Increasing evidence highlights the importance of biodiversity for the supply of ecosystem services and stability, making the ongoing loss of species a global concern (IPBES, 2019). Although a large part of the Earth’s biodiversity is literally hidden below ground, most research has focused on above-ground communities. We thus lack a comprehensive understanding of the effect of above–belowground interactions on ecosystem functioning, particularly in the context of accelerating biodiversity loss and climate change. The Symposium ‘Above- and belowground biodiversity for sustainable ecosystems’, supported by New Phytologist, provided a framework to connect researchers aiming to identify challenges and opportunities to drive this field of research forward. It took place in November 2019 at Agroscope, Zürich, Switzerland, and was organized in connection with a Pan-European BiodivERsA project (‘Digging Deeper’) on the conservation and sustainable management of biodiversity, focusing on soil biodiversity and agricultural diversification. In the following, we synthesize and elaborate on the most important challenges and research directions that were proposed during the Symposium as critical endeavors to gain a more holistic perception of the complex interactions between biodiversity, ecosystem functioning and factors of global change. Past empirical work has demonstrated that species loss can affect a wide range of ecosystem functions, including primary productivity (Van Der Heijden et al., 2008; Liang et al., 2016) and nutrient cycling (Philippot et al., 2013). Furthermore, the impact of biodiversity loss on the functioning of terrestrial ecosystems might even rival the effect of other drivers of global change, such as climate change (Hooper et al., 2012). Early biodiversity–ecosystem functioning (BEF) research has focused largely on the effect of biodiversity on individual ecosystem processes. However, assessing functions individually and disregarding trade-offs or synergies among functions might bias our understanding of ecosystem functioning (Byrnes et al., 2014). At the Symposium, various talks highlighted the importance of ecosystem multifunctionality (EMF), that is, the supply of multiple ecosystem functions simultaneously, and discussed its potential drivers. Such drivers are often ecosystems-specific, and Fernando Maestre (University of Alicante, Spain) showed that next to abiotic soil and climatic properties, functional plant diversity is a key factor for explaining EMF in drylands (Gross et al., 2017). Gaetan le Provost (Senckenberg Biodiversity and Climate Research Centre, Frankfurt am Main, Germany) further demonstrated that besides community traits and abiotic factors, landscape properties influence EMF as well. Ever since the emergence of the EMF concept, scientists have used a multitude of different approaches to represent the simultaneous supply of multiple ecosystem functions, making comparisons between studies difficult. Thus, there have been several efforts to review, standardize or give new directions in the use of EMF indices. At the Symposium, Eric Allan (University of Bern, Switzerland) presented the idea that by weighing distinct functions or groups of functions differently, an EMF index can be tailored to represent specific management priorities, which will inevitably affect the biodiversity–multifunctionality relationship (Allan et al., 2015). Additionally, Allan introduced concepts developed by Manning et al. (2018) highlighting the need to distinguish between ecosystem functions and ecosystem services, the latter referring exclusively to the benefits ecosystems provide to humans and not the underlying ecological processes. Although these recent developments have contributed to moving EMF research forward, two fundamental questions remain unanswered: how many functions or services need to be considered to appropriately gauge the multifunctionality of a given ecosystem; and which functions should an index implicitly contain to be a representative measure of overall ecosystem functioning? Addressing these questions would improve the pertinence of EMF studies and increase their reproducibility, which could further facilitate interpretations by the scientific community and eventually by policy-makers. BEF research has demonstrated that biotic interactions (e.g. competition and predation) are important driving forces for the assembly and functioning of soil communities (Soliveres et al., 2016). Multitrophic interactions thus probably have major implications for ecosystem functioning, particularly in the context of global change where biodiversity loss may occur across many taxonomic and functional groups of organisms. Yet, multitrophic research in terrestrial ecosystems remains in its infancy, and several speakers pointed out that most studies consider only one or a few individual groups of organisms when testing how ecosystems will respond to global change. For instance, Franciska de Vries (University of Amsterdam, the Netherlands) highlighted that most experiments looking at the response of soil ecosystems to drought have not included any plants, despite the evidence for cascading effects of drought on microbes and soil functioning through plant processes, particularly emphasizing the importance of root traits (Williams & Vries, 2020). Including plants, and more specifically plant traits, in global change experiments will be particularly important for gaining a better understanding of agro-ecosystems, where plant–soil feedbacks might be key to sustainable agricultural production (Bender & van der Heijden, 2015). However, although considering a multitrophic perspective is important, our basic understanding on the worldwide distribution, functioning