Long-term and realistic global change manipulations had low impact on diversity of soil biota in temperate heathland

Article metrics

Abstract

In a dry heathland ecosystem we manipulated temperature (warming), precipitation (drought) and atmospheric concentration of CO2 in a full-factorial experiment in order to investigate changes in below-ground biodiversity as a result of future climate change. We investigated the responses in community diversity of nematodes, enchytraeids, collembolans and oribatid mites at two and eight years of manipulations. We used a structural equation modelling (SEM) approach analyzing the three manipulations, soil moisture and temperature, and seven soil biological and chemical variables. The analysis revealed a persistent and positive effect of elevated CO2 on litter C:N ratio. After two years of treatment, the fungi to bacteria ratio was increased by warming, and the diversities within oribatid mites, collembolans and nematode groups were all affected by elevated CO2 mediated through increased litter C:N ratio. After eight years of treatment, however, the CO2-increased litter C:N ratio did not influence the diversity in any of the four fauna groups. The number of significant correlations between treatments, food source quality, and soil biota diversities was reduced from six to three after two and eight years, respectively. These results suggest a remarkable resilience within the soil biota against global climate change treatments in the long term.

Introduction

Human activity on our planet has caused global changes of atmospheric chemistry and climate. These changes are likely to amplify in the near future due to increasing concentrations of greenhouse gases such as carbon dioxide (CO2) in the atmosphere, and will result in increased temperature and altered precipitation patterns including more intense summer droughts over large parts of the Earth’s surface1,2. Such profound shifts of atmosphere chemistry and climate will fundamentally affect key drivers of ecosystem functioning and lead to changes in terrestrial ecosystems across the globe3,4. Each of these changes may affect ecosystem functioning, e.g. through CO2-stimulated photosynthesis and plant growth5, warming-induced increase in nutrient mineralization6,7 or drought-induced mortality and growth limitation of plants and animals8,9. The number of such single-factor studies has grown considerably, but studies considering simultaneous action of multiple climate change drivers are still relatively sparse. However, understanding their co-action on terrestrial ecosystems is of paramount importance in order to increase accuracy of global change predictions10,11. Aboveground responses to climate change have been more comprehensively studied in terrestrial ecosystems, whereas belowground organisms and processes have received less attention. There is an urgent need for obtaining insights into the mechanisms and patterns of belowground biota response to climate change, and especially quantifying combined action of different climate change agents.

Elevated CO2 levels are unlikely to have direct negative effects on belowground organisms, since soil biota are adapted to much higher CO2 concentrations in the soil pore space than found in the atmosphere. Most studies have shown that increased atmospheric CO2 concentrations affect soil biota positively12, probably due to increased primary production13, which would lead to increased litter input, root growth and root exudation, the major food sources for decomposer food webs14,15. Moreover, elevated CO2 improves the water use efficiency by plants and hence potentially increase the soil water content. Such conditions are also beneficial for most soil organisms16,17.

Increased temperature is likely to extend the growing season of plants in temperate regions with additional positive effects on root productivity, microbial biomass and activity11,18,19,20. Although such changes could potentially lead to increased food supply relevant for various soil fauna groups, most studies suggest that moderate temperature increase ascribed to climate change in the near future will have little impact on soil fauna12,21,22,23.

Increased occurrence and intensity of summer droughts may induce dramatic effects on various soil biota groups. This is especially true for soil invertebrates dependent on free water in the soil pores. Thus, enchytraeids are sensitive to decreased water content in soil, whereas microarthropods are probably more resistant to drought16,24,25,26,27,28. Apart from the direct effects of drying, negative effects of drought on plant growth and root exudation may also lead to reduced food availability for many decomposer organisms12.

Changes in the environmental conditions as described above may improve or worsen the chances for survival and growth of each species’ population in the given ecosystem and consequently change diversity status of the whole belowground community. There is growing evidence of the link between biodiversity of soil fauna and the level of ecosystem services delivery. This is particularly true for carbon sequestration and cycling29,30,31. It is therefore important from the perspective of belowground community functionality to evaluate the impact of climatic changes on soil fauna biodiversity12,32.

Recently, significant advances in our understanding of the causality of shifts in belowground ecological processes associated with environmental changes have been achieved by the use of Structural Equation Modeling (SEM) that allows parameterizing of complex models with many parameters and latent variables33,34,35. For example, the effects of drought, increased CO2 and nitrogen addition on soil food webs have been analyzed using SEM10,36.

In this study we quantified the simultaneous co-action of multiple environmental drivers usually associated with climate change on the structure and functionality of belowground fauna. In a field trial, we exposed a semi-natural Danish heathland to elevated atmospheric CO2 concentration (CO2), extended summer drought (drought) and increased temperature (warming) and their combinations in a full factorial experimental setup. The full combination of these treatments aimed to mimic climate predictions for Denmark year 2075 as closely as possible37. In addition to a comparison of un-treated control and the full combination of treatments, we used SEM analysis to investigate the effects of the climate treatments on the biodiversity (using the Shannon index) of various soil invertebrate groups, i.e. the interrelations between taxonomic or functional diversity of major soil invertebrate taxa (nematodes, enchytraeids, collembolans and oribatid mites), quality of their putative food sources and the main abiotic drivers (soil temperature and water content). We analyzed how the three manipulated factors have influenced biodiversity after two years of treatment, and after eight years of treatment, respectively. We hypothesized that (i) elevated CO2 would increase litter C:N ratio, fungi:bacteria ratio and microbial C:N ratio, which in turn would increase biodiversity of soil fauna through trophic interactions, (ii) reduction of soil water content would reduce biodiversity, and (iii) that moderately increased temperatures would have no direct effects on biodiversity of soil biota, but decrease the soil water content and therefore indirectly decrease biodiversity of soil invertebrates. Finally, (iv) we expected that climate change effects on soil biota were persistent and increasing with time, in particular for organisms with long generation times.

Results

Soil fauna communities

The nematode community was dominated by bacterial feeders (in particular belonging to the genera Acrobeloides, Cervidellus, Monhystrella, Plectus and Teratocephalus; data not shown) and plant parasitic species (especially the genera Paratylenchus and Helicotylenchus; data not shown). Fungal feeders (mostly the genera Aphelenchoides and Tylencholaimus; data not shown) and carni-omnivorous species (dominated by the genus Aporcelaimellus; data not shown) were less abundant (Table S1).

The enchytraeid community at the field site was dominated by Chamaedrilus chlorophilus which represented about 75% of the total population (Table S2). Three other species, Achaeta affinis, Oconnorella cambrensis and Enchytronia parva were also relatively abundant, and eight other species occurred sporadically.

