Changes in the intensity and extent of farmland have a profound impact on life on Earth. Although nature-rich farmlands have conservation value, their promotion competes with the protection and restoration of natural habitat. Restoring the productive capacity of existing farmland can help maximize opportunities for the restoration of both nature-rich farmlands and natural habitat. Changes in the intensity and extent of farmland have a profound impact on life on Earth. Although nature-rich farmlands have conservation value, their promotion competes with the protection and restoration of natural habitat. Restoring the productive capacity of existing farmland can help maximize opportunities for the restoration of both nature-rich farmlands and natural habitat. Depending on how exactly agriculture is defined, approximately half of Earth’s habitable surface is farmed. Farming remains responsible for an estimated 80% of tropical deforestation,1Kissinger G. Herold M. De Sy V. Drivers of Deforestation and Forest Degradation: A Synthesis Report for REDD+ Policymakers. Lexeme Consulting, 2012Google Scholar making food production the dominant threat to biodiversity.2Tilman D. Clark M. Williams D.R. Kimmel K. Polasky S. Packer C. Future threats to biodiversity and pathways to their prevention.Nature. 2017; 546: 73-81Crossref PubMed Scopus (438) Google Scholar Agriculture also drives over one-quarter of global greenhouse gas emissions, two-thirds of freshwater withdrawals, and a large share of acidifying and eutrophying pollution,3Poore J. Nemecek T. Reducing food’s environmental impacts through producers and consumers.Science. 2018; 360: 987-992Crossref PubMed Scopus (1324) Google Scholar the effects of which extend far beyond the boundaries of our farmed landscapes. How society chooses to manage agricultural landscapes will therefore have a profound impact on the future trajectory of life on Earth by determining not only how hospitable those already converted lands are but also the extent to which future habitat conversion can be avoided and historic conversion can be reversed. The UN’s Decade on Ecosystem Restoration, which runs from 2021 to 2030, aims to ramp up efforts to protect and restore ecosystems in order to help improve livelihoods and combat the linked climate and biodiversity crises. Farmlands are recognized by the Decade as one of seven core ecosystems for restoration; but what does it mean to restore an ecosystem as artificial as farmland, and how might we reconcile competition with the restoration of more natural ecosystems such as forests and peatlands? Farmlands encompass a huge variety of land-use patterns ranging from irrigated arable land dominated by one or a few crops to three-dimensional agroforests and near-natural grasslands. All incorporate at least some elements of natural ecosystems and so support at least some biodiversity. As a general rule, though, higher-yielding farmlands—where semi-natural habitats tend to be scarce and pests and weeds are strictly controlled—support fewer species in lesser numbers than lower-yielding farmlands. We might be tempted, therefore, to restore as much farmland as possible to a lower-yielding, more wildlife-friendly state to ensure that all our farmlands are fully shared with nature. Yet, to achieve the same biodiversity value as natural forest, farmlands need to host about as much tree cover as natural forests,4Karp D.S. Echeverri A. Zook J. Juarez P. Ke A. Krishnan J. Chan K.M.A. Frishkoff L.O. Remnant forest in Costa Rican working landscapes fosters bird communities that are indistinguishable from protected areas.J. Appl. Ecol. 2019; 56: 1839-1849Crossref Scopus (9) Google Scholar and even sympathetically managed farmlands impose an environmental opportunity cost in comparison with natural habitats.5Phalan B. Onial M. Balmford A. Green R.E. Reconciling food production and biodiversity conservation: land sharing and land sparing compared.Science. 2011; 333: 1289-1291Crossref PubMed Scopus (1045) Google Scholar For instance, natural habitat can host aboveground carbon stores around an order of magnitude higher than farmland,6Williams D.R. Phalan B. Feniuk C. Green R.E. Balmford A. Carbon storage and land-use strategies in agricultural landscapes across three continents.Curr. Biol. 2018; 28: 2500-2505.e4Abstract Full Text Full Text PDF PubMed Scopus (19) Google Scholar and many forest-dependent birds are absent from polycultural shaded coffee farms but persist in farms that integrate patches of regenerating natural forest.7Chandler R.B. King D.I. Raudales R. Trubey R. Chandler C. Chávez V.J. A small-scale land-sparing approach to conserving biological diversity in tropical agricultural landscapes: land sparing and land sharing.Conserv. Biol. 2013; 27: 785-795Crossref PubMed Scopus (92) Google Scholar From the coastal machairs of the northwest British Isles to