•Plant productivity is predicted to increase in northern latitudes as a result of climate warming; however, this may depend on whether biological nitrogen (N)-fixation also increases. We evaluated how the variation in temperature and light affects N-fixation by two boreal feather mosses, Pleurozium schreberi and Hylocomium splendens, which are the primary source of N-fixation in most boreal environments.
•We measured N-fixation rates 2 and 4 wk after exposure to a factorial combination of environments of normal, intermediate and high temperature (16.3, 22.0 and 30.3°C) and light (148.0, 295.7 and 517.3 μmol m−2 s−1).
•Our results showed that P. schreberi achieved higher N-fixation rates relative to H. splendens in response to warming treatments, but that the highest warming treatment eventually caused N-fixation to decline for both species. Light strongly interacted with warming treatments, having positive effects at low or intermediate temperatures and damaging effects at high temperatures.
•These results suggest that climate warming may increase N-fixation in boreal forests, but that increased shading by the forest canopy or the occurrence of extreme temperature events could limit increases. They also suggest that P. schreberi may become a larger source of N in boreal forests relative to H. splendens as climate warming progresses.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
An additional factor that may interact with temperature to control N-fixation by bryophytes is their exposure to different light intensities (Chapin et al., 1991; Liengen, 1999; Zielke et al., 2002). Biological N-fixation is a highly energy-demanding process and, as such, N-fixing organisms are often abundant in high-light environments (Vitousek et al., 2002; Houlton et al., 2008). Feather mosses, however, usually achieve their maximum biomass in relatively low-light environments beneath the canopy of boreal conifer species (Lindo & Gonzalez, 2010). Several studies have shown that feather moss N-fixation rates are negatively correlated with overstory productivity, suggesting that light availability may be one factor that limits their N-fixation rates (Zackrisson et al., 2004; Lagerström et al., 2007; DeLuca et al., 2008; Gundale et al., 2009, 2010). Although light may have a positive effect when temperatures are normal, increasing light intensity may have a negative effect on N-fixation rates when temperatures are extreme, by causing photodegradation, or by exacerbating heat or water stress to the feather moss or the cyanobiont (Hajek et al., 2009; Tobias & Niinemets, 2010). An understanding of the interactive effects of light and temperature on biological N-fixation is needed to predict and model how moderate or extreme climate change may influence N and C balances in a range of boreal environments and, further, to understand potential feedback effects of climate warming-induced changes in forest canopy shading on biological N inputs (Hungate et al., 2003; Sorensen & Michelsen, 2011).
In this study, we compared how N-fixation by two feather moss species, P. schreberi and H. splendens, responded to a factorial combination of three temperature and three light levels, with the levels representing a broad range of conditions experienced by boreal feather mosses during their growing season. In addition, we used a reciprocal transplant experiment to determine whether the N-fixation capacity of the two feather mosses was damaged or could rapidly recover following exposure to extreme environmental conditions. We tested the following hypotheses: (1) that N-fixation in the two feather mosses would be highly sensitive to temperature, with moderate temperature increases enhancing, and extreme temperature increases impairing, N-fixation rates; (2) that the effect of temperature on N-fixation rates would interact with light, with increasing light having a positive effect on N-fixation at normal and moderately elevated temperatures, and a negative effect at high temperatures; (3) that potential negative effects of extreme temperature and light on N-fixation would have damaging effects on the feather moss–cyanobacteria associations, meaning that negative effects would persist even after optimum conditions returned. Few data exist on how climate change may influence N-fixation in boreal forests; therefore, the testing of these three hypotheses will provide novel data that will greatly facilitate the modeling of net C and N balances of boreal ecosystems over long time-scales (Hungate et al., 2003; Turetsky, 2003).
