Ecosystems

, Volume 16, Issue 4, pp 590–603

Nitrogen Uptake by Trees and Mycorrhizal Fungi in a Successional Northern Temperate Forest: Insights from Multiple Isotopic Methods

Authors

    • University of Michigan Biological Station
    • Department of Ecology and Evolutionary BiologyUniversity of Michigan
  • Knute J. Nadelhoffer
    • University of Michigan Biological Station
    • Department of Ecology and Evolutionary BiologyUniversity of Michigan
  • James M. Le Moine
    • University of Michigan Biological Station
    • Department of Ecology and Evolutionary BiologyUniversity of Michigan
  • Linda T. A. van Diepen
    • Department of Natural Resources and the EnvironmentUniversity of New Hampshire
  • Jules K. Cooch
    • University of Michigan Biological Station
  • Nicholas J. Van Dyke
    • University of Michigan Biological Station
Article

DOI: 10.1007/s10021-012-9632-1

Cite this article as:
Nave, L.E., Nadelhoffer, K.J., Le Moine, J.M. et al. Ecosystems (2013) 16: 590. doi:10.1007/s10021-012-9632-1

Abstract

Forest succession may cause changes in nitrogen (N) availability, vegetation and fungal community composition that affect N uptake by trees and their mycorrhizal symbionts. Understanding how these changes affect the functioning of the mycorrhizal symbiosis is of interest to ecosystem ecology because of the fundamental roles mycorrhizae play in providing nutrition to trees and structuring forest ecosystems. We investigated changes in tree and mycorrhizal fungal community composition, the availability and uptake of N by trees and mycorrhizal fungi in a forest undergoing a successional transition (age-related loss of early successional tree taxa). In this system, 82–96% of mycorrhizal hyphae were ectomycorrhizal (EM). As biomass production of arbuscular mycorrhizal (AM) trees increased, AM hyphae comprised a significantly greater proportion of total fungal hyphae, and the EM contribution to the N requirement of EM-associated tree taxa declined from greater than 75% to less than 60%. Increasing N availability was associated with lower EM hyphal foraging and 15N tracer uptake, yet the EM-associated later-successional species Quercus rubra was nonetheless a stronger competitor for 15N than AM-associated Acer rubrum, likely due to the more extensive nature of the persistent EM hyphal network. These results indicate that successional increases in N availability and co-dominance by AM-associated trees have increased the importance of AM fungi in the mycorrhizal community, while down-regulating EM N acquisition and transfer processes. This work advances understanding of linkages between tree and fungal community composition, and indicates that successional changes in N availability may affect competition between tree taxa with divergent resource acquisition strategies.

Key words

mycorrhizahyphaecanopy15Ntreecompetition

Introduction

The effects of forest succession on nitrogen (N) availability and uptake by plants are important, because N limits plant productivity in temperate forests (Vitousek and Howarth 1991). Mycorrhizal symbionts play a major role in N acquisition for most temperate tree taxa (Courty and others 2010), but understanding of interactions between N availability, uptake by trees and mycorrhizal fungi during forest succession is limited. This knowledge gap partly results from a focus on severe disturbances and primary succession in studies of tree and mycorrhizal N cycling during succession, but also derives from the difficulty associated with measuring relevant processes (for example, the uptake and transfer of N from mycobionts to host plants) under field conditions. Because underlying feedbacks link the composition and functioning of plant and mycorrhizal communities (Francis and Read 1994; Read 1991; Rillig 2004), it is especially important to understand how forest succession (both above- and belowground) responds to and also influences the cycling of N through trees and mycorrhizal fungi.

Controlling influences of soil N availability and forest successional status on the composition of mycorrhizal fungal communities have been documented in a variety of temperate and boreal forest types (Cox and others 2010; Johnson and others 1991; Kranabetter and others 2005, 2009; Liu and others 2009; Wallander and others 2010; Visser 1995). On the other hand, studies that specifically integrate successional patterns in the composition of mycorrhizal fungal communities with their role in N cycling and plant nutrition tend to originate from forests undergoing major successional transitions such as stand replacement or even primary succession, rather than the more subtle transitions associated with tree species replacement during mid- to late-secondary succession. For example, Treseder and others (2004) documented a shift in abundance from arbuscular mycorrhizal (AM) to ectomycorrhizal (EM) fungi in boreal forests over time following severe, stand-replacing fire, corresponding to the accretion of soil organic matter stocks and N availability, whereas in a study of primary forest succession on glacial substrates the development of EM fungal communities was associated with an increasing contribution of mycorrhizal N uptake to plant N nutrition (Hobbie and others 2000). However, knowledge of changes in the mycorrhizal fungal community and its role in soil and plant N cycling derived from studies of stand replacement or primary succession may not be directly applicable to forests undergoing subtler, secondary successional transitions, due to the dramatically different states and trajectories of soil organic matter and nutrient accumulation during these stages of ecosystem development. Specifically, primary succession studies tend to support the hypothesis that EM fungi replace AM fungi over successional time, due to the increased importance of litter-bound and complex organic N sources that are considerably more accessible to EM fungi (Read 1991). In forests of more advanced age undergoing subtler successional transitions, the soil environment and the roles played by AM versus EM fungi in N cycling and plant uptake may be quite different.