the savannah-like dehesa of Iberia and the pastoral landscapes of Alpine Europe, the most wildlife-friendly farmlands are those that are hardly farmed at all and have a high proportion of semi-natural habitats, low stocking densities, and few chemical inputs. The catch? Without dedicated mitigation measures, middle-of-the-road socioeconomic development pathways suggest that around two-thirds more food will be consumed in 2050 than in 2010, leading to corresponding increases in cropland demand and other environmental pressures.8Springmann M. Clark M. Mason-D’Croz D. Wiebe K. Bodirsky B.L. Lassaletta L. de Vries W. Vermeulen S.J. Herrero M. Carlson K.M. et al.Options for keeping the food system within environmental limits.Nature. 2018; 562: 519-525Crossref PubMed Scopus (946) Google Scholar Ambitious food-waste reduction and dietary change, if successful, will reduce (though not eliminate) these demands, but the widespread adoption of high-nature farming systems with substantially reduced agricultural productivity pushes in the opposite direction. There is a tension between the essential food-producing function of farmlands on the one hand and their restoration to a much more nature-friendly state on the other. If nature is best served by the protection and restoration of natural habitat, then perhaps on-farm restoration efforts should focus only on those ecosystem elements that support current and future food production. By encouraging restoration interventions that, as far as possible, maintain or enhance yields, we minimize the environmental opportunity costs of producing the food we need and leave space for some farmlands to be fully restored to an unfarmed (or hardly farmed) state. In hard-nosed agronomic terms, there are plenty of good reasons to engage in farmland restoration. The most obvious object of such efforts is the soil on which the majority of our food production depends. An estimated 16% of soils have lifespans of less than 100 years, presenting an imminent threat to ongoing food production.9Evans D.L. Quinton J.N. Davies J.A.C. Zhao J. Govers G. Soil lifespans and how they can be extended by land use and management change.Environ. Res. Lett. 2020; 15: 0940b2Crossref Scopus (19) Google Scholar Pro-soil actions such as reduced tillage and cover cropping have variable impacts on yields in the short term10Tamburini G. Bommarco R. Wanger T.C. Kremen C. van der Heijden M.G.A. Liebman M. Hallin S. Agricultural diversification promotes multiple ecosystem services without compromising yield.Sci. Adv. 2020; 6: eaba1715Crossref PubMed Google Scholar but are essential for sustaining food production in the long term. The restoration of other specific components of natural ecosystems, such as pollinators and natural enemies, can deliver direct benefits to food production, at least in crops and landscapes where these components are currently limiting. Globally, pollinator-dependent crops show lower and more variable yield growth than other crops,11Garibaldi L.A. Aizen M.A. Klein A.M. Cunningham S.A. Harder L.D. Global growth and stability of agricultural yield decrease with pollinator dependence.Proc. Natl. Acad. Sci. USA. 2011; 108: 5909-5914Crossref PubMed Scopus (244) Google Scholar suggesting that pollen delivery does constrain crop yield. Trees can reduce wind erosion (and produce food in their own right), whereas nitrogen-fixing legumes can wholly or partially replace artificial fertilizers. Such restoration efforts are not necessarily targeted toward biodiversity conservation per se but instead make use of the services provided by so-called “functional biodiversity.” But to the extent that these interventions promote a more varied agricultural landscape, create additional uncultivated areas (buffering watercourses or supporting pollinators, for instance), or reduce chemical use, then wider conservation benefits are likely.10Tamburini G. Bommarco R. Wanger T.C. Kremen C. van der Heijden M.G.A. Liebman M. Hallin S. Agricultural diversification promotes multiple ecosystem services without compromising yield.Sci. Adv. 2020; 6: eaba1715Crossref PubMed Google Scholar Is food-focused restoration the only restoration we should be pursuing across our farmlands? I suggest not. Even if the majority of species are dependent on natural or semi-natural habitat,5Phalan B. Onial M. Balmford A. Green R.E. Reconciling food production and biodiversity conservation: land sharing and land sparing compared.Science. 