Materials and Methods
We conducted an experiment to determine the effect of temperature and light on N-fixation in two moss species, P. schreberi (Bridel) and H. splendens (Hedwig). Mosses were collected in August 2009 from Reivo Nature Reserve located in the northern boreal zone, near Arvidsjaur, Sweden (65°46′20.51″N, 19°6′26.23″E). The site is a late successional forest composed of coniferous trees (Picea abies and Pinus sylvestris), and an understory consisting mainly of ericaceous shrubs (Vaccinium vitis-idaea, V. myrtillus and Empetrum hermaphroditum) and a ground layer of feather mosses (P. schreberi and H. splendens). Typical of most boreal forests, background rates of N deposition are very low (< 2 kg ha−1 yr−1) in this region (Gundale et al., 2011). The mean July air temperature during the periods 1961–1985 and 1986–2010, recorded at the nearest weather station (Arjeplog, Sweden), increased from 12.9 to 13.7°C, respectively, and the mean daily high temperatures in July during these periods increased from 17.3 to 18.2°C, respectively. During this 48-yr period, daily high temperatures were 25°C or higher on 131 d, and 29°C or higher on 10 d. The highest daily maximum temperature recorded during this 48-yr period was 31.5°C (Swedish Meteorological Institute, SMHI). We collected mosses from multiple locations within this forest site during a cloudless day, and noon ceptometer (Sunflec CEP40, Delta-T Devices Ltd, Cambridge, UK) readings at the moss layer from these collection sites showed that light intensity varied between 136 and 263 μmol m−2 s−1. Light intensity at the same time without any tree cover was 1331 μmol m−2 s−1. Moss shoots of both species were collected and placed in a total of 216 black plastic pots measuring 7 cm × 7 cm. Given that H. splendens has approximately double the surface area and mass per shoot relative to P. schreberi, this resulted in c. 100 shoots of H. splendens and c. 200 shoots of P. schreberi per pot.
After transporting mosses to the laboratory, we placed the pots in a climate chamber (Karl Weiss, Giesser, Germany) for 1 wk, set at 16°C and a light intensity of 100 μmol m−2 s−1. During this period, we randomized and assigned each pot to one of nine treatments, consisting of a factorial cross of three air temperature levels (16.0, 22.0 and 30°C) and three light levels (150, 300 and 500 μmol m−2 s−1; OSRAM powerstar HQI-TS 400 W/D lighting system, Munich, Germany) with 12 replicate pots per treatment (i.e. 2 species × 3 light levels × 3 temperature levels × 12 replicates = 216 pots). We note that the highest level of light (500 μmol m−2 s−1) is not considered to be extreme in most plant studies, but is high relative to our field light intensity measurements, and relative to those that commonly occur in the understory of boreal forests. We also note that our intermediate temperature treatment represents the maximum predicted mean temperature increase in boreal ecosystems during the next century (IPPC, 2007). In addition, the extreme temperature treatment (30°C) represents a temperature extreme that has been reached very infrequently at our study location, but temperatures of this magnitude occur more frequently at lower latitude and at more continental locations within the boreal biome. The three temperature levels were established using three separate climate chambers, each set to one of the three different air temperature levels, but with the same relative humidity (70%).
The three light levels were established by creating light shields within each climate chamber that reduced light levels to the desired intensity. The shields consisted of 40 cm × 60 cm wooden frames covered by white fiber shading cloth, with the different light intensity treatments created by varying the number of cloth layers, with more layers allowing less light to pass through the cloth. Three separate shields for each light intensity level (i.e. subshields) were created in each chamber (i.e. 3 light levels × 3 subshields per chamber × 3 chambers = 27 shields), meaning that each individual shield covered four pots of each species. Each shield was placed 30 cm above each set of pots to allow air circulation beneath the shields. Each climate chamber was set to have identical diurnal photoperiods, consisting of 19 h of daylight followed by 5 h of night, and a 4°C night-time reduction in temperature. All mosses were watered twice daily throughout the duration of the experiment, once in the morning and once in the afternoon. This watering regime fully hydrated all mosses, although drying between watering periods was more extensive with increasing temperature. Temperature at the moss surface was measured every 2 h during the experiment using data loggers (Thermochron iButton, DS1921G; Maxim, Sunnyvale, CA, USA) and daytime light intensity was measured every 4 d beneath each shield using a ceptometer. These measurements showed that the actual mean temperatures at the moss surface of the three temperature treatments were 16.3, 22.0 and 30.3°C (i.e. within 0.3°C of the air temperature), and the actual mean light intensities of the three light treatments were 148.0, 295.7 and 517.3 μmol m−2 s−1. We hereafter use these actual values to describe the treatments.