In spite of the vital and long-acknowledged importance of mycorrhizal symbionts to plant N nutrition, there remains a need for studies that quantify basic symbiotic fluxes and their sources of variation over time and space. For example, how much of the N requirement of temperate trees can be attributed to mycorrhizal sources, and how do interactions between N availability and forest composition that arise during forest succession affect this quantity? 15N stable isotope methodologies are well-suited to such questions, because these methods are largely non-invasive with regards to the processes studied, and can be used to integrate knowledge of transfer processes derived from patterns of variation in 15N natural abundances with the results of 15N tracer applications that elucidate the movement of specific N compounds (Nadelhoffer and others 1999; Hobbie and Hobbie 2006). In this study, we combined isotopic methods (tracer and natural abundance) with measures of N availability, tree and fungal community composition to assess the impacts of ongoing changes in forest composition and N cycling on tree and mycorrhizal N uptake. Our core hypothesis was that increasing inorganic N availability would decrease the importance of EM N uptake during a successional transition, and that tree species replacements occurring during this successional transition would trigger corresponding shifts belowground within the mycorrhizal fungal community.

Methods

Site

This research is part of the Forest Accelerated Succession ExperimenT (FASET) at the University of Michigan Biological Station (UMBS), USA (45°35′N 84°43′W), where the mean annual temperature is 5.5°C and mean annual precipitation is 817 mm (including 294 cm snowfall). The study area covers approximately 140 ha on a high outwash plain and an adjacent gently sloping moraine. Soils are well- to excessively well-drained Haplorthods of the Rubicon, Blue Lake, or Cheboygan series, and have forest floors consisting of Oe horizons 1–3 cm thick overlying densely rooted, bioturbated AO horizons of 1–3 cm. AO horizons are underlain by an E horizon of 10–15 cm and a Bs horizon of sand with occasional gravel. About 53% of the fine root mass is located within the upper 20 cm of the soil profile. Forest floor C mass is 5–15 Mg C ha−1, and the mineral soil is about 95% sand and about 5% silt, with pH 4.5–5.5 in water. Soils have low fertility, with total N stocks to 40 cm depth of 2,000 kg ha−1, an average in situ net N-mineralization rate of 42 kg N ha−1 y−1, and less than 2% net nitrification (Nave and others 2009). The 90-year-old forest throughout the study area is dominated by bigtooth aspen (Populus grandidentata) at various stages of age-related or treatment-induced senescence, whereas other canopy species include red maple (Acer rubrum), red oak (Quercus rubra), white birch (Betula papyrifera), Eastern white pine (Pinus strobus), trembling aspen (Populus tremuloides), sugar maple (Acer saccharum), and American beech (Fagus grandifolia). Forest age, composition and disturbance history at the study site are representative of upland forests across the northern Great Lakes region, where aspen and birch-dominated hardwoods replaced pine-hemlock (Tsuga canadensis) forests following clearcutting and wildfires that ended in the early twentieth century (Gough and others 2010).

Experimental Treatment and Design

FASET consists of an ecosystem-scale, experimental forest manipulation (39 ha) paired with an approximately 100 ha control forest located 1–2 km away (Figure 1). At the far reaches of the control forest area are three, 2 ha replicate stands where the experimental treatment was applied to replicate the experimental treatment across site indices ranging from low to high. Experimental treatment (stem girdling of >6,700 mature, early successional Populus and Betula trees) occurred during 3 weeks in April–May 2008, and involved removing a 5–8 cm tall strip of cambium, phloem, and bark from each bole. This was accomplished by a team of chainsaw operators (who made two parallel cuts around the circumference of each bole) and a team of bark removers, who used crowbars to pry and peel the girdled bark strips from the trees. Stem girdling causes mortality of Populus trees within 1–3 years, and prevents the profuse root-sprouting (‘suckering’) behavior of this genus that is associated with cutting and felling the trees (Burns and Honkala 1990). Importantly, this experimental treatment is accelerating a successional transition that has been underway in the control forest for several years (Gough and others 2010), mimicking the impacts of pathogenic insects, fungi, or drought on the mortality rates of aging early successional species and releasing later-successional taxa (primarily Quercus and Acer). On-the-ground data collection occurs in permanent plots located along transects radiating out from two eddy-covariance meteorological towers—one which measures carbon, water, and energy exchange within a control forest footprint, and the other of which measures a footprint that has been subject to the experimental manipulation. Each tower is surrounded by an intensively measured 1.1-ha plot, which, relative to its surrounding measurement area (~100 ha for control and ~39 ha for the main treatment area) is of average soil fertility, aboveground biomass, and species composition. The remaining plots in each measurement area are smaller (0.1 ha) and are established at 100 m intervals along the transects extending out from the eddy covariance towers; several of these fall within the 2 ha treatment replicate areas though only one 0.1-ha plot sampled for this study was located in a treatment replicate area. Thus, the 3–7 plots sampled in each treatment for the various measurements in this study are essentially pseudo-replicates located within the main treatment and control areas; sampling across these largely pseudo-replicated plots nonetheless allows collection of data across ecosystem-level gradients in tree species composition and productivity noted in Table 1. For most results in this paper, we use these gradients to assess continuous variation in ecological properties and processes across the broader forest ecosystem, encompassing control and treatment plots distributed along a trajectory of successional change from light background mortality to rapidly accelerated succession. We make additional statistical comparisons between treatments, analyzing response data from pseudo-replicated control and treatment plots that were paired by species composition and aboveground biomass production before girdling using principal components analysis (PCA; see Nave and others 2011). Because this study presents results from a variety of different measurements from multiple collaborative projects ongoing within the large-scale experimental manipulation, the number of plots used for measurement and analysis varies between different parameters.
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-012-9632-1/MediaObjects/10021_2012_9632_Fig1_HTML.gif
Figure 1

Detailed view of the study area. Plots with unique, open symbols in the treatment forest (right) are paired according to species composition and aboveground biomass production with plots that have matching unique, closed symbols in the control forest (left). Stippled areas indicate locations of experimental treatment (stem girdling of all Populus and Betula trees in spring 2008). Note that the control forest includes 3 treatment replicate stands; one of these (located along D transect) contains a treatment plot visited for some of the data collection in this study.

