Introduction

Autotrophic plants have evolved a variety of different nutrient acquisition syndromes, and the relative viability of these syndromes appears to vary along environmental gradients. Relationships with climate have been documented, with ectomycorrhizal (ECM) plants usually dominant in climates that are unconducive to decomposition during much of year, and arbuscular mycorrhizas (AM) reported to be more prevalent in benign climates (Steidinger et al. 2019). There are also frequent reports of ECM association with soils that are poor in nitrogen (N) (Read 1991; Tsai et al. 2014; Rowlingson et al. 2022), or poor in nutrients in general (Wardle 1991; Niiyama et al. 1999; Corrales et al. 2018).

Although environment clearly influences fitness of different nutrient acquisition syndromes, these syndromes themselves can also alter local nutrient cycling. Several lines of evidence indicate AM and ECM associations can have different effects on soil nutrient dynamics. Some lineages of ECM fungi are able to extract nitrogen (N) and possibly other nutrients from soil organic matter (Read and Perez-Moreno 2003; Pellitier and Zak 2018; Romero-Olivares et al. 2021). This competition for N with saprophytes may slow decomposition (Gadgil and Gadgil 1971), increase the size of soil carbon C pools (Orwin et al. 2011; Averill et al. 2014), depress nutrient mineralization rates, and thus reduce availability of N (and potentially other nutrients) to plants that lack ECM associations (Corrales et al. 2016; Lin et al. 2017, 2018). This short-circuiting of the nitrogen cycle may thus set up a positive feedback mechanism favouring continued local dominance by ECM plants, resulting in sharp ecotones between AM and ECM-dominant vegetation in the absence of obvious discontinuities in the physical environment, and raising the possibility of mycorrhiza-mediated alternative stable states (Corrales et al. 2016; Lu and Hedin 2019; Lusk and Atatoa Carr 2023).

ECM effects on soil C and N dynamics are not uniformly observed, and it is not clear how these effects vary with environment or with plant taxonomy (Fernandez and Kennedy 2016). Data obtained by comparing soil chemistry under patches of AM and ECM trees in natural vegetation can be difficult to interpret, because of the possibility of sorting of AM and ECM species along pre-existing fine-scale environmental gradients (Nadelhoffer et al. 1983; Boerner and Koslowsky 1989; Hall et al. 2020). Comparisons of adjacent plantations of known history provide more incisive tests, and have frequently found lower N availability under ECM plantations than AM plantations (Lin et al. 2017). The ECM tree species included in the studies reviewed by Lin et al. (2017) belong predominantly to two northern-hemisphere families, Pinaceae and Fagaceae.

In the humid temperate forests of the southern hemisphere, Nothofagus (Nothofagaceae) is the only ECM genus that occurs widely in mature stands (Wardle 1991; Godoy and Marín 2019). Although the effects of Nothofagus on soil C and N dynamics have not been examined to date, its tendency to form monodominant stands (Wardle 1984) is reminiscent of the behaviour of some ECM tropical trees that appear to elevate soil C:N ratios and depress N mineralization (Corrales et al. 2016; Hall et al. 2020). Here we compare soil nutrient availability beneath replicated ~ 30-year-old stands of Chilean native Nothofagus and Cupressaceae (AM) planted on a relatively homogenous lowland site in south-central Chile. We asked (1) if ECM plantations of Nothofagus dominance has resulted in higher soil total C and C:N, and lower nutrient availability than beneath the adjacent AM plantations of Cupressaceae; and (2) if understorey floristic composition indicates different successional trajectories in the ECM and AM stands, foreshadowing the development of alternative stable states.