We identified about 20 species of Collembola (Table S3a,b). The community was dominated by euedaphic and hemiedaphic species such as Mesaphorura sp., Megalothorax minimus, Parisotoma notabilis, and Protaphorura sp. Larger epedaphic species (e.g. Tomocerus sp. and Orchesella sp.) did occur, but were much less abundant. Species richness of Collembola was similar in 2007 and 2013 (Table S3a,b).

About 35 species of oribatid species were recorded in 2007 (Table S4a). The most abundant species were found among the genera Galumna, Microppia, Scheloribates and Suctobelbella. Species richness of oribatid mites was higher in 2013 than in 2007, mainly due to the appearance of numerous oribatids from the Brachychthoniidae family (Table S4b).

Structural Equation Modeling

The significant relationships of the full model for the two years are shown graphically in Figs 1 and 2 and the summary statistics of the fitted structural equation models for the observations made two (October 2007) and eight years (April 2013) after the start of the experiment are shown in Table S5a and b, respectively. The fit of the SEM converged normally and the fitting properties were assessed using the standardized root mean square residual statistic (SRMR) which range from zero to one, where zero indicates a perfect fit and 0.08 is judged to be acceptable38. The SRMRs were estimated as 0.06 and 0.080 in 2007 and 2013 (N = 48 in both years), respectively, and consequently the fit was judged to be acceptable.

Figure 1: Structural equation model of effects of elevated CO2, drought and warming (exogenous factors; pink boxes) on physical soil parameters (blue boxes) and on litter C:N ratio, fungi:bacteria ratio, soil microbial C:N ratio and diversity (Shannon index) of Collembola, Oribatid and Nematoda communities (green ovals) in a Danish heathland exposed to two years of climate change manipulation.
figure1

Litter, microbial parameters and fauna was sampled in autumn 2007. Numbers on arrows are standardized regression coefficients. Regression coefficients marked with ** or *** are highly significant (P < 0.01 and P < 0.001, respectively) and coefficients marked with * are significant (0.01 < P < 0.05) with N = 48; black arrows indicate positive relationships, red arrows indicate negative relationships. Non-significant correlations are not shown (full model is shown in Table S5a).

Figure 2: Structural equation model of effects of elevated CO2, drought and warming (exogenous factors; pink boxes) on physical soil parameters (blue boxes) and on litter C:N ratio, fungi:bacteria ratio, soil microbial C:N ratio and diversity (Shannon index) of Collembola, Oribatida, Enchytraeidae and Nematoda communities (green ovals) in a Danish heathland exposed to eight years of climate change manipulation.
figure2

Litter, microbial parameters and fauna was sampled in April 2013. Numbers on arrows are standardized regression coefficients. Regression coefficients marked with ** or *** are highly significant (P < 0.01 and P < 0.001, respectively) and coefficients marked with * are significant (0.01 < P < 0.05) with N = 48; black arrows indicate positive relationships, red arrows indicate negative relationships. Non-significant correlations are not shown (full model is shown in Table S5b).

Since the model was based on prior ecological knowledge of the causal relationships we did not perform any model reduction of non-significant relationships. This choice was made in order not to perturb the a priori specified partitioning of variance in the structural equation model by committing type II errors due to insufficient sample sizes. In other words, if an arrow is falsely removed due to a relatively low sample size, then the variation explained by the removed arrow is unjustifiably “forced” into another causal pathway.

Soil water content and temperature

The three exogenous variables, i.e. the CO2, warming and drought treatments, generally had significant effects on the measured soil temperature and soil water content of the plots (Figs 1 and 2; Table S5a,b). As expected, the drought and warming treatments both had significant negative effects on soil water content in both years. The warming treatment had a positive effect on soil temperature, however, the effect was statistically significant only in 2007. CO2 had a significant positive influence on soil temperature in both years although the effect was not large (about 0.1 °C; Table S6). It is difficult to give a mechanistic explanation of this effect, but it is hardly a realistic consequence of climate change and perhaps rather an artefact of the experimental setup. For 2007 data, the model explained 89% of the variation in soil water content during drought, 32% of mean annual soil water content, and 52% of the variation in soil temperature (see R2-values in Table S5a). For 2013 data, the model explained 63% of the variation in soil water content during drought, 37% of mean annual soil water content, but only 13% of the variation in soil temperature (Table S5b).

Litter and microbial parameters

The litter C:N ratio was significantly increased by elevated CO2 (models explained about 20% of variation in both years), but was not affected by any other variable (Table S5a,b). The fungi:bacteria ratio was significantly correlated with soil temperature in 2007 with 17% of the variation explained by the model, but this correlation had disappeared in 2013. The microbial C:N ratio was not influenced by CO2, soil water content or soil temperature. The litter C:N was higher in October 2007 than in April 2013, probably due to the different seasons when samplings took place (Table S7). The fungi:bacteria ratio was lower in 2007 than in 2013 and was not correlated with CO2 or other variables (Figs 1 and 2; Table S7).

Biodiversity

The nematode functional diversity (i.e. the calculated Shannon index using feeding groups as categories) was significantly and positively correlated with litter C:N, and negatively correlated with microbial C:N in 2007. A substantial part of the variation in nematode functional diversity was explained by the model (39%). The positive correlation with litter C:N was mainly driven by increased abundance (data not shown) and proportion of carni-omnivorous nematodes at high litter C:N (Fig. 3) without an increase in the total nematode population (data not shown). The negative correlation between nematode functional diversity and microbial C:N in 2007 was not explained by the abundance of any particular feeding group (Table S1). In 2013, there was a significant correlation of nematode functional diversity with CO2, but correlations with litter C:N or microbial C:N had disappeared. Moreover, the model explained only 24% of the variation in nematode functional diversity, i.e. much less than in 2007, and the effect of CO2 was relatively small (Table S1).

Figure 3: Proportion of carni-omnivorous nematodes plotted against litter C:N ratio (October 2007).
figure3

Solid line indicates a significant positive correlation (P = 0.0007; N = 48).

In 2007, biodiversity of collembolans correlated with litter C:N, but this correlation was not present in 2013 (Figs 1 and 2). In 2013, we found a significant (positive) correlation between the biodiversity of collembolans and mean annual soil water content (Fig. 2). This effect seemed to be associated with a decrease of epedaphic species at the lower soil water contents, but the model explained only a minor part of the variation of collembolan diversity (16%; Table S5b).