Table 1

Plot-Level Characteristics and Sampling Design

Plot ID

Treatment

LAI

2007

LAI

2010

15N natural abundance

15N tracer addition

Hyphal biomass

Hyphal type

IERBs

Total

PoBe

Acer

QuPi

Total

PoBe

Acer

QuPi

m2

m−2

m2

m−2

%

%

m2

m−2

m2

m−2

%

%

C1 (I2)

Control

4.2

0.6

41

44

3.1

0.3

45

47

×

×

×

×

×

C2 (C9)

Control

5.7

3.2

45

0

3.4

1.8

46

3

×

    

C3 (F8)

Control

4.0

1.7

38

18

3.7

0.7

48

34

×

×

×

×

×

C4 (R60)

Control

3.4

1.5

24

31

3.2

1.0

26

42

×

 

×

×

×

C5 (C5)

Control

3.6

1.7

3

49

2.9

1.0

20

44

×

   

×

C6 (B4)

Control

3.7

1.9

28

19

2.9

1.2

40

19

×

  

×

×

C7 (Q5)

Control

3.6

1.2

25

42

2.6

0.6

26

48

×

×

×

×

×

C8 (Q7)

Control

3.3

1.3

25

35

2.7

1.4

22

27

×

  

×

×

G1 (D9)

Girdled

3.3

2.2

96

4

2.5

0.0

92

7

×

  

×

×

G2 (C3)

Girdled

5.8

1.2

55

24

2.4

0.1

51

46

×

×

×

×

×

G3 (B3)

Girdled

3.7

0.7

67

16

2.0

0.0

78

21

×

×

×

×

×

G4 (F60)

Girdled

3.6

1.3

32

32

2.4

0.4

42

44

×

 

×

×

×

G5 (D2)

Girdled

3.9

1.7

42

14

2.5

0.5

48

31

    

×

G6 (D1)

Girdled

3.5

1.6

23

32

2.1

0.2

41

48

×

×

×

 

×

G7 (A1)

Girdled

5.5

4.0

13

15

1.7

0.3

31

52

    

×

G8 (B1)

Girdled

2.8

1.4

3

48

1.7

0.2

7

80

    

×

Values in the Plot ID column denote the symbols used to show distribution of control versus girdled plots in most figures, whereas parenthetical plot IDs relate to site-specific plot names on the site map (Figure 1). Leaf area index (LAI) from 2007 and 2010 includes the total LAI (a metric of aboveground production), the LAI of early successional Populus and Betula species (PoBe; stem-girdled in 2008 in the girdled plots), and the proportions of total LAI represented by later-successional tree taxa (AM Acer; EM Quercus and Pinus QuPi). Other columns indicate the plots contributing data to this study.

Hyphal Biomass-Collection and Processing

We collected hyphal biomass using an ingrowth bag method similar to Wallander and others (2001). In each plot (8 total; see Table 1 for details), ingrowth bags were deployed in three randomly located clusters, each of which consisted of four bags inserted vertically into the top 10 cm of the soil profile (excluding the Oi horizon) within an area of ~1,500 cm2. One bag in each cluster was installed in soil isolated within a PVC tube (15 cm diameter, 30 cm depth) that had been driven into the ground to sever and exclude roots. By comparing hyphal biomass and 13C signatures of these ingrowth bags with those deployed outside of root exclusion tubes, we intended to constrain relative abundances of mycorrhizal versus saprotrophic hyphae, the latter of which should predominate inside the root exclusion tubes (Wallander and others 2001, 2004). Ingrowth bags (10 cm long, 2.5 cm diameter) were made of nylon mesh (SEFAR Inc., Depew, NY USA) of a pore size (50 μm) that excluded roots but permitted hyphal ingrowth. Bags were glued together with polyester resin and each contained 50–60 g of ashed, sieved sand (≥355 μm diameter). The ingrowth bags described in this analysis originate from three incubation periods (Oct 2009–May 2010, May 2010–Sep 2011, Sep 2011–Jul 2011), which varied according to logistical constraints and experimental design considerations. Upon each collection, bags from the deployment clusters were aggregated at the plot level to ensure sufficient hyphal biomass for processing and analysis. Ingrowth bags were kept frozen (−20°) until processing. Hyphae were extracted from the sand by flotation in water, and extracted hyphae were further cleaned of mineral contamination under a microscope with fine-tipped forceps. Hyphal biomass samples were frozen, freeze-dried, and weighed to the nearest 0.0001 g for eventual conversion to mass density values (kg ha−1).

Characterization of Hyphal Biomass

We used light microscopy (McGonigle and others 1990) to quantify proportions of septate versus aseptate hyphae contained in the ingrowth bags collected from each plot in May 2010. In brief, we prepared and viewed wet mount slides (100×) using rinse water from hyphal processing, to count 150 intersections. Because the presence of regularly spaced septae is a morphological trait that distinguishes EM and saprotrophic from AM hyphae (Smith and Read 2008), we assumed that aseptate hyphae were AM and septate hyphae were EM, pending the findings of biomass and 13C investigations into the abundance of saprotrophic hyphae (per Wallander and others 2001). Regarding our measurements of hyphal biomass, raw hyphal %C values indicated mineral contamination of hyphal samples (mean = 30% C) in spite of intensive efforts to remove adhering sand particles during ingrowth bag processing. We therefore applied a linear correction to hyphal biomass values based on knowledge that fungal biomass averages 43% C at our site (unpublished data from >200 sporocarps), such that each hyphal biomass value was scaled down in proportion to its amount of mineral contamination relative to the 43% C ‘clean’ criterion. This correction factor based on %C data was also applied to hyphal %N values, and corrected C and N concentrations in hyphal biomass were multiplied by production mass densities (kg ha−1 during the incubation periods) to determine hyphal production C and N increments for use in isotope N uptake and mass balance calculations. Analytical determinations were made on freeze-dried, ground samples using a Costech Analytical CHN analyzer (Costech Analytical, Valencia, CA, USA) coupled to a Finnigan Delta Plus XL isotope ratio mass spectrometer (Thermo Scientific, West Palm Beach, FL USA) at UMBS. The mass spectrometer was used to determine hyphal δ13C and δ15N, the latter of which was a component of both natural abundances and tracer approaches to assessing hyphal N cycling.