and management of individual groups of soil biota such as nematodes, earthworms, protists and arbuscular mycorrhizal fungi is just beginning to be revealed, as discussed by Tom Crowther (ETH Zürich, Switzerland), Nico Eisenhauer (Leipzig University, Germany), Florine Degrune (Freie Universität Berlin, Germany) and Maarja Öpik (University of Tartu, Estonia). Soil ecologists have been studying the effects of many drivers of global change, but typically only one or two factors at a time are considered. Matthias Rillig (Freie Universität Berlin, Germany) demonstrated that the negative effects of global change on soil properties, functions and the microbial community increased with a growing number of drivers, highlighting the necessity to shift the focus towards examining higher-order interactions in global change research (Rillig et al., 2019). These new insights inevitably give rise to further questions such as what controls the response of soil ecosystems to multiple drivers of global change, and whether above- and belowground biodiversity can dampen the impact of multiple global change stressors. The responses of soil biodiversity and associated functioning to global change, in particular the effect of drought, received a lot of attention at the Symposium. Richard Bardgett (University of Manchester, UK) emphasized the role of intrinsic and extrinsic factors in regulating the response of soils to extended periods of drought (Bardgett & Caruso, 2020). For example, fungal-based food webs were found to be more resistant to drought, thus mitigating nitrogen and carbon losses via soil food webs (de Vries et al., 2012). The susceptibility of microbial taxa to drought further depended on their different life strategies, particularly benefitting opportunistic taxa (de Vries et al., 2018). Moreover, the presence and diversity of plants could mediate the response of soil communities and functions to perturbations (Fry et al., 2018). Despite this growing effort to elucidate possible reactions of soil ecosystems to stressors, time series experiments and surveys investigating long-term responses to global change are scarce. It is now crucial to take the dimension of time into consideration in experiments and observation studies to be able to predict whether, when and to what extent ecosystems will be affected by altered conditions, as both Nico Eisenhauer and Richard Bardgett strongly emphasized. Agriculture is one of the main drivers of global biodiversity loss, due to land-use change and agricultural intensification. This creates an urgent need to manage agro-ecosystems in a more sustainable manner, whilst meeting the rising demand for agricultural products. Intensive management practices are associated with high environmental costs and often neglect the ecosystem services that agricultural land might provide (Bommarco et al., 2018). A meta-analysis presented by Sara Hallin (SLU, Uppsala, Sweden) suggested that exploiting agro-ecosystem services to a greater extent could safeguard food security in regions with large yield gaps, and increasing biodiversity has been shown to be a particularly promising tool to ecologically intensify agricultural production. Additionally, Claire Chenu (INRAE, AgroParisTech, Thiverval-Grignon, France) showed that agricultural diversification through agroforestry, inter- or cover-cropping can help to maintain and even increase soil organic carbon stocks and thus promote climate change mitigation and adaptability (Chenu et al., 2019). At the global scale, there is a clear trade-off between production and sustainability goals (Knapp & van der Heijden, 2018), as discussed by Marcel van der Heijden (Agroscope, Zürich, Switzerland). However, it may be possible to overcome, or at least to minimize, this issue by utilizing the food we produce more efficiently. For example, Urs Niggli (FiBL, Frick, Switzerland) stressed that moving towards more sustainable food production requires not only efforts toward ecological intensification, but also reductions in food waste and meat consumption. The international Symposium on ‘Above- and belowground biodiversity for sustainable ecosystems’ highlighted the need to embrace the complexity of ecosystems in order to predict and thus manage how global change can affect biodiversity and, in turn, overall ecosystem functioning. When evaluating how ecosystems respond to factors of global change, interactions between global change drivers as well as the biodiversity across different trophic levels must be considered. International and transdisciplinary collaborations will be key to overcoming the underlying difficulties in the study of complex environmental systems across time and space. In conclusion, meeting these challenges and addressing the multiple dimensions of EMF can enhance our ability to develop strategies to maintain functioning terrestrial ecosystems in a changing world. Meanwhile, political action is urgently required to put existing evidence into practice, with particular emphasis on agro-ecosystem design and management. We thank Agroscope and New Phytologist for hosting and organizing the Symposium, as well as all contributors for their active participation and lively discussions. Moreover, we thank Marcel van der Heijden for constructive comments on this report. We acknowledge funding through the 2015–2016 BiodivERsA COFUND call for research proposals, with the national funders Swiss National Science Foundation (grant 31BD30-172466), Deutsche Forschungsgemeinschaft (317895346), Swedish Research Council Formas (contract 2016-0194), Ministerio de Economía y Competitividad (Digging_Deeper, Ref. PCIN-2016-028) and Agence Nationale de la Recherche (ANR, France; grant ANR-16-EBI3-0004-01).