Introduction

Confronted with a world population growing at unprecedented pace, humankind strives to optimize agricultural practices to maximize yields (Smith et al. 2016). Following this increasing demand for food provisioning to humans and livestock as well as for natural resources in general, we can observe as a global trend that tracts of land are progressively converted into agriculturally managed systems that are dedicated to optimized production (Foley et al. 2005; Smith et al. 2016; Wheeler and Von Braun 2013). Thus, land-use intensification has long been identified as one of the major global change drivers, severely impacting terrestrial ecosystems worldwide (Maxwell et al. 2016; Newbold et al. 2015; Sala et al. 2000; Tilman et al. 2002). Such intensification processes include, for instance, a selection of high-yielding crop varieties and the restriction to a few top-performing forage species, the employment of heavy machinery for rigorous tillage, and short mowing intervals (Giller et al. 1997; Newbold et al. 2015; Tilman et al. 2002; Tsiafouli et al. 2015; Yagioka et al. 2015). This often implies a high use of mineral fertilizers and pesticides that further impair the integrity of such managed ecosystems (Marinari et al. 2000; Ramirez et al. 2010; Treseder 2008).

A growing body of literature questions the capability of these intensified systems to ensure food security in the long term, as they might be particularly susceptible to harvest losses due to climate extremes (Isbell et al. 2017; Lobell et al. 2008). Notably, such climate extremes are likely to become more frequent in Central Europe in the future, and climate change includes general warming trends of soil and air temperatures as well as altered seasonal precipitation patters leading to extended drought periods over summer, whereas spring and fall may receive considerably more rainfall (Doscher et al. 2002; IPCC 2007; Jacob and Podzun 1997; Kerr 2007; Meinke et al. 2010). Given that the strength of climate change effects may be mediated by land use (De Vries et al. 2012), this puts additional pressure on intensively managed, often low-diversity systems that might lack the ability to buffer or to recover after such disturbances in a reasonable timeframe (Lobell et al. 2008; Yachi and Loreau 1999). In this context, extensive management practices—both in grasslands as well as in farming systems—were identified as one promising avenue to mitigate detrimental climate change effects (De Vries et al. 2012; Isbell et al. 2017). Such extensive management practices largely refrain from using pesticides and/or mineral fertilizer, often representing more diverse mixtures or following specific crop rotation cycles that include green manuring (legumes) in every third year of the cycle. While extensive systems may lead to lower yields or slower pest control, their great benefits in the context of climate extremes are based on the fact that they include less disturbances and allow more complex and thus more diverse systems with a plethora of survival and coping strategies (e.g., functional redundancy or asynchrony) (Dı́az and Cabido 2001; Mazancourt et al. 2013) maintaining a multitude of ecosystem functions (Birkhofer et al. 2011; Goldenberg et al. 2018; Hautier et al. 2015; Isbell et al. 2015, 2017; Scialabba and Müller-Lindenlauf 2010).

The maintenance of the productivity and fertility of soils represents a key aspect of less-intensive and sustainable management of agricultural fields (Tilman et al. 2002). Soils harbor a significant portion of terrestrial biodiversity and are crucial for many ecosystem functions like nutrient cycling, crop yields, and pest control (Bardgett and van der Putten 2014; Brussaard et al. 2007; Wall et al. 2015). At the same time, soil biota are known to be highly susceptible to intensive agricultural management, thus putting their associated services at risk (Bengtsson et al. 2005; Culman et al. 2010; Tsiafouli et al. 2015). In addition, indirect effects on soil biota via changes in water-use efficiency, erosion, and overall soil quality were reported (Dendooven et al. 2012). Notably, land-use effects could also change the entire structure of soil food webs, e.g. by strongly decreasing the abundances of predators, which are known to be more vulnerable to disturbances than lower trophic levels (Murrell and Barton 2017; Voigt et al. 2007). As a consequence, predator–prey interactions that are crucial for biological pest control might be disrupted, subsequently impacting plant growth and crop yields. Organic farming was also shown to sustain high evenness of predators, hereby reducing the risk of pest outbreaks (Crowder et al. 2010; Murrell and Barton 2017; Wilby and Thomas 2002). Concurring climate change may superimpose such negative effects of intensive management on soil organisms with thus far widely unknown consequences.