In 2007, the biodiversity of oribatids was negatively correlated with litter C:N (Fig. 1), and the model explained 31% of the variation (Table S5a). The effect of litter C:N was quite strong with decreases of about 40% of biodiversity on average across the range of litter C:N values (Fig. 4). It was, however, not possible to identify particular species that were responsible for this relationship (Table S4a). In 2013 (Table S4b), we found no significant correlations between the biodiversity of oribatids and any of the explanatory variables (Fig. 2; Table S5b). The biodiversity of the enchytraeid community was not assessed in 2007, and in 2013 it did not correlate with any physical or biological parameters of the model (Fig. 2).

Figure 4: Shannon diversity index of oribatid mite communities in 2007 plotted against litter C:N ratio.
figure4

Solid line indicates a significant negative correlation (P = 0.0008; N = 48).

In addition to the SEM analysis we compared biodiversity of ambient plots and plots that had received the full combination of treatments for eight years (i.e. TDCO2). Using a one-way ANOVA we noted that TDCO2 did not differ significantly from ambient plots for any of the diversity indices after two years or after eight years of treatments (Fig. 5).

Figure 5: Shannon diversity indices (mean ± s.d.; N = 6) in October 2007 (upper panel) and April 2013 (lower panel) of Enchytraeidae (only in 2013), Collembola, Oribatida and functional diversity of Nematoda (based on feeding groups) for ambient plots (control; green columns) and plots receiving a combination of elevated CO2, increased drought and warming (TDCO2; orange columns).
figure5

No significant differences between treatments were found.

Discussion

It is quite unique to have detailed information on diversity within several groups of soil organisms rich in species as well as in individuals. To our knowledge it is the first time that data on diversity of nematodes, enchytraeids, collembolans, and oribatid mites are simultaneously analyzed in a multifactorial global change experiment. Moreover, by performing this detailed assessment of organism diversity twice (two and eight years after field manipulations started) we can analyze how stable the global change effects on soil biota are over a longer time scale. The fauna groups studied have small body sizes and relatively low mobility and we judge field plots to be sufficiently large to avoid noteworthy emi- or immigration and edge effects from neighbouring plots receiving a different treatment.

We stated earlier that each of the exogenous factors to some extent had unintentional influence on physical conditions of soil, hence precluding a traditional general linear model (nested) ANOVA approach where the treatments are assumed to be orthogonal. Indeed, we observed that experimental warming caused a significant (albeit not very pronounced) reduction of soil water content. The effect of warming causing decreased soil moisture was, however, apparently not related to increased temperature (and thus evaporation) per se, but probably more related to reduced dew formation during night-time due to the reflective curtains used for night-time passive warming. Lastly, FACE had a slight, but significant effect increasing the soil temperature. These observations further justify the SEM approach.

Other studies in which temperature have been manipulated more efficiently than in our study have demonstrated that warming can have influence – most often negative - on abundance and diversity of various soil fauna groups12,22,39. It has been proposed that warming favors short-lived and fast-reproducing species more than longer-lived species resulting in altered community composition and diversity40. The warming treatment in our study had limited effect on mean soil temperature (less than 1 °C on average) which may be a major reason why we did not observe direct effects of temperature on biodiversity. However, the SEM modeling indicated that experimental warming had an indirect negative effect on the biodiversity of collembolans through reduced soil water content.

The increase in the diversity and abundance of oribatid mites from October 2007 to April 2013 may be the result of seasonal fluctuations rather than a general increase of diversity over this time period. The high proportion of Brachychthoniidae and Oppiidae species in April 2013 may be due to massive hatching of their eggs during the well-known spring-peak of oribatid abundance41. However, the recorded differences in the overall abundance between two sampling occasions are within the known range of abundance and diversity variance across different seasons and years42,43.

Several other field studies similar to ours have shown that reduction of precipitation significantly decreased belowground biodiversity and abundance of soil fauna12, but some of these studies tended to have used more dramatic reductions of precipitation44,45 than we have. We note that drought intensity did not have direct influence on biodiversity when assessed months after the end of dry conditions, whereas we did observe an effect of mean annual soil water content on the diversity of collembolans. Our previous studies have shown that drought can cause substantial and acute negative effects on abundance of nematodes, collembolans and enchytraeids in our study area and similar ecosystems25,36,46,47. Biomass of enchytraeids, as determined shortly after experimental drought episodes, was reduced when water content (0–20 cm depth) was approaching 5 volume %, but the population recovered within months after the soil water content was restored25,36,48. The enchytraeid community at the study site was heavily dominated by C. chlorophilus, but about ten other species were also recorded. Despite severe reduction in numbers and biomass during droughts, we did not detect significant correlations between soil moisture and biodiversity of the enchytraeid community (neither soil water content during drought, nor annual mean water content). This may indicate that the enchytraeid community present at the study site is well adapted to occasional drought episodes, but it cannot be ruled out that some species preferring moist or wet conditions in the long run will become reduced in abundance or completely disappear in a drier environment as shown in another long-term climate change experiment9.

It has been reported that reduced stomatal conductance of plants under CO2 enrichment is most significant under drought regimes49 and that this may lead to a better water use efficiency and lessened depletion of soil water. In the present experiment we found no evidence that increased aerial CO2 concentrations led to increased soil moisture in the experimental plots. Thus, the decreased biodiversity of collembolans in dryer soils observed in 2013 was not related to CO2 enrichment, but rather related to legacy of the experimental drought periods and a small, but consistent, drying effect of the warming treatment. The positive correlation between soil water content and biodiversity of collembolans in 2013 was mainly driven by an increase in species richness (linear regression; P = 0.02; data not shown). The biodiversity of oribatids was not influenced by soil moisture in our experiment, perhaps due to their higher desiccation resistance as compared to other fauna groups26, and that our drought treatment was moderate. Studies in which drought was longer-lasting show that also oribatid communities can be severely impacted by drought and that recovery of biodiversity after severe drought incidents can take several years due to the relatively slow reproduction of these animals compared to other microarthropods such as collembolans44,45,50.

Despite the lack of an overall effect of the future climate scenario (i.e. TDCO2) on biodiversity, we were able to detect separate effects of increased CO2 (after two years of treatments) on biodiversity of oribatid mites and collembolans, and on functional diversity of nematodes. Increased atmospheric CO2 increased the C:N ratio of litter, and decreased the biodiversity of oribatids downstream, whereas diversity of collembolans increased. Elevated CO2 also stimulated abundance of long-lived and carni-omnivorous nematodes, and thus functional diversity of nematodes, likely due to an increased root biomass and root exudation which in turn increased the biomass of microbes and prey that this group of nematodes is feeding on. In 2013, elevated CO2 still affected nematode functional diversity, but, interestingly, at this point the CO2 effect was independent of litter C:N ratio. Nematodes are to a larger degree associated with the rhizosphere than oribatids, collembolans or enchytraeids and therefore more responsive to quantitative and qualitative root exudation CO2 responses. The litter independent CO2 effect on nematode functional diversity may therefore reflect increased root exudation at elevated CO2 as reported elsewhere51,52.

