Plant and Soil

, Volume 342, Issue 1, pp 141–148

Old trees contribute bio-available nitrogen through canopy bryophytes


    • Department of BiologyMcGill University
  • Jonathan A. Whiteley
    • Department of BiologyMcGill University
Regular Article

DOI: 10.1007/s11104-010-0678-6

Cite this article as:
Lindo, Z. & Whiteley, J.A. Plant Soil (2011) 342: 141. doi:10.1007/s11104-010-0678-6


Symbiotic cyanobacteria—bryophyte associations on the forest floor are shown to contribute significantly to stand-level nitrogen budgets through the process of biological nitrogen fixation (BNF), but few studies have considered the role of canopy bryophytes. Given the high biomass of epiphytic bryophytes in many tree species of the North American temperate rain forest, we suggest that canopy bryophytes may contribute substantially to stand-level N dynamics. We confirm the presence of cyanobacteria and measure rates of BNF at three heights (0, 15 and 30 m) in Sitka spruce trees across three watershed estuaries of Clayoquot Sound, British Columbia, Canada. This study is the first to report BNF by cyanobacteria associated with epiphytic and forest floor bryophytes in the coastal temperate rain forest of North America. Cyanobacteria density was significantly greater in epiphytic bryophytes compared to mosses on the forest floor, and rates of BNF were highest at 30 m in the canopy. The majority of total stand-level BNF (0.76 kg N · ha-1 · yr-1) occurs in the canopy, rather than on the forest floor (0.26 kg N · ha-1 · yr-1). We suggest that BNF by cyanobacterial-bryophyte associations in the canopy of coastal temperate rain forests is a unique source of ecosystem N, which is dependent on large, old trees with high epiphytic bryophyte biomass.


BryophyteBryosphereEpiphyticNitrogen fixationOld-growth forestsSymbiotic cyanobacteria


Observed declines in old-growth forest productivity and lower net primary productivity rates in older trees (Gower et al. 1996; Murty et al. 1996; Ryan et al. 1997) has led to a perception that old-growth forests are carbon-neutral, and no longer act as carbon sinks. Recent studies have challenged this perception (Carey et al. 2001; Luyssaert et al. 2008) demonstrating that old-growth forests can continue to accumulate biomass and sequester carbon beyond 800 years. In particular, the old-growth coniferous forests in the Pacific Northwest (PNW) of North America, which contain the tallest trees in the world (Lefsky 2010), and store more organic matter than any other forest type (Van Tuyl et al. 2005), exemplify the continued biomass and carbon accumulation (Luyssaert et al. 2008) that can occur in old-growth forests. Paradoxically, the productivity of most coniferous forests in the PNW is thought to be constrained by low nitrogen availability (Prescott et al. 1993, 2000; Twieg et al. 2009), as demonstrated by low nitrogen mineralization rates (Prescott et al. 1993, 2000), and increased growth rates under nitrogen fertilization (Blevins et al. 2006; Footen et al. 2009). Few explanations have been proposed for how old-growth forests of the PNW are able to maintain high productivity under such nutrient poor conditions.

Biological nitrogen fixation (BNF) performed by free-living and symbiotic bacteria is a major source of nitrogen (N) input to many natural systems (Cleveland et al. 1999; DeLuca et al. 2002; Turetsky 2003), and is the main pathway by which reactive N enters terrestrial ecosystems. Tree species hosting symbiotic associations with N-fixing bacteria are scarce in temperate zones (Houlton et al. 2008), but recently symbiotic cyanobacteria—bryophyte associations on the forest floor of boreal forest systems are shown to contribute significantly to whole forest-level and even global nitrogen budgets (DeLuca et al. 2002). While it is well established that the growth of many temperate and boreal forests is limited by N availability (Vitousek et al. 2002), it is suggested that BNF may provide a mechanism for sustaining long-term forest productivity by continuously supplementing available N at the stand-level.

