Introduction

Lichens are poikilohydric organisms that lack roots to absorb water and nutrients from their substrate. Thus, they are restricted to assimilating elements directly from atmospheric wet or dry deposition, or from stem flow (Nieboer et al., 1978). Deposition of nitrogen (N), in both oxidized (NO_x) and reduced (NH_y) forms, is increasing globally as a result of human activities (Galloway et al., 2008). For lichens, as for higher plants, N is an important nutrient that is involved in many processes for both the photobiont and the mycobiont. However, N can be a stressor if supplied in excess, and many studies have shown that N-sensitive lichens are replaced with more tolerant species in areas with high atmospheric N deposition (Söchting, 1995; Van Dobben & Ter Braak, 1998; Van Herk, 1999; Larsen Vilsholm et al., 2009). In addition to atmospheric N deposition, forest N fertilization represents a potential threat to lichens, as highlighted by the disappearance of the pendulous lichen Alectoria sarmentosa following 7–8 yr of fertilization in a Norway spruce forest (Hesselman, 1937).

It has been hypothesized that the sensitivity of lichens to high N deposition may depend on whether or not they can maintain a balanced carbon (C) to N stoichiometry between the symbiont partners when N is supplied in excess (Palmqvist, 2000; Palmqvist et al., 2008). In support of this hypothesis, the N-sensitive lichen Evernia prunastri had reduced mycobiont C-pools and mycobiont concentrations after 2 months of weekly N exposure in a study by Gaio-Oliveira et al. (2004). Other effects of increased N that might be related to an imbalance in resource investment include distorted coupling between thallus (hyphal) expansion and weight gain, as observed in N-fertilized Nephroma arcticum (Sundberg et al., 2001; Dahlman et al., 2002). In addition, the morphology and function of the lichen thallus might constrain how much N can be invested in algal cells in relation to fungal tissue. For example, the algal layer must be loosely packed to allow air circulation and reduce CO₂ diffusion resistance (Honegger, 1991), and avoid self-shading (Valladares et al., 1996). Tolerant lichens are clearly more capable of balancing responses to increases in N loads than sensitive lichens. For instance, in an experiment in which Platismatia glauca was fertilized daily for 3 months, a fivefold increase in thallus N was associated with a four- to fivefold increase in photobiont concentration, accompanied by increases in C-gain capacity, the soluble C pools and growth rate (Palmqvist & Dahlman, 2006). This is probably the reason why P. glauca has been found to be capable of surviving 15 yr of heavy forest fertilization (Dahlman et al., 2003).

Another factor that may contribute to the variations in sensitivity among lichens to increases in N availability is variation in their mechanisms for avoiding excessive uptake. Many lichens have evolved in N-poor ecosystems and thus have highly efficient N-uptake mechanisms, but may not have the ability to down-regulate them. In remote areas, the deposition consists of roughly equal proportions of NH₄⁺ and NO₃⁻. Lichens are able to take up both forms, although NH₄⁺ generally appears to be the preferred form (Dahlman et al., 2004; Palmqvist & Dahlman, 2006). Since NH₄⁺ may be more physiologically harmful than NO₃⁻ (Britto & Kronzucker, 2002), differences among species with respect to their N preference may also contribute to the differences in their N sensitivity. The difference in affinity between N forms has led to a considerable discussion, resulting in the idea that both the total quantity of deposited N and chemical form are important determinants for the ‘critical load’ (Bobbink et al., 2010) – defined as ‘a quantitative estimate of an exposure to one or more pollutants below which significant harmful effects on specified sensitive elements of the environment do not occur according to present knowledge’ (Nilsson & Grennfelt, 1998). A higher affinity for NH₄⁺ could thus explain why NH₄⁺ may often be more damaging to lichens than NO₃, as mentioned earlier (and hence its critical load lower)_. Empirical data suggest that the critical load for epiphytic lichens in boreal forests is c. 10–15 kg N ha⁻¹ yr⁻¹ (Bobbink et al., 2003), although lower critical loads (as low as 3 kg ha⁻¹ yr⁻¹) have been suggested (Fenn et al., 2008).