We detected no effects of CO2 on enchytraeid biodiversity although litter chemistry and other food-related parameters were clearly affected by CO2 with imprints on the physiology of enchytraeids at our field site36,46. Thus, a previous study of N-content of soil fauna at the same field experiment showed that elevated CO2 reduced enchytraeid nitrogen concentration, whereas collembolans were not affected, and oribatid mites had higher nitrogen concentration than at ambient CO236. This may reflect differences in the feeding biology between oribatid mites, enchytraeids and other soil fauna taxa, and indicates that aerial CO2 enrichment can have direct or indirect influence on animal nutritional physiology which may potentially manifest itself at higher levels of biological organization such as species abundance and diversity. This suggests that at elevated CO2, soil animals tend to be more nitrogen-limited and switch to alternative food resources which are N-enriched, e.g. shifting from direct detritus to predominantly fungal or bacterial feeding.

Despite the fact that we could identify significant separate effects of soil moisture and CO2 on soil animals, we saw no differences in biodiversity of any focal soil fauna taxon between ambient plots and plots receiving the full combination of treatments representing a likely climatic scenario for Denmark in year 2075. This reveals a pronounced resistance potential of belowground communities to the upcoming climatic shifts which might be explained by the high thermal and hygric buffering capacity of soil53 and redundancy of soil food webs32,54. Our study of effects of a persistent global change on soil biota over a period from few to many generations of the longest living members is quite unique. Reported recovery of soil fauna communities after disturbances show various levels of resilience ranging from decades after mining activity55 or years after extended drought50, to months following burning56,57,58 and moderate drought59,60.

The SEM performed on the same type of data obtained two and eight years after initiating global change manipulations showed that interactions from experimental manipulations via soil physical environment and litter chemistry to soil biota were markedly reduced over this period. The six significant interactions observed after two years were six years later reduced to only three significant interactions. The CO2 effect on litter C:N persisted for eight years of manipulation, however, litter C:N effect on diversity of oribatids, collembolans and nematodes after two years had disappeared after eight years. Eight years since the experiment started, the diversity of collembolans and nematodes depended directly on soil moisture and CO2 treatment, respectively. This was not what we expected and suggests that soil fauna has a marked resilience initially being strongly affected by abrupt environmental changes, and then later adjusting to the pre-manipulation state.

Although it could be argued that the present ecosystem did not have a high biodiversity to begin with, it is remarkable that species composition within four taxa belonging to the soil food web is only in the short term affected by a stable global change-induced shift in litter chemistry. When global change in this study did not have permanent effect on composition of the soil fauna, the important question that appears is: which factors are drivers of changes in soil fauna diversity? Although soil faunal diversity was only temporarily affected by the projected global change parameters, long-term global change did affect the functionality of the soil at the field site. For instance, labile carbon-induced soil organic matter decomposition (priming) was enhanced after long-term elevated CO2 treatment61. Hence, the resilience of soil biotic diversity should not lead us to conclude that in the future climate change scenario soil functionality will not be affected. Further, in a larger context, our study represents merely one particular future climate scenario in a local environment. We must therefore be cautious in attempting to extrapolate our findings to larger regions and climate change scenarios. However, the very clear result of our study suggests that belowground biodiversity can be quite resilient in responding to a realistic climatic scenario. Moreover, we detected separate effects of increased CO2 and soil moisture on the biodiversity of soil fauna, but these effects were not detected when the treatments were combined.

Methods

Site description and experimental design

The experimental site was a heath/grassland ecosystem dominated by grass (Deschampsia flexuosa) and an evergreen dwarf shrub (Calluna vulgaris). The soil was a coarse textured sandy arenosol from the Weichsel glaciation with only weak signs of podsolization. The annual mean temperature (1961–1990) of this area (Brandbjerg, Denmark; 55°30′ N, 11°80′ E), was 8.0 °C and the annual mean precipitation amounted to 613 mm37. The experiment was designed to mimic a realistic climate scenario for Denmark in year 2075, with elevated atmospheric CO2 (CO2), warming (T) and extended summer drought (D). Atmospheric CO2 concentration was increased during daytime hours in a Free Air CO2 Enrichment (FACE) setup (ambient (380 ppm) + 130 ppm = 510 ppm). Warming was done using automatic, passive night-time warming curtains, which overall for the period 2006–2013 increased mean monthly night-time temperatures in 20 cm height by 0.4 and 1.1 Kelvin during winter and summer months, respectively. In 5 cm soil depth, the corresponding increases were 0.17 and 0.35 Kelvin, respectively. Although these temperature increases were moderate it significantly increased the number of degree days in warmed plots (by 30 to 70 degree days in soil, and mostly during the growth season) during all experimental years (Fig. S1). The drought treatment was active during a 4–6 week campaign each year usually during early summer using automatic rainout curtains removing 5–11% of annual precipitation. The experimental setup was full factorial in a split-plot design, i.e. including all possible combinations of the three treatments. All combinations of treatments were replicated 6 times (i.e. 23 treatments × 6 replicates; N = 48). T and D treatments were nested within the CO2 treatment, i.e. six octagons (6.8 m diameter) were treated with elevated atmospheric CO2 while six octagons had ambient CO2 conditions (A). Each octagon was divided into four ‘slices’ (9.1 m2 per plot) to provide all eight treatment combinations in each pair of octagons (Fig. 6). See also Mikkelsen et al.37 for further details. We did not use the treatments as such, but the derived moisture and temperature effects in the SEM analysis (see later description of statistical analysis). The plots contained a mixture of patches with Deshcampsia and Calluna, but our investigations here consider only areas dominated by Deschampsia. In order to reduce edge effects as much as possible soil samples were taken at random, but at least 50 cm from the edge of the plot. The distances between the octagons were at least 2.5 times the octagon width to avoid CO2 contamination from the elevated to the ambient CO2 octagons. Boardwalks connected the octagons to avoid disturbance by trampling between the plots and a flexible boardwalk system in each octagon provided easy access and prevented trampling within the plots. Within each experimental octagon or plot, parameters were monitored to check the treatments and their effects on the physical and climatic conditions. Temperature data used in this study were measured in the soil (–5 cm) in each treatment plot every minute and averaged over hourly intervals. Soil moisture was measured by TDR at 0–20 cm soil depth providing half-hour averages.