Previous studies have shown that epiphytic cyanolichens (e.g. Lobaria oregana (Tuck.) Müll. Arg.) provide an important source of BNF (Pike 1978; Antoine 2004), but few studies have considered epiphytic bryophytes (but see Menge and Hedin 2009). Epiphytic bryophytes are a dominant form of biomass in PNW forests (McCune 1993; McCune et al. 1997). For example, Ellyson and Sillett (2003) observed an average biomass of 36.2 kg dry weight (dwt) of bryophytes in old-growth Sitka spruce trees in Washington State, compared to 10 kg dwt of lichen in the same trees. The presence of N-fixing cyanobacteria and rates of BNF within epiphytic bryophytes of PNW forests has previously not been confirmed nor measured. N-fixing cyanobacteria associated with bryophytes have provided missing values in boreal forest N budgets (DeLuca et al. 2002), and reveal a potential mechanism for maintaining forest productivity under seemingly N limited conditions. Given the high biomass of epiphytic bryophytes in many tree species of the temperate rain forest, we suggest that canopy bryophytes (the arboreal bryosphere sensu Lindo and Gonzalez 2010) may contribute substantially to stand-level N dynamics.

Materials and methods

We studied epiphytic bryophytes in the high canopy of a coastal temperate rain forest within three pristine watershed estuaries in the Clayoquot Sound UNESCO Biosphere Reserve on the west coast of Vancouver Island, British Columbia, Canada (49º2720′′N, 126º0194′′W). We selected 18 Sitka spruce (Picea sitchensis (Bong) Carr.) trees which were over 500 years old and had an average diameter at breast height of 1.52 m (± 0.3 SD) and 52.5 m (± 9.6 SD) average tree height. Sitka spruce is a co-dominant tree species in much of the coastal temperate rain forest of north-western North America, with average epiphytic bryophyte biomass estimated at 36.2 kg (dry weight) per tree (Ellyson and Sillett 2003).

Approximately 1 g dwt moss was collected into 50 ml optically clear polystyrene vials from the forest floor near the base of each tree (0 m) and from branches at 15 m, and 30 m, using single rope climbing techniques to access the canopy of six trees within each watershed. The dominant moss species in the canopy was Antitrichia curtipendula (Hedw.) Brid.; Rhytidiadelphus loreus (Hedw.) Warnst. was sampled as a comparative morphological form on the forest floor. Isothecium myosuroides Brid. was the subdominant species in the canopy, and the forest floor had high bryophyte diversity. Isothecium myosuroides and other moss species from the forest floor were also collected for comparison; these included Hylocomium splendens (Hedw.) Schimp., Rhytidiadelphus triquetrus (Hedw.) Warnst., Plagiothecium undulatum (Hedw.) Schimp., Dicranum fuscescens Turner, Kindbergia praelonga (Hedw.) Schimp. / oregana (Sull.) Ochyra.

BNF rates were estimated using acetylene reduction assay (ARA) (Hardy et al. 1968). Vials were capped with air-tight septa, and 10% of the head space was removed and replaced with acetylene gas. Controls consisted of moss without acetylene gas, and blank samples contained acetylene gas without moss. All samples were incubated for 24 h under greenhouse conditions (16°C, ambient light). Ethylene concentrations were determined by injecting a 1 ml subsample into a gas chromatograph (Shimadzu GC-2014), with a Carbosphere 80/100 packed column at 200°C, FID at 250°C, using He as a carrier at 30 ml/min. Raw values for ARA (μmol · ml-1 · d-1) were adjusted by subtracting blanks and controls, and standardized to per 100 g dwt moss.

The presence of cyanobacteria was confirmed by a novel protocol of disassociating the cyanobacteria from the moss using sonication. Approximately 100 mg dwt of each moss sample (two shoots ~7 cm in length from tip) was placed in 1 ml dH2O, and sonicated for 30 s at 25% amplitude (approx. 100 W power) using a Fisher Scientific Sonic Dismembrator 500. Two 10 μl subsamples of liquid were immediately transferred to a hemacytometer and scanned using fluorescence microscopy with a Texas Red filter under 200x magnification. The hemacytometer is a calibrated, gridded slide which standardizes the volume of liquid used to count cell densities. The total number of vegetative cells and heterocysts were recorded, and identified as the genus Scytonema (Rippka et al. 1979). Count data were converted to number of cyanobacterial cells per mg dwt moss.

