Human activities are resulting in global change. These can be localised changes, such as destroying a forest, or more global ones, such as changing of the climate through increased atmospheric emissions. Most people are aware that our activities can negatively affect animals like elephants or bears, or that our actions are causing invasive species like Japanese knotweed to spread. However, soil harbours some of the highest diversity on the planet, and yet we know surprisingly little about how soil organisms might be impacted by our actions. In recent years, ecologists have been conducting a number of reviews to assess how humans impact biodiversity, typically the number of species, by amalgamating individual studies from different locations. The results of these reviews have allowed general conclusions to be made, which can be applied across the globe and ideally result in conservation actions. However, these reviews have generally looked at biodiversity of aboveground animals and plants, largely ignoring organisms in the soil, such as earthworms. Yet, soil, and the organisms within it, are critical for life on land and support many ecosystem services that are essential for human well-being. For example, earthworms have been to shown to increase crop production and help regulate water movement into the soil. Many organisms within the soil are also involved in the carbon cycle, by breaking down litter material from the soil surface and storing carbon in the soil. While a number of previous reviews have been conducted on the impacts of human activities on soil communities, most have focussed on a single species group, or identified only a couple of human impacts. However, as all organisms are subjected to multiple human impacts, and these are likely to affect organisms in a variety of ways, as well as influence the interactions among organisms, a comprehensive review which encompasses a multitude of organisms and human impacts is necessary.

In this study, we propose to systematically review the literature to determine how human activities impact soil organisms. For this review, we will use a meta-analytical approach - analysing the results of previously published studies to obtain an overall indication of the direction and the magnitude of the effect. The published literature we collate will not be restricted to a specific group of soil fauna, and we will attempt to include a wide variety of species. Similarly, we will use data from studies investigating a wide range of human impacts including land use change, habitat fragmentation, pollution, invasive species, and climate change. We expect to find that, on average, all soil organisms will be negatively affected by the human impacts, but with some human impacts, such as land use change and invasive species, having a greater effect. However, we anticipate that within each of the human impacts, such as among different land use types, the results might vary. The proposed meta-analysis will allow us to better understand and manage human impacts on soil biodiversity.

Human impacts are causing an unprecedented change of biodiversity, at global and local scales. To quantify the nature and degree of the biodiversity change, there have been a number of meta-analysis studies investigating the effects of global change drivers, e.g., land use, climate and pollution (e.g., Murphy and Romanuk 2013, Mantyka-pringle et al. 2011). However, these studies include few primary literature studies of soil biodiversity. Soil biodiversity is exceptionally important for a variety of ecosystem services that are critical for human wellbeing (Wall et al. 2015, Bardgett and van der Putten 2014). Yet, we know surprisingly little about how soil organisms may respond to changing environmental conditions (Bardgett and van der Putten 2014). Although studies have investigated the impact of global change drivers on soil biodiversity, they lack sufficient depth in the number of drivers and/or taxa included (Blankinship et al. 2011, Mazor et al. 2018). Global change consists of a multitude of drivers, and the degree of the change is likely to vary across taxa or body size groups. Organisms in soil span a huge gradient of body size (Veresoglou et al. 2015), which is linked to their microhabitat requirements, dispersal capabilities and reproductive rates, all of which will influence their response to global change drivers similar to that seen in aboveground organisms (Brook et al. 2008). Additionally, the previous focus on aboveground organisms has resulted in a bias to certain global change drivers, as a result of the bias in the primary literature. For example, climate change and land use change is more often studied, whilst pollution is typically understudied as a global change driver (Bernhardt et al. 2017, Mazor et al. 2018).

Building on previous studies, we will conduct a meta-analysis to compare the effects of global change drivers (land use change and intensification, habitat fragmentation/loss, climate change, invasive species, pollution, and nutrient enrichment) on soil fauna (micro- to macro-invertebrates). An advantage of using a meta-analytical approach is that results from globally distributed, small-scale experiments and observations can be combined to create more generalisable results (Koricheva et al. 2013). In addition, the amalgamation of the individual studies eventually spans a variety of biomes, countries and gradients of global change drivers, and thus can answer questions that the individual studies by themselves can not. By creating such a dataset to conduct this meta-analysis, we will also be creating a resource that can be used to answer additional questions in the future; for example, identifying particularly sensitive regions, or an above-/belowground comparison. Increasing the ease of access to soil biodiversity data will likely result in these taxa being included in further analyses. For example, the TRY database, which collates functional traits of plants from across the globe, has now been used in over 150 further publications (https://www.try-db.org).

We will address three general questions with this meta-analysis (Fig. 1): What are the relative impacts of different drivers of global change on soil fauna? Is there variation in the response of different soil fauna groups? Does the classical grouping of soil fauna based on body size adequately represent their response to global change drivers?

Figure 1.  

Conceptual framework of the meta-analysis.

The project will address a hierarchy of hypotheses, looking at the overall impacts across and within global change drivers.

First-level hypotheses

Second-level hypotheses

Following a literature search, using and expanding upon previously successful search terms (discussed below; Suppl. material 1), titles and abstracts will be screened and irrelevant papers removed. The Covidence software will be used for screening titles and entire papers, which will allow multiple team members to screen papers simultaneously, whilst also facilitating testing the repeatability and reliability of the screening process. Suitability of the remaining papers will then be assessed by screening the main text. Papers will be suitable if they

1. investigated the effect of at least one global change driver on at least one group of soil fauna and
2. contain at least one reference/undisturbed site (i.e. control), and one site impacted by the global change driver.

