Introduction

Food and land systems are key to food security and wellbeing and are increasingly regarded as a key driver of environmental impacts and a major contributor to global environmental change (Hoek et al. 2021; Willett et al. 2019). Land-use change, biodiversity loss, freshwater use, atmospheric greenhouse gas (GHG) emissions, and nitrogen (N) and phosphorus (P) use have all surged due to agricultural expansion and intensification (Campbell et al. 2017; Foley et al. 2011; IPCC 2019; Sukhdev 2018). Global demand for agricultural goods is expected to increase further with population, income growth, and dietary shifts (Crist et al. 2017; FAO 2018; Pereira et al. 2020; Willett et al. 2019). Recent studies, therefore, warn against the continuation of a business-as-usual trajectory of agricultural and land-use management, calling for a system transformation to ensure a sustainable trajectory for humanity (Springmann et al. 2018; Willett et al. 2019; Clark et al. 2020; Steiner et al. 2020; Fanzo 2021; FOLU 2019, 2021; Hebinck et al. 2021; Rockström et al. 2020; Webb et al. 2020; Oliver et al. 2018).

While a food and land system transformation has been defined and modelled at the global level (Searchinger et al. 2018; Willett et al. 2019; FOLU 2019, 2021; Steiner et al. 2020), there is an urgent need to elaborate on what this would entail at the national level, particularly given the diverse starting points and roles of different countries and regions in a globalised agri-food system and to manage for the power dynamics and imbalances that exist within it (Allen and Wilson 2008; Howard 2021; Pereira et al. 2020; Steiner et al. 2020). The pathway towards food and land system transformation at the national level is not clearly defined (Sukhdev et al. 2016; Béné et al. 2019b). There is a need for national-scale frameworks to identify the sustainability transitions required for system transformation and prioritise solutions that can deliver the best outcomes to meet United Nations Sustainable Development Goals (SDGs) (UN 2015) that consider the complexity of the underpinning socio-ecological, socio-technical, and political-economic systems (Oliver et al. 2018; Pereira et al. 2020; Fesenfeld et al. 2022; FOLU 2021). Concurrently achieving all SDGs within the 2030 timeframe presents many challenges and requires managing the tensions between development, the environment, and the inherent trade-offs between SDGs (Bryan et al. 2019; Griggs et al. 2017; Orbons et al. 2024; Pradhan et al. 2017).

Progressing the global sustainability agenda requires successful national-scale implementation of solutions (Gao and Bryan 2017). There are many competing narratives as to what constitutes a sustainable food and land system and what the optimal mix of solutions is for achieving system-level sustainability (Springmann et al. 2018; Searchinger et al. 2018; Roe et al. 2021; Mosnier et al. 2022; CSIRO 2023; NFF 2019; McRobert et al. 2022; Béné et al. 2019a). Several studies have highlighted the need to move beyond a focus on productivity or single-paradigm approaches (Allen and Wilson 2008; Howard 2021, 2023; Dornelles et al. 2022; Lindgren et al. 2018; Sukhdev 2018; Faulkner 1944), suggesting a shift to a systems approach in defining and measuring sustainability to account for regional variations at the national and sub-national scale (Fanzo et al. 2021; Hebinck et al. 2021). Global scale frameworks have been developed for establishing and monitoring progress towards indicators (e.g., Fanzo et al. 2021; Hebinck et al. 2021; Jones et al. 2016; Stefanovic et al. 2020; Willett et al. 2019), for supporting decision-making and the implementation of solutions for system change (Béné et al. 2019a; Silva et al. 2022; TEEB 2018). Australian-specific sustainability frameworks identifying indicators and roadmaps for a sustainable food and agriculture system have been developed (McRobert et al. 2022; CSIRO 2023), but lack focus on which solutions should be prioritised and their potential synergies and trade-offs, which are critical for supporting strong governance, decision-making, and negotiations between stakeholders (Hebinck et al. 2021; Oliver et al. 2018).

The sustainable transition of food and land systems requires the rapid implementation and scaling up of a broad suite of behaviour-oriented (demand-side) and technology-driven (supply-side) solutions as well as alternative paradigms such as agroecology (Béné et al. 2019a; Röös et al. 2017; Herrero et al. 2020; Wezel et al. 2014, 2009; Wezel and Soldat 2009; Gordon et al. 2023). Explicit procedures for assessing solutions are required (Tàbara et al. 2017); however, solution prioritisation depends heavily on indicators selected for assessing system sustainability (Garnett 2014). The diverse views on choice and weighting of indicators diverge even more at regional and national scales (Bennett et al. 2021). To manage for this at the national and sub-national scale, to capture the local specificity of food and land systems, local contexts and stakeholders should guide the development of locally relevant indicators and solutions (Bandari et al. 2022; Béné et al. 2019a; Moallemi et al. 2020, 2021; Szetey et al. 2021).

Frameworks are required to prioritise solutions and focus the development of integrated models for scenario analysis, to inform policy and highlight knowledge and technology gaps (Nilsson et al. 2016, 2018). However, to adequately capture the complexity of sustainable transitions and support effective adoption, frameworks must account for the diversity of stakeholders across the food and land system from land-use practitioners and civil society to national governments and private sector and navigate the competing dimensions of food and land system sustainability and the complex interdependencies (and trade-offs) between SDGs (Béné et al. 2019b; Oliver et al. 2018; Pereira et al. 2020).