Materials and methods

Study site

The plantations we studied were established between 1989 and 1995 by Prof. Claudio Donoso (Universidad Austral de Chile) at Senderos del Bosque (39.867 S, 73.157 W) on the outskirts of Valdivia in south-central Chile. A suite of small plantations of Chilean native tree species were established at between 60 and 100 m asl on a west-facing slope. Slopes were mostly in the 15–20° range. The red clay soils at the site belong to Los Ulmos soil series (CIREN 1999), range in depth from 50 to 100 cm, and are developed over Permian-Triassic schists. Valdivia has an oceanic temperate climate, with a mean annual temperature of 11.7 °C and a mean annual precipitation of 1,750 mm (Dirección Meteorológica de Chile 2001). Seasonal thermal amplitude is low: the mean daily high of the warmest month (February) is 23.8° C, and the mean daily low of the coldest month (July) is 4.0 °C. Seasonal variation in precipitation is much more pronounced, with > 60% falling during the winter months of May to August. Summer is the driest season, both January and February averaging < 50 mm. Snowfalls are rare.

The vegetation present on the site at the time plantations were established had been severely degraded by a long history of wood-cutting, leaving a low cover (2–3 m tall) of the bamboo Chusquea quila (Poaceae) and shrubs such as Ugni molinae (Myrtaceae) and Gaultheria spp. (Ericaceae), with scattered emergent trees up to 10 m tall, mostly Laurelia sempervirens (Atherospermaceae) and Gevuina avellana (Proteaceae). In less disturbed native forest stands on similar sites near Valdivia, the main canopy species are the evergreens Aextoxicon punctatum (Aextoxicaceae), Eucryphia cordifolia (Cunoniaceae), L. sempervirens, Luma apiculata (Myrtaceae) and Podocarpus saligna (Podocarpaceae), with emergent deciduous Nothofagus obliqua (Nothofagaceae) and evergreen N. dombeyi.

Establishment of plantations

A mosaic of 0.3 to 0.5 ha plantations of a range of Chilean native timber species was established on the site. Individual plantations were sited opportunistically on gentle spurs, avoiding gullies. These included four Nothofagus spp. stands (ECM) and three Cupressaceae stands (AM) that we later used for our study (Table 1; Fig. 1). The planted Nothofagus species were N. dombeyi and N. obliqua, and the Cupressaceae were Austrocedrus chilensis and Fitzroya cupressoides. Small gaps were opened in the shrub cover to enable planting of seedlings on a 2 m x 2 m grid, equivalent to a density of 2,500 plants ha−1, in accordance with standard Chilean forestry practice (Donoso et al. 1998). Bare-rooted two-year-old seedlings were planted; seedlings of Nothofagus were two years old and typically 60–80 cm tall, and those of Cupressaceae were three years old and 30–40 cm tall. Six of the plantations were monospecific, the seventh being a mixture of two Nothofagus species (Table 1). Plantations were not thinned, pruned, or weeded; and no fertilizer was applied.

Table 1 Plantations established at Senderos del Bosque that were used in the present study
Full size table
Fig. 1
figure 1

Spatial arrangement of plantations of Cupressaceae (Austrocedrus chilensis, Fitzroya cupressoides) and Nothofagus (N. dombeyi, N. obliqua) at Senderos del Bosque, Chile. See Table 1 for explanations of plantation names

Full size image

Although non-target tree species had become abundant in some plantations by the time of sampling, they did not compromise our intended comparison of AM- versus ECM-dominant stands. Growth of residual L. sempervirens, and new recruitment of fast-growing species such as Embothrium coccineum (Proteaceae), resulted in considerable contributions of non-target species to the basal area of some plantations (Table 1). Most notably, basal area of the planted species (F. cupressoides) on plot FITCUP2 was exceeded by that of L. sempervirens. However, ectomycorrhizal trees (Nothofagus) contributed negligible basal area to the three Cupressaceae plantations (Table 1). AM tree species contributed > 85% to the basal areas of all three of these plantations, non-mycorrhizal Proteaceae comprising the remainder. Similarly, ECM species contributed at least 68% of the basal area of all four Nothofagus plantations (Table 1). AM Chusquea bamboos were present in the understories of all stands; although we did not quantify their abundance, visual estimates suggested up to c. 20% cover of Chusquea spp. in the Cupressaceae stands, and somewhat less in the Nothofagus stands.