Foliar Biomass and Chemistry

We obtained canopy sun leaf samples (≥3 leaves from 2–3 dominant trees per plot) using shotguns in August of each year, sampling the 6 tracer plots (see “15N tracer—application and recovery calculations” section) in all years (2008–2010) and 2 additional plots in the pre-labeling years of 2008–2009 (the 1.1 ha plots described in “Experimental treatment and design” section). Plot-level foliar biomass production for the tree taxa of interest to this analysis was estimated as the quantity of litterfall (kg ha−1 y−1), averaged across 3 litter traps per plot (20 in each 1.1-ha plot). To assess interspecific competition based on determinations of 15N tracer recovery (see section below) in foliage, we used published allometric equations (Perala and Alban 1993) to calculate the foliar biomass of individual trees of each species from dbh measurements. Foliar samples were oven-dried (60°), ground in a ball mill, and analyzed at UMBS for %C, %N, δ13C, and δ15N.

Soil N Availability

We used ion-exchange resin bags (IERBs) to measure seasonally integrated \( {\text{NH}}_{4}^{ + } \) availability from June–September 2010 in 14 of the 16 PCA-paired control–treatment plots (see Experimental treatment and design section). Construction and deployment methods appear elsewhere (Nave and others 2011), but in brief we deployed 9 IERBs per plot beneath intact flaps of forest floor. Each IERB was a nylon foot stocking (MacPherson Leather, Seattle, WA, USA) containing around 30 mL of Dowex Marathon MR-3 mixed bed IER beads (Dow Chemical, Midland, MI, USA), packed into a PVC ring (5 cm diameter, 2 cm height). After collection, IERBs were extracted with 2 M LiCl, and the extract solutions analyzed for [NH4-N] on a SmartChem 200 (Westco Scientific Instruments, Brookfield, CT, USA) using the EPA 350.1 method. Ammonium-N concentrations were scaled by extract volume, resin mass, and PVC ring area to calculate an index of \( {\text{NH}}_{4}^{ + } \) availability across plots.

15N Methods—Natural Abundances

We applied the analytical model of Hobbie and Hobbie (2006) to estimate the fraction of the N requirement of EM-associated canopy trees that is derived from hyphal N transfer (f) across a range of plots in which we conducted natural abundances surveys before initiation of labeling experiments. The model is based on the observation that EM-associated plants tend to have more depleted δ15N as they derive increasing proportions of their N from EM hyphal N sources, and that sporocarps produced by these mycorrhizal symbioses become correspondingly enriched in 15N as the plant tissue becomes depleted (Nadelhoffer and others 1996; Hobbie and Hobbie 2006). The mechanism for this isotopic pattern likely relates to discrimination against 15N during synthesis of N transfer compounds within the EM hyphae, such that the N transferred to the host (for example, glutamine) is depleted in 15N and the residual, enriched N within the hyphae is allocated to sporocarp production. The model uses three simultaneous equations with four unknown values to constrain parameters including f, and requires as inputs the δ15N of EM-associated plant tissue (foliage), EM fungal sporocarps, the pool of N available for root and hyphal uptake, and the proportional discrimination against 15N (Δ) during assembly of N transfer compounds in the hyphae. We set Δ = 9‰ (Hobbie and Hobbie 2006), and assumed that the δ15N of the available N pool (1.6‰) was similar to ingrowth bag hyphae, saprotrophic sporocarps, and our measurements of filter paper packs after colonization by fungal hyphae during 8-wk soil incubations in summer 2008 (Hendricks and others 1997). These three substrates showed similar δ15N in their respective years of sampling (see Table 2), as did EM-associated tree foliage and EM fungal sporocarps across years, the latter of which were collected during July–September of each year, identified to genus and averaged across genera within each plot. To best constrain the contribution of hyphal transfer to the N requirement of EM trees using the available data, we calculated f for the 8 plots sampled in both 2008 and 2009, using 2-year average values for EM-associated foliage and sporocarp δ15N (labeling occurred on a subset of the plots in 2010).
Table 2

Mass Production and δ13C of Fungal Hyphae Collected from Ingrowth Bags During 2 Incubations (2010 and 2011 Deployments)

Bag type

Hyphal mass (mg bag−1)

SE

Hyphal δ13C (‰)

SE

Roots present

11.5a

2.0

−26.70a

0.10

Roots restricted

4.5b

1.0

−26.91a

0.11

Significantly different mean mass and 13C values for root-present versus root-restricted hyphal bags are indicated by superscripts.

