Work programme including proposed research methods
We propose to use an existing field experiment (see below) to mechanistically link two of the most prominent global-change drivers in temperate grasslands, land-use intensity and climate change, to changes in above-belowground multitrophic interaction networks and ecosystem multifunctionality. Instead of merely describing how communities change, we will obtain mechanistic insights on how these changes translate into altered ecosystem processes and multifunctionality – which is what society ultimately depends on.
We will investigate how land-use intensity and climate change jointly impact above-belowground communities and ecosystem processes. This is important because both stressors affect individual- to community-level aspects of biology and ecology and are likely to interact, i.e. to influence each other’s impact (Oliver and Morecroft 2014). We will obtain a mechanistic understanding of these joint impacts by using two state-of-the-art molecular methods to assess how multitrophic interactions change and combine the resulting information on altered interactions and basal resources with a powerful method for calculating energy flux through food webs. We will use the Global Change Experimental Facility (GCEF, Helmholtz-Zentrum für Umweltforschung – UFZ, Bad Lauchstädt, Fig. 1a), established in 2012 (Schädler et al. 2019), which combines a land-use intensity treatment with a climate-change treatment in a temperate-grassland context. The GCEF features two climate levels (current vs. future climate – the latter with shifted precipitation patterns, ~ 20% reduction in summer and ~ 10% increase in spring and autumn and increased temperature – 0.55°C increased daily mean temperature, increased minimum temperature, longer frost-free periods and more growing degree days), enabled via automatic roofs and irrigation systems and five land-use types spanning common management options for croplands and grasslands. Each of the 10 main plots (80 x 24 m) represents either current or future climate and consists of five randomly-ordered plots (16 x 24 m) for each of the land-use types (Schädler et al. 2019). We will test the overarching hypothesis that land-use intensity and climate change magnify each other’s impact on above-belowground communities and functioning. We will use the climate treatment together with three land-use types: intensively-used meadow (sown with 5 species/cultivars in 2013, moderately NPK fertilised, mown four times a year), extensively-used meadow (sown with a 56-species mix, mown twice a year) and extensively-used pasture (as ext. meadow, but grazed three times a year by ~ 20 sheep for 24 h).
The GCEF was set up to investigate how climate change and land-use intensity interactively impact ecosystem functioning via above-belowground biodiversity. GCEF research has shown that climate and land-use intensification independently and interactively modify aboveground arthropod species richness, abundance and community composition (Sohlström et al. 2022). The two global-change drivers were found to alter soil-fauna biomass via different pathways (climate change via body-size reductions, land-use intensity via density reductions; Yin et al. (2020)). Despite this previous research, we know little about: i) how these community changes translate into altered ecosystem processes and multifunctionality and ii) how above- and belowground systems interact in their response to global change. To answer these questions, we propose to use the GCEF to simultaneously sample above- and belowground invertebrate communities (WP1 and OB1), use gut-content metabarcoding (Sousa et al. 2019) and CSIA-AA (Pollierer et al. 2019) for assessing trophic interactions (WP2 and OB2) and establish plot-level food webs, calculate above-belowground energy flux, ecosystem processes, energy channels (Potapov 2022) and multifunctionality to unravel the mechanisms of global-change effects on grassland communities and ecosystems (WP3 and OB3; Barnes et al. (2018), Potapov et al. (2019), Jochum and Eisenhauer (2021)).
WP1 – Above-belowground invertebrate communities
In WP1, our aim is to simultaneously assess the responses of above- and belowground invertebrate communities in temperate grasslands to the joint influence of land-use intensity and climate change (OB1). To achieve this goal, we (PhD & student helpers) will sample above- and belowground invertebrates in 30 GCEF plots (10 main plots, three land-use plots each: intensive and extensive meadow, extensive pasture) in two consecutive years (1 and 2, see Fig. 1b). Aboveground fauna will be suction-sampled and belowground fauna will be extracted from soil cores or using mustard extraction directly from the ground (earthworms). Suction samplers are effective tools to quantitatively sample aboveground grassland arthropods. We will use the ecoVac (ecoTech, Bonn, Germany) device which has been successfully used in temperate grasslands before (Sohlström et al. 2022). On each plot, in each year, we will sample one randomly-selected 45-cm-diameter area for aboveground (vegetation and ground) arthropods. When approaching the sampling spot, we will put a modified circular laundry basket (45 cm diameter, IKEA Fyllen, 50 cm high, Polyester) on the ground to prevent mobile arthropods from escaping. Subsequently, we will suction-sample the vegetation and ground inside the basket for four minutes and afterwards empty out the upper cage (Jochum et al. 2022, Sohlström et al. 2022). Suction samples will be cooled in the field, frozen at -20°C as soon as possible and subsequently transferred to 70% ethanol (spiders will be placed in molecular-grade ethanol for subsequent molecular analyses, i.e. to avoid DNA breakdown) for later identification and analyses. We will measure sward height to be used as a correction parameter for aboveground arthropod suction sampling.