Although many studies have shown that lichen growth appears to be co-limited by N and light once they are wet and metabolically active (Crittenden et al., 1994; Sundberg et al., 2001; Palmqvist et al., 2008), recent studies have shown that lichens can also be limited by phosphorus (P) (Benner & Vitousek, 2007; McCune & Caldwell, 2009; Hogan et al., 2010a,b). These findings suggest that P or other elements may become limiting when N is supplied in excess and that this might constrain the ability of lichens to avoid toxic N concentrations through increased growth.

To date, there have been no experimental studies to our knowledge that have examined epiphytic lichen responses to N deposition under natural conditions over more than a few months. Earlier studies of lichen N responses have been either short-term (Loppi & Frati, 2006; Palmqvist & Dahlman, 2006) – sometimes involving unrealistically high N doses (Dahlman et al., 2003) or applications of N to the forest floor instead of directly to the lichens (Hesselman, 1937) – or restricted to studies of mat-forming terricolous lichens (Fremstad et al., 2005). Thus, the paucity of information prompted us to conduct a large-scale, long-term study of the effects of realistically simulating atmospheric N deposition on epiphytic lichens. Here we present the results from a whole-tree fertilization experiment, in which N deposition at loads ranging from 0.6 to 50 kg ha⁻¹ yr⁻¹ has been simulated by the daily spraying of the trees with low N-concentration solutions during the summer and autumn for 3 yr in an area with low background N deposition (2 kg N ha⁻¹ yr⁻¹; Forsum et al., 2006). The specific aim of this study was to follow the initial responses, over 3 yr, of the apparently less N-sensitive foliose lichen P. glauca, and the more N-sensitive pendulous lichen, A. sarmentosa, when exposed to an N-deposition gradient in their natural low-N environment (a boreal Norway spruce-dominated forest). We examined whether the lichens displayed different N assimilation patterns and subsequent increases in N concentration following the simulated N deposition. We also wanted to determine if changes could be related to different affinities for the two N sources, NH₄⁺ and NO₃⁻, that could support different critical loads for these two N forms. Previous studies have indicated that when N uptake by P. glauca increases, most of the increased supply is allocated to the photobiont, the concentration of which thus increases. We hypothesized that A. sarmentosa is less successful in increasing its photobiont concentration because of morphological constraints imposed by its growth form. Therefore, A. sarmentosa may be more sensitive to increased N loads than P. glauca.

Materials and Methods

Field site and treatment units

A relatively open (basal area 24 m² ha⁻¹) old growth forest stand, described in detail by Tiren (1937), dominated by Norway spruce (Picea abies) (10–15 m high and 15–30 cm diameter in breast height) at Kulbäcksliden (64°12′N, 19°33′E), was chosen for the experiment. The site (Fig. 1) is part of The Unit for Field-based Forest Research, The Swedish University of Agricultural Sciences (Vindeln, Sweden), and is equipped with electricity (220 V) and a 700-m-long water pipe to the nearest road from where water could be pumped. Sixteen trees taller than 8 m with a rich lichen flora were selected as the treatment units (Figs 1, 2a). Wooden towers (6.5 m high and 4 × 4 m square) made of untreated wood, placed on 1-m-thick concrete blocks partly submerged in the humic soil layer, were built around each tree (Fig. 2a). The towers were each equipped with a circular irrigation-fertilization tube with 10 small sprinklers on top of the construction (Fig. 2b). Each tube was connected to a 100-m-long pipe and a separate 1000 l treatment-solution tank. The small sprinklers were adjusted to gently spray an evenly dispersed mist towards the tree crown. Each tower also had two platforms 4 m from the ground to allow sampling of native lichens. One tree was randomly selected as a ‘dry’ control, which received no irrigation or fertilization, but was still surrounded by a wooden tower. Data from this tree will not be presented in this study. The remaining 15 trees were divided into three groups based on lichen abundance (species and biomass), followed by stratified randomization for five N-concentration treatments within each abundance group (Fig. 1). This created a block design with the trees acting as replicate units.