The effect of height on ARA and cyanobacterial abundance was analysed using a generalized linear model with Poisson distributed errors. The relationship between cyanobacterial abundance and rates of acetylene reduction was explored using Spearman’s rank correlation, followed by linear regression. ARA values were converted to BNF rates using a molar ratio of 3 moles ethylene produced for each mole of N2 fixed (DeLuca et al. 2002; Lagerström et al. 2007; Zehr and Montoya 2007). We estimate annual BNF at the tree and stand-level based on 36.2 kg dwt epiphytic bryophytes per tree (Ellyson and Sillett 2003), average stand density of 400 stems per hectare (Ministry of Forests 2001), and the number of above-zero degree days as recorded by dataloggers placed in the canopy at sample locations. Stems per hectare measures in the Cedar-Western Hemlock very wet maritime (CWHvm) biogeoclimatic zone for the area range from 720–1,515, but include many younger trees without extensive epiphytic bryophytes, thus we use a conservative estimate of 400 stems per hectare.


Preliminary tests did not reveal any significant effect of watershed or moss species on either ARA rates or cyanobacterial abundance. Collection height of the moss sample had a significant effect on rates of ARA (F2,34 = 5.239, p = 0.010), whereby the average 30 m sample ARA value was significantly greater than at 15 m (Fisher LSD p = 0.015) and 0 m (Fisher LSD p = 0.005) (Fig. 1a). The total number of cyanobacterial cells was highly variable among samples, but significantly greater in 30 m moss samples compared to 15 m or forest floor samples (F2,34 = 3.326, p = 0.048; Fisher LDS posthoc test: p = 0.055, p = 0.021, respectively) (Fig. 1b). There was a significant, positive relationship between the total number of cyanobacterial cells and ARA rates (R2 = 0.268, F1,52 = 19.08, p < 0.001), described by the equation ARA (μmol · day-1) = 0.004 + 5.3 × 10-6 cells) (Fig. 1c). The cyanobacteria observed in all cases were within the genus Scytonema (Rippka et al. 1979). BNF rates from the canopy are estimated at 1.25 g N · tree-1· yr-1 for individual trees, and 0.50 kg N · ha-1 · yr-1 at the stand-level, while BNF at the forest floor is 0.26 kg N · ha-1 · yr-1 at the stand-level, based on 80% cover by bryophytes.
Fig. 1

Biological nitrogen fixation by cyanobacteria (Scytonema sp.) associated with bryophytes on 500 year old Sitka spruce trees in the temperate rain forest. a Nitrogen fixation rates, as measured by acetylene reduction assay, and b the number of cyanobacterial cells in floor and epiphytic bryophytes, increases with height in canopy. C) Nitrogen fixation (ARA) increases with the number of cyanobacterial cells [ARA (μmol · day-1) = 0.004 + 5.3 × 10-6cells] (95% limits shown) and is greater in epiphytic bryophytes than on the forest floor


This is the first study to confirm the presence of cyanobacteria associated with epiphytic bryophytes, and quantify rates of BNF within the arboreal bryosphere of a PNW temperate rain forest. We show that the vast majority of total stand-level BNF (0.76 kg N · ha-1 · yr-1) occurs in the canopy, rather than on the forest floor. Annual rates of BNF in bryophyte systems vary widely depending on number of growing days, and successional stage of the forest (Zackrisson et al. 2004; Menge and Hedin 2009). Our values are intermediate to those reported for late (1.7 kg N · ha-1 · yr-1) (DeLuca et al. 2002) versus early (0.5 kg N · ha-1 · yr-1) (Zackrisson et al. 2004) successional stands of the Boreal forest, and similar to lower end estimates of BNF associated with bryophytes in a temperate New Zealand forest (range: 0.7 and 10 kg N · ha-1 · yr-1) (Menge and Hedin 2009).