15N tracer—Application and Recovery Calculations

We added aqueous 15NH4Cl (20 mg 15N m−2) to the forest floors of 6 paired plots (3 control; 3 treatment) in 2010 using backpack sprayers. The tracer addition was distributed over three events (May, June, July) to isotopically enrich the \( {\text{NH}}_{4}^{ + } \) pool and allow insights into the cycling, uptake and distribution of this and other actively cycling N compounds during the course of the growing season. For foliar and hyphal biomass pools sampled in 2010, tracer recoveries were determined using 2008–2009 pre-label averages of the 15N natural abundance, and 2010 values for the N pool size of each biomass pool. We used mass balancing to quantify the amount of added 15N that was recovered in individual biomass pools, based on the following equation (Nadelhoffer and others 1999):
$$ ^{{15}} N_{{{\text{rec}}}} = m_{{{\text{pool}}}} \cdot\left[ {\left( {{\text{atom}}\% ^{{15}} {\text{N}}_{{{\text{pool}}}} - {\text{atom}}\% ^{{15}} {\text{N}}_{{{\text{ref}}}} } \right) \div \left( {{\text{atom}}\% ^{{15}} {\text{N}}_{{{\text{tracer}}}} - {\text{atom}}\% ^{{15}} N_{{{\text{ref}}}} } \right)} \right] $$
(1)
where 15Nrec is the mass of 15N tracer recovered in the labeled pool, mpool is the mass (pool size) of the labeled N pool, atom%15Npool is the atom percent 15N in the labeled N pool, atom%15Nref is the atom percent 15N in the pre-labeled pool, and atom%15Ntracer is the atom percent 15N of the applied tracer. We used equation (1) to calculate the mass of 15N tracer recovered in biomass pools and subsequently to express this mass as a proportion (%) of the applied label, to understand ecosystem-scale movements of actively cycling N during the 2010 season. We examined variation in 15N label recovery in fungal hyphae as well as foliage of the dominant tree species, the latter of which were considered individually and on a whole-canopy basis to understand both interspecific variation in N uptake and the functioning of the canopy at stand-to-ecosystem scales. We contextualized our measurements of 15N label recovery in these biomass N pools by quantifying the δ15N of IERB extract solutions, using acidified filter diffusions and mass spectrometric analysis of prepared filters to obtain raw δ15N values of the forest floor \( {\text{NH}}_{4}^{ + } \) adsorbed onto the resins. This procedure was performed by the Cornell Isotope Laboratory (COIL) according to methods described in Stephan and Kavanaugh (2009). In brief, we composited aliquots of the individual IERB extracts from each plot into an aggregate sample for each of the 6 labeled plots; after adding MgO to the aggregate samples to volatilize \( {\text{NH}}_{4}^{ + } \) these samples were incubated in 90 ml polypropylene vials for 7 days at 30° on an orbital shaker. The \( {\text{NH}}_{4}^{ + } \) diffusing into the vial headspace was trapped on an acidified glass-fiber filter and subsequently analyzed on an isotope ratio mass spectrometer at COIL. Then, rather than performing mass-balance calculations, we assessed isotopic enrichment in these IERB extracts using the raw δ15N values returned by the lab because: (a) pre-labellng δ15N values for the IERB extracts were not specifically known, and (b) the computation of an N pool size from IERB extracts is of no additional value, as IERBs are useful more for indices of N availability but of uncertain application to ecosystem budgets or mass balances.

Data Analysis

We conducted parametric statistical tests (t test, ANOVA, simple linear regression) of the data using SigmaPlot (SYSTAT Software, San Jose, CA USA). We tested each response parameter distribution to determine whether it met crucial parametric assumptions (we considered modest deviations from normality acceptable and used ln-transformation as necessary to normalize highly non-normal variables), and set P < 0.05 as the threshold for accepting test results as significant. We established our permanent sampling plots as the statistical replicates in all tests except the analysis of 15N tracer uptake by trees, which we assumed to be competing as individuals and therefore appropriate units of replication.

Results

Fungal Hyphae—Characterization and Functional Relationships with Tree Taxa

Across all plots, hyphal ingrowth bags deployed inside of root exclusion tubes contained significantly less hyphal biomass than ingrowth bags deployed in unrestricted soil, but the 13C signatures of hyphae from the two deployment methods did not differ (Table 3). Septate, EM hyphae constituted 82–96% of total hyphae across plots, whereas the remaining 4–18% of hyphae were aseptate (AM hyphae; Figure 2). Across plots, the relative abundance of aseptate hyphae increased significantly with increasing leaf production by AM-associated tree species (Figure 2). 15N natural abundances patterns indicated that along a gradient of increasing AM codominance, the contribution of EM N transfer to the N requirement of EM-associated tree taxa declined from >75 to <60% (Figure 3).
Table 3

Site-Level Summary Statistics for the 15N Signatures of Ecosystem Pools Used for Calculating the Contribution of Hyphal N Transfer to the N Requirement of EM-Associated Tree Species

Pool

2008

2009

δ15N

σ

n

δ15N

σ

n

Ingrowth hyphae

1.6

1.2

13

SAP sporocarps and NIFsa

1.5

1.1

12

EM fungal sporocarps

5.9

1.3

8

6.9

1.5

13

EM tree species foliageb

−4.8

0.7

8

−4.0

1.0

13

aNIF, nitrogen immobilization filter, a cellulose card deployed in the top 10 cm of soil, used to measure δ15N of the N translocated from surrounding soil by colonizing (probably saprotrophic) hyphae (Hendricks and others 1997).

bEM foliage: mass-weighted average δ15N of litterfall for all EM-forming tree taxa in each plot Means, standard deviations, and sample sizes refer to the number of plots in which samples were collected in each year.

https://static-content.springer.com/image/art%3A10.1007%2Fs10021-012-9632-1/MediaObjects/10021_2012_9632_Fig2_HTML.gif
Figure 2

Proportions of AM (symbols below) versus EM (symbols above) hyphae in slide counts as a function of leaf production by AM-associated tree taxa (Acer spp.). Points correspond to individual plots sampled in 2009; “C” labels designate control plots, whereas in experimental “G” plots all Populus and Betula were girdled (see Table 1 for more plot information). Note the break in the y-axis scale; note also that because the best-fit lines are applied to data points which are the inverse of one another, the absolute r2 and P values are the same for the proportion of either hyphal type as a function of AM-associated tree production.

https://static-content.springer.com/image/art%3A10.1007%2Fs10021-012-9632-1/MediaObjects/10021_2012_9632_Fig3_HTML.gif
Figure 3

Relationship between leaf production by AM-associated tree taxa and f, the proportional contribution of EM hyphal N transfer to the foliar N requirement of EM-associated tree species. Points represent means of 2008 and 2009 plot-level data; “C” labels designate control plots, whereas in experimental “G” plots all Populus and Betula were girdled (see Table 1 for more plot information).