Amongst the different groups of soil organisms, free-living nematodes occur in exceptionally high numbers and represent one of the most diverse metazoan groups in terrestrial ecosystems (Bernard 1992). They have a special position within the soil fauna for several reasons: first, they cover all major trophic groups of the soil food web ranging from detritivores to herbivores and predators (Ferris 2010; Yeates et al. 1993), which makes them particularly suitable for studying global change effects on different trophic levels within the same faunistic group. Second, they are useful bioindicators with their functional shifts providing valuable information on the state of an ecosystem, and thus allowing inferences to other biotic groups and soil health (Ferris et al. 2001). Third, herbivorous nematodes are known as important agricultural pests (Evans et al. 1993), making knowledge about their performance under different land-use and climate scenarios highly relevant for farmers. While nematodes are often comparatively unaffected by climate warming (Cesarz et al. 2015; De Long et al. 2016; Yeates et al. 2002), different land-use practices were shown to strongly impact nematode functional groups and thus nematode community structure (Postma-Blaauw et al. 2010; Tsiafouli et al. 2015; Yeates et al. 1999). As nematodes cover a wide range of life-history and feeding strategies, the different functional groups show vastly different responses to environmental disturbances and global change: while opportunistic nematodes, that are known as typical r-strategists or colonizers, might be able to benefit from enhanced plant growth and higher nutrient availability under land-use intensification, the more sensitive nematode groups consisting of K-strategists (i.e., persisters) may decline, resulting in simplified and less structured soil food webs (Ferris et al. 2001). Such shifts can be captured by nematode functional indices that allow drawing general conclusions about the state of the ecosystem (Bongers 1990). Our current knowledge of how interactive global change drivers (e.g., climate change and land use) affect soil nematode communities in managed agricultural systems remains scarce though.

Here, we investigate the effects of climate change in different land-use scenarios (managed grasslands and farmlands) on the diversity and abundance of free-living soil nematode functional groups and nematode functional indices. The study was conducted within the Global Change Experimental Facility (GCEF) in Bad Lauchstädt, Germany. With this large-scale experimental platform, climatic conditions predicted for the region for ~ 2070–2100 (referred hereafter as ‘future climate’) are simulated by altered temperature (ambient versus ambient +0.6 °C) and precipitation regimes (ambient versus ~ 20% reduction in summer, ~ 10% addition in spring and fall, respectively) using fully automatized fold-out roofs (Schädler et al. 2019). The manipulations reflect realistic scenarios that allow for inter-annual variability instead of highly controlled conditions. The climate treatments are crossed with five different land-use scenarios in a split-plot design (Fig. S1a) that all represent common local management practices (including a specific plant species pool, crop rotations, and management strategies/intervals). By covering two highly relevant seasons for soil processes in two consecutive years (spring and fall in 2015 and 2016), we were able to obtain a comprehensive picture of soil nematode responses to interacting global change drivers across seasons.

We hypothesized that (1) soil nematode diversity will be lowest at high levels of land-use intensity (i.e., in conventional farmlands and intensively used grasslands). Furthermore, (2) climate change effects on soil nematode communities will be most pronounced on those intensively managed land-use types, whereas the organic farmland and the two extensive grasslands might be able to buffer the detrimental effects of future climatic conditions due to a more sustainable, soil-protecting crop rotation in the case of the farmland and higher biodiversity in case of the grasslands. (3) Furthermore, we expected different responses within the nematode fauna to interacting global change drivers. While opportunistic nematodes might benefit from increased resource availability under intensive management (i.e., fertilization, high plant biomass) and ambient conditions, higher trophic groups in particular may decline under intensive management and future climatic conditions, thus leading to simplified, less structured communities.

Methods

Study design

The study site is located at the field research station of the Helmholtz-Centre for Environmental Research (UFZ) in Bad Lauchstädt, Germany (51° 22′ 60 N, 11° 50′ 60 E, 118 m a.s.l.), and was formerly used as an arable field (last crop cultivation in 2012). The site is located in the Central German dry area (Querfurter Platte) and characterized by a mean annual temperature of 8.9 °C (1896–2013) and a mean annual precipitation of 489 mm (1896–2013). The soil is a Haplic Chernozem with a humus layer reaching down to more than 40 cm depth. This highly fertile soil type was developed upon carbonatic loess substrates (around 70% silt and 20% clay content). The soil is known for its high water-retention capacity (nearly reaching the mean annual precipitation), ensuring comparatively low susceptibility to drought stress (Altermann et al. 2005). Within the upper 15 cm, pH values ranged from 5.8 to 7.5, total carbon and total nitrogen content varied between 1.71–2.09% and 0.15–0.18%, respectively.