There are several considerations in our stand-level rates of bryophyte-associated BNF. Specifically, we rely on values from the literature for ARA conversion factors to N fixation, and for epiphytic biomass from nearby Washington State. Ellyson and Sillett (2003) estimated that old growth Sitka spruce trees carry an average epiphytic bryophyte biomass of 36.2 kg (dwt) per tree. However, while epiphytic bryophyte distribution patterns are similar on other tree species (Sillett and Rambo 2000), epiphytic biomass estimates are often lower in other co-dominant tree species (e.g. 18–27 kg / tree for Douglas fir) (Pike et al. 1977; Clement and Shaw 1999). However, our values may also be conservative estimates for several reasons. While the active season (above zero degree days) in our study is longer than in the Boreal forest, we need to consider that BNF during our sampling may have been limited by warmer temperatures and lower moisture availability during summer months. Summer BNF rates are typically the lowest of the active season. Favilli and Messini (1990) found that summer BNF rates were reduced by 82% compared to spring rates. Similarly, DeLuca et al. (2002) showed a bimodal peak in annual BNF rate; BNF peaked early in the season after snow melt, and again later in the summer, likely after the first rainfall, but BNF rates were reduced during the mid-summer months. The mechanism underlying the decrease in summer BNF rates in these studies were not identified but are possibly due to moisture limitation or temperatures above 30°C (Solheim and Zielke 2002; Gentili et al. 2005; Gundale et al. 2009). Temperature increases with height in the canopy (average yearly temp: 30 m = 7.90°C versus 15 m = 7.54°C), which may increase rates of BNF over the year, but mid-summer temperatures within the upper canopy can also reach 35°C, whereby BNF activity may be significantly reduced (Gentili et al. 2005). Additionally, mosses in arboreal habitats are on average 3 times drier than on the forest floor (Lindo and Winchester 2009). Thus canopy rates of BNF sampled during summer months may underestimate yearly BNF activity.

Stoichiometrically, the molar conversion ratio of acetylene reduction to nitrogen fixed is 4:1, but quantitatively most studies have found a 3:1 molar ratio for a wide range of moss species (DeLuca et al. 2002; Lagerström et al. 2007, ); occasionally studies have calculated a much lower ratio (e.g. 0.25:1 cited by Menge and Hedin 2009). As such we use the commonly reported 3:1 molar ratio which gives a conservative estimate of potential N fixation in this system (lower ratios would suggest greater amounts of N-fixation).

Nevertheless, both ARA rates and cyanobacterial abundance were significantly greater in the canopy than on the forest floor, and as epiphyte biomass generally increases with tree diameter at breast height (Hsu et al. 2002; Diaz et al. 2010), we posit that the oldest and largest trees of the forest are contributing the most to stand-level bryosphere BNF. Epiphytic mosses are considered an indicator of old-growth forests in temperate rain forests of PNW (Sillett 1995; Sillett and Rambo 2000; Newmaster et al. 2003). Younger, second growth trees do not harbour the high epiphytic biomass, and the trees samples in this study are estimated at over 500 years old. Moss species found in the canopy of this study (A. curtipendula and I. myosuroides) are typical old-growth associates, declining in abundance adjacent to forest gaps caused by forest harvesting (Hazell and Gustafsson 1999), and extirpated within clear-cuts (Rosso et al. 2001). Old-growth forest stands promote habitat and environmental conditions conducive to bryophyte growth (high humidity, low wind, moderate light) both in the canopy, and on the forest floor. In the same forest system as this study, Newmaster et al. (2003) found that forest floor bryophyte cover was reduced by nearly 1/3 in young, second growth forests compared to old-growth forests. Thus it appears that not only forest stand and tree age, but also stand continuity and integrity, are important correlates of bryophyte biomass and contributions to BNF.

Canopy contributions to BNF by cyanobacteria is dependent on the availability of appropriate habitat, namely epiphytic bryophytes, and establishment by colonizing cyanobacteria. Within our site, abundances of cyanobacteria may differ between canopy and forest floor systems due to differences in wind-mediated dispersal and colonization rates of cyanobacteria. This is the first study to document Scytonema associated with BNF in bryophytes, which may be physiologically and behaviourally adapted to canopy habitats. Scytonema cells are surrounded by a thick shealth containing the pigment scytonemin, which absorbs and protects against UV radiation (Stal 2006), and most species are categorized as aerially dispersing (Komárek et al. 2003). Airborne transport is an important mechanism of dispersal for terrestrial cyanobacteria (Marshall and Chalmers 1997), and may occur via dispersal vectors such as moss propagules. For instance, Lindo (2010) showed that high amounts (5,755 kg/ha) of bryophyte litter originate, circulate and are retained within canopy habitats of temperate rain forests.

Epiphytic BNF rates are associated with greater cyanobacterial abundance in the canopy than on the forest floor, but rates of BNF differ between 15 m and 30 m despite high cyanobacterial densities at both heights. The exact mechanisms to explain these patterns are not yet identified, but may be related to increased light availability (Vitousek and Howarth 1991; McCune et al. 1997), greater P availability (Smith 1992; Liengen 1999; Lindo and Winchester 2007) or higher colonization rates of aerial dispersing cyanobacteria in the upper canopy.

