Agriculture is the most dominant and widespread land use globally (Ramankutty et al., 2018), and a third of it occurs in areas of high biodiversity conservation priority (Hoang et al., 2023). The conversion of natural ecosystems to farmland and the intensification of existing agricultural systems are the leading drivers of global terrestrial biodiversity loss (Maxwell et al., 2016; IPBES, 2019; Jaureguiberry et al., 2022). Bending the curve of biodiversity loss necessitates tackling its root causes; thus, entailing major food system transformations that include both demand-and supply-side efforts (Díaz et al., 2019; Leclère et al., 2020; Williams et al., 2021).

The erosion of biodiversity within farmland can result in negative feedback on agricultural yields (Burian et al., 2024). The productivity, sustainability, and resilience of food production systems are underpinned by biodiversity and associated ecosystem services, such as pollination, natural pest suppression, and nutrient cycling (Dainese et al., 2019; IPBES, 2019). There is ample evidence to suggest that biodiversity loss in agricultural landscapes increases risks to food production and human well-being (Foley et al., 2011; Pretty, 2018; Garibaldi et al., 2021; Burian et al., 2024; Maney et al., 2024).

However, despite the inextricable, multifaceted link between providing food security for people and conserving and restoring biodiversity, the two challenges have historically been separated within academia, policy, and practice (Glamann et al., 2017; IPBES, 2024). This is particularly acute in India, where biodiversity features little at the food-water nexus (Martin et al., 2024). Food system and nature conservation programmes or policies are often poorly integrated, and by focussing on a singular aspect the potential for trade-offs and negative repercussions is high (Liu et al., 2018; Meyfroidt et al., 2022; IPBES, 2024).

At the global policy level, there have been increased efforts to better integrate the twin challenges. For example, addressing food production and consumption is central to the Kunming-Montreal Global Biodiversity Framework (GBF), biodiversity is acknowledged to play a key role in supporting the United Nations Sustainable Development Goals, and the interlinkages between food and biodiversity were recently assessed in the IPBES Nexus Assessment (IPBES, 2024). However, operationalising the integration of the two challenges and associated transformative change necessitates a far better understanding of the complexities of implementation in different socio-ecological and socio-political contexts (Delabre et al., 2021).

Navigating conflicts between agriculture and conservation is particularly challenging in India, with agriculture representing the largest threat to terrestrial vertebrates (Chaudhary et al., 2022). India has an exceptionally high agricultural land use footprint (Roy et al., 2015), which has led to a low connectivity between natural ecosystems and mixed effectiveness of the national protected area network (Ghosh-Harihar et al., 2019; Sengupta et al., 2024).

Conversely, without careful planning, a drastic expansion of the protected area network could threaten food security (Mehrabi et al., 2018). India has the highest number of undernourished people globally, and if current trends continue, 60% of India’s population is likely to experience severe calorie-, digestible protein-and fat deficits by 2050 (Ritchie et al., 2018). The excessive and non-targeted use of agrichemicals and extensive irrigation have been supported by government subsidies, reflecting powerful political-economic interests, including those of corporate agri-food industries (Pingali et al., 2019). This has not only led to a significant erosion of ecosystem services and to a heightened susceptibility to environmental shocks, but also to a pervasive agrarian crisis (Vasavi, 2009; Rahman, 2015; Veluguri et al., 2021; Chaudhary and Krishna, 2024).

In response, numerous grass-roots movements and government-led programmes have emerged across India to re-design agricultural systems for enhanced livelihoods, sustainability and equity. The largest of these is the Zero Budget Natural Farming (ZBNF) programme (also termed Andhra Pradesh Community Managed Natural Farming, APCNF; Supplementary Figure 1) in the southern state Andhra Pradesh. ZBNF entails regenerating biotic interactions underpinning yield-supporting ecosystem services, using locally available, non-synthetic ingredients to make bio-inoculants (Bharucha et al., 2020; Smith et al., 2020; Supplementary Figure 1a). Notably, ZBNF represents one of the largest agroecological transitions globally (Veluguri et al., 2021). The government of Andhra Pradesh is aiming to roll out ZBNF to all six million farmer households in the state, spread across almost 9 million hectares, by 2030 (RySS, 2024), with the hope of improving farmer well-being and the economic sustainability of farms. Despite the central role of biodiversity in ZBNF, and ZBNF being acknowledged as highly relevant to global biodiversity conservation (Sutherland et al., 2020), the relationship between ZBNF and biodiversity conservation has received almost no research or policy interest (but see Berger et al., 2024).