In order to ensure an equal influence of random effects to all treatments, we randomly repositioned pots beneath each subshield every day, and between each of the subshields for each light intensity level every 4 d, which ensured that each pot remained independent from one another. In order to avoid any confounding influence of individual climate chambers on the feather mosses, three times during the course of the experiment, we rotated pots among the three growth chambers, resetting the temperature in each chamber to correspond with the incoming set of pots.
We estimated N-fixation rates 1 d before the start of the experiment (time 0), and 2 and 4 wk after the treatments had started, using the acetylene reduction method (Schöllhorn & Burris, 1967). This method involves incubating moss samples for 24 h in a headspace of acetylene, which is converted to ethylene by the nitrogenase enzyme (Schöllhorn & Burris, 1967). The incubation was performed by removing 10 shoots of P. schreberi or five shoots of H. splendens (which have an approximately equal mass) from each pot, and placing each set of shoots in a 22-ml gas chromatograph (GC) vial, misting the shoots with deionized water, sealing the vials with a septa cap and replacing 10% of the headspace with acetylene using a syringe (Gundale et al., 2010). The addition of acetylene to each tube was staggered such that each tube could be analyzed on the GC exactly 24 h later, which is an adequately short period to prevent CO2 consumption by the mosses from limiting nitrogenase activity. After acetylene injection, each sample tube was placed next to the pot in which the moss shoots originated, so that acetylene reduction was influenced directly by the same light and temperature treatments. Samples were analyzed for total ethylene production on a Perkin-Elmer, Clarus 500 GC, with Turbomatrix 40 headspace injector (Waltham, Massachusetts, USA). The samples were injected isothermically onto a 30-m Perkin-Elmer Elite-Alumina capillary column (inside diameter, 0.053 mm). In addition to estimating ethylene concentrations in moss sample tubes, we also estimated ethylene concentrations in acetylene-only tubes and moss-only tubes (i.e. no acetylene injected) to ensure that ethylene measured in sample tubes was derived via the nitrogenase enzyme. Following analysis, all moss samples were dried at 65°C for 48 h and weighed. Analytical-grade ethylene standards were used to estimate the ethylene concentration in each sample tube. These estimates were then converted to μmol g−1 d−1 using the universal gas law and the moss dry mass. These values were further converted to μg N fixed per moss mass using a ratio of 3 mol of ethylene reduced per mol N and the molar mass of N. This ratio has been shown previously, using 15N techniques, to be appropriate for both P. schreberi and H. splendens (DeLuca et al., 2002; Lagerström et al., 2007).
Chlorophyll fluorescence measurements
At week four, we used measurements of chlorophyll fluorescence to study the effects of light and temperature on the efficiency of the photosynthetic light reaction (photochemistry), which is indicative of the physiological health of a plant in response to environmental conditions and stress (Maxwell & Johnson, 2000). For this analysis, we removed one moss shoot from each pot and combined the four shoots of each species occurring beneath each individual light shield, yielding three composite samples for each treatment (i.e. n = 3). Following collection, each sample was placed in a dark box (20°C) for 4 h to limit enzymatic activity and to dark adapt the plants. Fluorescence was measured using a laboratory-based fluorometer (IMAGING-PAM MINI-Version, Heinz Walz GmbH, Effeltrich, Germany). The fluorometer had a sufficiently large field to place the moss shoots next to each other in a single layer and to measure specific areas within the measurement field. We selected two standard polygonal areas for all measurements (one area near the tip of the shoot and one at 1 cm), and averaged the fluorometric readings of these two areas. A detailed explanation of the calculations of the fluorescence parameters is given in Krause & Weis (1991). We first measured the maximal potential quantum yield of photosystem (PS) II (Fv/Fm) using the saturation pulse method (Schreiber et al., 1995), whereby fluorescence is first measured in the dark (Fo) and then again following a pulse of saturated light (Fm). From this, Fv was determined as Fm – Fo (Krause & Weis, 1991). Values of Fv/Fm indicate the relative portion of absorbed light energy that can be used for photochemistry, with higher values signifying healthier plants and greater potential photosynthetic rates. The moss shoots were then illuminated with actinic light (76 μmol m−2 s−1) for 120 s and given another pulse of saturating light to determine the effective PS II quantum yield (φp), which represents the proportion of absorbed light that is actually used for photochemistry (i.e. photochemical quenching). Values of φp are typically lower than Fv/Fm because of the induction of nonphotochemical quenching. Both regulatory (NPQ) and nonregulatory (NO) nonphotochemical quenching were also measured at this time. Regulatory NPQ refers to the safe dissipation of excess light energy through the xanthophyll cycle (Demmig-Adams & Adams, 1996), whereas a high NO value indicates that light levels are in excess of those that can be dissipated through NPQ, thus leading to the photodamage of tissues. These measurements were not able to separate the influence of the feather moss and cyanobacteria PSs; however, microscopic images show that cyanobacterial colonies generally cover a very small portion of the feather moss surface (DeLuca et al., 2002; Gundale et al., 2011), suggesting that cyanobacteria have a minor influence on these measurements.