Figure 6
figure6

(a) The layout of a single block of the experiment. The direction of extension of the curtains that provided the night-time warming (T) and drought (D) treatments is shown on the upper octagon. One octagon is at ambient [CO2] (380 ppm) while the other received elevated [CO2] (510 ppm) from the Free Air Carbon Enrichment (FACE) system. (b) The arrangement of the field experiment at Brandbjerg, Denmark. The octagons are labelled 1 through 12, with consecutive pairs belonging to the same experimental block. Octagons 2, 4, 5, 8, 10 and 12 received elevated [CO2] and are circled. The boardwalks are shown by solid lines and the locations of two meteorological stations are shown as M1 and M2. The two rectangles represent buildings that housed computers, control systems and field laboratories.

Explanatory variables

We employed SEM allowing parameterizing relations between environmental parameters and biodiversity of various soil animal groups. Thus, two of the three applied treatments (warming and drought) were not entered as categorical variables; instead we analyzed how treatments influenced soil temperature and moisture directly and used these parameters as explanatory variables. As discussed above, the applied treatments (exogenic factors) sometimes had indirect effects on physical conditions in the soil (e.g. CO2 increased soil temperature; warming decreased soil water content) hence confounding a traditional general linear model ANOVA approach. We therefore calculated plot-specific mean soil temperature (5 cm depth) during the 12 month period preceding sampling of biotic parameters in October, 2007, and in April, 2013, respectively. Likewise, plot-specific mean soil water contents (mean SWC) were calculated for the same periods. Finally, mean soil water content during the last week of the latest applied drought period before sampling of biotic parameters (June 15–22, 2007, and June 4–11, 2012, respectively; mean drought SWC) were used as a measure of the intensity of the latest drought application. A summary of these explanatory variables are shown in Supplementary Information Table S6.

Measurements of biotic components

Measurements of several biotic components were made in October 2007, about two years after climate treatments began in October 2005. Measurements were repeated in April 2013, eight years after treatments began. Some of these results have been published elsewhere (see Supplementary Information), but a brief description of the sampling methodology follows here.

Nematodes

Nematodes were extracted from between 10 and 40 g (fresh weight) of soil by a modified Baermann tray extraction method62. Samples were extracted for 48–72 h, and nematodes were, based on the morphology of mouthparts and the oesophagus63, then classified to different trophic groups (predacious, omnivorous (these two groups were grouped as carni-omnivorous), herbivorous, fungivorous and bacterivorous) and counted at x40 magnification using a dissecting microscope. Based on the relative abundances of these trophic groups we calculated the Shannon index as a measure of “functional diversity” of the nematode community.

Enchytraeids

Enchytraeids were sampled using a cylindrical soil corer with an inner diameter of 5.5 cm to a depth of 9 cm. Each soil core was divided into layers 3 cm thick with a knife. The samples were kept in plastic beakers at 5 °C until extraction which was initiated within two weeks after collection. The extraction was a modified version of O’Connor’s wet funnel extraction64 performed at room temperature over 6 hours. Extracted enchytraeids were identified in vivo based on the key and the handling procedure described in Schmelz and Collado65. Based on the relative abundances of species we calculated the Shannon index for enchytraeids.

Microarthropods

Microarthropods were sampled using a cylindrical soil corer with the inner diameter of 6 cm to a depth of 10 cm. Each soil core was divided in the field into two layers, each 5 cm thick with a knife. The samples were kept in closed plastic cylinders at 5 °C until extraction which was initiated within two weeks after collection. They were extracted in a high gradient extraction apparatus (MacFadyen type), where the temperature in the upper compartment increased stepwise from 25 °C to 50 °C within seven days while the temperature at the lower compartment remained constant at 3 °C66. The microarthropods were collected in benzoic acid and subsequently conserved and stored in glycerol until identification. Collembolans were identified to species level according to Fjellberg67,68. Oribatid mites were identified to species or genus level as described in detail by Zaitsev et al.69. Based on the relative abundances of species we calculated the Shannon index for Collembola and Oribatida, respectively.

Fungi:bacteria ratio, litter and microbial C:N ratio

Responses in biodiversity of fauna was related to the quality of the litter input and microbial biomass by assessing the fungi:bacteria ratio, and the C:N ratio of the soil microbial biomass and litter in each plot. Fungi:bacteria ratio was based on copy numbers of the internal transcribed spacer (ITS) region of fungal ribosome-encoding genes, and the bacterial 16S rRNA gene, respectively, as quantified by qPCR. Details of the method are published elsewhere70. Soil microbial C:N ratio was determined as follows. Fresh soil from each plot was extracted directly, and a parallel soil sample after vacuum-incubation, with chloroform for 24 hours71,72. A digestion with H2SeO3, H2SO4 and H2O2 was followed by the NH4+ analysis (indophenol-blue reaction). Total microbial N (MicN) was calculated as N in the fumigated sample minus N in the non-fumigated sample, using 0.4 as the extractability factor73. Analysis for carbon (dissolved organic, DOC) was done using a Shimadzu TOC 5000A analyzer (Shimadzu, Kyoto, Japan). Total microbial C (MicC) was calculated as DOC in the fumigated sample minus DOC in the non-fumigated sample, using 0.45 as the extractability factor74. Microbial C:N ratio was MicC divided by MicN. Plant litter was sampled in September-October 2007 and in June 2013 in mixed vegetation. The air dried litter was analyzed for total C and N by Dumas combustion (1020 °C) on an elemental analyzer (EA Flash 2000, Thermo Fisher Scientific, Milan Italy).

Structural equation modeling

In structural equation models (SEM) a hypothesis of the causal relationships between the measured variables in a system is fitted to observed data34. The main use of an SEM is to quantify which of possible direct or indirect causal pathways that is most important in regulating a variable, and whether the influences of different factors on a variable are positive or negative. Generally, a SEM does not prove the hypothesized causal relationships, but it is possible to test if specific causal relationships are supported by data. The measured abiotic and biotic variables were arranged in a causal network according to prior knowledge of soil ecological cause-effect relationships (Figs 1 and 2; Table S5a,b). The prior ecological knowledge was specified by applying generally established ecological theory that for many of the specific causal relationships had been studied at the experimental site36,46. Since the measurements made in 2007 and in 2013, and the number of replicates (N = 48), were largely the same, we had the possibility to compare and infer from data the appearance of resilience over time. Some of the variables were log-transformed in order to achieve an approximate normal distribution and some of the variables were rescaled so that all variables had an approximately equal scale. From graphical inspection of the distribution of the variable, it was judged whether or not to log-transform the variable. After rescaling, the variables were fitted to the structural equation model using the software package lavaan75 with all the nodes and edges of the prior model. In order to increase precision and take into account the effect of covariance between soil temperature and soil water content, the measured environmental states of the plots were used to predict soil fauna diversity rather than the exogenous treatment levels (0 or 1) of temperature and drought.