The log-proportional change between the mean diversity of the control and the impacted sites (log-response ratio) will be calculated for each global change driver and taxa group in the study. We will focus on species richness (the most commonly reported diversity measure) but will likely have sufficient data to investigate other measures that are often reported in the soil literature (e.g., abundance, biomass, Shannon diversity index).

To be able to answer the second-level hypotheses (i.e. changes within each of the global change drivers), each impacted site will have additional information extracted from the paper, which will also be used to inform the classification within the global change drivers. For example, for effect sizes from papers on land use, we will capture information on which land use the reference and impacted sites are (e.g., pastures, croplands) and other disturbances within the land use category that might impact soil fauna diversity (e.g., ploughing and tillage practices). For non-categorical global change drivers, such as pollutants or fire, we will capture rates, intensity or amounts of the global change driver.

As it is likely that factors unrelated to the global change drivers or the soil fauna will influence the magnitude of the response, we will also extract additional information that can be used as covariates in the meta-analysis. For example, we will test whether the effect sizes differ between observational or experimental (but non-manipulated communities) designs, and between biomes or regions in which the original study was conducted. Where possible, we will also capture geographic coordinates of the study sites. This will enable us to match the effect sizes to external data layers, such as those relating to soil parameters (using SoilGrids, with a resolution of 250 m). There are many covariates that may influence the response of soil fauna to global change drivers. These will be discussed in the initial workshop organised as part of this project. Following discussions, the structure of the database with be finalised, to ensure all potentially important variables are captured.

Analyses will first focus on the entire dataset, using the global change drivers as predictors for the soil taxa diversity measures (species richness, Shannon index, abundance and biomass). In subsequent models, detailed analyses will be performed using hierarchical categories within each global change driver (e.g., to examine the effects of different types of land use on soil fauna), as effects may vary within a driver. Mixed effects models (using the R package ‘metafor’), with random effects to account for non-independence of observations, will be used throughout.

All members of the project team have previously been or are currently involved in similar meta-analysis projects, and thus have preliminary work that can be used in this project. HRPP, EKC, and VJB have previously compiled search terms that identify soil organisms within the literature. This will be modified for this project in order to successfully capture all soil fauna. HRPP, LB, OF, NE and VJB have search terms that identify some of the global change drivers. These search terms will be modified and added to ensuring adequate representation of the global change drivers that will be studied in this project. Finally, HRPP, LB, EKC and OF have bibliographic information on datasets that encompass aspects of the soil fauna and global change drivers that this meta-analysis will deal with. For example, HRPP has bibliography data for papers that measured earthworm diversity in different land uses. This bibliographic information can be used to ensure that we are capturing the literature we expect to with our search terms.

HRPP, LB, EKC, OF and NE have previously done meta-analytical analyses, and thus have knowledge, information and R code for such analyses. In particular, we will modify previously used R code to increase efficiency. All code used in this project will be made publically available via GitHub (www.github.com).

All code will be made publically available via GitHub (https://github.com). Data will be made publically available on the iDiv portal (https://idata.idiv.de), but could also be made available on other platforms, such as the Natural History Museums’ data portal (http://data.nhm.ac.uk).

At least one high impact paper will result from this project. Leipzig University and iDiv will pay for all papers to be fully open-access. Press releases will accompany all papers, assisted by media teams at our institutes. All team members will present results in talks at conferences and seminars.

We will showcase the project at outreach events at iDiv (e.g., ‘Science Notes’, ‘Long Night of Science’) and the Natural History Museum, London (e.g., ‘Lates’). Blogs of team members will also communicate information about the project and results to scientific and non-scientific audiences.

The project will start in September 2018 (after the student assistant position has been filled). The literature search will be performed immediately. From personal experience (previous meta-analyses) and a preliminary literature search, we estimate that obtaining data will take 12 months (Screening of titles and abstracts: 10-15 days; Screening of main text: 30-60 days; Data extraction: 28 weeks), with the student assistant helping for 11 months (part-time, in accordance with German student employment law). Analysis will begin in July 2019 (before all data extraction is complete), and the first manuscript will be submitted at the end of the project (March 2020).

British Ecological Society - Large Research Grant (2018)

The effects of global change on soil faunal communities: a meta-analytic approach

The work will be conducted at the German Centre for Integrative Biodiversity Research (iDiv) Halle-Jena-Leipzig. There are further researchers within the institute that have a great deal of experience conducting meta-analysis (e.g., Dr. Dylan Craven, Dr. Simone Cesarz). A number of lab groups have a focus on soil and soil organisms (e.g., led by Prof. Dr. Nico Eisenhauer, Prof. Dr. Francois Buscot, Prof. Dr. Kirsten Küsel, and Prof. Dr. Nicole van Dam). We are in contact with researchers all over the world who have large datasets on soil faunal communities. iDiv also runs a synthesis centre (sDiv), where many high-profile researchers visit for working groups. iDiv has a High Performance Computer cluster (HPC), that are available if the models require more processing power. iDiv has a media and communications team, to assist with publicity of publications or events.

No conflicts of interest