Andra Pradesh is one of India’s major rice-producers, and over 60% of its population depends on agriculture for their livelihood (Department of Agriculture and Farmers Welfare, 2023). It also contains numerous sites of national conservation importance because they represent rare natural habitats, provide key ecosystem services, and/or have a high diversity of threatened species (Srivathsa et al., 2023). Consequently, Andhra Pradesh encompasses areas where agricultural commodities (in particular, rice) are grown in areas of high conservation priority (Hoang et al., 2023). Agricultural encroachment into natural ecosystems has continued over the last few decades (Hansen et al., 2013; Roy et al., 2015; Pendrill et al., 2022; Potapov et al., 2022; Department of Agriculture and Farmers Welfare, 2023; Meng et al., 2023). Despite this, agriculture and the need for transformative change in food systems features little in state-level biodiversity strategies (AP State Biodiversity Board Vision Plan, 2024).

Thus, food and biodiversity challenges are strongly interlinked in Andhra Pradesh, however, the research, management and governance of them has largely been separate. Finding pathways to overcome this demands stakeholder perspectives on the barriers and enablers within relevant sectors and institutions, such as government, farming, business, academia, and conservation organisations. Our study is built on a stakeholder consultation with the overarching goal of identifying how sufficient food for people can be obtained while conserving and restoring biodiversity in Andhra Pradesh. Specifically, we present policy-, practitioner-, and researcher-led visions of what this goal would look like, what the key knowledge gaps are that impede achieving this vision, and what the top priorities for policy and action are. This consultation is the first of its kind in this context and, given Andhra Pradesh’s unique agricultural context, represents a globally highly relevant case study.

We agreed on a vision that entails a widespread adoption of farming practices that do not erode farmland biodiversity (i.e., diversity of wild species residing in and around agricultural land). This includes ZBNF, but more generally, we envisioned a few central principles and features, such as restored soil health, greater resource (including water and nitrogen use) efficiency without loss of agricultural yields, no or low levels of agrichemical inputs, and heightened ecological functioning. In line with the principles of agroecology (FAO, 2018; Barrios et al., 2020; Sietz et al., 2022) we envisaged a re-focus away from the singular objective of maximised yields to multifunctional agricultural systems that perform well along multiple dimensions. We stress the importance of strong support for and integration of biodiversity and associated ecosystem functions and services (such as pollination, nutrient cycling, and pest control) in food production systems, while acknowledging that not all species utilising agricultural landscapes contribute to food production and actions aimed to benefit those species should not compromise food security.

We also agreed that high levels of agrobiodiversity (i.e., a large variety of domesticated species used for food and agriculture; Jago et al., 2024) will be critical for a sustainable future (Supplementary Figure 1b), especially with regards to climate resilience and food security and sovereignty. We envisioned diversified cropping systems, especially those involving indigenous crop varieties, to play a central role in maintaining indigenous knowledge systems and overcoming undernutrition and micronutrient malnutrition in Andhra Pradesh, given that a high crop diversity frequently, albeit not necessarily, improves the dietary diversity of smallholders (Sibhatu et al., 2015).

Our visions slightly diverged regarding the role of emerging technologies and artificial intelligence. Whilst some envisioned that future food demand will purely be met through ZBNF and other agroecological farming practices, others had a more technocratic vision entailing new biotechnologies, including genetic modification, and precision agriculture. The latter harnesses the power of technologies such as drones, hyperspectral cameras, and thermal sensors to boost the productivity and sustainability of agriculture (Gebbers and Adamchuk, 2010; Cisternas et al., 2020). Nonetheless, we agreed that the future lies in finding the right balance, with measures tailored to each region’s context.

Overall, given that agroecological approaches can facilitate transformative change of food systems and optimise interlinked ecological, socioeconomic, and political processes (Wezel et al., 2020), we agreed that ZBNF and other agroecological farming practices are essential elements for achieving the vision of sustainable food and agriculture in Andhra Pradesh by 2040.

We agreed on a vision that entails high landscape heterogeneity which fosters species persistence and ample ecological corridors connecting natural ecosystems, including those protected. The quality and permeability of agricultural landscapes tends to strongly influence the effectiveness of area-based conservation efforts (Fletcher et al., 2024). Hence, holistic, integrated landscape-scale conservation planning, where efforts to boost farmland biodiversity are aligned with conservation strategies focused on natural ecosystems, are needed in Andhra Pradesh. We agreed that identifying priority sites for wildlife movement, habitat potential for endangered and endemic species, and restoring natural vegetation patches within the agricultural matrix will be key to reaching the biodiversity goals of sustainable agriculture in the state. We stress that specific strategies must be attuned to local contexts, with different land system objectives carefully balanced. In some instances, identifying strategies that optimise ecosystem service provisioning will be at the forefront, whereas in other instances identifying ways of channelling wildlife through the matrix to minimise human-wildlife conflict will be critical. Moreover, we envisioned land-use planning to be a highly democratic, equity-focussed, transparent, and inclusive process, involving marginalised rural and forest-dependent people, including indigenous communities.