The Global Change Experimental Facility was established in 2013 to study the interactive effects of climate change (including elevated temperature and changes in precipitation patterns) and land-use intensity on managed terrestrial ecosystems using realistic scenarios. Each land-use scenario represents a common local management type (rather than an experimental gradient) that entails several facets like plant diversity and community structure, management intervals, crop rotation, and fertilization, for which the responses to climate change can be assessed (see Table S1 for an overview of the land-use history of the site). The climate treatments were first applied in 2014 (spring 2014: start of temperature treatment; summer 2014: start of precipitation treatment). The experiment consists of 50 plots arranged in ten mainplots (Fig. S1a). The two experimental treatments are implemented in a split-plot design with the climate treatment carried out at the mainplot level (n = 10) and the land-use treatment implemented at the plot level (n = 50), randomly arranged within the mainplots. This results in five replicates of each combined experimental treatment (climate x land use). Thus, for each land-use type there are five plots with future climatic conditions and five plots with ambient climatic conditions that serve as a climate control. A relevant spatial scale is realized by a large plot size of 16 × 24 m that allows to employ equipment commonly used in the agricultural sector. Each plot has a buffer zone of 2 m to the eastern and western sides and 4.5 m to the southern and northern sides. Soil sampling took place in the inner plot area (15 × 12 m) on a specific transect for soil measurements (Fig. S1b). Five of the mainplots have ambient climatic conditions with steel constructions (5 m height) to control for potential infrastructure effects, such as microclimatic effects. The other five mainplots represent future climatic conditions, which means that their roof constructions include an irrigation system and mobile roof and side panels that can be closed via rain sensors and timers. The climate treatments were chosen based on a consensus scenario across several dynamic models for Central Germany for 2070-2100 [COSMO-CLM (Rockel et al. 2008), REMO (Jacob and Podzun 1997), and RCAO (Doscher et al. 2002)], which include higher inter-annual rainfall variability with longer drought periods over summer and increased precipitation in spring and fall. Furthermore, the mean annual temperature is predicted to rise up to 3 °C. The treatments are applied as realistic climate scenarios, which include natural variability instead of highly controlled conditions. Thus, the climate treatments are applied relative to the ambient conditions and allow for inter-annual variability. Over summer (June–August), the climate treatment includes a reduction of precipitation by 20% (closing of roof and side panels via rain sensors). In spring (March–May) and fall (September–November), the precipitation is increased by 10% by the use of an irrigation system that uses water from a large rain water reservoir (Fig. S2). Passive warming overnight, which increases air and soil temperature on average by ~ 0.6 °C, is realized during the active roofing phase (Table S2). In 2015, the roofing phase started on February 25th and ended on December 11th (roofs were closed 79% of night time, deviations were due to frost and wind). In 2016, the roofing phase started on March 23rd and ended on November 23rd (roofs were closed 82% of night time).

Within each mainplot, the five plots represent the five different land-use treatments, with different levels of land-use intensity: (1) conventional farmland; (2) organic farmland; (3) intensively mown grassland; (4) extensively mown grassland; and (5) extensively grazed grassland. The conventional and organic farmlands are managed according to commonly used management practices in the region. For the conventional farmland, this includes a crop rotation consisting of winter rape, winter wheat, and winter barley. Mineral fertilization as well as herbicides, fungicides, and insecticides are applied at realistic rates and frequencies. The organic farmland has significantly stricter regulations concerning fertilization (only allowing legumes for nitrogen fixation in every third year, and the application of rock phosphate (P–Ca–Mg) and patent kali (K–Mg–S) in the first year of the crop cycle) and pesticides (mainly mechanical weed control using harrows). A bipartite crop sequence is applied, in which alfalfa and white clover are intended to be planted in the first and fourth year of the cycle. The other crop cycles also include winter wheat and winter barley as on the conventional farmland. In addition, the three grasslands represent common management types in the region. The intensively managed grassland consists of typical forage species and is fertilized with mineral fertilizer and frequently mown (three times in 2015, four times in 2016), whereas the extensively managed grasslands are either moderately mown (twice in 2015 and 2016, no fertilization) or moderately grazed by sheep (twice in 2015, three times in 2016, by a group of 20 sheep grazing on each plot for 24 h). After each mowing event, the plant biomass is removed from the plots, as it would normally happen during hay harvest. On the extensively mown grassland, the mown biomass was left on the plots for some days to allow natural sowing. While the extensively used grasslands contain 53.1 ± 2.0 plant species per 9 m2, the intensively used grassland has 10.1 ± 3.6 plant species per 9 m2 (mean ± SD for 2015–2017). For details on the plant species pools, see Data S1 in the supplementary information.