Irrigation-fertilization

Nitrogen was added in the form of NH₄NO₃ at five concentrations: 0.04, 0.41, 0.81, 1.63 and 3.2 mM. The lowest concentration was the same as that in local rainwater (Forsum et al., 2006); it was not expected to cause any treatment effect and was applied to trace N uptake under normal supply conditions. For tracing, every N solution was spiked with ¹⁵N in the form of ¹⁵NH₄¹⁴NO₃, ¹⁴NH₄¹⁵NO₃ (0.002 mM ¹⁵N) or ¹⁵NH₄¹⁵NO₃ (0.004 mM ¹⁵N) as detailed in Fig. 1. The fertilizer (17–18 l) was administered daily at 06:00 h (GSM + 1) in three 2 min pulses with 10 min intervals (using separate pumps and tubing for each tree) from 7 June to 27 September 2006, 15 June to 21 September 2007, and 19 June to 1 October 2008. The ‘growing season’ in this part of Sweden starts by the end of May or early June and continues up until late September or early October. All other months can have freezing temperatures and with sub-zero temperatures dominating from November to April. Our 4 month treatment period each year thus coincides with the active growth period for the forest vegetation. For practical reasons, the irrigation could not get started until the road conditions allowed for the heavy transport of water, meaning the first weeks of June, and we had to empty tanks, tubing and pumps before winter conditions set in. The N applications were controlled by an automated system consisting of a PLC205 D2-240 AutomationDirect DirectLOGIC Modular Programmable Logic Controller with Directsoft software programming version 2.0c from Koyo Electronics Industries Company Ltd, Tokyo, Japan. The treatment mimicked a mixture of dry and wet deposition via condensed fog, so the lichens were less wet following treatment than after rain (Fig. 3). All trees were treated until 1800 l of the solution had been dispersed each year, equivalent to deposition of 0.6, 6, 12.5, 25 and 50 kg N ha⁻¹ during the c. 4 month treatment period each year. In the first year (2006), the NH₄NO₃ was dissolved in artificial rainwater: 8.8 mg l⁻¹ K₂CO₃, 4.6 mg l⁻¹ Na₂CO₃, 5 mg l⁻¹ CaCO₃, 4.4 mg l⁻¹ NaH₂PO₄, 0.25 mg l⁻¹ Fe₂SO₄7H₂O, and 0.6 mg l⁻¹ MnSO₄H₂O (Tamm, 1953). However, analyses of the local water source revealed that it had similar concentrations of K, Na, Ca, Na, P, Fe, S and Mn, so these nutrients were omitted from the artificial rainwater in the subsequent 2 yr (2007 and 2008). The water originated from the Vindeln community source, which was transported to Kulbäcksliden in clean 1000 l tanks and pumped to 15 tanks (one for each tree) at the site twice each year (in early June and early August). At the site, all tanks were thoroughly rinsed before each treatment year and covered with several layers of black plastic and dark tarpaulins to avoid contamination from algae or cyanobacteria.

Chemical analyses

Chlorophyll pigments were quantified after extracting 10 (± 0.01) mg portions of homogenized lichen powder in 1.5 ml MgCO₃-saturated dimethyl sulfoxide (DMSO) at 60°C for 40 min (Palmqvist & Sundberg, 2001). The total N concentration and the isotopic ratio of ¹⁵N : ¹⁴N were analyzed by a certified laboratory (Colorado Plateau Stable Isotope Laboratory, CPSIL, Northern Arizona University, AZ, USA) using an elemental analyzer-isotopic ratio mass spectrometer (EA-IRMS). N uptake was calculated using the following equation:

(Eqn 1)

where At%_s is the atomic percentage of ¹⁵N in the sample, At%_c is the mean atomic percentage of ¹⁵N, total N_s is the total N concentration of the sample (g DW), and fraction ¹⁵N corrects for the amount of labeled N in the fertilizer, as defined earlier in the text.

The P concentration was also analyzed by CPSIL. The samples were extracted using a microwave digest in nitric acid under controlled temperature and pressure conditions, and analyzed using a Lachat Instruments QuikChem 8000 Series FIA+ equipped with a XYZ autosampler.