Soil analyses

In 2021 we obtained soil samples from beneath each of the seven plantations, to assess similarity of total soil nutrient capital and pH. Given that we did not have initial soil data from 30 years previously, we assume that soils were overall similar among stands with AM trees and with ECM trees before plantations. The long axis of most plantations ran approximately east to west; soil was sampled on three north-south transects randomly spaced between 20 and 40 m apart along this east-west axis. At four randomly spaced points on each transect, a 20 mm diameter soil augur was used to extract the uppermost 120 mm of the mineral soil. Soil from these 12 points was homogenized and submitted as a single sample from each plantation.

Soils were dried at 60° C for 72 h, and sent for analysis at Hill Laboratories, Hamilton, New Zealand. pH was measured potentiometrically in a 1:2 (v/v) soil:water slurry. Organic C and total N were estimated by Dumas combustion. Total P was estimated by nitric/hydrochloric digestion, followed by inductively coupled plasma-optical emission spectroscopy.

Mineral nutrient availability

Plant Root Simulator (PRS®) probes (Western Ag Innovations Inc., Saskatoon, Canada) were used to measure ion availability in the mineral soil beneath each of the seven plantations. The probes consist of 10 cm2 ion exchange resin membranes held in plastic supports that are inserted vertically into the soil. Anion probes have a positively-charged membrane to attract and adsorb all negatively-charged anions, and cation probes a negatively charged membrane to adsorb all positively-charged ions.

PRS® probes were buried for two eight-week periods. The first burial period (15/08/21 to 14/10/21) was intended to represent early growing season conditions when mineralization rates are likely to be temperature-limited. The second burial period (14/01/22 to 11/03/22) was intended to represent late growing season conditions when mineralization rates are likely to be limited primarily by water availability, as the summer months of January and February are also the driest months in Valdivia.

A pair of PRS® probes (anion + cation) was buried at each of the same sampling points where soil samples were taken. This meant four pairs of probes were buried along each of three transects in each plantation during each burial period. Probes were inserted into slots cut in the soil with a garden knife, leaving the c. 1 cm of the plastic support protruding from the soil surface. Soil was tamped closely around the probe, to ensure soil contact with the ion exchange membrane. As the ion exchange membrane begins 6 cm below the top of the plastic support, most or all of the membrane was in contact with the mineral soil, rather than the surface organic horizons. Brightly-coloured flagging was attached to the top of each probe, to facilitate retrieval eight weeks later.

We successfully retrieved 330 (98.2%) of 336 buried probes. After retrieval, probes were cleaned with a nailbrush and ultrapure water, and submitted to WesternAg Innovations for analysis. The three pairs of probes inserted on each transect were treated as a single sample: they were submitted in the same bag to the laboratory, where they were eluted together. This meant four estimates of the availability of each ion were obtained from each plantation during each burial period.

Understorey composition

Understorey composition was surveyed on the same plots used for measurement of stand basal area (Table 1). We recorded frequencies of arborescent species ≥ 100 cm tall and < 5.0 cm diameter, excluding Chusquea bamboos.

Statistics

All statistical analysis was performed in R (R core team 2022). The overall effect of dominant taxon (Nothofagus vs. Cupressaceae) and season (spring vs. summer) on mineral nutrient availability was visualised using principal components analysis (function prcomp with scale = TRUE), and tested using redundancy analysis (rda in package vegan, Oksanen et al. 2019). As redundancy analysis did not account for non-independence of transects within stands, we took the geometric mean value of the pooled probes for each of three transects within each stand at each season for this analysis. As environmental differences associated with the spatial arrangement of our stands could also potentially influence nutrient availability, we also tested the effect of spatial distance using RDA. As we found no significant effect of spatial distance (P = 0.134), this variable was not explored further in our analyses or interpretations.

After finding overall significant effects of taxon and season in redundancy analysis, we tested the significance of their effects on availability of each nutrient. A linear mixed effect model with a random term was used to account for the three replicate samples within each combination of taxon and season. “Taxon” was treated as a categorical variable (Nothofagus vs. Cupressaceae). We used model simplification, using likelihood ratio tests to determine the significance of each term (function anova on lme models) for each soil nutrient. As the residuals of many of the soil measurements were non-normally distributed, we rank-transformed all variables before tests.