Nematode sampling and analyses


Soil samples were taken on a transect within the inner plot area in May and October 2015 and 2016 using a steel corer (1 cm diameter; 15 cm deep). Seven subsamples per plot were homogenized, sieved at 2 mm, and stored at 4 °C. Nematode extraction was conducted with a modified Baermann method (Ruess 1995). Approximately 25 g of soil per plot were transferred to plastic vessels with a milk filter and a fine gauze (200 µm) at the bottom and placed in water-filled funnels. More water was added to saturate the soil samples and to ensure a connected water column throughout the sample and the funnel. Hence, nematodes migrated from the soil through the milk filter and the gauze into the water column and gravitationally settled at the bottom of a closed tube connected to the funnel. After 72 h at 20 °C, the nematodes were transferred to a 4% formaldehyde solution. All nematodes per sample were counted at 100 × magnification using a Leica DMI 4000B light microscope. Identification was conducted at 400 × magnification. For identification, sediment material from the bottom of each sample vial was extracted with a 2 ml plastic pipette and examined in temporary-mounted microscope slides. At least 100 well-preserved specimens (if available in the sample) were randomly selected and identified to genus (adults and most of the juveniles) or family level (juveniles), following Bongers (1988). Nematode taxa were then arranged into trophic groups (bacteria, fungal, and plant feeders, omnivores, and predators) (Okada et al. 2005; Yeates et al. 1993). Due to low densities, omnivorous and predatory nematodes were grouped into a combined feeding type for most analyses (but see Figs. S3–4 and Table S3 for treatment effects on individual groups). This is justified by the fact that both groups often respond similar to environmental disturbances, as they mainly consist of K-strategists and represent higher trophic levels (Bongers and Ferris 1999; Porazinska et al. 2003). In addition, the ratio between microbivores (bacteria feeders and fungal feeders) and plant feeders was calculated to evaluate the net effect of the nematode community on plant productivity. Ratios above 1.0 indicate that beneficial effects are stronger than negative effects (Eisenhauer et al. 2011; Wasilewska 1997). Nematodes were ordered according to the colonization-persistence gradient (c–p values) (Bongers 1990; Bongers and Bongers 1998) based on their life-history strategy (i.e., r- or K-strategists). Nematodes included in the cp-1 category are characterized by their short generation cycles and high fecundity. They mainly feed on bacteria. Nematodes assigned to the cp-2 class have longer generation times, lower fecundity, and consist of bacterivores and fungivores (Ferris et al. 2001). Both cp-1 and cp-2 groups are categorized as r-strategists. Nematodes belonging to cp-3 to cp-5 classes are K-strategist with longer generation times, higher trophic feeding levels and increasing sensitivity against disturbances (Ferris et al. 2001). The c–p scaling was used to calculate the Maturity Index (MI) of nematode communities according to the following formula:

$$MI = \mathop \sum \limits_{i = 1}^{n} \left( {\begin{array}{*{20}c} n \\ k \\ \end{array} } \right) v \left( i \right) \times f\left( i \right)$$

where v(i) is the c–p value of taxon i and f(i) is the frequency of that taxon in a sample. MI is used to assess environmental disturbance, soil health, and as an indicator of overall soil food-web complexity (Bongers 1990; Bongers and Bongers 1998). Furthermore, nematode taxa were assigned to functional guilds according to Ferris et al. (2001), which then served as a basis to calculate the Enrichment Index, Channel Index, and Structure Index to assess ecosystem functions. Functional guilds refer to the following trophic groups: bacterial feeders (BaX), fungal feeders (FuX), omnivores (OmX), and carnivores (CaX). The associated numbers (i.e., the x of the respective trophic group) refer to the c–p values mentioned above. The Enrichment Index (EI) indicates the responsiveness of the opportunistic bacteria (Ba1 and Ba2) and fungal feeders (Fu2) to food-web enrichment (Ferris et al. 2001) and is calculated as follows:

$${\text{EI}} = 100 \times \left[ {\frac{e}{e + b}} \right],$$

where e is the weighted frequency of Ba1 and Fu2 and b is the weighted frequencies of Ba2 and Fu2 nematodes (Ferris et al. 2001). The Channel Index (CI) reflects the nature of decomposition channels through the soil food web. High values indicate a predominant decomposition pathway of organic matter dominated by fungal-feeding nematodes, whereas low values refer to bacterial-dominated decomposition pathways (Ferris et al. 2001):

$${\text{CI}} = 100 \times \left[ {0.8 \times \frac{{{\text{Fu}}_{2} }}{{3.2 \times {\text{Ba}}_{1} + 0.8 \times {\text{Fu}}_{2} }}} \right],$$

where 0.8 and 3.2 represent the enrichment weightings for Fu2 and Ba1 nematodes (Ferris et al. 2001). The Structure Index (SI) provides information about the complexity of the soil food web. High values indicate structured soil food webs with higher interconnectance between multiple trophic levels, implying ecosystem stability, while low values indicate simplified soil food web and reflect environmental disturbance (Ferris et al. 2001):

$${\text{SI}} = 100 \times \left[ {\frac{s}{s + b}} \right],$$

where s is calculated as the weighted frequencies of Ba3–Ba4, Fu3–Fu4, Ca3–Ca5, and Om3–Om5 nematodes, and b represents the weighted frequencies of Ba2 and Fu2 nematodes (Ferris et al. 2001).

In addition, nematode taxon richness, Shannon diversity index, and Pielou’s evenness (referred hereafter as taxon richness, Shannon diversity and evenness) were calculated using the R package “vegan” (Oksanen et al. 2013).