Additional Information

How to cite this article: Holmstrup, M. et al. Long-term and realistic global change manipulations had low impact on diversity of soil biota in temperate heathland. Sci. Rep. 7, 41388; doi: 10.1038/srep41388 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. 1

    IPCC. Working Group I Contribution to the IPCC Fifth Assessment Report Climate Change 2013: The Physical Science Basis. Summary for Policymakers. (IPCC, 2013).

  2. 2

    Orlowsky, B. & Seneviratne, S. I. Global changes in extreme events: regional and seasonal dimension. Climatic Change 110, 669–696, doi: 10.1007/s10584-011-0122-9 (2012).

  3. 3

    Parmesan, C. & Yohe, G. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42, doi: 10.1038/nature01286 (2003).

  4. 4

    Walther, G. R. et al. Ecological responses to recent climate change. Nature 416, 389–395, doi: 10.1038/416389a (2002).

  5. 5

    Luo, Y. Q., Hui, D. F. & Zhang, D. Q. Elevated CO2 stimulates net accumulations of carbon and nitrogen in land ecosystems: A meta-analysis. Ecology 87, 53–63, doi: 10.1890/04-1724 (2006).

  6. 6

    Emmett, B. A. et al. The response of soil processes to climate change: Results from manipulation studies of shrublands across an environmental gradient. Ecosystems 7, 625–637, doi: 10.1007/s10021-004-0220-x (2004).

  7. 7

    Björsne, A.-K., Rutting, T. & Ambus, P. Combined climate factors alleviate changes in gross soil nitrogen dynamics in heathlands. Biogeochemistry 120, 191–201, doi: 10.1007/s10533-014-9990-1 (2014).

  8. 8

    Peñuelas, J. et al. Response of plant species richness and primary productivity in shrublands along a north-south gradient in Europe to seven years of experimental warming and drought: reductions in primary productivity in the heat and drought year of 2003. Global Change Biology 13, 2563–2581, doi: 10.1111/j.1365-2486.2007.01464.x (2007).

  9. 9

    Holmstrup, M. et al. Increased frequency of drought reduces species richness of enchytraeid communities in both wet and dry heathland soils. Soil Biology & Biochemistry 53, 43–49, doi: 10.1016/j.soilbio.2012.05.001 (2012).

  10. 10

    Eisenhauer, N., Cesarz, S., Koller, R., Worm, K. & Reich, P. B. Global change belowground: impacts of elevated CO2, nitrogen, and summer drought on soil food webs and biodiversity. Global Change Biology 18, 435–447, doi: 10.1111/j.1365-2486.2011.02555.x (2012).

  11. 11

    Larsen, K. S. et al. Reduced N cycling in response to elevated CO2, warming, and drought in a Danish heathland: Synthesizing results of the CLIMAITE project after two years of treatments. Global Change Biology 17, 1884–1899, doi: 10.1111/j.1365-2486.2010.02351.x (2011).

  12. 12

    Blankinship, J. C., Niklaus, P. A. & Hungate, B. A. A meta-analysis of responses of soil biota to global change. Oecologia 165, 553–565, doi: 10.1007/s00442-011-1909-0 (2011).

  13. 13

    Amthor, J. S. Effects of atmospheric CO2 concentration on wheat yield: review of results from experiments using various approaches to control CO2 concentration. Field Crops Research 73, 1–34 (2001).

  14. 14

    Arnone, J. A. et al. Dynamics of root systems in native grasslands: effects of elevated atmospheric CO2 . New Phytologist 147, 73–86 (2000).

  15. 15

    Couteaux, M. M. & Bolger, T. Interactions between atmospheric CO2 enrichment and soil fauna. Plant and Soil 224, 123–134 (2000).

  16. 16

    Rapoport, E. & Tschapek, M. Soil water and soil fauna. Revue d'Écologie et Biologie du Sol 4, 1–58 (1967).

  17. 17

    Niklaus, P. A. et al. Six years of in situ CO2 enrichment evoke changes in soil structure and soil biota of nutrient-poor grassland. Global Change Biology 9, 585–600 (2003).

  18. 18

    Beier, C., Schmidt, I. K. & Kristensen, H. L. Effects of climate and ecosystem disturbances on biogeochemical cycling in a semi-natural terrestrial ecosystem. Water, Air & Soil Pollution Focus 4, 191–206 (2004).

  19. 19

    Prieto, P. et al. Changes in the onset of spring growth in shrubland species in response to experimental warming along a north-south gradient in Europe. Global Ecology and Biogeography 18, 473–484, doi: 10.1111/j.1466-8238.2009.00460.x (2009).

  20. 20

    Arndal, M. F. et al. Net root growth and nutrient acquisition in response to predicted climate change in two contrasting heathland species. Plant and Soil 369, 615–629, doi: 10.1007/s11104-013-1601-8 (2013).

  21. 21

    Hågvar, S. & Klanderud, K. Effect of simulated environmental change on alpine soil arthropods. Global Change Biology 15, 2972–2980, doi: 10.1111/j.1365-2486.2009.01926.x (2009).

  22. 22

    Haimi, J. et al. Impacts of elevated CO2 and temperature on the soil fauna of boreal forests. Applied Soil Ecology 30, 104–112, doi: 10.1016/j.apsoil.2005.02.006 (2005).

  23. 23

    Kardol, P., Reynolds, W. N., Norby, R. J. & Classen, A. T. Climate change effects on soil microarthropod abundance and community structure. Applied Soil Ecology 47, 37–44, doi: 10.1016/j.apsoil.2010.11.001 (2011).

  24. 24

    Bardgett, R. D. The Biology of Soil. (Oxford University Press Inc., 2005).

  25. 25

    Maraldo, K. et al. The counteracting effects of elevated atmospheric CO2 concentrations and drought episodes: Studies of enchytraeid communities in a dry heathland. Soil Biology & Biochemistry 42, 1958–1966, doi: 10.1016/j.soilbio.2010.07.015 (2010).

  26. 26

    Wauthy, G. & Vannier, G. Application of an analytical model measuring tolerance of an oribatid community (Acari) to continuously increasing drought. Experimental & Applied Acarology 5, 137–150, doi: 10.1007/bf02053823 (1988).