Reciprocal transplant experiment
Four weeks after the start of the experiment, we conducted a reciprocal transplant experiment to determine whether treatments had damaged the capacity of the feather moss associations to fix N, or whether N-fixation could rapidly adjust to the new light and temperature environments. This experiment utilized the four most extreme temperature (16.3 and 30.3°C) and light (148.0 and 517.3 μmol m−2 s−1) treatment combinations (low-temperature, low-light; low-temperature, high-light; high-temperature, low-light; high-temperature, high-light). For both of the moss species, three pots originating from each of these four environments were moved to each of the other extreme light and temperature environments, so that a full cross of temperature and light of the origin and destination environment was made (n = 3). The pots were placed in their new environments for 7 d, and biological N-fixation was measured using the method described above.
We evaluated the effect of light and temperature on N-fixation rates at times 0, 2 and 4 wk after the start of the experiment using a repeated-measures, four-factor-nested ANOVA with time, temperature, light and species as main factors, with species nested within light, light nested within temperature, and light, temperature and species nested within time. As this analysis revealed significant effects of time, we followed this analysis with three-factor ANOVAs at each time, with species and light nested within temperature as main fixed factors (n = 12). Each of the four fluorescence variables was also compared using a two-factor ANOVA with the same main factors and interactions as used for the N-fixation data, but with three replicates per treatment (n = 3). For the reciprocal transplant experiment, we used four-factor ANOVAs for each species separately, with temperature and light of the original and destination environments serving as fixed factors (n = 3). For all analyses, we used an α level of 0.05 as a significance threshold, evaluated data for assumptions of normality and homoscedasticity, and transformed data (by loge(X + 1)) as necessary to satisfy these assumptions. Post-hoc Tukey’s tests were performed to assess pair-wise differences when significant ANOVA effects were found. All data analyses were conducted with SPSS (version 19.0) statistical software (Armonk, New York, USA).
The initial repeated-measures ANOVA showed significant main effects of species, temperature, light and time, as well as numerous interactions between these variables (Supporting Information Table S1). This initial ANOVA was followed by individual ANOVAs at each time (Table 1), which showed that P. schreberi had 6.7%, 36.4% and 17.1% higher N-fixation rates than H. splendens at times 0, 2 and 4 wk, respectively (Fig. 1). This analysis also showed that no unintended initial differences between the light or temperature treatments were present before the start of the experiment (time 0), but that many significant main and interactive effects of temperature and light emerged during the course of the experiment (Table 1, Fig. 1). Two weeks after the start of the experiment, significant main effects of temperature, temperature by species interaction and temperature by light interaction were present (Table 1). Both species showed higher N-fixation rates in response to the intermediate warming treatment relative to the low-temperature treatment, whereas extreme warming also caused P. schreberi N-fixation rates (but not H. splendens) to increase relative to the low-temperature treatment, but not as high as in the intermediate temperature treatment (Fig. 1b,e). The significant interactive effect of light and temperature at week 2 was driven by a positive effect of light on N-fixation at low temperatures for P. schreberi (Fig. 1b), and a negative effect of increasing light on both species at the highest temperature (Fig. 1e).