System-level change requires transdisciplinary collaboration across a broad range of stakeholders. This enables more diverse views and values, and minimises the risk of unforeseeable consequences and/or trade-offs, to more comprehensively reflect available knowledge and conceptualise novel sustainability innovations (Schneider et al. 2021; Moallemi et al. 2020; Nielsen et al. 2020; Mauser et al. 2013; Pereira et al. 2020). Iterative and collaborative processes that integrate knowledge and stakeholders from diverse domains are known as co-production or co-creation (Mauser et al. 2013; Reed et al. 2022; Wyborn et al. 2019) and result in context-specific knowledge which can be used to underpin the development of locally contextualised sustainability pathways (Chambers et al. 2021; Mauser et al. 2013; Norström et al. 2020). The value of co-produced knowledge is well established in the field of sustainability science (Jassanoff 2004; Moallemi et al. 2020, 2021) leading to mutually reinforcing and reciprocal outcomes that represent more inclusive, legitimised, impactful, and systemic change for local contexts (Jassanoff 2004; Norström et al. 2020; Schneider et al. 2021; Wyborn et al. 2019). As such, adopting co-production methods can improve the integration of environmental, social, economic, political and cultural factors into conceptualising system sustainability, and support navigating synergies and trade-offs in a just, transparent, and efficient manner (Béné et al. 2019b; Chambers et al. 2021; Moallemi et al. 2021, 2022).

In this study, we bring together a diverse range of stakeholders to co-produce an extensive suite of nationally relevant SDG-aligned sustainability indicators and solutions for the Australian food and land sector. We then apply an adapted interaction mapping method (Nilsson et al. 2016) to rapidly assess the relationship between nationally relevant solutions and indicators, and global SDGs. We demonstrate the value of this framework for identifying ‘win–win’ sustainability solutions that can progress multiple indicators and SDGs at the same time, and identify solutions with trade-offs (i.e., solution–indicator interactions with negative causal relationships) (Allen et al. 2019; Griggs et al. 2017; Hopkins et al. 2021). Our approach also identifies solutions with impacts that lack consensus and gaps in indicators and SDGs where a few solutions are currently known or available for the local context, highlighting priorities for future research and investment. This study is a targeted contribution to the broader body of work required to enable the sustainable transformation of the Australian food and land system.

Methods

Study area: the Australian food and land system

Australia is a significant global food and fibre producer, particularly for key commodities, such as beef, sheep and wool, dairy, wheat, wine, and cotton (DFAT 2020) (Fig. 1). The economic and social importance of agriculture is juxtaposed by its significant negative impacts on the environment, most notably on biodiversity, water availability and quality, and GHG emissions (Turner et al. 2018). Australian agriculture is export-oriented, with 72% of the total annual value of agricultural production going to exports. This accounted for 12% of goods and services exports and 1.9% of Australia’s GDP in 2021. Agriculture currently accounts for 55% of Australian land use (excluding timber production), 74% of extracted water (ABARES 2022; ABS 2020-21), and 17.5% of GHG emissions (DCCEEW 2023).

Fig. 1
figure 1

Agricultural land use map for Australia, displaying areas under extensive grazing, intensive grazing, cropping and horticulture as well as urban areas, with State and territory administrative boundaries overlayed. The bar graphs display the number of individuals involved in this project within each stakeholder groups for each major city co-production workshop. Land-use data for Australia are taken from the Australian Land Use Map 2010. Stakeholder data were collected during workshops in Phase 1 of this study

Full size image

There are several sustainability narratives promoted for the future of Australian agriculture, which reflect elements of the global food system transformation (Béné et al. 2019a; Grundy et al. 2016). These range from encouraging technologies and farming practices that can reduce resource use, GHG emissions and depletion of soils (Turner et al. 2016), stricter conservation and carbon sequestration priorities and funded initiatives (Bryan et al. 2014, 2016b), and facilitating shifts towards sustainable diets and food waste reduction (Geyik et al. 2022; Springmann et al. 2018; Willett et al. 2019). At the same time, there are a number of potentially conflicting socio-economic priorities such as calls for continued growth in the value and volume of exports (NFF 2019), concerns around food system resilience and nutrition security due to the rising costs of fresh produce (Ridoutt et al. 2017), and concerns around farmer welfare and vitality in regional areas (NFF 2019).

Theory of change

Several narratives and pathways for a sustainable Australian food and land system have been described (Béné et al. 2019a; Bryan et al. 2016a; CSIRO 2023; Gao and Bryan 2017; Grundy et al. 2016; NFF 2019). Successful transition requires buy-in from stakeholders across a system, achievable through a co-production methodology. We adopt the framing of Scoones et al. (2020) to conceptualise the system change required, shifting from its current unsustainable state to the desired state using the SDGs as the framework for achieving the sustainable transition of the Australian food and land system. Our study makes a targeted contribution to this broader theory of change by improving our locally specific knowledge of solutions and sustainability indicators for the transition of the Australian food and land system. Likewise, this approach builds our understanding of the solution–indicator relationships to guide decision-making. We propose that the expected value of information (EVOI) (Dakins 1999) of solution–indicator relationships derived from co-production processes, where the expected increase in the value of information is likely associated with quality (Tàbara et al. 2017) and type of information (Davis et al. 2019) balanced with the quantity of information relevant to the decision process (Dakins 1999), can support a well-informed procedure to identify sustainability transition pathways for Australia.

A knowledge co-production framework for identifying indicators and prioritising solutions

The knowledge co-production framework for identifying indicators and prioritising solutions applied in this study (Fig. 2) is characterised by three distinct phases: Phase 1 development of an indicators and solutions database with stakeholders; Phase 2 refinement of the solutions and indicators database; and Phase 3 mapping of solutions–indicator interactions using an adapted interaction mapping method (Nilsson et al. 2016) to determine win–win solutions, trade-offs, and gaps in current availability of solutions to support the achievement of different indicators and SDGs.