2011; 333: 1289-1291Crossref PubMed Scopus (1045) Google Scholar some species undoubtedly do better in farmed landscapes than unfarmed ones. Many of these are, admittedly, winners of the Anthropocene and so not necessarily a priority for conservation. But, for example, a suite of globally threatened birds—including bustards, cranes, waders, and ground-nesting passerines—are intimately associated with low-impact, “traditional” farming systems.12Wright H.L. Lake I.R. Dolman P.M. Agriculture—a key element for conservation in the developing world.Conserv. Lett. 2012; 5: 11-19Crossref Scopus (108) Google Scholar Prior to the advent of farming, such species were presumably rare and confined to areas subject to the natural disturbances for which farming now acts as a proxy. Although these natural disturbance processes could be restored, wildlife populations that have come to rely on farmland could fail to colonize, or even recognize, these wilder spaces. And if “farmland” species achieve lower densities in rewilded areas, these areas would need to cover a relatively large area to ensure viable populations in the absence of high-nature farming systems. Restoring traditional, low-input farming systems that incorporate semi-natural habitats is, therefore, an important tool in our conservation inventory. These high-nature farmlands support species of regional and global importance but are threatened by both abandonment and intensification. Any conservation strategy that is focused only on natural habitat would risk condemning these farmlands and the species they support.13Feniuk C. Balmford A. Green R.E. Land sparing to make space for species dependent on natural habitats and high nature value farmland.Proc. Biol. Sci. 2019; 286: 20191483PubMed Google Scholar There are other good reasons to promote at least some nature-focused restoration in farmlands. Sympathetically managed farmland can buffer and connect spared areas and support widespread and common species that foster human connections to nature. In Britain, skylarks—whose song has inspired countless works of art and literature—are most frequently encountered in farmland. Although the species would probably persist if forced to retreat to saltmarshes and semi-natural grasslands, such a range reduction would most likely limit opportunities for future generations to be inspired by the skylark’s song. Farmland restoration should also target agricultural pollutants, which are responsible for around one-third of terrestrial acidification and three-quarters of eutrophication.3Poore J. Nemecek T. Reducing food’s environmental impacts through producers and consumers.Science. 2018; 360: 987-992Crossref PubMed Scopus (1324) Google Scholar Modern chemical-intensive farming systems are often seen as synonymous with high yields and are pitted against lower-yielding, chemical-free alternatives. But agrochemical use is often highly inefficient, and targeted reductions in insecticide use, for example, can be achieved without any detectable impact on animal pest burdens.14Muneret L. Mitchell M. Seufert V. Aviron S. Djoudi E.A. Petillon J. Plantegenest M. Thiery D. Rusch A. Evidence that organic farming promotes pest control.Nat. Sustain. 2018; 1: 361-368Crossref Scopus (71) Google Scholar In other cases, it might be worth paying modest yield penalties in order to limit the negative burden imposed by agriculture on adjacent (and even distant) environments.15Schulte L.A. Niemi J. Helmers M.J. Liebman M. Arbuckle J.G. James D.E. Kolka R.K. O’Neal M.E. Tomer M.D. Tyndall J.C. et al.Prairie strips improve biodiversity and the delivery of multiple ecosystem services from corn-soybean croplands.Proc. Natl. Acad. Sci. USA. 2017; 114: 11247-11252Crossref PubMed Scopus (142) Google Scholar Many of these sympathetic farming practices are relatively low cost, and some can even increase profits by reducing inputs. Similarly, some of these interventions incur virtually no yield penalty; unsown skylark plots, for example, provide nesting opportunities at the expense of <0.5% of the cropped area. But the most nature-friendly systems, which are needed to support the most sensitive and threatened farmland species, are likely to entail large reductions in outputs and profits (Figure 1); we should focus support for such restoration efforts where it really counts. Although the restoration of high-nature farming systems and that of natural habitats are not mutually exclusive, they are in competition. Given a finite area of land, the more we choose to reduce yields through on-farm conservation, the less slack we leave for reverting farmlands back to natural habitat. This is true whether or not food demand goes up (alongside rising populations and incomes) or down (through waste reduction and dietary changes). The relative importance of these competing restoration strategies depends on conservation objectives. Although evidence continues to stack up in favor of a land-sparing approach (i.e., prioritizing the protection and restoration of natural habitat),5Phalan B. Onial M. Balmford A. Green R.E. Reconciling food production and biodiversity conservation: land sharing and land sparing compared.Science. 