We used Permutational multivariate analysis of variance (PERMANOVA, vegan package) to test for differences in understorey composition between Nothofagus and Cupressaceae plantations. Non-metric multidimensional scaling (NMDS, vegan package) was used to provide a visual representation of variation in understorey composition.

Results

Soil analyses

Plantations of Cupressaceae and Nothofagus did not differ significantly in soil C, N or P content (Table 2). According to the criteria described by Blakemore et al. (1987) for similar south-temperate ecosystems in New Zealand, total N levels beneath the plantations were moderate to high; although C:N ratios were also rather high. Total P levels were low to moderate according to the same criteria. Average soil pH beneath Nothofagus stands (5.2) was significantly lower than that measured beneath Cupressaceae stands (5.7).

Table 2 Overall chemical properties of mineral soil (top 120 mm) beneath each of the seven plantations sampled at Senderos del Bosque, Chile. Plantations of Cupressaceae and Nothofagus differed significantly only in pH
Full size table

Mineral nutrient availability measured with PRS® probes

We removed cadmium from our dataset, as availability of this element was generally below the method detection limit. The remaining ions measured by the probes were ammonium (NH4+), nitrate (NO3−), calcium (Ca2+), magnesium (Mg2+), potassium (K+), phosphate (PO43−), iron (Fe2+/Fe3+), manganese (Mn2+), copper (Cu2+), zinc (Zn2+), boron (B3+), sulphate (SO42−), lead (Pb2+) and aluminium (Al3+).

Redundancy analysis showed significant effects of taxon (Cupressaceae vs. Nothofagus: F1,11 = 2.56, P = 0.046) and season (F1,11 = 4.70, P = 0.005) on mineral nutrient availability, the two constrained axes explaining 39.8% of variance. This was reflected in the PCA analysis (Fig. 2), which showed a strong separation by season. One of the four Nothofagus stands (NOTSPP) clustered with Cupressaceae stands in the ordination, while the other three Nothofagus stands clustered together. Neither N nor P availability contributed strongly to either of the first two principal components (Table 3). Two variables contributed strongly to the second component: Ca2+ (0.550) and Mn2+ (-0.506).

Fig. 2
figure 2

PCA ordination of nutrient availability data obtained from soil beneath seven plantations at Senderos del Bosque, Chile. PRS® probes were used to measure availability of the following mineral nutrients over two 8-week periods during the 2021-22 growing season: NH4+, NO3−, Ca, Mg, K, P, Fe, Mn, Cu, Zn, B, S, Pb and Al. PC1 represents 37.4 and PC2 17.3% of the variance, all other axes explain < 10% of variance

Full size image
Table 3 Loadings of each soil nutrient on the first two PCA axes, which explained 67.4% of total variation in nutrient availability beneath Nothofagus and Cupressaceae plantations at Senderos, Chile
Full size table

In univariate analysis, season was retained as a significant predictor in the best models for all soil variables except Mn and B (Table 4). Dominant taxon was only retained in the best model for three soil variables: Ca, Mn, and as part of a significant interaction with season for NH4+. Ca availability was higher beneath Cupressaceae stands, and Mn was higher beneath Nothofagus stands (Fig. 3). Availability of both Ca and Mn was higher in spring than in summer (Fig. 3), as was true of most other ions (Table S1).

Table 4 Significance of variables retained in the best model for each soil nutrient
Full size table
Fig. 3
figure 3

Significant responses of soil nutrients to dominant taxon in plantations at Senderos del Bosque, Chile. Units are µg adsorbed by the 10 cm2 ion exchange membrane of PRS probes during an 8-week burial period

Full size image

Understorey composition

Understorey composition differed rather idiosyncratically across individual plantations (Table S2). Despite a lack of overlap between convex hulls fitted to the NMDS ordination (Fig. 4), PERMANOVA did not show a significant difference between average understorey composition of Nothofagus and Cupressaceae plantations (R2 = 0.24, F = 1.57, P = 0.183). The commonest understorey species were (in descending order) Luma apiculata, Embothrium coccineum (Proteaceae), Eucryphia cordifolia and Gaultheria mucronata; these four species accounting for > 54% of all saplings on the seven plots (Table S2).