Fig. 2
figure 2

Visual representation of the process used to develop the knowledge co-production framework for identifying indicators and prioritising solutions. Coloured squares represent stakeholder engagement. Along the pathway, light blue represents processes, deep blue represents inputs, and purple represents outputs and goals. Icons are designed by Freepik from Flaticon. The number of participants engaged during national workshops (Phase 1) and expert analysis (Phase 3) are indicated (n =)

Full size image

Phase 1: Developing a preliminary list of indicators and solutions

We undertook a review of academic and grey literature to identify an extensive preliminary list of supply-side (i.e., practice change/technological) and demand-side (i.e., behaviour change) solutions, focusing on major reports by the Food and Land Use Coalition (FOLU) (2019), Project Drawdown (2019), World Resources Institute (Searchinger et al. 2018); Beyond Zero Emissions (Longmire et al. 2014); and the EAT-Lancet Commission (Willett et al. 2019). Candidate supply-side solutions included land-use and management practices spanning sustainable intensification, agroecological, and conservation and circular economy paradigms, breakthrough technologies, alternative proteins, nature-based solutions, energy decarbonisation, and carbon sequestration. Candidate demand-side solutions included reducing food and fibre waste and loss and dietary shifts. Supply-side solutions included spanned the agricultural production stage as well as key upstream industries that supply goods and services to agricultural producers, such as water, fertilisers, pesticides, animal feeds, and energy (electricity and fuel) (Gao and Bryan 2017). Supply chain solutions, although critically important to the sustainable transition of the food and land system (Ahumada and Villalobos 2009; FAO 2018; FOLU 2021; Poore and Nemecek 2018; Steiner et al. 2020), were outside the scope of this study.

A preliminary list of indicators was developed in parallel following the same approach to capture the diversity of economic, socio-cultural, and environmental criteria relevant to the Australian food and land system. These indicators drew on well-established international frameworks such as the Economics of Ecosystems and Biodiversity (TEEB) agri-food (2018), the System of Environmental Economic Accounting (SEEA) (United Nations et al. 2014), and FOLU (2019).

In 2019, we conducted a series of stakeholder engagement workshops convening 164 stakeholders from 100 + organisations over seven workshops in capital cities across Australia. Workshop participants included a mix of agricultural industry representatives, federal and state government, finance and investment, Aboriginal and/or Torres Strait Islander peoples, landowners, natural resource managers, research/advisory and development organisations, and sustainable agriculture consultants (Fig. 1). We aimed for an even gender representation but did not request participants to self-identify gender during the participation process. Across all workshops, policy makers from state and federal government agencies made up 26.2% of participants, agricultural representatives and landholder 23.8%, environmental organisations and natural resource managers 21.9%, researchers 14.6%, finance and business 11.6%, and Aboriginal and/or Torres Strait Islander peoples 1.8% (Fig. 1).

Workshop participants were presented with the preliminary solutions in the first session, and the preliminary indicators in the following session. In both sessions, they were invited to review the lists independently and provide feedback using sticky notes against a feedback matrix capturing what they liked, wanted to add, wanted to remove, or had questions or comments about. Sticky notes were analysed using Nvivo13 (2020, R1) (Lumivero 2020) to produce word clouds highlighting priority indicators and solutions as identified by stakeholders (Figure S1 Supplementary Information). We synthesised these workshop outputs to form a preliminary database of co-produced solutions and indicators (see Supplementary Information for further details on the co-production workshop process).

Phase 2: Refinement of indicators and solutions

We refined the solutions and indicators database using ad-hoc expert consultation and further literature reviews. The indicators were mapped to relevant SDGs and described (Table 1). Indicators were arranged under the SDGs Wedding Cake framework (Sukhdev and Rockström 2016) as a way of conceptualising how healthy and sustainable food directly or indirectly connects all SDGs by encompassing the social, economic, and ecological aspects of the SDGs (Table 1). Drawing on established frameworks such as FOLU (2019) and outputs from the workshops, 19 high-level solution categories were established to enable grouping of like solutions for the Australian context (Table 2).

Table 1 Co-produced solutions database
Full size table
Table 2 Co-produced indicators database*
Full size table

Each solution was mapped to the appropriate category from the FOLU ‘10 critical transitions’ framework (FOLU 2019). Adopting the FOLU framework cuts through ideological or paradigmatic approaches to land management and land use and enables solutions from across multiple paradigms to be included simultaneously in the database. For each solution, we finalised the framing in terms of actual land-use or practice change that can be modelled; for example, virtual fencing is not identified as a standalone solution—rather as an enabling technology for managing grazing pressure (a practice defined with a specific bundle of assumptions). Each solution was then allocated to a high-level solution category with similar solutions, for example protect and restore nature (Table 2).

Phase 3: Mapping solution–indicator interactions and SDGs

To ensure alignment with global goals, the final list of selected indicators was mapped to 9 relevant SDGs: Zero hunger (SDG2); Good health and wellbeing (SDG3); Clean water and sanitation (SDG 6); Affordable and clean energy (SDG7); Decent work and economic growth (SDG8); Responsible consumption and production (SDG12) Climate change (SDG13); Life under water (SDG14); and Life on land (SDG15).

We then applied a team coding approach of an adapted interaction mapping method using a 7-point scale (Nilsson et al. 2016) to encode the strength and direction of each solution–indicator interaction (Fig. 3). We defined a synergy as a positive causal relationship between a solution and indicator, with three levels of interaction, and a trade-off as a negative causal relationship between a solution and an indicator, with three levels of interactions (Allen et al. 2019; Hopkins et al. 2021; Nilsson et al. 2018). Levels of synergy and trade-off are defined in Fig. 3 (Nilsson et al. 2016 , 2018). To simplify the mapping process and to ensure consistency, coders were asked to focus on the direct co-benefits/trade-offs rather than the secondary or indirect co-benefits/trade-offs of each solution–indicator interaction. It was outside the scope of this research to consider the multiple non-linear, irreversible, and cumulative processes that may exist between solution–indicator relationships.