Table 1. The results (F and P values) from three-factor ANOVAs (n = 12) evaluating the effects of species (Pleurozium schreberi and Hylocomium splendens), temperature (16.3, 22.0 and 30.3°C), light (148.0, 295.7 and 517.3 μmol m−2 s−1) and the interactive effects of these factors on biological nitrogen (N)-fixation rates (μg N g−1 moss mass d−1) at three separate times during a 4-wk incubation experiment
Significant effects (α = 0.05) are highlighted in bold.
Species × temperature
Species × light
Temperature × light
Species × temperature × light
At 4 wk, all of these same significant effects persisted, and an additional main effect of light was detected, as well as a significant light by temperature by species interaction (Table 1). These main and interactive light effects on N-fixation were driven by a positive effect of light at the lowest temperature for both species, a positive effect of light for P. schreberi at the intermediate temperature and a negative effect at the highest temperature for both species (Fig. 1c,f). For all four fluorescence parameters (Fv/Fm, φp, NPQ and NO), which were also measured at week 4, we found a significant effect of temperature, but no significant main light, species effects or any interactive effects (Table 2). These significant effects were driven by a significant decrease in the maximum potential and actual quantum yield (Fv/Fm and φp, respectively), a decrease in protective nonphotochemical quenching (NPQ), and an increase in damaging nonphotochemical quenching (NO) in response to the extreme warming treatment for both species (Fig. 2).
Table 2. The results (F and P values) from a three-factor ANOVA (n = 3) evaluating the effects of species (Pleurozium schreberi and Hylocomium splendens), temperature (16.3, 22.0 and 30.3°C), light (148.0, 295.7 and 517.3 μmol m−2 s−1) and the interactive effects of these factors on the potential quantum yield of photosystem II (Fv/Fm), actual photosystem II quantum yield (φp), regulatory quenching (NPQ) and nonregulatory quenching (NO) at the end of a 4-wk incubation experiment
Significant effects (α = 0.05) are highlighted in bold.
Species × temperature
Species × light
Temperature × light
Species × temperature × light
In the reciprocal transplant experiment, we found numerous significant and near-significant effects (P < 0.1) of the original and destination environments, and interactions between these factors (Table S2). For both moss species, original environment effects on N-fixation rates were similar to those found during the 4-wk measurement of the main experiment (Figs 1, 3). Effects of the destination environment included main effects of temperature for both species, of light for P. schreberi, but not for H. splendens, and of temperature by light interactive effects for both species (Table S2). In several cases, N-fixation rates in the two species were also affected by interactions between the original and destination environment. For P. schreberi, a four-way interaction between the original and destination temperature and light environments was found (Table S2). This occurred because there were always differences between the destination environments as a result of a negative temperature by light interaction (Fig. 3a–c), except when originating from the high-temperature, high-light environment, in which case N-fixation rates never recovered (Fig. 3d). For H. splendens, a three-way interaction between the original temperature and destination temperature and light environments occurred (Table S2). This occurred because significant differences always emerged between the destination environments as a result of the occurrence of either temperature or light effects, except when originating from either of the high-temperature environments, in which case N-fixation rates never recovered (Fig. 3g,h).
Temperature and light experiment
Our aim was to understand the interactive effects of temperature and light on biological N-fixation in two common boreal feather mosses, which is relevant for an understanding of how climate change may alter biological N inputs in boreal ecosystems. Our data showed that both temperature and light can individually and interactively have strong effects on N-fixation rates in both feather moss species, and that the two feather moss species differ in their responses to these factors (Fig. 1). For P. schreberi, N-fixation rates positively responded to intermediate warming (22.0°C) throughout the duration of the experiment, and temporarily increased in response to extreme warming (30.3°C) (Fig. 1b,c). By contrast, the only positive N-fixation response shown by H. splendens was a temporary increase (i.e. week 2) in the intermediate warming treatment (Fig. 1e,f). Long-term exposure (i.e. 4 wk) to the extreme warming treatment negatively affected the N-fixation rates of both species (Fig. 1c,f). These data support our first hypothesis, that the N-fixation rates of both species are sensitive to temperature, and, for the first time, show that N-fixation by H. splendens is less tolerant of climate warming relative to P. schreberi. One mechanism that could explain the different responses of the two feather mosses is their tolerance of desiccation. Although we watered mosses twice a day to minimize the effect of temperature-associated water loss, mosses growing at higher temperatures unavoidably experienced a higher degree of drying between watering events. Several studies have shown that moisture can strongly control N-fixation rates in feather mosses (Gundale et al., 2009; Jackson et al., 2011; Stewart et al., 2011a,b) and, further, that P. schreberi is more dominant in dry and exposed sites relative to H. splendens (Anderson et al., 1995; Zielke et al., 2002; Gundale et al., 2010). Pleurozium schreberi may be able to provide a more optimal environment for its cyanobionts under drier conditions, which, in turn, could explain its higher N-fixation rates in warmer environments relative to H. splendens. Research has also shown that H. splendens and P. schreberi associate with different cyanobiont communities (Zackrisson et al., 2009; Ininbergs et al., 2011), and that different cyanobionts have different temperature optima for N-fixation (Gentili et al., 2005). The higher N-fixation rates achieved by P. schreberi relative to H. splendens with warming may therefore be a result of physiological differences between the two feather mosses and between their cyanobionts.