To quantitatively sample soil fauna at peak biomass, and in concert with the simultaneous aboveground-fauna assessment, we will take one large (20 cm diameter) and one small (5 cm diameter) soil core, both to a depth of 10 cm, to adequately assess soil macro- and mesofauna, respectively (Potapov et al. 2022). All cores will be taken inside the suction-sampled area right after suction sampling. Soil cores will be stored in sealed plastic containers and transported to Leipzig (Eisenhauer lab, iDiv), where we will use heat to quantitatively extract belowground fauna using modified Macfadyen and Kempson heat extraction schemes (8 d, starting at 20°C with 5°C steps/24 h, until going from 50°C - 60°C at day 7; Birkhofer et al. (2017)). After extraction, soil fauna will be stored in 70% ethanol for later identification and analyses (staphylinids will be placed in molecular-grade ethanol and be frozen at -20°C because of subsequent molecular analyses). To complement soil-fauna data from these soil cores, we will quantitatively sample earthworms from deeper layers by twice applying 2 l of mustard solution (with 5 min break) to the 20-cm soil-core holes and collecting earthworms until 5 min after the application (Eisenhauer et al. 2008). Additionally, soil nematodes will be extracted by wet extraction (additional core; 2-cm diameter, 10 cm depth). In addition to this fauna sampling, we will take pictures of the vegetation inside the suction cage – and of the direct surroundings in a standardised way, to later assess dominant plant species in this exact patch of the plot. Overall, we estimate fieldwork to take 5 days per year for the team (PhD, student helpers). To complement the data collected here, we will have access to GCEF joint data on plant communities, soil abiotics and microclimate.
Macro- and mesofauna will will be pre-sorted to order or higher-level taxon by the PhD student and student helpers. All taxa, including nematodes and earthworms, will then be identified to species- or functional-group level by specialists to ascribe them to trophic species for food-web construction. We will measure body lengths of five individuals per species/highest identification level and plot to calculate body masses with allometric regressions (Sohlström et al. 2018). We will calculate plot-level taxon richness, individual density (ind/m²), fresh biomass (mg/m²) and community composition, split into above- and belowground communities and for the overall community. We will analyse the impact of climate, land use and their interaction on above-belowground community properties (HA). We expect above- and belowground communities and different community properties to respond differently to the stressors (Barnes et al. 2014, Jochum et al. 2017a, Le Provost et al. 2021, Jochum et al. 2021b). More specifically, we expect aboveground communities (H1.1; Le Provost et al. (2021)) and higher trophic levels to respond most strongly (H1.2; Brose et al. (2012)) and density to respond more strongly than species richness (H1.3; Barnes et al. (2014), Jochum et al. (2021b), Jochum et al. (2022)). We expect land use to more-strongly affect communities via community composition and climate change to more strongly affect their body-size structure (H1.4; Yin et al. (2020), Sohlström et al. (2022)). Structural-equation models will assess joint impacts of climate and land use on above-belowground interdependencies. We expect decoupling between above-belowground communities and their properties (e.g., diversity of above- and belowground taxa) in response to combined stressors (H1.5; Manning et al. (2015), Ochoa-Hueso et al. (2020)).
WP2 – Molecular analyses of trophic interactions
To assess global-change impacts on multitrophic interactions and energy channels (OB2), we will use two state-of-the-art methods, gut-content metabarcoding for assessing short-term, direct trophic interactions (Who eats whom?) and compound-specific isotope analysis of amino acids (CSIA-AA) to quantify the relative dietary contribution of long-term basal resources, such as bacteria, fungi and plants, for consumers (... and how much?). The little information we have on global-change impacts on grassland-interaction networks is commonly constrained to plant-pollinator (Weiner et al. 2014) or plant-herbivore networks (Neff et al. 2021), with higher trophic levels typically neglected. However, to understand how global change alters ecosystems, these higher trophic levels and their community-wide impacts are crucial (Brose et al. 2012, Jochum et al. 2012) and ignoring them could result in misinterpreting the overall patterns (Jochum et al. 2021a). Consequently, both molecular methods will be used on above- and belowground predators. As we need time to extract and sort the invertebrate samples and run the molecular analyses, we will focus these analyses on samples taken in year 1 and later expand the obtained information on feeding relationships and basal resources to the community data from both years in WP3. We know from two previous PhD projects that spiders and staphylinid beetles are the most common above- and belowground predators in the GCEF grasslands, respectively. We will focus on these taxa, present across all plots and important for subsequent flux calculations, to assess how the treatments change their trophic interactions and basal resources.