Fig. 3
figure 3

Rapid interaction mapping coding classes: adapted 7-point scale (Nilsson et al. 2016)

Full size image

The relationship between a solution and each indicator is inherently complex. For example, the impact of solutions can vary by context (e.g., spatial location, scale and temporal dimensions), by the specific way in which the solution is designed and implemented, and by the people or organisations adopting these solutions. To manage for this complexity, three teams of expert coders (herein referred to as coders), and a subsection of stakeholders and researchers from Phase 1, coded every interaction between the 78 solutions and 18 sustainability indicators. The first team was made up of a food system researcher and a conservation scientist, the second team was made up of two practitioners with expertise in climate and food systems, and the third team was a single practitioner with expertise in climate and food systems. Coding teams may have their own biases due to their knowledge base and expertise. To overcome these potential sources of uncertainty, all coding teams initially reviewed the same sub-sample of solution–indicator interactions and intercoder reliability was assessed using the Fleiss’ Kappa statistic (Fleiss 1971) through the raters package in R (Quatto and Ripamonti 2022; Team 2020). Disagreements were discussed to reach an agreement across all sub-sample solution–indicator interactions prior to progressing to screening all interactions. We aimed for agreement as the intent of the study was to enable the use of the rapid coding method to prioritise solutions based on a consensus relationship between solutions and indicators.

For each solution–indicator interaction, an average rating was determined. Results were interrogated by individual solution, solution category, indicator, and by SDG to determine the likely performance of a solution in achieving a sustainable transition of the Australian food and land system that aligns with global goals (Supplementary Information 3). To represent empirical uncertainty at a solution–indicator level, for each solution–indicator interaction, we calculated the difference between the maximum and minimum level assignment on the 7-point scale (Fig. 3) across the three coders. The average level of (dis)agreement between coders is expressed as the difference in levels (e.g., 0, 1, 2, and 3 levels) between coders on the strength and direction of the solution–indicator interaction. Strong consensus was defined as between 0 and 1 levels of (dis)agreement between coders, and weak consensus was defined as between 3 and 4 levels of (dis)agreement between coders. For relationships where coding teams greatly disagreed (by a level of 3 or 4, for example neutral vs. major synergy), expert input was gained to reach a consensus on the solution–indicator interactions to add an additional layer of rigour to the interaction codes (Hill et al. 1997, 2005). We sought expert input through a multidisciplinary team of 13 scientists with expertise across environmental science, climate, food and agricultural systems, ecosystem and forest science, public health and nutrition, systems modelling, energy, and sustainability who provided additional ratings for solution–indicator relationships with weak consensus.

Results

Co-production and mapping of the solutions and indicator database

Phase 1 (database development) workshop series resulted in: a total of 496 comments on solutions and 478 comments on indicators (add, modify, remove, like); 86 substantive comments on wording of or gaps in preliminary solutions were considered; 375 comments and questions on indicators; 24 wording changes to existing solutions suggested; and 7 new solutions. Phase 2, refinement of the framework, resulted in 18 co-produced sustainability indicators that mapped to 9 SDG domains (Table 1), and 78 co-produced sustainability solutions categorised into 19 broader solution categories and mapped to the FOLU ‘critical transition’ categories (FOLU 2019) to support the sustainable transition of the Australian food and land system (Table 2). Supplementary Information 1 provides detailed descriptions, examples, and references for each solution.

In Phase 3 (mapping solution–indicator interactions), a total of 1440 solution–indicator interaction pairs were mapped. Across all solutions–indicators assessed a Fleiss’ Kappa value of 0.6 indicated ‘moderate’ inter-rater agreement. This level of agreement was deemed acceptable due to the complexity of the mapping process (Cohen 1960; Fleiss and Cohen 1973; McHugh 2012). There were nine solution–indicator interactions where strong disagreement occurred between coding teams (difference of 4 levels) and 57 solution–indicator interactions where disagreement occurred (difference of 3 levels). Detailed results for the Kappa analysis are located in Supplementary Information 2.

Here, we focus on results from Phase 3, summarised by solution category to explore locally relevant solutions to meet (or hinder) the achievement of sustainability indicators nationally, and more broadly contribute to global SDGs (Fig. 4). We highlight solution categories with strong consensus between coders [between 0 and 1 levels of (dis)agreement] and those that show most promise for achieving multiple indicators. Key trade-offs and gaps that require consideration to meet sustainability goals will be highlighted. Figure 4 provides a summary of the spread of synergies and trade-offs for solutions across indicators, mapped to SDGs. Of the total solutions identified, 39.7% were found to have major synergies with the achievement of SDGs, and 16.7% to have associated trade-offs. See Supplementary Information 2: Extended results for detailed results for all synergies and trade-offs.

Fig. 4
figure 4

Interaction matrix for 78 solutions (left-side column, individual solutions; right-side column solution categories) and 18 individual indicators mapped to 9 SDGs and the 3 SDG wedding cake categories: society, economy, and biosphere (top row). Each solution–indicator interaction is assessed using the adapted 7-point scale (Nilsson et al. 2016) (bottom row) by the degree to which each solution–indicator interaction achieves each indicator/SDG (rows) and is likely to affect the achievement of other SDGs (columns). The colours represent the 7-point scale (bottom row), from major trade-off (darkest orange) to neutral (white) to major synergy (darkest blue); i.e., the darkest row/column intersections are those with the strongest influence (either positive or negative) for achieving an indicator/SDG (column label). Solution–indicator interactions that were assessed as having three levels of uncertainty are highlighted in a hatched pattern

Full size image

Priority ‘win–win’ solutions

The solution categories ‘Protecting and restoring nature’, ‘Circular economy and energy decarbonisation’, and ‘Increased crop productivity’ displayed synergies across diverse sustainability indicators. These solution categories cumulatively represented 59.6% of major synergies and 30.2% of moderate synergies, across 10 and 15 indicators, respectively. In total, ‘Increased crop productivity’ had the greatest number of synergies (minor, moderate, and major) across indicators, representing 10.9% of total coded synergies. ‘Increased crop productivity’ also had the highest number of minor synergies (11.4%) between 8 solutions and 16 indicators mapping to 7 SDGs; however, coders had strong disagreement over these interactions. This solution category was also found to create the conditions to deliver the highest number of socio-economic and health co-benefits (Table 3).