N Availability, Mycorrhizal and Canopy Functioning

The experimental treatment (stem girdling all Populus and Betula individuals) simultaneously increased forest floor \( {\text{NH}}_{4}^{ + } \) availability (Figure 4) and altered the N uptake dynamics of trees and mycorrhizal fungi. Across the six 15N-labeled plots, the proportion of \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) tracer recovered in hyphae was negatively correlated with forest floor \( {\text{NH}}_{4}^{ + } \) availability (Figure 5). This relationship between forest floor \( {\text{NH}}_{4}^{ + } \) availability and the recovery of 15N in hyphae was paralleled by a similar (negative) relationship between \( {\text{NH}}_{4}^{ + } \) availability and the ingrowth N increment of hyphae (Figure 6a), whereas in contrast there was a positive relationship between \( {\text{NH}}_{4}^{ + } \) availability and the δ15N of \( {\text{NH}}_{4}^{ + } \) (Figure 6b). Experimental treatment significantly decreased the mean 15N tracer uptake across the three dominant species (ANOVA, P < 0.05); however, this overall effect of treatment on the canopy was associated with significant interspecific variation (P < 0.001) and a treatment × species interaction (P < 0.001). Specifically, the three dominant species comprising the canopy exhibited counteracting treatment effects (Figure 7), with P. grandidentata showing a major decline in 15N tracer uptake due to stem girdling that was partially compensated by significantly increased tracer recoveries in the foliage of Q. rubra and A. rubrum in the girdled plots. In control plots, the foliage of P. grandidentata and Q. rubra contained a similar amount of the added 15N, a quantity which was significantly greater than that recovered in A. rubrum. In the treatment plots, however, all three species showed significantly different levels of 15N recovery (Q. rubra > A. rubrum > P. grandidentata).
https://static-content.springer.com/image/art%3A10.1007%2Fs10021-012-9632-1/MediaObjects/10021_2012_9632_Fig4_HTML.gif
Figure 4

Integrated forest floor \( {\text{NH}}_{4}^{ + } \) availability, June–September 2010, as measured with ion-exchange resins in the control forest and the treatment forest subjected to stem girdling of early successional tree species. Plotted values are means and standard errors of n = 7 plots in each forest, with superscripts denoting significantly different groups.

https://static-content.springer.com/image/art%3A10.1007%2Fs10021-012-9632-1/MediaObjects/10021_2012_9632_Fig5_HTML.gif
Figure 5

Relationship between forest floor \( {\text{NH}}_{4}^{ + } \) availability and the recovery of \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) tracer in hyphae. Each point is the mean value for one labeled plot during the season of 15N tracer addition (2010); “C” labels designate control plots whereas in experimental “G” plots all Populus and Betula were girdled (see Table 1 for more plot information).Values for r2 and P correspond to a simple linear regression test relating \( {\text{NH}}_{4}^{ + } \) availability and 15N recovery in hyphae.

https://static-content.springer.com/image/art%3A10.1007%2Fs10021-012-9632-1/MediaObjects/10021_2012_9632_Fig6_HTML.gif
Figure 6

Relationships between forest floor \( {\text{NH}}_{4}^{ + } \) availability and the N increment of ingrowth hyphae (atop) and the 15N signature of forest floor \( {\text{NH}}_{4}^{ + } \) (bbottom). Each point is the mean value for one labeled plot during the season of 15N tracer addition (2010); “C” labels designate control plots whereas in experimental “G” plots all Populus and Betula were girdled (see Table 1 for more plot information).Values for r2 and P correspond to simple linear regression tests relating \( {\text{NH}}_{4}^{ + } \) availability to the respective response parameters.

https://static-content.springer.com/image/art%3A10.1007%2Fs10021-012-9632-1/MediaObjects/10021_2012_9632_Fig7_HTML.gif
Figure 7

Recovery of 15N in tree foliage following \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) tracer addition in 2010, expressed as the percent of added 15N recovered in the foliar N pools of the three dominant species. Plotted values represent means and standard errors for 4–9 individuals of each species, and taxonomic codes are ACRU Acer rubrum, POGR Populus grandidentata, and QURU Quercus rubra. The left panel (a) shows data for trees from control plots, whereas the right panel (b) corresponds to trees from girdled plots. Note that treatment effects were significant for all species; superscripts therefore denote significant differences between species within each treatment.

Discussion

Hyphal Ingrowth Bags-Interpretation of Methods

Variation in hyphal biomass production, morphology and δ13C, as driven by canopy composition and root restriction tubes provides significant insights into the fungal community composition of this study system. On basic principle (Wallander and others 2001), the much smaller amount of hyphal production inside of root restriction tubes suggests that these hyphae may have been saprotrophic. Nevertheless, these root-restricted hyphae had very similar δ13C to the hyphae from ingrowth bags deployed in unrestricted soil (range of −26 to −27‰), approximating C pools derived from recent photosynthate such as tree foliage, roots, and EM sporocarps (δ13C ≈ −24 to −29‰ at UMBS). In contrast, known saprotrophic and pathogenic fungi at UMBS have δ13C ≈ −19 to −23‰. Thus, our ingrowth bags were primarily colonized by mycorrhizal as opposed to saprotrophic hyphae, which is similar to the studies that established this method (Wallander and others 2001, 2004); for further interpretations, we consider the hyphae recovered from ingrowth bags to be mycorrhizal (EM and/or AM).

