or the decision to submit the work for publication.

Acknowledgements

JMA received funding from the European Union’s Horizon 2020 research and innovation program (grant no. 678841). JMA and TWNW received funding from the Swiss National Science Foundation (grant no. 31,003 A-176044). KG is funded by the Swiss National Science Foundation (grant no. PZ00P2-174047). SB and the central Alps experiment were funded by the Swiss National Science Foundation (grant no. 31,003 A-173210). TM received funding from the French National Research Agency (ANR-20-CE02-0021).

Senior Editor

  1. Meredith C Schuman, University of Zurich, Switzerland

Reviewing Editor

  1. Bernhard Schmid, University of Zurich, Switzerland

Reviewers

  1. Bernhard Schmid, University of Zurich, Switzerland
  2. Xin Jing
  3. Biao Zhu, Peking University, China

Publication history

  1. Preprint posted: July 22, 2021 (view preprint)
  2. Received: March 11, 2022
  3. Accepted: May 6, 2022
  4. Accepted Manuscript published: May 12, 2022 (version 1)
  5. Version of Record published: June 13, 2022 (version 2)

Copyright

© 2022, Walker et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 787
    Page views
  • 237
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Tom WN Walker
  2. Konstantin Gavazov
  3. Thomas Guillaume
  4. Thibault Lambert
  5. Pierre Mariotte
  6. Devin Routh
  7. Constant Signarbieux
  8. Sebastián Block
  9. Tamara Münkemüller
  10. Hanna Nomoto
  11. Thomas W Crowther
  12. Andreas Richter
  13. Alexandre Buttler
  14. Jake M Alexander
(2022)
Lowland plant arrival in alpine ecosystems facilitates a decrease in soil carbon content under experimental climate warming
eLife 11:e78555.
https://doi.org/10.7554/eLife.78555

Categories and tags

Further reading

    1. Ecology
    2. Evolutionary Biology

    Multiple preferred escape trajectories are explained by a geometric model incorporating prey’s turn and predator attack endpoint

    Yuuki Kawabata, Hideyuki Akada ... Paolo Domenici
    Research Article Updated

    The escape trajectory (ET) of prey – measured as the angle relative to the predator’s approach path – plays a major role in avoiding predation. Previous geometric models predict a single ET; however, many species show highly variable ETs with multiple preferred directions. Although such a high ET variability may confer unpredictability to avoid predation, the reasons why animals prefer specific multiple ETs remain unclear. Here, we constructed a novel geometric model that incorporates the time required for prey to turn and the predator’s position at the end of its attack. The optimal ET was determined by maximizing the time difference of arrival at the edge of the safety zone between the prey and predator. By fitting the model to the experimental data of fish Pagrus major, we show that the model can clearly explain the observed multiple preferred ETs. By changing the parameters of the same model within a realistic range, we were able to produce various patterns of ETs empirically observed in other species (e.g., insects and frogs): a single preferred ET and multiple preferred ETs at small (20–50°) and large (150–180°) angles from the predator. Our results open new avenues of investigation for understanding how animals choose their ETs from behavioral and neurosensory perspectives.

    1. Ecology
    2. Evolutionary Biology

    The deep-rooted origin of disulfide-rich spider venom toxins

    Naeem Yusuf Shaikh, Kartik Sunagar
    Research Article Updated

    Spider venoms are a complex concoction of enzymes, polyamines, inorganic salts, and disulfide-rich peptides (DRPs). Although DRPs are widely distributed and abundant, their bevolutionary origin has remained elusive. This knowledge gap stems from the extensive molecular divergence of DRPs and a lack of sequence and structural data from diverse lineages. By evaluating DRPs under a comprehensive phylogenetic, structural and evolutionary framework, we have not only identified 78 novel spider toxin superfamilies but also provided the first evidence for their common origin. We trace the origin of these toxin superfamilies to a primordial knot – which we name ‘Adi Shakti’, after the creator of the Universe according to Hindu mythology – 375 MYA in the common ancestor of Araneomorphae and Mygalomorphae. As the lineages under evaluation constitute nearly 60% of extant spiders, our findings provide fascinating insights into the early evolution and diversification of the spider venom arsenal. Reliance on a single molecular toxin scaffold by nearly all spiders is in complete contrast to most other venomous animals that have recruited into their venoms diverse toxins with independent origins. By comparatively evaluating the molecular evolutionary histories of araneomorph and mygalomorph spider venom toxins, we highlight their contrasting evolutionary diversification rates. Our results also suggest that venom deployment (e.g. prey capture or self-defense) influences evolutionary diversification of DRP toxin superfamilies.

    1. Ecology
    2. Epidemiology and Global Health

    Antibiotic Resistance: A mobile target

    Carolina Oliveira de Santana, Pieter Spealman, Gabriel G Perron
    Insight

    The global spread of antibiotic resistance could be due to a number of factors, and not just the overuse of antibiotics in agriculture and medicine as previously thought.