Statistical analysis

Generalized mixed-effects models (GLMMs) were used to analyze the effects of climate, land use, season (spring and fall 2015 and 2016), and their interactions on nematode density (i.e., total number of nematode individuals in the community), taxon richness, and trophic groups (i.e., percentage of individuals in each trophic group). Taxon richness and density were modelled with Poisson distribution (link = ”log"), while the trophic groups were modelled with binomial distribution (link = ”“logit”). Linear mixed-effects models (LMMs) were used to assess the effects of climate, land use, season (spring and fall 2015 and 2016), and their interactions on Shannon diversity, evenness, Enrichment Index, Structure Index, Channel Index, Maturity Index, and the ratio between microbivores and plant feeders. The random intercept of all models was structured with plots nested within mainplots, nested within year (year as a categorical factor). Model assumptions of both LMMs and GLMMs were checked by visually inspecting the residual plots for homogeneity and the quantile–quantile plots for normality. There was no correlation between the residuals and the fitted parameters of the model. To meet the assumptions of the model, the microbivore–plant feeder ratio had to be log transformed. GLMMs were also used to assess the effects of climate, land use, and their interactions on nematode functional guilds and c–p groups (Table S3) using binomial distribution (link =  “logit”). The random intercept of the model was structured with plots nested within mainplots, nested within season, nested within year (year as a categorical factor). LMMs and GLMMs were performed using the “lme4” package (Bates et al. 2014). All statistical analyses were conducted using R version 3.4.2 (R Core Team 2017).

Results

Total nematode density and taxonomic diversity metrics

Overall, the total density of nematodes was significantly increased under future climatic conditions (+16.8%; p < 0.05) and varied significantly across land-use types (CF = 18.85 ± 1.17 ind/g DW soil (all values are presented ± SE), EP = 9.86 ± 0.51 ind/g DW soil; p < 0.001; Fig. 1a; Table 1, see Table S4 for an overview of the mean values for each response variable). According to a significant interaction of land use and season (p < 0.001), we found in spring that highest nematode densities were reached in conventional farmlands (19.19 ± 1.73 ind/g DW soil) and lowest densities in both extensively used grassland treatments (EM = 9.74 ± 0.88 ind/g DW soil; EP = 9.77 ± 0.76 ind/g DW soil). In fall, the extensive grasslands still showed the lowest densities (EM = 10.60 ± 0.68 ind/g DW soil; EP = 9.95 ± 0.72 ind/g DW soil), but densities in the intensively used grasslands (23.17 ± 2.58 ind/g DW soil) exceeded those in conventional farmlands (18.51 ± 1.60 ind/g DW soil). Furthermore, nematode density was significantly affected by the interaction of climate and season (p < 0.05): while future climate increased densities in both seasons, the strongest effects were detected in fall (+7.3% in spring; +26.0% in fall). The lowest densities were found in spring under ambient climatic conditions (12.82 ± 0.86 ind/g DW soil).

Fig. 1
figure 1

Effects of climate (C), land use (L), and season (S) on nematode density and taxonomic diversity metrics a total nematode density per g dry weight soil; b nematode taxon richness; c Shannon diversity; and d evenness. Error bars represent ± standard error. Black = ambient climate; grey = future climate. CF = conventional farming (brown); OF = organic farming (red); IM = intensively used meadow (yellow); EM = extensively used meadow (green); EP = extensively used pasture (blue). Data from spring (May) and fall (October) 2015 and 2016 is included. (*)p < 0.1; *p < 0.05; **p < 0.01; ***p < 0.001 (colour figure online)

Full size image
Table 1 Chi-squared values (χ2) of the generalized mixed-effects models for the effects of climate, land use, season (spring and fall 2015 and 2016), and their interaction on soil nematode variables
Full size table

Taxon richness was significantly affected by seasonal changes with overall lower levels in fall (− 9.6%; p < 0.01; Fig. 1b). In addition, land use significantly influenced nematode richness (p < 0.01): the highest numbers of nematode taxa were found in extensively grazed grasslands (20.75 ± 0.57 taxa on average), whereas richness was particularly low in intensively used grasslands (18.45 ± 0.62 taxa on average). For Shannon diversity and evenness (Fig. 1c, d), we found a significant interaction between land use and season (Shannon diversity H′: p < 0.001; Evenness J′: p < 0.001). They were both particularly low in intensively used grasslands, with land-use effects being most pronounced in fall (H′ = 1.61 ± 0.07; J′ = 0.23 ± 0.008). The highest Shannon diversity values were reached in extensively grazed plots in spring (H′ = 2.68 ± 0.04), whereas greatest evenness was found in organic farmlands in fall (J′ = 0.31 ± 0.014). In addition, climate and season interactively affected evenness (p < 0.05) by showing the lowest evenness scores under future climate particularly in fall (− 4.7%, J′ = 0.28 ± 0.005), whereas climate effects were minimal to slightly positive in spring (+1.6%).