Gut-content metabarcoding (Sousa et al. 2019) will be used to detect changes in predator-prey interactions. We use this method on spiders (aboveground) and staphylinids (belowground), taken from suction and heat-extraction samples, to test how their trophic interactions change with land-use intensity and climate change. We will use whole-body DNA extraction linked with predator-specific blocking primers during PCR to increase prey-DNA amplification and subsequent detection of prey species. Such primers are available for spiders (Lafage et al. 2020) and we have tested them successfully (EcoStressWeb). Blocking primers for staphylinids will be designed together with collaborators. As a back-up solution, we could use beetle regurgitates for gut-content metabarcoding as has successfully been done before (Tiede et al. 2016). On each plot (n = 30), we will use the five most common spider and staphylinid species, respectively and analyse one multiple-individual sample (30 plots*2 consumers*5 mix-samples = 300 metabarcoding samples, Fig. 1c). Metabarcoding will be done by a company specialised in high-throughput sample processing and sequencing, including specimen preparation, DNA extraction and amplification, marker gene library preparation and sequencing. We will obtain plot-level data regarding above- and belowground consumer prey spectra in response to two stressors. This will allow us to analyse if and how trophic interactions change (HB). We expect land-use intensity and climate change to interactively affect the number of trophic interactions (diversity of prey items), with a reduced number of trophic links for predators under future-climate conditions (H2.1) and higher-intensity land use (H2.2) and the effect of one stressor being dependent on the other (H2.3; Lindmark et al. (2019), Åkesson et al. (2021)). We expect changes in trophic interactions to be more pronounced for above- than belowground consumers (H2.4; Le Provost et al. (2021)). Finally, to improve our understanding of the mechanisms behind global-change impacts on altered trophic links, we will quantify how well consumer prey choice matches prey relative abundance patterns (links to most abundant prey) or body-size changes (optimal prey size), based on data from WP1 (Gauzens et al. (2021)). Here, we expect both altered body size and shifted relative abundance patterns to impact consumer trophic links, with land use having a stronger impact via community composition and climate via body-size changes (H2.5; Yin et al. (2020), Sohlström et al. (2022)).
Compound-specific isotope analysis of amino acids (Pollierer et al. 2019) will be used to assess global-change driven shifts in the long-term relative-importance of basal resources, such as plants, fungi and bacteria. The method is based on the fact that essential amino acids are synthesised by these producers and then recycled up the food chain. Since producer-derived amino acids differ in their carbon (C) isotopic signatures given differing biosynthetic pathways, we can use the isotopic signatures of essential amino acids from consumer-body tissue to indirectly assess the relative importance of basal resources for specific consumers (Pollierer et al. 2019). Furthermore, nitrogen (N) isotopic signatures of “trophic” and “source” amino acids, which are either strongly or very slightly enriched via trophic transfer, can be used to assess consumer trophic positions (Pollierer et al. 2019). We will use predators as indicators for the communities’ overall dependence on the three basal resources. We will analyse multi-individual samples for each of the five most common spider and staphylinid species per plot (300 CSIA-AA samples). The PhD student will acquire the skills to perform these analyses and analyse these samples supervised by one of the collaborators. Generally, we expect global change to shift basal-resource dependence more strongly in below- than aboveground consumers, as we expect the aboveground system to remain strongly based on the plants as their main basal resource (H2.6). We expect the belowground communities to be shifted towards a higher dependence on bacteria as basal resource under higher land-use intensity (H2.7; Bloor et al. (2021)) and on fungi under future-climate conditions (H2.8; De Vries et al. (2012)). We will combine the results from WP2 into one paper on multiple-stressor impacts on above-belowground multitrophic interactions and basal-resource dependency (HB).
WP3 – Food webs, energy flux and ecosystem multifunctionality
We know that global change impacts communities and ecosystem processes, but we lack the mechanistic understanding of how the two things are linked. In WP3, we aim to establish plot-level food webs to calculate above-belowground energy flux, ecosystem processes (e.g., predation, herbivory, decomposition), energy channels (plant-, bacteria-, fungi-based) and multifunctionality and unravel the mechanisms of global-change effects on grassland communities and ecosystems (OB3). We will combine the community data from WP1 (both years) with the trophic information from WP2 (based on sampling in year 1) to establish plot-specific food webs for both years. More specifically, we will use a combination of taxonomic and functional information together with novel, plot-specific data on predator-prey interactions from WP2 to build food webs (see Hines et al. (2019), similar approach without molecular methods). Hence, rather than using a metaweb-subweb approach, we will have plot-specific data on body sizes (WP1), trophic interactions and relative importance of basal resources (both WP2) that we can use with the taxonomic information to build plot-specific food webs. These food webs will be used to calculate plot-level energy fluxes providing information on ecosystem processes, energy channels (Potapov 2022) and multifunctionality.