Results

The first summer (2006) was significantly drier than the following two summers, with respect to both the total precipitation and the number of rainy days (Fig. 4). In addition to the natural precipitation, which varied from 210 (2006) to 341 (2008) mm, the irrigation-fertilization corresponded to an additional 110 mm each year, although this was more similar to dew condensation or fog than to actual precipitation (Fig. 3). Monthly mean temperatures for July and August were highest in 2006 and lowest in 2008 (Fig. 4), with 2007 having close to ‘normal’ temperatures for these two months (based on data collected between 1961 and 1990; Swedish Meteorological and Hydrological Institute).

The thallus N concentration increased significantly in both lichen species with increasing N deposition (Fig. 5a,b), and the differences in thallus N concentration between the treatment (T) levels increased over time (Y) (there was a significant T × Y interaction; Table 1). The thallus N concentration varied between 5 and 6 mg g⁻¹ DW for both species after the first year at the lowest N deposition (Fig. 5a,b), while after the third year, the N concentration had increased to 10–12 mg g⁻¹ DW in A. sarmentosa and to c. 14 mg g⁻¹ DW in P. glauca at the highest N deposition (Fig. 5a,b). There were significant differences in N concentration between the control and both of the two highest N doses in all years for both species, except for A. sarmentosa in the third year where only the highest doses had a significant effect. By contrast, neither the 6 nor 12.5 kg doses induced significant differences in thallus N concentrations from the control in any of the three years.

The increase in thallus N concentration was attributed to the fertilizer application, and (accordingly) there were strong linear relationships between the N concentration and uptake of both NH₄⁺ and NO₃⁻, which were independent of year (A. sarmentosa, F = 354, P < 0.001; P. glauca, F = 610, P < 0.001, Fig. 6). There was no significant difference in the uptake of the two N forms for either species. The uptake of N from the fertilizer was low in the spruce needles, < 1 mg g⁻¹ DW in the current-year needles for all three years (data not shown) and there was no difference between the NH₄⁺ and NO₃⁻ uptake. The needle N concentration (mean values for each tree and year) varied between 6.3 and 9.8 mg g⁻¹.

The increase in thallus N concentration also resulted in an increased Chla concentration, which is indicative of an increased photobiont concentration in the thallus, and consistent with results of previous N fertilization experiments with lichens. When related to the accumulated N load over the 3 yr, there was a curvilinear relationship between N exposure and Chla concentration in both species (Fig. 7). In A. sarmentosa the thallus appears to have been saturated with photobiont cells at a Chla concentration of 1.5 mg g⁻¹ DW (Fig. 7a), with no further increase beyond an accumulated N load of 50 kg N ha⁻¹. In P. glauca, the Chla concentration increased to 2 mg g⁻¹ DW at the highest accumulated N load, that is, 100–150 kg N ha⁻¹ (Fig. 7b).

The P concentration was significantly decreased by N fertilization in A. sarmentosa (T × Y interaction), while in P. glauca the P concentration decreased over time independently of N treatment (Table 1, Fig. 5c,d).

The combined changes in N and P resulted in a significantly increased N : P ratio. for P. glauca there was a significant response already at the 12.5 kg N ha⁻¹ load during the second year, while for A. sarmentosa it was only significant at the 50 kg N ha⁻¹ load. The N : P ratio ranged from < 5 after the first year at the lowest N deposition to > 20 at the highest N deposition by the third year in P. glauca, and from < 8 to > 20 in A. sarmentosa (Fig. 5e,f). It should be noted that while the increase in N concentration slowed between the second and the third year, the N : P ratio continued to increase at the same rate throughout the 3 yr period.