In support of our second hypothesis, we found that variation in light showed strong interactive effects with temperature on N-fixation rates for both feather moss species (Fig. 1b,c,f). For P. schreberi, increasing light had a positive effect on N-fixation at both low and intermediate temperatures (16.3 and 22.0°C) at both sampling periods (Fig. 1b,c), whereas, for H. splendens, increasing light only had a positive effect during the fourth week, and only in the lowest temperature treatment (Fig. 1f). At both the middle and end of the experiment, both species were negatively affected by increasing light at high temperatures. These data suggest that P. schreberi is able to achieve higher N-fixation rates relative to H. splendens in more exposed environments, whereas both species are impaired by high light in the presence of extreme temperature conditions. Given that N-fixation is an energy-demanding process and therefore highly reliant on photosynthesis (Vitousek et al., 2002), we further investigated whether the response of photosynthetic light use (photochemistry) of the feather mosses to variations in temperature and light could explain the response of N-fixation. The data showed that photochemistry by the feather mosses was not affected by light intensity, but that extreme warming contributed to reduced potential and actual photosynthetic light use of both feather moss species (as indicated by low Fv/Fm and φp), diminished their leaf photoprotection capacity and ultimately increased potential damage to their PSs (as indicated by relatively low NPQ and high NO; Fig. 2). Although it is not yet known whether the photosynthate used for N-fixation is derived from the feather moss or the cyanobiont (DeLuca et al., 2002; Turetsky, 2003), these data demonstrate, for the first time, that N-fixation by the cyanobiont and photochemistry of the feather mosses respond independently to temperature and light stress, and suggest that N-fixation by the cyanobiont is not strongly coupled to feather moss photosynthesis. A plausible alternative explanation for the strong interactive effect of temperature and light on N-fixation by both species is that excess light may have resulted in inner leaf warming, causing the cyanobiont’s microenvironment within the feather moss leaves to warm above ambient temperatures (Niyogi, 1999; Ruban et al., 2007). Given that the nitrogenase enzyme reaches its maximum efficiency near 25°C (Vitousek et al., 2002; Houlton et al., 2008), this mechanism could explain why increasing light intensity had a positive effect on N-fixation at temperatures below 25°C, and a negative effect in the extreme warming treatment for both species.
These results provide valuable insights into how mean annual temperature increases may affect N-fixation in boreal forests. Global climate models predict 2–5°C warming in northern latitudes during the next century (IPPC, 2007). Our data suggest that, even the maximum projected increase in mean annual temperature (i.e. represented by our intermediate temperature treatment), is likely to have a direct positive effect on N-fixation in both feather moss species, particularly for P. schreberi. However, the positive effect of light on N-fixation under normal and moderately elevated temperatures suggests that changes in forest canopy cover may strongly influence N-fixation rates of the two feather moss species. Several models predict that net primary productivity in boreal forests will increase in response to climate warming (Keeling et al., 1996; Cramer et al., 2001), which is likely to correspond to higher stand-level leaf areas (Ise & Moorcroft, 2010), which would reduce light availability to the feather moss layer, and thereby reduce their N-fixation rates. Although vascular plant productivity in boreal forests is not thought to be directly dependent on biological N-fixation over short time-scales (Zackrisson et al., 2004; DeLuca et al., 2008), over longer time-scales reduced biological N inputs as a result of canopy shading may limit N availability to vascular plants, and thereby constrain vascular plant responses to climate warming (Hungate et al., 2003; Reich et al., 2006; Gerber et al., 2010). This interpretation is consistent with research in the sub-Arctic, showing that long-term experimental warming enhances vascular plant shading of bryophytes, thereby reducing their annual N-fixation inputs per unit area (Sorensen & Michelsen, 2011). Although this potential feedback requires further investigation in boreal forests, our data are the first to suggest that climate change-induced increases in tree productivity may constrain how N-fixation by boreal bryophytes responds to climate warming.