  27. 27

    Vannier, G. The porosphere as an ecological medium emphasized in Professor Ghilarov’s work on soil animal adaptations. Biology and Fertility of Soils 3, 39–44 (1987).

  28. 28

    Holmstrup, M. et al. Soil microarthropods are only weakly impacted after 13 years of repeated drought treatment in wet and dry heathland soils. Soil Biology & Biochemistry 66, 110–118, doi: 10.1016/j.soilbio.2013.06.023 (2013).

  29. 29

    Nielsen, U. N., Ayres, E., Wall, D. H. & Bardgett, R. D. Soil biodiversity and carbon cycling: a review and synthesis of studies examining diversity-function relationships. European Journal of Soil Science 62, 105–116, doi: 10.1111/j.1365-2389.2010.01314.x (2011).

  30. 30

    Wall, D. H. et al. Global decomposition experiment shows soil animal impacts on decomposition are climate-dependent. Global Change Biology 14, 2661–2677, doi: 10.1111/j.1365-2486.2008.01672.x (2008).

  31. 31

    de Vries, F. T. et al. Soil food web properties explain ecosystem services across European land use systems. Proceedings of the National Academy of Sciences of the United States of America 110, 14296–14301, doi: 10.1073/pnas.1305198110 (2013).

  32. 32

    Bardgett, R. D. & van der Putten, W. H. Belowground biodiversity and ecosystem functioning. Nature 515, 505–511, doi: 10.1038/nature13855 (2014).

  33. 33

    Damgaard, C., Strandberg, M., Kristiansen, S. M., Nielsen, K. E. & Bak, J. L. Is Erica tetralix abundance on wet heathlands controlled by nitrogen deposition or soil acidification? Environmental Pollution 184, 1–8, doi: 10.1016/j.envpol.2013.07.047 (2014).

  34. 34

    Grace, J. B., Anderson, T. M., Olff, H. & Scheiner, S. M. On the specification of structural equation models for ecological systems. Ecological Monographs 80, 67–87, doi: 10.1890/09-0464.1 (2010).

  35. 35

    Grace, J. B., Harrison, S. & Damschen, E. I. Local richness along gradients in the Siskiyou herb flora: R. H. Whittaker revisited. Ecology 92, 108–120, doi: 10.1890/09-2137.1 (2011).

  36. 36

    Vestergård, M., Dyrnum, K., Michelsen, A., Damgaard, C. & Holmstrup, M. Long-term multifactorial climate change impacts on mesofaunal biomass and nitrogen content. Applied Soil Ecology 92, 54–63 (2015).

  37. 37

    Mikkelsen, T. N. et al. Experimental design of multifactor climate change experiments with elevated CO2, warming and drought: the CLIMAITE project. Functional Ecology 22, 185–195, doi: 10.1111/j.1365-2435.2007.01362.x (2008).

  38. 38

    Hooper, D., Coughlan, J. & Mullen, M. R. Structural equation modelling: guidelines for determining model fit. The Electronic Journal of Business Research Methods 6, 53–60 (2008).

  39. 39

    Briones, M. J. I., Ineson, P. & Heinemeyer, A. Predicting potential impacts of climate change on the geographical distribution of enchytraeids: a meta-analysis approach. Global Change Biology 13, 2252–2269 (2007).

  40. 40

    Lindo, Z. Warming favours small-bodied organisms through enhanced reproduction and compositional shifts in belowground systems. Soil Biology & Biochemistry 91, 271–278 (2015).

  41. 41

    Krivolutsky, D. A. Oribatid mites 224 pp. (Nauka Publishers, Moscow, 1995) (In Russian).

  42. 42

    Lindo, Z. & Winchester, N. N. Oribatid mite communities and foliar litter decomposition in canopy suspended soils and forest floor habitats of western redcedar forests, Vancouver Island, Canada. Soil Biology & Biochemistry 39, 2957–2966, doi: 10.1016/j.soilbio.2007.06.009 (2007).

  43. 43

    Gergócs, V., Garamvölgyi, A., Homoródi, R. & Hufnagel, L. Seasonal change of oribatid mite communities (Acari, Oribatida) in three different types of microhabitats in an oak forest. Applied Ecology and Environmental Research 9, 181–195 (2011).

  44. 44

    Lindberg, N. & Bengtsson, J. Population responses of oribatid mites and collembolans after drought. Applied Soil Ecology 28, 163–174 (2005).

  45. 45

    Lindberg, N., Bengtsson, J. B. & Persson, T. Effects of experimental irrigation and drought on the composition and diversity of soil fauna in a coniferous stand. Journal of Applied Ecology 39, 924–936 (2002).

  46. 46

    Andresen, L. C. et al. Organic matter flow in the food web at a temperate heath under multifactorial climate change. Rapid Communications in Mass Spectrometry 25, 1485–1496, doi: 10.1002/rcm.4907 (2011).

  47. 47

    Stevnbak, K. et al. Suppression of soil decomposers and promotion of long-lived, root herbivorous nematodes by climate change. European Journal of Soil Biology 52, 1–7, doi: 10.1016/j.ejsobi.2012.04.001 (2012).

  48. 48

    Holmstrup, M. et al. Responses of enchytraeids to increased temperature, drought and atmospheric CO2: Results of an eight-year field experiment in dry heathland. European Journal of Soil Biology 70, 15–22, doi: 10.1016/j.ejsobi.2015.06.004 (2015).

  49. 49

    Field, C. B., Jackson, R. B. & Mooney, H. A. Stomatal responses to increased CO2: implications from the plant to the global scale. Plant Cell and Environment 18, 1214–1225, doi: 10.1111/j.1365-3040.1995.tb00630.x (1995).

  50. 50

    Lindberg, N. & Bengtsson, J. Recovery of forest soil fauna diversity and composition after repeated summer droughts. Oikos 114, 494–506, doi: 10.1111/j.2006.0030-1299.14396.x (2006).

  51. 51

    Pendall, E., Mosier, A. R. & Morgan, J. A. Rhizodeposition stimulated by elevated CO2 in a semiarid grassland. New Phytologist 162, 447–458, doi: 10.1111/j.1469-8137.2004.01054.x (2004).

  52. 52

    Phillips, R. P., Finzi, A. C. & Bernhardt, E. S. Enhanced root exudation induces microbial feedbacks to N cycling in a pine forest under long-term CO2 fumigation. Ecology Letters 14, 187–194, doi: 10.1111/j.1461-0248.2010.01570.x (2011).