Nematode functional diversity metrics

The Enrichment Index significantly varied between seasons with higher enrichment in fall (+23.5%; p < 0.001; Fig. 2a). The Structure Index was significantly affected by the land-use types (p < 0.001), clearly separating farmlands (low structure; CF = 15.35 ± 1.91 and OF = 19.75 ± 2.33) and grasslands (high structure; EM = 40.72 ± 2.94, EP = 39.99 ± 2.66, and IM = 48.34 ± 3.29) into two groups (Fig. 2b). The same patterns were also reflected by the Maturity Index (p < 0.001), which was substantially lower in farmlands (CF = 2.00 ± 0.02; OF = 1.99 ± 0.02) than in grasslands (EM = 2.23 ± 0.05; EP = 2.21 ± 0.03; IM = 2.29 ± 0.04). In addition, the Maturity Index was significantly lower in fall (− 2.6%; p < 0.05; Fig. 2c). The Channel Index was strongly driven by seasonal patterns (p < 0.001) that indicate fungal-dominated decomposition pathways in spring (+26.9%, high values), whereas bacteria-driven decomposition pathways were more dominant in fall (low values; Fig. 2d).

Fig. 2
figure 2

Effects of climate (C), land use (L), and season (S) on nematode functional diversity metrics a Enrichment Index; b Structure Index; c Maturity Index; d Channel Index; and e microbivores/plant-feeder ratio, the black line represents the ratio being at the critical value of 1.0—lower ratios indicate negative soil effects on plant productivity. Error bars represent ± standard error. Black = ambient climate; grey = future climate. CF = conventional farming (brown); OF = organic farming (red); IM = intensively used meadow (yellow); EM = extensively used meadow (green); EP = extensively used pasture (blue). Data from spring (May) and fall (October) 2015 and 2016 is included. (*)p < 0.1; *p < 0.05; ***p < 0.001 (colour figure online)

Full size image

Nematode trophic groups

Bacteria-feeding nematodes were significantly affected by an interaction between land use and season (p < 0.001). They showed the highest densities in conventional (CF = 33.2% bacterial feeders) and organic farming plots (OF = 33.1% bacterial feeders) in spring, with particularly low levels in intensively used meadow plots in fall (IM = 10.1% bacterial feeders). Climate and season also interactively affected bacteria-feeding nematodes (p < 0.01): future climate had negative impacts in fall (− 5.9%), but negligible effects in spring (− 0.4%; Fig. 3a). The abundances of fungal-feeding nematodes were driven by a three-way interaction between climate, land use, and season (p < 0.001). They were most abundant in conventional and organic farming plots under future climatic conditions in fall (CF = 42.4% fungal feeders; OF = 42.6% fungal feeders), whereas the intensively used meadow plots had the lowest abundances under future climate in fall (IM = 10.1% fungal feeders). The positive effects of future climate were most pronounced in extensively mown grasslands in spring (EM: +30.0%) and in conventional and organic farming plots in fall (CF: +37.3%; OF: +71.0%; Fig. 3b). In addition, plant-feeding nematodes were affected by a significant three-way interaction between climate, land use, and season (p < 0.001). They were most abundant in intensively used grassland plots in both seasons with particularly high levels in fall (future climate: IM = 76.0% plant feeders, ambient climate: IM = 75.7% plant feeders). The lowest abundances were reached on organic farmlands in spring under ambient climate (OF = 21.9% plant feeders), and the strongest negative climate effects were found on conventional farming plots in fall (− 33.3%). However, future climate had a comparatively strong positive effect on plant-feeding nematodes in organic farmlands in spring (+31.3%) and in extensively mown grassland plots in fall (+51.2%; Fig. 3c). Omnivorous and predatory nematodes were affected by the interaction between land use and season (p < 0.001). They generally had very low abundances, being particularly rare in conventional farming plots in spring (CF = 1.5% omnivores and predators). The highest levels were reached in grassland plots, particularly in extensively mown grasslands in spring (EM = 5.5% omnivores and predators; Fig. 3d).

Fig. 3
figure 3

Effects of climate (C), land use (L), and season (S) on nematode trophic groups a bacteria-feeding nematodes; b fungal-feeding nematodes; c plant-feeding nematodes; and d omnivorous and predatory nematodes. Error bars represent ± standard error. Black = ambient climate; grey = future climate CF = conventional farming (brown); OF = organic farming (red); IM = intensively used meadow (yellow); EM = extensively used meadow (green); EP = extensively used pasture (blue). Data from spring (May) and fall (October) 2015 and 2016 is included. (*)p < 0.1; *p < 0.05; **p < 0.01; ***p < 0.001 (colour figure online)

Full size image

The ratio between microbivores (bacteria- and fungal-feeding nematodes) and plant-feeding nematodes was significantly influenced by the interactive effects of land use and season (p < 0.001). Highest levels were reached in conventional farmlands in spring (CF = 3.67 ± 0.54) compared to particularly low values in intensively managed grasslands in fall (IM = 0.28 ± 0.03). Only the intensively managed grasslands were found below the critical value of 1.0 in both seasons, indicating overall negative soil effects on plant productivity. For all other land-use types, values were above 1.0, and thus beneficial soil effects on plants can be assumed (Fig. 2e).