2011; 333: 1289-1291Crossref PubMed Scopus (1045) Google Scholar this does not rule out some land sharing in some places. Mixed approaches such as described by Feniuk et al.13Feniuk C. Balmford A. Green R.E. Land sparing to make space for species dependent on natural habitats and high nature value farmland.Proc. Biol. Sci. 2019; 286: 20191483PubMed Google Scholar will be necessary whenever decision makers are interested in both species dependent on natural habitat and those dependent on high-nature farming. The development of regional restoration strategies should identify targets and opportunities for different forms of restoration while accounting for local conservation objectives and socio-economic considerations. The restoration of high-nature farming might be prioritized in regions with strong cultural and historical links to such farming systems (e.g., Scottish crofting), where existing populations of sensitive farmland-dependent species are in need of bolstering (e.g., cirl buntings [Emberiza cirlus] in southwest England), or where higher-yielding farming systems are economically unviable (e.g., in flood-prone or otherwise marginal areas). In practical terms, on-farm conservation is likely to encounter fewer social barriers than wholesale reversion to natural habitat, though monitoring and enforcement are perhaps more challenging because of the diffuse nature of on-farm interventions. In contrast, the restoration of natural habitat might be more appropriate on larger land holdings, in areas where the environmental costs of continued farming are high (e.g., on peat soils or steep slopes), and where augmenting the existing habitat network (e.g., to connect fragmented heathlands, grasslands, or woodlands) is likely to increase particular ecological value. Any such exercise should engage local communities and build trust among stakeholders and in delivery mechanisms. As alluded to above, there are three key levers that could be used to create sufficient slack within food systems to allow nature-focused restoration (Figure 2). First, the potential area of land needed to meet food demand can be minimized by improved crop and animal breeding, the closing of yield gaps, and the encouragement of practices that maintain or enhance yields (see “restoring the productive capacity of farmland”). But to be clear, although yields are important, they should not be blindly pursued without consideration of negative environmental externalities and long-term sustainability. Second, overall demand for food production can be minimized by the reduction of food waste, which currently results in the loss of around one-third of the food we produce.8Springmann M. Clark M. Mason-D’Croz D. Wiebe K. Bodirsky B.L. Lassaletta L. de Vries W. Vermeulen S.J. Herrero M. Carlson K.M. et al.Options for keeping the food system within environmental limits.Nature. 2018; 562: 519-525Crossref PubMed Scopus (946) Google Scholar Third and finally, dietary change can reduce the land burden associated with meeting our nutritional requirements. Plant-based diets have the potential to reduce food’s land use by up to three-quarters,3Poore J. Nemecek T. Reducing food’s environmental impacts through producers and consumers.Science. 2018; 360: 987-992Crossref PubMed Scopus (1324) Google Scholar easing pressure on both feed-producing croplands (which could instead be used to produce human-edible crops or released for restoration) and fodder-producing grasslands (many of which are located in once-forested regions). All three levers have a role to play in bringing the food system back within environmental limits.7Chandler R.B. King D.I. Raudales R. Trubey R. Chandler C. Chávez V.J. A small-scale land-sparing approach to conserving biological diversity in tropical agricultural landscapes: land sparing and land sharing.Conserv. Biol. 2013; 27: 785-795Crossref PubMed Scopus (92) Google Scholar Importantly, though, these levers alone are insufficient. Critics of the land-sparing approach rightly point out that yield growth risks simply inducing additional agricultural production rather than sparing natural habitat. Equally, dietary change and waste reduction don’t automatically result in restoration; farmers who find that their products are no longer in demand could simply switch to producing alternative non-food commodities. Strong policies that target the linked food and land systems are needed to ensure that the three levers of yield growth, waste reduction, and dietary change realize their full restoration potential. To conclude, farmland restoration should not be considered in isolation from the restoration of other ecosystems. Restoration efforts that boost agricultural productivity can potentially spare land, but we have choices to make about whether this slack is used for the restoration of high-nature farming or natural habitat. The Decade on Restoration should consider these linkages alongside the demand-side levers of food-waste reduction and dietary change.