We also highlight that, in comparison to other Indian states, the forest restoration potential is high in Andhra Pradesh (Gopalakrishna et al., 2022) but stress that careful consideration must be given towards how competing demands for agriculture and forest restoration will best be met, envisioning, based on recent scenario-modelling (Mehrabi et al., 2018; Wolff et al., 2018), that this will likely involve mosaics of forests and agriculture. Nonetheless, Andhra Pradesh hosts a diverse set of natural vegetation types (Reddy, 2010) and which of these should be the focus of conservation efforts must vary by region. For example, in the Eastern Ghats of northern Andhra Pradesh moist deciduous forests must be better protected (Reddy, 2010), whereas in the South, open habitats, such as grasslands and shrublands, are predominant (Reddy, 2010), but these are currently neglected at the state and national level (Reddy, 2010; Sengupta et al., 2024). Efforts to restore natural ecosystems must be evidence-based. We stressed that timber plantations, especially when entailing single, non-native tree species, have extremely limited biodiversity value and should not contribute towards meeting area-based conservation targets. Overall, we envisaged a move away from poorly designed and executed tree-planting programmes and from a protected area network that undermines the rights of indigenous communities and is not ecologically representative.

Each of our perceptions of the adequacy and effectiveness of current area-based and species-centred conservation efforts in Andhra Pradesh dictated how we envisioned the idealised future for 2040. Some of us, predominantly those with an agronomy background, did not necessarily envision any changes from now, likely primarily valuing biodiversity for its contribution to food production. Those of us with an arguably more multidimensional view of nature and its relationships with people (see Mace, 2014), in particular conservation scientists and practitioners, considered current efforts wholly inadequate and envisioned a step-change in the quantity and quality of Andhra Pradesh’s protected area network, the amount of funding for conservation and restoration efforts, and the training and support available to practitioners. We highlight that the biodiversity conservation capacity in Andhra Pradesh, especially of the Forest Department, is presently exceptionally low in comparison to other Indian states and regions.

The third aspect of the vision was to achieve social equity, including gender and caste equity, with the existing economic, political, and sociocultural patterns that are driving ecological and social degradation disrupted. We envisioned a food system sitting within wider transformed economies, where structural barriers that promote unsustainable land systems are dismantled. This includes ending harmful agricultural subsidies such as those for agrichemical inputs, groundwater extraction, and seeds of climatically unsuitable crop varieties. Conversely, we envisaged enabling policies and initiatives to be nurtured, such as improvements in infrastructure, value chains, and social organisations. We also envisioned low levels of post-harvest crop losses and food waste, and behavioral change with respect to food consumption, health, and nutrition. Substantial dietary behavior changes are needed to deliver desirable human nutrition and environmental outcomes in Andhra Pradesh, where this entails lower intakes of foods such as rice, sugar, and poultry, and higher intakes of fruits, vegetables, and pulses (Chaudhary and Krishna, 2021). Thus, we imagined increased incentives for growing and consuming more diverse, locally appropriate, and nutrient-rich crops. Furthermore, we envisioned that power and policy asymmetries will be addressed, and participatory and transparent decision-making frameworks created, with the values and knowledge systems of a diverse set of actors (including tribal farmers, women, and youth) incorporated.

Agrichemicals, fossil fuel-based energy, and credits to conventional agriculture are still subsidised in Andhra Pradesh (Department of Fertilizers, 2024; Reddy, 2010). Phasing them out and redirecting finance to activities compatible with our vision will be critical through, for example, payment for ecosystem service (PES), certification, and agri-environment schemes. At present, the ZBNF programme does not involve financial support to farmers, and thus innovative financing mechanisms, such as sustainable certification schemes and carbon and biodiversity markets, could accelerate the transition and enhance the financial sustainability of the ZBNF programme.

Certification standards and PES schemes can be effective at delivering desirable ecological and social outcomes when carefully designed and implemented (Lambin et al., 2014; Tscharntke et al., 2014; DeFries et al., 2017). They require strong institutions, and we believe that the capacity in Andhra Pradesh to conduct detailed valuations of incremental changes in ecosystem services and biodiversity needs to be improved in order for these policy instruments to be effectively employed. Nonetheless, tools for evaluating ecosystem services (such as the TEEBAgriFood Evaluation Framework and TESSA - Toolkit for Ecosystem Services Site-based Assessment) are now widely accessible and relatively easy to implement, and technological advances in biodiversity monitoring hold promise (Besson et al., 2022).