  53. 53

    Zaitsev, A. S., Chauvat, M., Pflug, A. & Wolters, V. Oribatid mite diversity and community dynamics in a spruce chronosequence. Soil Biology & Biochemistry 34, 1919–1927, doi: 10.1016/s0038-0717(02)00208-0 (2002).

  54. 54

    Wolters, V. Biodiversity of soil animals and its function. European Journal of Soil Biology 37, 221–227, doi: 10.1016/s1164-5563(01)01088-3 (2001).

  55. 55

    Rufaut, C. G., Clearwater, S. & Craw, D. Recolonisation and recovery of soil invertebrate assemblages at an inactive coal mine in southern New Zealand. New Zealand Natural Sciences 35, 17–30 (2010).

  56. 56

    Abbott, I., Burbidge, T., Strehlow, K., Mellican, A. & Wills, A. Logging and burning impacts on cockroaches, crickets and grasshoppers, and spiders in Jarrah forest, Western Australia. Forest Ecology and Management 174, 383–399, doi: 10.1016/s0378-1127(02)00058-0 (2003).

  57. 57

    Zaitsev, A. S., Gongalsky, K. B., Malmstrom, A., Persson, T. & Bengtsson, J. Why are forest fires generally neglected in soil fauna research? A mini-review. Applied Soil Ecology 98, 261–271, doi: 10.1016/j.apsoil.2015.10.012 (2016).

  58. 58

    Podgaiski, L. R. et al. Burning effects on detritivory and litter decay in Campos grasslands. Austral Ecology 39, 686–695, doi: 10.1111/aec.12132 (2014).

  59. 59

    de Vries, F. T. et al. Legacy effects of drought on plant growth and the soil food web. Oecologia 170, 821–833, doi: 10.1007/s00442-012-2331-y (2012).

  60. 60

    Maraldo, K. & Holmstrup, M. Recovery of enchytraeid populations after severe drought events. Applied Soil Ecology 42, 227–235, doi: 10.1016/j.apsoil.2009.04.004 (2009).

  61. 61

    Vestergård, M., Reinsch, S., Bengtson, P., Ambus, P. & Christensen, S. Enhanced priming of old, not new soil carbon at elevated atmospheric CO2 . Soil Biology & Biochemistry 100, 140–148, doi: 10.1016/j.soilbio.2016.06.010 (2016).

  62. 62

    Georgieva, S., Christensen, S. & Stevnbak, K. Nematode succession and microfauna-microorganism interactions during root residue decomposition. Soil Biology & Biochemistry 37, 1763–1774, doi: 10.1016/j.soilbio.2005.02.010 (2005).

  63. 63

    Yeates, G. W., Bongers, T., Degoede, R. G. M., Freckman, D. W. & Georgieva, S. S. Feeding habits in soil nematode families and genera - an outline for soil ecologists. Journal of Nematology 25, 315–331 (1993).

  64. 64

    O’Connor, F. Extraction of enchytraeid worms from a coniferous forest soil. Nature 175, 815–816 (1955).

  65. 65

    Schmelz, R. M. & Collado, R. A guide to European terrestrial and freshwater species of Enchytraeidae (Oligochaeta). Soil Organisms 82, 1–176 (2010).

  66. 66

    Krogh, P. & Pedersen, M. Ecological effects assessment of industrial sludge for microarthropods and decomposition in a spruce plantation. Ecotoxicology and Environmental Safety 36, 162–168 (1997).

  67. 67

    Fjellberg, A. The Collembola of Fennoscandia and Denmark. Part II: Entomobryomorpha and Symphypleona. Vol. 42 (Brill, 2007).

  68. 68

    Fjellberg, A. The Collembola of Fennoscandia and Denmark. Part I: Poduromorpha. Vol. 35 (Brill, 1998).

  69. 69

    Zaitsev, A. S., Wolters, V., Waldhardt, R. & Dauber, J. Long-term succession of oribatid mites after conversion of croplands to grasslands. Applied Soil Ecology 34, 230–239, doi: 10.1016/j.apsoil.2006.01.005 (2006).

  70. 70

    Haugwitz, M. S. et al. Soil microorganisms respond to five years of climate change manipulations and elevated atmospheric CO2 in a temperate heath ecosystem. Plant and Soil 374, 211–222, doi: 10.1007/s11104-013-1855-1 (2014).

  71. 71

    Brookes, P. C., Landman, A., Pruden, G. & Jenkinson, D. S. Chloroform fumigation and the release of soil-nitrogen - a rapid direct extraction method to measure microbial biomass nitrogen in soil. Soil Biology & Biochemistry 17, 837–842, doi: 10.1016/0038-0717(85)90144-0 (1985).

  72. 72

    Joergensen, R. G. & Mueller, T. The fumigation-extraction method to estimate soil microbial biomass: Calibration of the k EN value. Soil Biology & Biochemistry 28, 33–37, doi: 10.1016/0038-0717(95)00101-8 (1996).

  73. 73

    Schmidt, I. K., Jonasson, S. & Michelsen, A. Mineralization and microbial immobilization of N and P in arctic soils in relation to season, temperature and nutrient amendment. Applied Soil Ecology 11, 147–160, doi: 10.1016/s0929-1393(98)00147-4 (1999).

  74. 74

    Schmidt, I. K. et al. Long-term manipulation of the microbes and microfauna of two subarctic heaths by addition of fungicide, bactericide, carbon and fertilizer. Soil Biology & Biochemistry 32, 707–720, doi: 10.1016/s0038-0717(99)00207-2 (2000).

  75. 75

    Rosseel, Y. lavaan: An R package for Structural Equation Modeling. Journal of Statistical Software 48, 1–36 (2012).

Download references

Acknowledgements

The present study was supported by The Villum Foundation. MH was supported by a Skou senior fellowship at Aarhus Institute of Advanced Studies. ASZ was supported by RSF grant 14-14-00894. The authors wish to thank all the technicians associated to the CLIMAITE consortium for excellent assistance and commitment to the project.

Author information

C.B., I.K.S., T.N.M., P.A., K.P., A.M., A.P., S.C. and M.H. designed the experiment. M.H., I.K.S., M.F.A., T.N.M., K.S.L., A.M., L.C.A., M.H.a., L.B., A.S.Z., S.G., M.D., M.V. and S.C. performed the experimental work. C.D. and M.H. analyzed the data. M.H., C.D., M.V., A.S.Z. and S.C. wrote the main manuscript text. All authors reviewed the manuscript.

Correspondence to Martin Holmstrup.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Supplementary information

Rights and permissions

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.