Fig. 4
figure 4

Non-metric multidimensional scaling analysis of understorey composition beneath plantations of native ectomycorrhizal Nothofagus species (red) and arbuscular mycorrhizal Cupressaceae (black). The low stress value of 0.047 indicates a satisfactory representation in two dimensions

Full size image

Discussion

We found that mineral nutrient availability beneath plantations of Chilean Cupressaceae and Nothofagus had become stoichiometrically distinct within 30 years (Fig. 2). Lower Ca availability in the soil A horizon beneath Nothofagus stands might reflect leaching of base cations as a result of acidification (Binkley and Valentine 1991), though it is unclear why similar losses of K and Mg were not apparent (Table 2). Similarly, a study of 50-yr old plantations reported significantly higher Ca availability beneath AM Fraxinus than beneath ECM Picea (Binkley and Valentine 1991), and a comparison of 34-yr old plantations of four conifer species found Ca availability to be highest beneath the only AM species, Thuja plicata (Prescott et al. 2000). Alternatively, differences in Ca availability beneath Cupressaceae and Nothofagus stands could be actively driven by differences in Ca cycling. Kiilsgaard et al. (1987) found that Ca concentrations in leaf litter of AM North American conifers (Cupressaceae and Taxodiaceae) were two- to three-fold higher than those of ECM Pinaceae.

Our finding of higher Mn availability beneath Nothofagus stands than beneath AM-dominant stands (Fig. 3) may reflect Mn peroxidase activity associated with ECM fungi. Mn peroxidase is one of the main ligninolytic enzymes produced by Basidiomycetes (Hofrichter 2002), and some ECM lineages including Cortinarius have retained genes coding for Mn peroxidase production (Chen et al. 2001; Pellitier and Zak 2018). Mn peroxidase oxidizes the manganese (II) ions (Mn2+) in wood and soil organic matter to highly reactive Mn3+, which is stabilized by fungal chelators such as oxalic acid. Chelated Mn3 + in turn attacks phenolic lignin structures (Hofrichter 2002). Mn peroxidase may be one of the primary drivers of the organic N economy that often develops in ECM forests (Lindahl and Tunlid 2015; Sterkenburg et al. 2018; Pellitier and Zak 2021).

Despite overall differences in the stoichiometry of soil nutrient availability, and probable evidence of elevated Mn peroxidase activity beneath Nothofagus stands, we found no clear evidence of differences in soil C and N dynamics beneath AM and ECM stands. There was no evidence of a depression of mineral N availability beneath Nothofagus stands (Table 3; Fig. 3), or of elevated C:N ratios (Table 2). Furthermore, understorey composition showed little evidence of divergent successional trajectories beneath Nothofagus and Cupressaceae plantations (Fig. 4, Table S2). Ericaceous shrubs are common in the understories of Nothofagus forest on low-N sites New Zealand (Wardle 1984; Rowlingson et al. 2022), and at high elevations and high latitudes in Chile (Cuevas 2000; Romanyà et al. 2005; Fuentes-Ramirez et al. 2020). This may reflect the ability of ericoid mycorrhizal fungi to access organic N sources (Read and Perez-Moreno 2003), in common with some ECM fungi. Although one species of Ericaceae (Gaultheria mucronata) was well-represented in understories at Senderos, more individuals of this species were recorded beneath AM Cupressaceae stands than beneath ECM Nothofagus (Table S2). Plantation understorey composition at Senderos thus did not appear to portend development of alternative stable states beneath AM vs. ECM plantations.

The limited evidence of an organic N economy at Senderos del Bosque might reflect the relatively N-rich character of the soil there. Modelling by Orwin et al. (2011) predicted nutrient immobilisation due to ECM mining of nutrients from soil organic matter on sites with low fertility, but not on sites with higher initial nutrient levels – a prediction upheld by subsequent empirical work (Mayer et al. 2023). Under this scenario, we would expect that Nothofagus will be transient on N-rich soils, undergoing successional replacement by more shade-tolerant trees. There is some evidence that ECM fungal lineages with low peroxidase activity (e.g., Cenococcum) impose lower carbon costs on their host plants than lineages that are more capable of accessing organic N sources using peroxidases (e.g. Cortinarius, Ramaria) (Hupperts and Lilleskov 2022). Host plants may therefore allocate carbon preferentially to fungi that mine N from soil organic matter only when N is in short supply. Although the elevated levels of Mn beneath Nothofagus stands at Senderos del Bosque are consistent with the presence of ECM fungi capable of enzymatic liberation of N from soil organic matter, lineages well-endowed with this capacity may not be the dominant component of local ECM communities.