In support of our third hypothesis, the reciprocal transplant experiment showed that the N-fixation of P. schreberi originating from the high-temperature, high-light environment, and of H. splendens originating from either of the high-temperature environments (i.e. low and high light), did not recover after being placed for 1 wk in more hospitable environments (Fig. 3d,g,h); however, the N-fixation rates of mosses originating from any of the other environments were responsive to light or temperature variability in their new environments. These data suggest that the negative response of N-fixation to these extreme environments is not caused by direct kinetic control on the nitrogenase enzyme or by experimental artifacts of increased warming during the acetylene incubation, but, rather, that extreme temperatures cause cellular damage that reduces the capacity of the feather moss associations to fix N. It is well known that cyanobacteria can rapidly enter dormancy in response to environmental stress, which causes their N-fixation activity to markedly decline (Dodds et al., 1995). However, it is also well known that cyanobacteria can exit dormancy and repair their biochemistry within hours following the removal of stress (Dodds et al., 1995; Dadheech, 2010), which suggests that the persistent reduction in N-fixation rates shown by the two feather moss species after being removed from these stressful environments was probably a result of permanent cellular damage to either the feather moss or their cyanobionts (Brown, 1995; Guschina & Harwood, 2006). Climate models predict that extreme temperature events will become more intense and frequent in the next century in northern latitudes (Easterling et al., 2000; IPPC, 2007; Brown et al., 2008). Although further work is needed to understand the recovery rate of N-fixation following extreme climatic events, our data provide a first-time evaluation of the temperature and light thresholds at which the N-fixation capacity of the two feather mosses is likely to be damaged. The data also show that N-fixation by H. splendens is more sensitive than that by P. schreberi to extreme temperature events and, further, that the response of both species to warming is probably dependent on the light intensity in the local environments in which they occur.
Our experiments provide a first-time evaluation of the interactive effect of temperature and light on N-fixation by boreal feather moss associations, which is needed to effectively model boreal N and C balances in response to climate change (Hungate et al., 2003; Jain et al., 2009; Gerber et al., 2010; Zaehle et al., 2010; Esser et al., 2011). The data suggest that maximum predicted increases in mean annual temperatures (i.e. 5°C; IPPC, 2007) during the next century are likely to directly increase N-fixation rates by the two feather moss species, meaning that increased biological N inputs may satisfy the higher N requirements of a more productive vascular plant community in warmer climates. However, the data also show that N-fixation by the two feather moss species is highly sensitive to light availability, which could mean that climate-induced increases in vascular plant productivity will limit biological N-fixation because of greater shading, as has been shown in Arctic ecosystems (Sorensen & Michelsen, 2011), but never before in boreal ecosystems. Secondly, the data show that exposure to extreme temperature events, especially in environments with high light intensity, substantially damages the N-fixation capacity of the two feather moss species. This suggests that predicted increases in the frequency and intensity of extreme temperature events in the next century (Easterling et al., 2000; IPPC, 2007; Brown et al., 2008) could reduce biological N inputs in some boreal environments. Finally, our data show, for the first time, that the N-fixation rates of the two feather moss species differ in their response to climate warming, with P. schreberi achieving higher N-fixation rates than H. splendens in warmer environments. This suggests that P. schreberi may become a larger source of N-fixation relative to H. splendens as climate change progresses, and that changes in the relative abundance of the two species may be an important mechanism through which N-fixation responds to climate change.
We thank E. Ögren for use of the fluorometer. We thank K. Palmqvist, S. Bokhorst and S. Keel for helpful comments on an earlier draft of the manuscript. We also acknowledge funding from the Swedish Research Council Formas.