Linkages Between Tree and Mycorrhizal Community Composition

Across all control and treatment plots where the proportion of AM (aseptate) versus EM (septate) hyphae was measured, greater than 80% of hyphae were EM, and variation in the amount of leaf production by EM-associated tree species did not affect this metric of fungal community composition. Conversely, as leaf production by AM-associated trees increased across the range of plots, AM hyphal relative abundance increased (Figure 2). In the context of spatial patterns of aboveground woody plant succession at our site, this close relationship between the composition of the mycorrhizal fungal community and the tree community is a logical result congruent with the mortality of two EM-associated genera (Populus and Betula) and partial replacement by AM-associated Acer spp. Other studies investigating shifts in the relative abundance of AM versus EM fungi during forest succession have documented the opposite pattern—that EM fungi become more prevalent during succession because the accumulation of surface litter and soil organic matter favors their enzymatic capabilities over AM fungi, which are mostly restricted to accessing inorganic nutrients (Read 1991; Piotrowski and others 2008; Treseder and others 2004). Therefore, the different trajectory of succession at our site reinforces other evidence (Nave and others 2011; Nave and others, in preparation) that successional changes in soil organic matter and nitrogen dynamics are following a trajectory of increasing fertility. These findings from FASET are furthermore supported by landscape-level successional patterns in the area—at slightly older forests occurring less than 10 km away on moraines, Acer spp. are the principal canopy dominants, AM hyphae represent greater than 92% of all mycorrhizal fungal hyphae, and soil N availability is much higher (van Diepen and others 2010; Zogg and others 1996). Lastly, strong co-variance between the amount of Acer spp. in the canopy and the abundance of non-native earthworms in area forests provides an additional feedback to interactions between soil organic matter/nutrient dynamics and mycorrhizal community composition, as earthworms accelerate the breakdown of surface litter and its mixing with mineral soil both in our system and in other studies (Crumsey, in preparation; Bohlen and others 2004; Hobbie and others 2006). This additional influence of bioturbating macrofauna was not quantified as an explicit component of this work, but can clearly act as a belowground mechanism that reinforces successional transition towards AM-associated trees and fungi at our site.

Successional Changes in Mycorrhizal Function

As the plots within our control and treatment forests move at varying successional paces (some experimentally accelerated) towards increased Acer co-dominance, concurrent changes in tree species composition and N availability are altering fungal community composition and hyphal N uptake, and through these changes decreasing the functional role of the yet-dominant EM network as a C-consuming, N-acquiring link in the biogeochemistry of this forest. This functional shift is exemplified by the significant decline in the fraction of the EM-associated tree species’ N requirement (f) supplied by EM hyphae with increasing co-dominance of AM-associated taxa (Acer spp.; Figure 3). As AM-associated trees and fungi increase during this period of successional change, simultaneous increases in inorganic N availability reported here (Figure 4) and elsewhere (Nave and others 2011) cause EM tree species to reduce their dependence on EM hyphal N uptake, presumably through shedding of mycorrhizal associations or decreased C allocation to extra-radical hyphae, favoring instead the direct uptake of inorganic N forms directly through roots.

Our combination of 15N natural abundance and tracer work increases understanding of the role EM fungi play in host N nutrition during forest succession, and along gradients in inorganic N availability. Although the analytical model used to calculate f (Hobbie and Hobbie 2006) demonstrates a reasonable pattern of EM functioning that is consistent with changes in tree and fungal community composition and N availability during succession, it does not definitively establish a belowground mechanism for the EM trees’ decreased reliance on hyphal N uptake. However, the complementary labeling approach provides this mechanism, indicating that EM hyphae take up less \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) as the availability of \( {\text{NH}}_{4}^{ + } \) increases (Figure 5). Superficially, this result could appear to be a simple effect of \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) tracer dilution within the native soil \( {\text{NH}}_{4}^{ + } \) pool, such that there were effectively fewer \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) molecules available to hyphae in soils with larger \( {\text{NH}}_{4}^{ + } \) pools. On the contrary, \( ^{ 1 5} {\text{N - NH}}_{4}^{ + } \) data reveal that \( {\text{NH}}_{4}^{ + } \) was in fact more 15N-enriched in stands with greater \( {\text{NH}}_{4}^{ + } \) availability (Figure 6b), indicating that this pool dilution effect did not occur and suggesting that low rates of underlying uptake processes (for example, roots and hyphae) were driving high levels of enriched \( {\text{NH}}_{4}^{ + } \) availability. In turn, this mechanism is reinforced by the lower amount of hyphal \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) tracer recovery in plots with higher \( {\text{NH}}_{4}^{ + } \) availability, which was due to an underlying negative feedback between \( {\text{NH}}_{4}^{ + } \) availability and the size of the hyphal N pool (Figure 6a). Because the hyphal N pool size as quantified by ingrowth bags is a production metric that integrates rates of hyphal biomass growth and N assimilation, we suggest here that it is an approximation of N foraging activity by the extra-radical hyphae, some of which explore considerable volumes of soil if N or other resources are highly limited (Agerer 2001). In contrast, when N availability is high, hyphal production and other aspects of mycorrhizal functioning are typically downregulated (Brzostek and Finzi 2011; Treseder 2004; Wallander 1995; van Diepen and others 2007). In this case, the combined isotopic approaches converge to yield the insight that as soil inorganic N availability increases, EM hyphae take up less inorganic N and contribute less to the N nutrition of host trees. Because the 15N tracer was added to the soil surface throughout the 2010 season concurrently with measurements of soil N availability and hyphal ingrowth, the tight linkage between \( {\text{NH}}_{4}^{ + } \) availability and \( ^{ 1 5} {\text{NH}}_{4}^{ + } \) recovery in hyphae indicates that both ion-exchange and hyphal ingrowth methods measure ecologically meaningful, mutually consistent quantities, providing important validation of their relevance to biogeochemical processes occurring in undisturbed soil. Furthermore, the result that the negative impact of N availability on hyphal N foraging was apparent as continuous variation rather than a categorical treatment effect highlights the fact that Populus and Betula mortality is proceeding at a range of paces across the plots visited for this study; the experimental treatment of stem girdling thus works in concert with ongoing background senescence to represent the impact of an acute agent on ‘natural’ succession quite well, for example, an insect defoliation causing carbohydrate starvation of already stressed and senescing trees. Long-term declines in Populus and Betula observed on control plots (Table 1; see also Gough and others 2010) indicate a baseline level of mortality among these aging trees; stem girdling exacerbates background senescence with the net result being a patchy, temporally distributed mortality event rather than a sudden, complete loss of the taxa, with correspondingly subtle effects on biogeochemical processes.