Discussion

We studied the responses of the ecologically important group of soil nematodes in two highly relevant seasons for soil processes in two consecutive years, which allowed us to obtain a comprehensive picture of their structural and functional shifts in response to interacting global change drivers across seasons. Our study revealed that especially land-use induced complex shifts in soil nematode communities with strong seasonal dynamics. The climate treatment as well as the interaction between climate and land use exerted weaker effects, but were also dependent on the season. Notably, we could show that higher nematode densities under the combination of future climate and intensive land use were realized at the expense of higher shares of opportunists and potential pest species (plant-feeding nematodes), which coincided with a less diverse and less structured community. These results suggest that the combination of intensive land use under future climatic conditions may shape soil food webs in a way that compromises plant growth and other important ecosystem functions.

In general, total nematode densities increased under future climatic conditions. While climate was often shown to have only weak effects on nematode densities (Cesarz et al. 2015; De Long et al. 2016; Siebert et al. 2019a), the positive effects we found in our study contradict a large body of literature reporting negative effects under warmed and/or drought conditions (Dong et al. 2013; Kardol et al. 2010; Yan et al. 2017). Intensive land-use management also led to higher nematode densities, which was surprising as well, given the known susceptibility of soil nematodes to land-use intensification that was confirmed by many previous studies (Dong et al. 2013; Mulder et al. 2003; Okada and Harada 2007; Thomas 1978). A closer look at the feeding groups and functional guilds revealed complex shifts in community composition that were driven by the two interacting global change drivers, likely changing the functioning of communities and highlighting the importance of multifactorial investigations. Based on the multitude of trophic levels and life-history strategies covered by nematodes, we found varying responses with distinct temporal patterns. For instance, nematode indices that reflect the overall state of the ecosystem (Ferris et al. 2001) revealed complex seasonal dynamics with generally enriched and bacteria dominated food webs in fall. Such enrichment is usually caused by high mortality rates after unfavorable conditions (i.e., disturbances) that opportunistic species can benefit from (Ferris et al. 1996, 2001). This is particularly true for species in the fast bacterial decomposition channel that efficiently use easily degradable substrates and rapidly increase in numbers as soon as food availability allows, whereas the slower fungal decomposition remains dependent on more complex material with high cellulose and lignin content (Bongers 1990; Bongers and Bongers 1998; Bongers and Ferris 1999; Ferris et al. 2001). Indeed, our results for the highly opportunistic Ba1-nematodes confirm higher numbers in fall compared to the spring samplings (Fig. S4a). In addition, data on microbial activity shows significantly higher activity levels in fall compared to spring (Table S5, Fig. S5), which likely provided the basis for the rise in opportunistic bacterial-feeding nematodes. This pattern may be a consequence of the combined effects of the climate treatment over summer (June–August) when soil organisms were exposed to 20% less precipitation and ~ 0.6 °C soil and air warming (Fig. S2). Based on soil moisture data for spring, summer, and fall (Table S6, Fig. S6), the severe environmental conditions in summer may still be reflected by a highly disturbed nematode community in our sampling campaign in fall, even though the precipitation treatment was then already changed to an increase in rainfall. Similar results are reported in a study by Kardol et al. (2010), in which it was assumed that 10% soil water content marks a critical threshold for nematode survival and reproduction, rapidly becoming detrimental when approaching 5% (Fig. S6). Thus, a combination of air and soil warming and reduced precipitation may affect the community structure of nematodes in a way that cannot be compensated during more favorable conditions in fall (Bakonyi et al. 2007; Colagiero 2011; Dong et al. 2013; Yan et al. 2017).