Table 3 Tabular summary of solutions–indicator interactions for each solution category
Full size table

The solution category ‘Protecting and restoring nature’ had the highest number of major synergies (34.6%, Table 3) towards achieving 7 sustainability indicators mapping to SDG6, SDG12, SDG14, and SDG15, with strong consensus between coders. Within this solution category, SDG15 and SDG6 had the highest number of individual solutions coded as major synergies, with 50.0% and 27.8% of solutions, respectively. These solutions included conservation and restoration activities such as expanding protected areas and improving their management and connectivity, minimising runoff, fire risk management, wetland conservation and the rehabilitation of floodplains, waterways, and riparian areas. Interestingly, no major synergies and only 18.8% of total synergies for this solution category were linked to the achievement of SDG13.

The solution category ‘Circular economy and energy decarbonisation’ had the second highest number of major synergies (17.3%, Table 3) towards achieving 5 indicators mapping to SDG7, SDG12, and SDG13, with strong consensus between coders. These solutions were on-farm practice changes that would see shifts in fertiliser and feedstock requirements and energy production and use. A higher number of moderate and minor synergies were coded for this solution category: 28.9% of total synergies were moderate with over half of these moderate synergies aiding the achievement of SDG12; and 57.9% of total synergies were minor synergies creating the conditions for achieving all 9 SDGs. Proportionally, more minor synergies creating the conditions for the achievement of SDG8 and SDG12 were identified, suggesting economic and resource use efficiency co-benefits associated with this solution category. Only 1 solution–indicator interaction (on-farm energy efficiency) for this solution category was found to contribute towards the achievement of human and ecosystem health.

Key trade-offs

There were 13 solutions in total that worked in both synergy and trade-off with various sustainability indicators. Only 1.9% of all solution–indicator interactions were identified as trade-offs limiting options on another indictor, with 76.9% of total trade-offs limiting the achievement of SDG15, SDG6, and SDG8 (Table 4). Within these SDGs, indicators ‘Biodiversity’, ‘Soil’, ‘Water sustainability’, ‘Water efficiency’, and ‘Regional development’ cumulatively accounted for 69.2% of total trade-offs. Only 1 moderate trade-off was identified (Table 3): ‘Shifting to monogastric production’ clashed with the achievement of the ‘Animal welfare’ indicator; however, this solution was also coded with major synergies for achieving emission reductions and moderate synergies with the efficient use of natural resources and improving productivity.

Table 4 Tabular summary of solution–indicator interactions across SDGs
Full size table

Despite major synergies for reducing carbon emissions and improving soil health, the solution category ‘Carbon sequestration’ had the highest number (44.0% of total, Table 3) of minor trade-offs. Solutions for bioenergy feedstock production, carbon plantations, and carbon capture and storage were perceived to limit the achievement of human and ecosystem health and sustainability indicators. The solution ‘Bioenergy carbon capture and storage (BECCS)’ exhibited the greatest number of minor trade-offs (24.0%) limiting the achievement of human and ecosystem health and sustainability indicators. No trade-offs were coded for SDG6 and SDG7 (Table 4), and no major trade-offs were identified among solution–indicator interactions (impossible to achieve other indicators) (Table 3).

Identifying indicators and SDGs with limited solutions with major synergies

The indicator ‘Renewable energy’ was coded with < 1% of total synergies and < 1% of synergies across all levels of interaction. Very few major synergies were coded for achieving indicators ‘Carbon sequestration’, ‘Animal welfare’, ‘Water efficiency’, and ‘Soil health’. No solutions were coded with major synergies for the achievement of SDG2, SDG3, and SDG8. Table 4 provides a summary of the spread of synergies and trade-offs across SDGs to demonstrate coverage and gaps. These gaps and under-representation in delivering indicators and SDGs need careful consideration—they may be a product of methodological limitations or may highlight key challenges within the system or opportunities for innovation.

Discussion

We have developed a framework that draws on a diverse group of stakeholders and transdisciplinary experts to identify the strength and directionality of relationships across an extensive suite of solutions and locally relevant national-scale indicators. Through this co-production process ‘Protecting and restoring nature’, ‘Circular economy and energy decarbonisation’ and ‘Increase crop productivity’ emerged as priority ‘win–win’ solution categories with the highest potential for capturing synergies and avoiding trade-offs. The solution category ‘Carbon sequestration’ emerged with the highest number of trade-offs for the achievement of human and ecosystem health and sustainability indicators and gaps were identified for achieving ‘Renewable energy’, ‘Carbon sequestration’, ‘Animal welfare’, ‘Water efficiency’, and ‘Soil health’ sustainability indicators. These findings and their likely implications are discussed below.

Solutions and indicators for food and land system sustainability

To maximise synergies across multiple indicators, we found that solution categories ‘Protecting and restoring nature’, ‘Circular economy and energy decarbonisation’, and ‘Increase crop productivity’ hold the greatest number of individual solutions (with strong consensus) inextricably linked to meeting the greatest number of indicators at the national scale, and for achieving multiple SDGs. We considered these as ‘win–win’ solutions with very few co-occurring minor trade-offs and suggest prioritising these solution types for inclusion in future modelling efforts will support the identification of robust pathways towards a sustainable Australian food and land-use system. We suggest that the greatest opportunity for innovation lies in the gap identified in the current set of solutions for achieving SDG7 at a national level. Identifying and/or developing new solutions to meet this gap could be considered a priority for future research to support Australia in achieving these SDGs.