This study did not identify all potential sources of ecosystem N. Free-living N-fixing cyanobacteria have been observed in both canopy systems (the phylloplane) and on the forest floor (Jones 1982; Vitousek et al. 2002; Brunner and Kimmins 2003; Burgoyne and DeLuca 2009; Perez et al. 2010), however, BNF rates for cyanobacteria symbiotically associated with bryophytes are typically several-fold higher than for the same free-living cyanobacteria (Adams and Duggan 2008). Biological N fixation is significantly greater per unit area (or mass) for cyanolichens than cyanobacteria associated with bryophytes (Gavazov et al. 2010), and high stand-level BNF contributions of the cyanolichen Lobaria oregana has been documented at other PNW forest sites (Antoine 2004). Input of BNF from cyanolichens is not considered a major input in our sites, as bryophytes are dominant in the canopy and epiphytic cyanolichen biomass is relatively low, which suggests that epiphytic BNF by cyanobacterial-bryophyte associations is an important source of ecosystem N in coastal temperate rain forests.

Few studies have explored nutrient dynamics in arboreal systems (but see Vance and Nadkarni 1990; Nadkarni and Matelson 1991; Clarke et al. 1998; Nadkarni et al. 2002; Heitz et al. 2002; Lindo and Winchester 2007), and nutrient dynamics within arboreal habitats is poorly understood. Edmonds et al. (1991) deduced that canopy BNF in temperate forests may account for disparity in nutrient budgets derived from the chemical analysis of precipitation, throughfall and stemflow. The preponderance of BNF in the canopy may also explain the greater nutrient concentrations in suspended soil underlying epiphytic bryophytes compared to forest floor soils (Lindo and Winchester 2007).

However, the fate of the cyanobacterial fixed N is unknown. Bentley and Carpenter (1984) found that only 10–25% of cyanobacterial BNF was transferred to their host plant. At the tree level, uptake of canopy-derived nitrogen may take several forms: excess N may accumulate in suspended soils underlying arboreal moss mats (Lindo and Winchester 2007) and be taken-up by adventitious roots from the host trees (Nadkarni 1994; Sillett and Bailey 2003), filter down to the forest floor through stemflow and throughput (Edmonds et al. 1991), or fall to the forest floor via windthrow of whole trees, branches or individual moss mats. BNF in the canopy may ultimately contribute to the overall functioning and productivity of temperate rain forests by supplementing available nitrogen to the forest floor. As such, we suggest that intact old-growth temperate rain forests are important for the maintenance of ecosystem processes such as BNF, and that the loss of old-growth trees and associated epiphytic bryophytes, may have a disproportionate impact on forest N dynamics.

Old-growth forests promote environmental conditions conducive to bryophyte growth (high humidity, low wind, moderate light), and create habitat for high levels of bryophyte biomass and associated N-fixing cyanobacteria. Menge and Hedin (2009) suggest that epiphytic BNF may create a positive N feedback loop, as supplemental N from canopy sources to the forest floor could indirectly increase epiphytic biomass and consequently epiphytic BNF. However, it is more likely that increased N deposition arising from canopy sources would trigger an ecosystem-level feedback and reduce N fixation rates on the forest floor. This ecosystem-level regulation of nitrogen availability is consistent with DeLuca et al. (2008) who found lower N fixation in forest floors where high throughfall of N was being deposited, and may also explain the low rates of N fixation on the forest floor in this study despite coastal temperate rain forests having high forest floor bryophyte biomass.


This research was funded by a grant from the Natural Sciences and Engineering Research Council of Canada to Z. Lindo. A grant from Clayoquot Biosphere Trust to N. Winchester provided site access. We thank K. Jordan for logistical help in the field, A. MacKinnon, A. Gonzalez, T. DeLuca, J. Campbell, C. Prescott, M. Gundale, and anonymous reviewers for helpful comments on previous versions of this manuscript. A. Gonzalez provided laboratory space and equipment. Z. Lindo and J. Whiteley designed the experiment, Z. Lindo conducted the sampling, sample processing and data analyses. Z. Lindo and J. Whiteley wrote the manuscript.

Copyright information

© Springer Science+Business Media B.V. 2011