PES and certification schemes themselves must be designed in a manner that does not reinforce existing power asymmetries and injustices. For example, farmers with small landholdings and/or without formalized tenure arrangements should not be excluded. Furthermore, small-scale enterprises, cooperatives, and grassroots community organisations (including Farmer Producer Organisations) play a central role in supporting farmers and in implementing PES schemes. They should thus be supported financially and via capacity building. Conversely, the disproportional political-economic power of agribusinesses creates lock-ins of current unsustainable conditions and must be dismantled.

Andhra Pradesh’s food systems are susceptible to numerous interacting shocks and stressors, and diversification along the entire value chain is important to improve their ecological, social, and economic resilience.

Global climatic changes are increasing the susceptibility, coverage, frequency, and severity of environmental hazards to food production. Climate change is likely to substantially alter land suitability for agriculture in the future, where currently productive areas might fail to produce enough food in the future (Jia et al., 2019). It is already having adverse effects on agriculture in Andhra Pradesh (Rao et al., 2017; Mishra and Aadhar, 2021), and numerous districts in the state are highly vulnerable to increasing temperatures, changes in rainfall patterns, and increased occurrence of extreme weather events (Rao et al., 2017). ZBNF and other diversification approaches likely reduce climate change vulnerability, however, a holistic approach for resilience and risk reduction is needed, ideally harmonised at the national level. While 80% of farms already have some form of insurance in India (GIZ, 2021), the system is beset with several problems and necessitates reforming (Mahul et al., 2012; Gulati et al., 2018). As a first step, cross-sectoral meetings convened by the public policy think-tank NITI Aayog, government ministries (Ministry of Agriculture and Farmers’ Welfare, Ministry of Environment, Forest and Climate Change, and Ministry of Earth Sciences), the statutory insurance body (Insurance Regulatory and Development Authority), state governments, and other key stakeholders, such as insurance companies, will be key to developing a policy framework and operational guidelines for insurance schemes.

Adopting agroecological and biodiversity-friendly practices can pose a risk to farmers, with benefits frequently only materialising a few years post intervention and productivity outcomes being highly context-dependent (Tamburini et al., 2020; Rasmussen et al., 2024). The risk of transitional dynamics and unwanted outcomes due to context-dependency should not be borne by the farmers. RySS has conducted experiments on their own fields (Bharucha et al., 2020; Duddigan et al., 2023) and encourages farmers to initially adopt ZBNF on only part of their land, however, landholdings are small and many biophysical processes that drive yield benefits operate at larger scales (Ghazoul et al., 2009; Edwards and Laurance, 2012; Tscharntke et al., 2014). Thus, a more comprehensive de-risking framework, where farmers are compensated for yield losses in the early years, is needed. Social safety nets and household income diversification may additionally improve rural households’ food security (Hertel et al., 2023).

Concentration of power in supply chains, including trade, and volatile market prices pose a threat to economic resilience (Naylor and Falcon, 2010). In general, purchase programmes and price guarantee policies hold promise, but given that each supply chain has unique characteristics, solutions must be locally-grounded, and commodity-specific.

At present, ZBNF produce does not universally fetch premium prices. There are some localised NGO-led efforts to create ZBNF-specific markets (e.g., by the Centre for Sustainable Agriculture, Jattu Trust, and Rainforest Alliance), but coordinated government and business support is urgently needed for this to scale at the state-level and to create a ZBNF brand value. Value-addition through processing would likely enhance ZBNF farmers’ (and farmers employing other biodiversity-friendly practices) access to competitive regional and national markets. Tilting inter-state trade agreements in favour of ZBNF farmers will also be crucial, as will improved transparency in supply chains.

Furthermore, physical infrastructure needs to be improved to facilitate ZBNF farmers’ access to markets, and, more fundamentally, to reduce post-harvest crop losses. Similarly, demand-side interventions tackling food waste will be critical since food losses as a whole represent a substantial sustainability challenge in Andhra Pradesh (Jha et al., 2015; Agarwal et al., 2021). However, devising evidence-based policies is hampered by a limited research base on the effectiveness of different interventions in India (Agarwal et al., 2021). A better understanding of the demographic and sociopolitical factors shaping food consumption is needed (Pandey et al., 2020) in order to identify crucial levers to reduce food waste and to transition to sustainable diets more broadly.