Solution–indicator interactions are complex (Nilsson et al. 2016, 2018; Bandari et al. 2022; Griggs et al. 2017; Grundy et al. 2016; Pradhan et al. 2017; van Soest et al. 2019), and thus, we suggest that solutions or solution categories with conflict between major synergies and trade-offs such as ‘Carbon sequestration’, ‘Shifting towards healthy and sustainable diets’, ‘Novel sources of protein’, and ‘Livestock productivity’ are also important to feature in future work as they provide the greatest insights into the key sustainability challenges (Hebinck et al. 2021; Zurek et al. 2021). Quantifying the impacts of priority solutions where conflict between synergies and trade-offs occur will provide critical insights into the magnitude of effect across various SDG domains. Quantifying and modelling these impacts will enable us to explore challenging questions such as ‘do the carbon sequestration benefits of a solution outweigh the biodiversity impacts’, or ‘how comfortable are we (as a society) to increase livestock productivity with certain solutions that compromise on animal welfare’. The achievement of so-called ‘win–win’ solutions will be enabled or accompanied by difficult societal choices or trade-offs (necessary burden shifting), and this must be clearly communicated (Béné et al. 2019a).

Harnessing synergies and overcoming trade-offs and gaps

Identifying priority solutions is not as simple as identifying major synergies. Our results draw attention to the paradigmatic dichotomy of producing less or producing better (Gerber et al. 2013; Steinfeld and Gerber 2010), and the importance of looking beyond the scope of a single indicator to evaluate a solution. Our results suggest that despite the broad range of important co-benefits to people and ecosystems derived from ‘Protecting and restoring nature’ solutions (Keith et al. 2021; Miralles-Wilhelm 2021; Seddon et al. 2020), these solutions were not viewed by experts as major contributors at scale to the achievement of climate change mitigation for the Australian food and land system, and should not be viewed as a substitute for the rapid decarbonisation of the entire economy (Seddon et al. 2021). As such, priority solutions must also be contextualised by the sustainability goals co-produced by stakeholders (Bandari et al. 2022; Moallemi et al. 2021; Szetey et al. 2021) and informed by the intended scale of application (Gao and Bryan 2017; Nilsson et al. 2016).

Triggering and accelerating change across the food and land system requires identifying points in the system where targeted solutions can enable positive feedback loops and activate positive tipping points (Fesenfeld et al. 2022; FOLU 2021; Pereira et al. 2020). FOLU (2021) proposed a framework for identifying early signs of positive tipping points and suggest solutions and sequencing that hold the greatest potential for triggering these positive feedback loops. Like FOLU (2021), our results indicate that the solution category ‘Protecting and restoring nature’ holds many solutions that may underpin the sustainability transitions required. We suggest that our rapid approach to identifying ‘win–win’ solutions (and trade-offs) supports early identification of solutions that may trigger such positive tipping points. Sukhdev and Rockström (2016) demonstrated the hierarchy of SDGs for the food system, arguing that solutions that support conserving the biosphere (SDGs 6, 13, 14, and 15) underpin the success of achieving all other SDGs. As such, prioritising solution categories ‘Conserving land’ (8 solutions) and ‘Conserving oceans’ (4 solutions), solutions with multiple synergistic solution–indicator relationships and no identified trade-offs across biosphere SDGs (Fig. 4), may support the sustainability transitions required for Australia.

Each solution is subject to complicated power relationships, temporal dimensions, multiple cascading effects, and/or feedbacks and requires different options for institutional and societal innovation (Fesenfeld et al. 2022; FOLU 2019, 2021; Howard 2021; Pereira et al. 2020; Steiner et al. 2020). FOLU (2019) identified key actor groups and their roles in the transition of the food and land system: government, business, farmers, investors, financial institutions, participants in multilateral processes and multi-stakeholder partnerships, and civil society. There are underlying actions required by these specific actors to enable transitions across different solution domains with many complex interacting relationships. For example, solutions in the ‘critical transition’ domain ‘protect and restore nature’ require government to establish and enforce policy, regulation, and incentive schemes, while business and suppliers must establish transparent supply chains to enable these solutions to be realised successfully. Whereas solutions in the ‘critical transition’ domain ‘scale productive and regenerative agriculture’ require government and business to establish and scale payments for ecosystem services and improve training and access to technologies. Likewise, business and investors should invest in sustainable supply chains and deploy innovation financing (FOLU 2019, 2021). While land managers and farmers are responsible for the implementation of many of the solutions identified in this study, the pace and scale of uptake and implementation will be largely determined by national and local policy, regulation, financing, and consumer demand. The role of each actor group in enabling these key transitions should be the focus of future research.

Highly optimistic global pathways often entail several assumptions (e.g., BECCS, afforestation) that may be at odds with the local sustainability context (Stoy et al. 2018). Without careful consideration and prioritisation of research and actions to attend to these trade-offs, or without systematic oversight across the complexity of solution–indicator interactions, we run the risk of encountering unintended or unanticipated consequences of implementing solutions (Zurek et al. 2021) to achieve myopic or singular indicators. Our analysis identified key trade-offs for solution category ‘Carbon sequestration’ across multiple indicators and SDGs, indicating that solutions identified within these categories may have several risks and/or limitations, and require further exploration and deliberation with key stakeholders across sectors before they are considered for modelling and implementation. The solution ‘BECCS’, a Carbon Dioxide Removal (CDR) technology, had the highest number of trade-offs, spread across several indicators. Studies have quantified these trade-offs demonstrating that although BECCS provides the opportunity for ambitious levels of carbon sequestration there are risks to ecosystem services, threats to biodiversity, and social and economic implications of displaced food production (e.g. Withey et al. 2019; Cobo et al. 2022; Stoy et al. 2018) which are highly context and site specific (Donnison et al. 2020). Compared to other studies (De Neve and Sachs 2020; Ioannou et al. 2023; Mainali et al. 2018; Nilsson et al. 2016; Pradhan et al. 2017; Stoy et al. 2018), a few trade-offs were identified. However, a majority of these studies focused on synergies and trade-offs between SDG pairs rather than solutions to achieve them. Exploring this space for a suite of land management and land-use practice solutions against SDGs revealed fewer trade-offs for the Australian context; however, due to the qualitative nature of this study, the magnitude of these trade-offs remains to be determined which will play an integral role in decision-making and policy agenda setting.