Another significant community alteration could be observed in response to intensive grassland management. Whereas bacteria feeders and fungal feeders showed low densities on this land-use type, plant feeders particularly thrived in intensive meadows, reaching by far their highest abundances (especially plant feeders within the cp2 and cp3 classes, Fig. S3, Table S3). As many plant-feeding nematodes are classified as agricultural pests, which can impair plant growth above- and belowground, subsequently putting stable crop yields at risk (Evans et al. 1993; Mueller et al. 2016; Nicol et al. 2011), it is critical to understand their population dynamics under different land-use scenarios. Ratio analyses indicated that the high numbers of plant-feeding nematodes in intensively used grasslands may in fact present a threat to plant productivity (Eisenhauer et al. 2011; Wasilewska 1997)—a finding that could be alarming for farmers and practitioners. Even without considering the impact of concurrent global change factors, it is estimated that about 10% of the world’s crop production is reduced by the activities of plant-feeding nematodes (Smiley and Nicol 2009; Whitehead 1997). To prevent yield losses by nematode species feeding on plant tissues and harming their roots (Neher 2010), the control by natural antagonists, including predatory nematodes, can be among the most effective and sustainable measures (Neher 2010) as a well-structured soil food web may be able to regulate the rise of opportunistic species and pest nematodes (Doran and Parkin 1994; Ferris et al. 2001; Kennedy and Smith 1995). However, we are aware of the fact that differences in plant biomass among the land-use types (see for example Siebert et al. (2019b)) may also be accountable for varying abundances of plant-feeding nematodes and may induce bottom–up effects on the nematode community. However, as we currently lack an assessment of herbivory pressure that would allow us to draw strong conclusions about top–down effects of plant-feeding nematodes on plant growth, we need to be cautious when discussing potential effects of rises in plant-feeding nematodes. Future research can address this knowledge gap by performing plant–soil feedback experiments, manipulating the nematode community structure in controlled mesocosm studies, or by determining herbivory pressure through energy flux calculations (Barnes et al. 2018).

Notably, we detected large differences in the structure of nematode communities between land-use types, mainly separating poorly structured communities in farmlands from those in grasslands. In grasslands, we also found more mature nematode communities with higher abundances of predators and omnivorous species. This is in line with the general expectation that land-use types associated with frequent fertilization, ploughing, and highly productive monocultures are leading to simplified communities with a dominance of opportunistic taxa that can withstand such conditions (Bongers and Bongers 1998; Dmowska and Kozlowska 1988). There is strong evidence that complex, highly structured, and mature nematode assemblages are more capable to ensure the stable provisioning of diverse ecosystem functions under environmental stress and to recover after disturbances, e.g., due to functional redundancy and/or higher connectance (Ferris et al. 2001; Menge 1995; Strong 1992; Wardle et al. 1995; Yeates and Wardle 1996). To facilitate the continued existence of such highly structured soil food webs and diverse soil nematode communities under environmental disturbances, the role of plant diversity is often highlighted as one avenue to stabilize communities (Hautier et al. 2014; Neher 2010; Wagner et al. 2015), which might then again be able to prevent the rise of plant-feeding, potential pest species (Ferris et al. 2001). However, the nematode communities in the two extensively managed grasslands were not the most mature and structured, but were excelled by the intensive grasslands. Moreover, we could not find clear evidence for buffering effects of land-use types comprising higher plant diversity under future climatic conditions, thus failing to confirm our second hypothesis. In contrast to other studies, which also reported a decrease in the structure of nematode assemblages based on climate treatments (Porazinska et al. 1999; Thakur et al. 2014), these effects were only marginal in our study.

In accordance with many previous studies (Bongers and Bongers 1998; Briones et al. 1997; Dong et al. 2013; Porazinska et al. 1999; Sánchez-Moreno et al. 2006; Yeates and Bird 1994; Zhang et al. 2012), our findings showed that global change represents a threat to soil biodiversity. Relating to a recent study by Li et al. (2019), in which the authors only found climate effects in natural habitats, we can provide additional evidence that nematode communities are also affected in agricultural systems. Nematode diversity metrics (both taxon richness and Shannon diversity) indicate that while extensive management fosters diverse communities, diversity declines under intensive management. In our study, the extensive grasslands displayed the highest diversity levels, whereas organic farmlands showed the greatest evenness. These findings highlight the value of less disturbed systems that preserve microhabitats in the soil and, therefore, allow more sensitive, often K-strategic taxa within the omnivorous and predatory groups to survive (Bongers and Bongers 1998; Dong et al. 2013). Indeed, our results of the respective feeding groups show that higher trophic levels such as predators and omnivores were almost absent in farmlands, but benefitted from extensive land-use management. This underlines the great susceptibility of higher trophic levels, whose long generation times and low fecundity are often incompatible with intensive management practices (Bongers and Bongers 1998; Ferris et al. 2001; Thakur et al. 2014). As a consequence of this strong reduction in omnivores and predators, top–down regulation of lower trophic levels (e.g., opportunists and plant feeders) may be impaired (de Ruiter et al. 1995; Ferris et al. 2001), with heretofore unknown consequences for crop yields (Evans et al. 1993; Mueller et al. 2016; Nicol et al. 2011).

In conclusion, we found the nematode community to be often affected by complex interactions between both global change drivers and seasonal dynamics, while being particularly sensitive to land-use practices. High enrichment levels in fall are indicative of persisting effects of the climatic conditions over summer that may have long-lasting impacts on nematode communities. The high nematode densities we found under the combination of future climatic conditions and conventional farming coincided with a less diverse, structurally simplified community. Future studies should explore the ecological consequences of these shifts in nematode community composition, because they could cause changes in multiple ecosystem functions and reduce the ability of nematode communities to withstand environmental stress.