Discussion

Many species of lichen have disappeared from areas that have been subject to high N loads for long periods of time (Hesselman, 1937; Söchting, 1995; Van Dobben & Ter Braak, 1998; Van Herk, 1999; Dahlman et al., 2003; Larsen Vilsholm et al., 2009). By contrast, some lichens have been reported to grow better following high doses of N fertilization in short-term studies (< 5 months) (Palmqvist & Dahlman, 2006). Nitrogen may therefore be either deleterious for the lichen symbiosis or promote growth, depending on the species and the conditions in the environment where the N is deposited. The mechanisms behind changes in species composition following N deposition thus need to be studied, while considering both of these aspects of lichen ecology. In this study we compared N uptake and allocation of the pendulous N-sensitive lichen A. sarmentosa and the foliose, less N-sensitive lichen P. glauca during 3 yr of simulated N deposition. The simulated N deposition increased the N concentration in both species, indicating that these epiphytic lichens, like previously studied terricolous lichens, lack mechanisms to down-regulate their N uptake with increasing N load (Hyvärinen & Crittenden, 1998). This is not unexpected since the majority of lichens have evolved in N-poor ecosystems (Palmqvist et al., 2002) and might therefore (in a similar way to plants of such habitats) be adapted to low-resource environments, be nutrient-conservative and take up high amounts of resources whenever available (Chapin, 1980). The thallus N concentration had increased in both species to c. 12–14 mg g⁻¹ DW after the third year at the highest N loads (Fig. 5), which for P. glauca was similar to the N concentration previously reported for this species after 15 yr of intensive fertilization (Dahlman et al., 2003). Both lichens further displayed a smaller increase in N concentration during the last year (2008), compared with the first two years (Fig. 5). This suggests that the lichens in our study may be able to adapt to the increased N supply either through uptake regulation, or through increased growth, as discussed later.

Previous studies have shown a preference for NH₄⁺ over NO₃⁻ uptake for a range of lichens (Dahlman et al., 2004; Palmqvist & Dahlman, 2006). This might, however, be related to low light (since there is a higher energy cost when assimilating NO₃⁻ than when assimilating NH₄⁺) and high N doses. At the ecologically relevant N concentrations and doses used in the present study there was no significant difference in uptake of the two N forms by either lichen species, regardless of thallus N concentration and accumulated N load (Fig. 6). This is also in accordance with the results of an Antarctic study showing equal uptake of both NH₄⁺ over NO₃⁻ for Usnea spacelata during natural conditions (Crittenden, 1998). Hence, our results indicate that in the case of P. glauca and A. sarmentosa there is no need to differentiate between NH₄⁺ and NO₃⁻ when determining critical N loads, as there was no support for species-specific differences in the physiological capacity of these lichens to take up different relative amounts of NH₄⁺ and NO₃⁻ that could influence their N sensitivity.

The increase in Chla with increased N load observed in both lichen species (Fig. 7) supports earlier studies both within and across species (Palmqvist et al., 2002, 2008). We have also found a good correlation between Chla and photobiont concentration for A. sarmentosa and P. glauca (O. Johansson et al., pers. comm.). Chla is also positively correlated with the photosynthetic capacity of lichens (Palmqvist et al., 2002, 2008), and can be used as an indirect measure of photobiont biomass (Raven et al., 1990). Hence, both photobiont biomass and net CO₂ gain may have increased in the two lichens in response to the increased N load. This is consistent with the hypothesis that tolerance to increased N may be coupled with the ability to maintain balanced carbon (C) to N stoichiometry between the symbiont partners when excess N is supplied (Dahlman et al., 2003). Provided that N is the most limiting resource, a balanced C to N stoichiometry would, according to this hypothesis, be maintained by the investment of the extra N into more photobiont cells, thereby increasing net CO₂ gain and potentially lowering the tissue N concentration through increased growth. However, the two species also displayed somewhat different responses with respect to Chla investment. In P. glauca, the Chla concentration increased linearly with increasing N load up to an accumulated N load of 150 kg N ha⁻¹, showing little sign of saturation at a Chla concentration of 1.8–2 mg g⁻¹ DW; still well below the 4.6 mg g⁻¹ DW found in thalli after 15 yr of N fertilization by Dahlman et al. (2003). The Chla concentration also initially increased with N load in A. sarmentosa, but was apparently already saturated at c. 1.5 mg g⁻¹ DW and at accumulated N loads above 50 kg N ha⁻¹ (Fig. 7). This apparent saturation for photobiont density in the thallus may be caused by continued thallus expansion or arrestation of photobiont division (Hill, 1993). The latter possibility is supported by the observations that the thallus N concentration was increased in both species at the 25 and 50 kg N ha⁻¹ yr⁻¹ loads, indicating that at these loads the N deposition and uptake were greater than the lichens could dilute through growth (Fig. 5). However, Chla saturation may also result from morphological and/or developmental constraints. In addition to allowing adequate gas diffusion and reducing self-shading (Honegger, 1991; Valladares et al., 1996), the thallus structure of the filamentous A. sarmentosa imposes an additional limitation compared with the foliose structure of P. glauca. In the former, the photobiont cells are situated inside the cortex towards the filament cavity, where the cells are normally discontinuous and clumped (Galloe, 1950; Hawksworth, 1972). Hence, the volume of the cavity surrounded by the fungal hyphae sets limits for algal expansion in the absence of accompanying fungal growth along the length axis.