The greatest challenges and opportunities for the Australian context lie in the gap identified for achieving the ‘Renewable energy’ indicator. Many on-farm renewable energy opportunities exist; however, policy and funding schemes are required to support and enable innovation in this space (Chel and Kaushik 2011), likewise for the transition to renewable energy across the food and land sector for Australia (NFF 2019). Limited infrastructure in rural areas is a barrier to the adoption of renewable energy and electrification solutions (e.g., Karakaya and Sriwannawit 2015). This finding may also reflect the exclusion of transportation and refrigeration of goods/products (i.e., supply chain solutions from farm to fork) from the scope of this study, where further opportunities for decarbonisation exist through the adoption of renewable energy solutions (AEC 2022). It has been estimated that the food system is responsible for ~ 18 Gt CO2e/yr, amounting to 34% of global anthropogenic GHG emissions. On average, 71% (55–77%) of emissions in the food system come from agricultural production, suggesting that significant efforts are required in this domain (Crippa et al. 2021). However, solutions to reduce supply chain inefficiencies also offer great opportunity for decarbonisation and should be considered in future studies as they can make an important contribution to delivering system transition (Poore and Nemecek 2018; Steiner et al. 2020). Very few solutions were identified that are inextricably linked to the achievement of SDG2, SDG3, and SDG8. Achieving these SDGs with the current set of identified solutions for the Australian food and land system occurs primarily as a co-benefit. This is likely the product of solutions focusing on land use and practice change at the national scale, presenting an opportunity for innovation to include solutions that directly improve human health and wellbeing and regional livelihoods.

The intrinsic value of knowledge co-production for the food and land system

The complexity and uncertainty of system transition can be further intensified by adopting disciplinary silo approaches for designing pathways based on formalised methodologies that are less sensitive to cultural values, human preferences, and social complexities. Designing pathways for the food and land system requires transdisciplinary approaches that bridge scientific findings with stakeholder knowledge of local contexts and enable knowledge co-production. Given the large complexity and non-linearity in the dynamics of the food and land system and interaction with other systems, it is not possible to forecast the entire suite of potentially transformative sustainability solutions required to achieve the SDGs (Tàbara et al. 2018). The co-production frameworks such as the one presented in this study serve as a foundational tool for embedding knowledge from stakeholders from across a system in establishing indicators to measure sustainability and uncovering solutions to achieve them (Moallemi et al. 2021). Additionally, it addresses the inadequacies of the knowledge-deficit model by focusing on the value of co-produced knowledge as a preliminary step towards enabling a transformation oriented, locally specific knowledge production process (Tàbara et al. 2017).

Sustainability goals that are prioritised in a co-production process draw on the plurality of different visions for the future, local specificalities, and cultural narratives that various stakeholders present (Chabay et al. 2021; Jasanoff and Kim 2015; Szetey et al. 2021). Likewise, a set of solutions for achieving goals needs to be co-developed. Stakeholders will more likely support planned solutions if there is a perceived link to their social identities (Chabay et al. 2021) and they reflect local knowledge (Manzo and Perkins 2006). Constructive dialogues with diverse stakeholder groups about solutions can help in just, equitable and publicly supported implementations that are widely supported (Chabay et al. 2021). Stakeholders can understand the surrounding cultural and political context and define what ‘critical’ solutions will be convincing and actionable on the ground.

Our framework provides a structured, systematic, and meaningful approach identifying locally relevant indicators and solutions to enable locally specific and system appropriate modelling that explores national scale contributions to global SDGs. To this end, this framework could be used by a broad range of stakeholders seeking to establish the greatest opportunities for climate change mitigation and simultaneously meeting environmental and socio-economic goals, and identifying the key trade-offs and gaps that must be navigated. We suggest that this framework is a useful, transdisciplinary tool that can be successfully applied at the national and sub-national level for identifying and prioritising key solutions to achieve locally relevant and contextualised solutions for sustainable transitions of the food and land-use system.

Limitations and future research

We identify three key limitations. The first is methodological, where the co-production (Phase 1) of the solutions and indicators database are subject to participant bias. Despite best efforts for representation across sectors and management of power dynamics between different actor types (Bandari et al. 2022; Moallemi et al. 2021; Szetey et al. 2021), we cannot be certain that an even representation of stakeholders and opinions was achieved during Phase 1. Barriers and challenges to knowledge inclusion, exchange, and transmission may have created some biases in our results (Schiller-Merkens and Machin 2023). During Phase 3, the coding of synergies and trade-offs was undertaken by a small subset of stakeholders and researchers, as such it is limited in its representation of the diverse perspectives and knowledge of the large number of participants and stakeholder types involved in the co-production process in Phase 1. Likewise, bias and varying degrees of confidence in the encoding of synergies and trade-offs (Bandari et al. 2022), lack of research or content knowledge for some solution–indicator areas, and varied hypotheses between coders in terms of their assumed strengths in solution–indicator interactions or indirect impacts (Phase 3) also had an impact on the results. At the individual solution level, some solutions were grouped together to reduce the number of coded relationships. For example, the solution ‘Regenerative agriculture’ was used in this study to describe multiple practices that underpin a regenerative or agroecological approach to land management. This reductionist method results in the nuances of various land management practices that can be applied within the regenerative approach being overlooked, and the co-benefits potentially being under- or over-estimated by coders.