A second possibility is that organic nutrient use by ectomycorrhizal fungi requires a more extended period to effect significant N immobilization, and hence influence stand dynamics. Trenching experiments suggest that switches from organic to mineral N economies can occur within 12 months in some situations (Gadgil and Gadgil 1971, 1975), but less is known about switches in the opposite direction. Comparisons of small, ~ 90 year-old forestry plots in the Morton Arboretum in Illinois found that topsoil C:N ratios were significantly higher beneath ECM trees than beneath AM species (Midgley and Sims 2020); however, ammonium levels and N mineralization were higher beneath ECM trees, counter to what would be expected if ECM fungi were competing for N with saprophytes. It may be relevant that the soils studied by Midgley and Sims (2020) had relatively high total N (0.44–0.54%) and fairly low C:N ratios (12.9–14.9). Like our study site at Senderos del Bosque, the Morton Herbarium might thus represent another situation where development of a strong organic N economy is unlikely (Orwin et al. 2011).

A third possibility is that leaf trait differences between the target species masked the impact of ECM fungi on C and N cycling. We did not measure leaf nutrient concentrations at our study sites, but sympatric comparisons elsewhere in Chile show higher leaf N in Nothofagus species than in Austrocedrus or Fitzroya (Lusk and Matus 2000; Lusk et al. 2003). Any N-depletion of organic matter by ECM fungi may therefore have been offset by higher initial litter N concentrations in the Nothofagus plantations, resulting in similar overall decomposition and mineralization rates in the two stand types. This third possibility seems the most promising explanation, as it reconciles the evidence of organic matter degradation by ECM fungi (elevated Mn availability beneath Nothofagus plantations) with the similarity of soil total C and N levels beneath the two stand types (Table 2).

The lower pH beneath Nothofagus stands (Table 2) may reflect a general tendency of ectomycorrhizal associations to acidify soils. Comparative studies of plantations have reported that pH is generally lower beneath ectomycorrhizal Pinaceae and Fagaceae than under neighbouring stands of arbuscular mycorrhizal Acer, Fraxinus and Thuja (Binkley and Valentine 1991; Prescott et al. 2000; Augusto et al. 2002), although France et al. (1989) reported the opposite pattern on a site subject to acid deposition from upwind industries. The generality of the effect of mycorrhizal associations on soil pH was confirmed by a comparative study of 30 tree species in an arboretum, reporting averages of 6.3 vs. 5.6 beneath AM and ECM species, respectively (Midgley and Sims 2020).

In conclusion, we found little evidence that the establishment of AM and ECM stands at Senderos del Bosque will give rise to alternative stable states. Although elevated Mn levels beneath Nothofagus plantations were consistent with organic matter degradation by ECM fungi (Fig. 3), the similarity of N availability and organic C beneath Nothofagus and Cupressaceae plantations may indicate that any impact on decomposition and mineralization rates was offset by countervailing leaf trait differences between the dominant tree species. The understories of the ~ 30-yr old Nothofagus plantations at Senderos del Bosque are populated largely by more shade-tolerant AM and non-mycorrhizal tree species (Table 1) that would eventually replace the Nothofagus overstorey if it were left unharvested – as is seen in the normal course of successional development in lowland south-central Chile (Veblen et al. 1996). An organic N economy is likely to have greater influence at higher elevations and higher latitudes in Chile, where ericaceous shrubs often grow in the understories of Nothofagus pumilio stands, on soils with higher C:N ratios (Cuevas 2000; Romanyà et al. 2005; Fuentes-Ramirez et al. 2020).