Successional Changes in Canopy Function

15N tracer recovery results confirm prior findings (using other methods) that successional increases in \( {\text{NH}}_{4}^{ + } \) availability are primarily due to decreased N uptake by senescing Populus trees (Nave and others 2011). In fact, experimentally girdled Populus showed no evidence of 15N tracer uptake in 2010, indicating that the N requirement of foliage produced after treatment (spring 2008) was met entirely by stored N reserves. Considering total N uptake by the canopy as a whole in the context of N leaching losses, Nave and others (2011) concluded that proliferation of leaf area by later-successional species acted as a mechanism to promote complete ecosystem N retention during the first 2 years of experimental treatment and Populus/Betula mortality. As evidenced by the 15N tracer results (Figure 7), A. rubrum and Q. rubra were both important components of this compensatory response, with greater amounts of 15N recovery likely corresponding to greater rates of N uptake in response to elevated N availability. Across both treatments, greater 15N tracer recovery in foliage of Q. rubra than A. rubrum may result from the former taxon being integrated with a more extensive hyphal network (>80% of all ingrowth hyphae were EM) or a more active rhizosphere (Phillips and Fahey 2006), both of which may be expected to provide greater supply of N than the limited amount of AM hyphae present in this soil. Ectomycorrhizal rhizospheres and hyphae (and accordingly the canopies that support them) have greater ability than AM-associated species to access organic, often enzymatically expensive (C-demanding) N forms (Brzostek and Finzi 2011; Gallet-Budynek and others 2009). Although this capacity for assimilation of complex N forms provides Q. rubra a competitive advantage over A. rubrum, it also places the result that increasing inorganic N availability causes down-regulation of EM hyphal foraging, 15N uptake and N transfer to EM host trees in a logical C–N cost-benefit framework (Bidartondo and others 2001; Phillips and Fahey 2008). Specifically, with energetically cheap, easily accessible inorganic N forms present in elevated supply in soil, it is plausible that EM-associated trees may redirect C formerly allocated to EM processes (for example, production of extra-radical hyphae and enzymes) to other sinks, such as aboveground growth or root biomass production. If this mechanism linking successional changes in N availability and EM N uptake is indeed at play, the trajectory of N cycling in this successional system will remain a critical driver of mycorrhizal functioning. Going forward, we hypothesize that continued increases in Acer co-dominance and a shift in the mycorrhizal community from EM towards AM dominance will feed back to soil N availability through litterfall processes, as the fast-turnover leaf and root litter produced by Acer spp. (Ferrari 1999; Langley and Hungate 2003) become larger components of above- and belowground litter inputs. If true, this hypothesis will further de-emphasize the importance of mycorrhizal N uptake for Quercus and other EM-associated trees at this site.

Conclusions

The mortality of EM-associated early successional tree taxa at our site, and the partial replacement of these trees by ascendant AM-associated taxa (Acer spp.) led to net replacement of EM fungal hyphae by AM fungal hyphae. Concurrently with these changes in tree and fungal community composition, increased inorganic N availability led to down-regulation of EM hyphal foraging and N uptake, decreasing the role that EM fungi play in the N nutrition of EM-associated trees. This work advances the understanding of linkages between tree and fungal community composition, how these ecosystem parameters change in response to succession and N availability, and the consequences of such changes for EM functioning and the N nutrition of EM-associated trees. Although the above- and belowground components of this forest ecosystem are yet EM-dominated, the continued progression of successional change in species composition and N cycling will influence mechanisms of N uptake among competing trees with functionally different mycorrhizal associations. Forthcoming work from this study that documents successional changes in soil organic matter dynamics and nutrient cycling, coupled with broader investigation of mycorrhizal functioning during forest succession, will continue to reveal a more complete picture of the linkages between these critical ecosystem processes.

Acknowledgments

This research is supported by awards from the National Science Foundation (DEB-0911461) and U.S. Department of Energy (DE-FC02-06ER64158). The authors thank R. M. Miller for assistance with hyphal ingrowth methodology, and the following individuals for assistance with fieldwork, sample preparation, and analysis: A. Bajcz, A. Baldick, C. Bogdan, A. Brenske, A. DeGabriele, A. Do, Z. Fortier, A. Gold, M. Grant, J. Halick, S. Liao, K. McClure, J. Pollack, K. Sparks, S. Sheperd, F. Soper, C. Vogel, S. Webster, B. Weiler, S. Yassine, and Z. Zeneberg. Lastly, the authors are grateful to two anonymous reviewers who provided helpful feedback that improved this paper from its manuscript form.

Copyright information

© Springer Science+Business Media New York 2013