There is also the possibility that other nutrients (e.g. P, magnesium or iron) may have become limiting for continued photobiont investments and/or the thallus growth. A likely candidate is P, concentrations of which decreased in A. sarmentosa during the third year of treatment (Fig. 5). Phosphorus limitation has previously been shown for epiphytic lichens in a Hawaiian rainforest, especially for N-fixing species (Benner & Vitousek, 2007; Benner et al., 2007), in Cladonia portentosa in heathlands and moorlands in the UK (Hogan et al., 2010a) and in L. pulmonaria, growth of which doubled following P additions in a study by McCune & Caldwell (2009).

Although coniferous trees may take up N via the canopy (Gaige et al., 2007), the spruce needles in this study displayed very low N uptake compared with the two lichen species. The low concentrations of ¹⁵N label found in needles may originate from canopy uptake or from soil N uptake by the trees. Independent of the route for tree N uptake, it highlights that tree-living lichens can be important in boreal forest N cycling by efficiently absorbing N deposited in the tree canopy and thereby altering precipitation chemistry.

In conclusion, our data indicate that there is a critical N load of 10–15 kg N ha⁻¹ yr⁻¹ as in Bobbink et al. (2003) for the epiphytic lichens also in our study, since the only observed significant response to N deposition of 12.5 kg N ha⁻¹ yr⁻¹ or less was an increased N : P ratio in P. glauca. However, the linear responses of chlorophyll concentrations to accumulated N load that we detected suggest that lower loads may have significant effects over the longer term. A mechanism that could allow lichens to tolerate increases in N concentration, within certain limits, is that the additional N may be invested in photobiont tissue, thereby increasing net C gain and growth. However, the substantial uptake of both NH₄⁺ and NO₃⁻ observed in lichens subjected to all treatments, and the significant increase in the N concentration following the two highest N-deposition treatments, indicate that neither of the two lichen species down-regulated their N uptake in response to the N load. Moreover, the photobiont concentration and the concomitant hyphal growth can only increase linearly with increasing N until limited by morphological constraints, water, light or other nutrients. The filamentous A. sarmentosa lichen reached Chla saturation more rapidly than the foliose P. glauca, providing a mechanistic reason for the sensitivity of the former species detected in earlier N fertilization studies.

Acknowledgements

We would like to thank the staff and researchers associated with The Unit for Field-based Forest Research, The Swedish University of Agricultural Sciences (Vindeln, Sweden), including Prof. Tomas Lundmark for ideas, technical and economic support; Tomas Hörnlund, Gunnar Karlsson and Jon Sandqvist for help with the construction of the towers; Jan Parsby, for help with setting up the automatic irrigation system; and Prof. Sune Linder for inspiration. Field assistance was provided by staff and PhD students from EMG: Ann Sehlstedt, Anna Cabrajic, Johan Asplund, Jonas Höglund, Olle Palmqvist, and Linus Jansson. This work was funded by grants to K.P. from FORMAS, Sweden (24.0795/97), and the Kempe Foundation, Sweden; to O.J. from the Björkman Foundation; and to A.N. from the Swedish Clean Air Program (SCARP).