Based on the pre-sorting and expert identification of all invertebrates (WP1), we will group them into trophically-consistent feeding types, such as carnivores, herbivores, detritivores or omnivores, within taxonomic orders. Omnivores will be further specified into carnivore-herbivores, carnivore-detritivores, herbivore-detritivores or generalistic omnivores consuming animal, plant and detritus resources (Lind et al. 2015, Barnes et al. 2020). We use this combined taxonomic-functional approach, because feeding relationships in grasslands are highly phylogenetically conserved and taxonomic groups provide information on the trophic behaviour and prey vulnerability, for example, due to morphological traits, such as sclerotisation (Lind et al. 2015, Potapov et al. 2018, Barnes et al. 2020). These taxonomic-functional entities will be the trophic nodes in the food webs and energy-flux calculation. We will calculate energy flux through plot-level food webs and use the resulting fluxes to quantify processes, energy channels and multifunctionality. Energy flux will be calculated following a well-established routine, the adapted food-web energetics approach (Barnes et al. 2018, Gauzens et al. 2018, Jochum and Eisenhauer 2021, Jochum et al. 2021a). It will be based on the community and trophic data assembled in WP1 and WP2, providing plot- and taxon-specific individual density and body size, as well as trophic-link information for spiders and staphylinid beetles. We will calculate individual-level metabolic rates based on taxonomy, body size (from WP1) and plot-level temperature (plot-level soil- and vegetation-level temperature available from GCEF) using well-established regressions for arthropods (Ehnes et al. 2011). On each plot, we will scale the data obtained from different above- and belowground sampling techniques (Fig. 1b) to one square metre. Subsequently, for each of the food-web nodes present in a plot, we will sum up the individual metabolic rates of the respective individuals resulting in the metabolic demand of the whole node (Barnes et al. 2014, Jochum et al. 2021a). We will assign link-level assimilation efficiencies, based on the respective resource type and ambient temperature (Lang et al. 2017). We will apply consumer preferences taking into account the relative abundance of prey types (WP1), as well as information on relative importance of basal resources (from WP2). Subsequently, we will use the “fluxweb” R package (Gauzens et al. 2018) to calculate energy flux through the plot-level food webs, resulting in flux estimations for each consumer-resource interaction, i.e. every single food-web link.
Based on this information, we will quantify ecosystem processes, such as primary and intraguild predation, herbivory and decomposition as all fluxes going out of the respective resource types, for example, out of autotroph nodes for herbivory. We will quantify energy channels by summing up all fluxes upstream of plants, bacteria and fungi, respectively. These assessments are based on established routines (Barnes et al. 2018, Jochum et al. 2021a, Potapov 2022). We will use flux-based ecosystem processes (predation, herbivory etc.) to calculate trophic multifunctionality (Potapov et al. 2019) with the averaging and multiple-threshold approaches (Byrnes et al. 2013). Subsequently, the calculated ecosystem processes, energy channels and multifunctionality will be related back to the underlying climate and land-use treatments (HC). We will identify differences and similarities between treatment impacts on the belowground and aboveground parts of the response variables (belowground vs. aboveground ecosystem processes, channels, multifunctionality) and on energy fluxes across the above-belowground boundary (e.g., predation of aboveground predators on predominantly-belowground prey).
Generally, we expect higher land-use intensity and future climate to reduce total energy flux, but more severely above than below the ground (H3.1; Le Provost et al. (2021)). Single ecosystem processes will respond to the treatments, with, for example, increased herbivory at high land-use intensity (H3.2; Barnes et al. (2020)). We expect functions associated with higher trophic levels, for example, predation, to be more-heavily affected than those associated with lower trophic levels (H3.3; Potapov et al. (2019)). The bacterial energy channel will likely be strengthened under high land-use intensity (Bloor et al. 2021) and the fungal channel will be strengthened under future climate conditions (De Vries et al. 2012) (H3.4). Finally, we hypothesise land-use intensity and climate change to reduce above- and belowground ecosystem multifunctionality (H3.5; Allan et al. (2015)), mediated via several community properties. We will analyse how average body mass, total biomass, taxon richness, food-web properties and taxonomic composition mechanistically drive energy flux, ecosystem processes and multifunctionality (HD; Barnes et al. (2018), Jochum et al. (2021a), Wan et al. (2022)). The results on land-use and climate impacts on grassland above-belowground energy flux, ecosystem processes and multifunctionality will be summarised in one paper.
In summary, we will obtain fundamentally-novel information on how land-use intensity and climate change jointly alter above-belowground communities and trophic interactions and how this mechanistically modulates above-belowground ecosystem functions, energy channels and multifunctionality. Our results will offer unprecedented insights into how exactly multiple global-change drivers alter ecosystem performance, which is important to understand, manage and mitigate global-change impacts on temperate grasslands.