The second limitation is the absence of quantified impacts associated with each solution–indicator interaction. This limitation constrains the use of the framework for prioritisation of solutions rather than implementation, as the magnitude of impact (both positive and negative) and feasibility of implementing solutions or achieving against an indicator remain unknown. In contrast, the absence of these details ensures a rapid and low-cost approach for prioritising solutions for future modelling exercises. Quantitative modelling of impacts (or technical potential) often entails a significant level of abstraction and is usually only possible for a subset of the total number of solutions considered in qualitative narratives. It is important for the qualitative mapping process to encompass all solutions. Although it is likely (but not certain) that experts would have implicitly incorporated feasibility considerations during the coding process, we suggest that additional screening is required based on an appropriate feasibility framework (e.g., Brutschin et al. 2021; Nielsen et al. 2020) to account for the technological, economic, behavioural, cultural, and social feasibility to be modelled for the Australian context and to determine the availability of quality data. Despite these limitations, we have developed a framework that allows for the successful integration of stakeholder co-produced and transdisciplinary expert knowledge to identify the strength and directionality of relationships across an extensive suite of solutions and indicators which would otherwise be unachievable through quantitative modelling.

The third is the absence of any consideration of temporal dimensions (and feedbacks) associated with each solution. We did not explicitly consider the multiple non-linear, irreversible, and cumulative processes which could be used to support the identification of conditions for positive tipping points (Fazey et al. 2018; Fesenfeld et al. 2022; FOLU 2021; Pahl-Wostl 2009; Tàbara et al. 2018). It could be suggested that ‘win–win’ solutions with no identified trade-offs may in fact enable the conditions for positive tipping points to occur and should be considered for further research in exploratory modelling exercises. One such example is the solution ‘Regenerative Agriculture’, a bundle of solutions that exist within an alternative paradigm of socio-ecological dynamics, which has been recognised by some for their transformative potential for food production and ecosystem repair (Gordon et al. 2023; Massy 2013).

Future work could include the translation of the co-produced indicators into specific measurable targets that would enable monitoring progress towards the co-produced indicators of a sustainable food and land system for Australia. The next phase of research should bring together the same diversity of stakeholders to undertake a participatory visioning and backcasting process, underpinned by the indicators and solutions database, to explore possible pathways for achieving a single desired future for the Australian food and land system and articulating the steps required to realise the desired future (Bibri and Krogstie 2019; Ebolor 2023; Kanter et al. 2016; Quist and Vergragt 2006). This should include examining the key drivers and mechanisms for accelerating system-level change, the risks, uncertainties, and alternative pathways to achieving these transitions (Fazey et al. 2018), and illuminate opportunities to harness positive feedbacks and balance negative feedbacks (Fesenfeld et al. 2022). The insights gained during this process could be analysed using quantitative modelling tools underpinned by the research presented here to support stakeholder discussions and provide tools for decision and policy makers. This would successfully situate the findings of this targeted research in a broader co-production process that mobilises local knowledge from multi-level interactions to inform system change (Pahl-Wostl 2009).

Policy insights

The transition to a sustainable Australian food and land system requires appropriate governance, policy and market mechanisms and sufficient investment and funding to ensure the pace and scale of change required to meet SDGs within the 2030 timeframe. Rapid change can only occur if we recognise the interlinkages across different systems. Each system is inherently complex and interrelated with much political debate surrounding the need and approach. As such, any system change must be underpinned by early and systematic interventions that promote synergies while minimising trade-offs and spill overs across different economic sectors and systems (Moallemi et al. 2020; Soergel et al. 2021).

While this work is intended to prioritise solutions for data assembly and model development, qualitative insights can be gained by policy and decision-makers. For example, policies that incentivise sequestration must account for the potential trade-offs with food and water security, energy, soil health, biodiversity, and socio-economic impacts to communities (Climate Change Authority 2023; Stoy et al. 2018). Likewise, our framework could support the rapid identification of potential positive tipping points for sustainable transitions of food and land systems (Fesenfeld et al. 2022; FOLU 2021; Steiner et al. 2020), and also provide the foundation for targeted research contextualised by local knowledge and for advancing policy based on best practice science.

Conclusion

Co-production frameworks such as the one presented in this paper can be a first but important step towards engaging stakeholders in thinking and planning for sustainable food and land systems. Co-produced knowledge is also a critical early step in ensuring that pathways for transition are positioned for successful adoption among stakeholders. This study has provided a systematic overview of the likely synergies and trade-offs across a carefully selected suite of sustainability solutions for achieving national-scale sustainability indicators specific to the Australian food and land system that can scale to meet SDGs. It has advanced our understanding of likely ‘win–win’ solutions, identifying ‘Protecting and restoring nature’, ‘Circular economy and energy decarbonisation’, and ‘Increase crop productivity’ as priority solution categories for capturing synergies and minimising trade-offs. Likewise, it has highlighted some key gaps and trade-offs that exist in meeting sustainability indicators for the Australian food and land system. The solution category ‘Carbon sequestration’ emerged with the highest number of minor trade-offs with individual sustainability indictors ‘Water efficiency’, ‘Water sustainability’, ‘Biodiversity’, and ‘Soil health’, and a key gap was identified in solutions available for achieving the indicator ‘Renewable energy’. Our findings can directly inform data assembly and quantitative modelling for sustainable food and land systems for the Australian context and facilitate future stakeholder dialogues by transparently reporting on key trade-offs, gaps, and 'win–win’ solutions.