Agroforestry is regarded as a pivotal climate mitigation and adaptation strategy with positive livelihood outcomes that can also reduce deforestation (Teo et al., 2025). However, farmland and agroforestry trees are rapidly disappearing across India, including in Andhra Pradesh (Brandt et al., 2024; Supplementary Figure 1d). Reducing structural differences and increasing resource similarity between natural ecosystems and adjacent agricultural landscapes can increase species’ effective area of habitat, ecological connectivity, and population sizes (Kremen and Merenlender, 2018; Fletcher et al., 2024). Agroforestry may be one strategy to deliver this in certain contexts (Ferreira et al., 2020), i.e., in regions that would be naturally tree-dense rather than grassland-dominated. Other strategies may include setting parcels of land aside for ecosystem restoration, but any such efforts must be community-led and extreme care must be taken to ensure that they do not result in land dispossession and are ecologically appropriate.

Conservation models focussed on private lands are in their infancy in India compared to other nations (Drescher and Brenner, 2018; Mariyam et al., 2021). However, OECMs (“other effective area-based conservation measures”) being critical to meeting India’s area-based conservation targets in line with GBF (Sengupta et al., 2024) represents an opportunity to up-scale private land conservation models (surrounding and connecting protected areas). To reduce conflict and transfer conservation benefits to local communities, innovative land system policies and incentive schemes are needed (Anand et al., 2010; Karanth and Karanth, 2012; Ghosh-Harihar et al., 2019). These must be co-designed with land-owners and managers to be effective, as the willingness to participate in such schemes varies by socioeconomic group (Mariyam et al., 2021).

Landscape-level zonation and prioritisation exercises can be useful tools for guiding land management and policymaking by highlighting the relative importance of different areas towards meeting biodiversity, ecosystem service, food security, human well-being, and other land system goals. However, priority mapping, especially at the global scale, has received ample criticism, both on ecological and equity grounds (Wyborn and Evans, 2021; Fleischman et al., 2022; Schultz et al., 2022). Nonetheless, we believe that priority maps for Andhra Pradesh would represent valuable, structured decision-support frameworks that can help identify where synergies as well as trade-offs between different land system objectives are likely to be particularly pronounced. A comparison to national priority maps may further guide activities in Andhra Pradesh (see Srivathsa et al., 2023). However, these maps must be locally-grounded, incorporating fine-scale (field) data, and involving organisations that represent farmers and other land managers in the process.

Building on these maps, context-specific policies should be devised that are harmonised at the state-level and are aligned with national strategies. At present, different institutions and government departments operate in silos, and local and decentralised forms of governance are not well-integrated (Bharucha et al., 2020). The public sector agricultural extension services, particularly those supported by the Andhra Pradesh Department of Agriculture, ensure continuity and convergence in programmes and facilitate adaptive governance (Singh and Burman, 2019; Raina et al., 2022). Thus, they must thus be strengthened. Furthermore, the footprint of individual conservation interventions is small, and in order for these to scale, greater collaboration in integration of programmes is needed between conservation organisations.

Landscape-scale policymaking and planning must be sensitive to the context-dependency of the effectiveness of agricultural and conservation interventions. For example, ZBNF’s impact on agricultural yields appears to depend on the crop type, biophysical conditions, completeness of adoption, and the counterfactual farming system it aims to replace (Bharucha et al., 2020; Duddigan et al., 2023; GIST Impact Report, 2023; Duddigan et al., 2024; but see Berger et al., 2024). Furthermore, the impact of interventions frequently exceeds the direct local land footprint, with both positive and negative spillover effects being common. Thus, interventions are less likely to have unwanted outcomes and land-use planning is more effective when measured or conducted across scales, including at larger (e.g., state-wide) scales that can capture more complex land use dynamics (Meyfroidt et al., 2022; zu Ermgassen et al., 2024).

To transition to the shared vision through transformative, multi-scalar policy change, we considered the types of governance needed and how policy enforcement could be equitable and inclusive. We agreed on a widely shared notion of policy choices, action items, and the problems associated with ineffective governance, and we consequently suggest five policy actions (Table 2), under the governance and actors framing of environmental governance (Ostrom, 1999). Simply put they are: support farmers; incentivize reduction in agrichemical use; engage with industry and economy; promote investment in biodiversity-friendly farming; and promote interdisciplinary science and innovation. Each of these priorities together aids in a transition from the current high-agrichemical use to sustainable food production systems through environmental governance strategies that are a mix of regulatory, market and agent-focused instruments, community-based, and hybrid governance.