Meeting food-security targets within environmental limits requires integrating genetic advances with data-driven design and adaptive management of crop populations as engineered systems. Herein, we propose a framework for AI-designed crop populations, in which AI-enabled approaches coordinate above- and belowground architecture with management to improve productivity, resource-use efficiency, and climate resilience. AI can act as an architect by coupling phenomics with process-based crop models to optimize multiobjective population designs and identify locally tailored configurations. It can also act as a regulator by integrating sensing, model-based prediction, and environmental feedback to guide in-season adaptation under climate variability. By linking trait innovation with population-level interactions and adaptive regulation, this framework offers a potentially transferable route for shifting the yield-efficiency frontier for sustainable intensification.

Decades of cultivation and the often exclusive use of mineral fertilisers as a substitute for organic inputs have reduced the soil organic carbon (SOC) content of agricultural soils, meaning they now represent a potential sink for carbon sequestration to mitigate climate change and improve soil function. As well as being a legacy of management, SOC will also be dependent on local scale climate, topography, and soil properties; accounting for this local context is important when benchmarking fields and quantifying the potential for additional carbon sequestration. We developed a landscape-scale methodology, using a handheld infrared device, for baselining SOC stocks in the top 30 cm across a 45,000 ha farm cluster in the UK. The cluster is exploring opportunities for landscape-scale environmental improvement with a focus on natural flood protection and water pollution reduction through conversion of arable land to permanent grassland. We used the baseline data to estimate additional benefits of arable reversion for soil carbon sequestration. Because all the farms in the cluster share the same pedoclimatic conditions, variance in SOC at the field scale could be confidently attributed to differences in soil type and land use. Average SOC stocks in arable and permanent pasture fields were 103.9 and 140.3 Mg C ha−1, respectively. Variance in %SOC was modelled using soil series, sample depth, land use, and clay content, and fields were benchmarked based on deviation from the expected value. The fields with the largest SOC stocks were identified and used as references to predict future potential sequestration. The conversion of arable land to permanent pasture resulted in a predicted average uplift in SOC of 55.0 Mg C ha−1. Our landscape-scale methodology provides robust evidence on current and future carbon stocks for public subsidy schemes and natural capital markets that account for local constraints and opportunities.

Background Grassland reseeding typically requires intensive tillage. This disrupts soil nutrient dynamics, especially under varying drainage conditions.Methods This study evaluated the combined effects of tillage legacy, drainage and soil depth on key soil properties within a long-term grassland experiment. Treatments compared two no-tillage durations (5 years of no tillage, 5YNT; 30 years of no tillage, 30YNT), two drainage systems (drained and undrained) and two depths (0-10 and 10-30 cm) following reseeding. Total carbon (TC), total nitrogen (TN), total phosphorus (TP) and pH were measured.Results Significant three-way interactions were detected for all variables. At 0-10 cm, undrained 30YNT plots showed 69% greater TC and 66.9% higher TP than the lowest values recorded in 5YNT x drained combinations at 10-30 cm. TN followed a similar pattern, with substantial enrichment under long-term no-tillage. Reseeding reduced TC, TN and TP within 5 years, particularly in surface soils, with the largest proportional losses in undrained plots where nutrient concentrations had been high prior to tillage.Conclusions Findings highlight the importance of conserving long-term no-tillage systems to enhance nutrient retention and promote sustainable grassland productivity. Conventional tillage of nutrient-rich pastures should be avoided to prevent major nutrient losses.

Achieving stable and sustainable soybean production under increasing climate variability and soil degradation remains a global challenge. Straw mulching is promoted to increase soybean seed yield in arid and semi-arid agricultural systems, but its long-term impacts on soil fertility and yield stability remain poorly quantified. We conducted a long-term field experiment, involving three treatments: straw removing (SR), straw mulching (SM), and straw crushing (SC). SM increased soil enzymes activity and improved topsoil nutrients. Among the three treatments, the SM exhibited the highest mean weight diameter (2.12), while the lowest soil solid phase proportion (49.09%). SM resulted in the longest chlorophyll retention duration in soybean leaves (135.55 d), followed by the SC (120.81 d) and SR (95.25 d). Furthermore, at the R1 stage, the SM exhibited the highest leaf area index (LAI) and biomass, both of which showed a significant positive correlation with seed yield. Compared with SR and SC, SM increased seed yield, yield stability, and yield sustainability by an average of 17.76, 73.64, and 15.42%, respectively. Long-term retention of crop residues represents a scalable, low-input strategy to rebuild soil fertility, buffer climatic stress, and secure yield stability – contributing to global goals for sustainable agriculture and food security.

The land-sharing versus land-sparing debate represents a critical juncture in agricultural policy development. However, applying either of these approaches uniformly at a national scale has been challenged suggesting that more effective strategies may require a context-dependent mix of methods. This study evaluates plausible strategies of land sparing and land sharing at regional scale in Great Britain using the Long-Term Large Scale integrated modelling framework. We consider these strategies in various combinations to get national scale outcomes for nutrient losses to freshwater and agricultural productivity. By simulating various land-use configurations across 11 International Territorial Level regions, we generated over 1.79 trillion scenarios with differing regional distributions of arable and semi-natural land. We used multiple objective optimization to find an optimal solution set. Our analysis identified 24,412 Pareto-optimal solutions that also improved on business-as-usual. The Pareto-optimal solutions all favoured combining land-sparing and land-sharing approaches. These optimized scenarios achieved increases of up to 9.7 % in livestock calories and 5.2 % in crop calories, while reducing phosphorus losses by 6.9 % and nitrate losses by 11.9 % in comparison to a business-as-usual scenario. Our findings demonstrate that spatially differentiated land-use strategies tailored to regional characteristics outperform uniform national sharing or sparing approaches. However, these modest improvements suggest that transformative change will require complementary innovations beyond land allocation strategies alone. This approach advances landscape planning from binary sharing-sparing debates towards a multidimensional optimization of food production and environmental quality that acknowledges the inherent complexity of dynamic landscapes while supporting evidence-based agricultural policy development.

Context: Rothamsted Research's Park Grass Experiment, established in 1856, is the longest-running grassland study globally. Naturally regenerating grassland swards are grown in plots with varying applications of fertiliser including ammonium sulphate and sodium nitrate (at varying application rates), organic fertiliser, minerals (K, Mg, Na, P), and lime, which is mown twice a year. As the world's most widely produced crop, grass is predominantly used to feed ruminants, however, the nutritional properties and carrying capacities of these plots have not previously been quantified. Objective: The objective of this study was to characterise the nutritional profile of forage gathered from the Park Grass plots from 1860 to 2020 and the ruminant carrying capacity that the plots would support. The study further aimed to explore the trade-offs between productivity, forage nutritional quality, and biodiversity. Method: Dried PGE herbage samples were taken from the Rothamsted sample archive at decade intervals from 1860 to 2020, representing a range of plot treatments. Proximate analysis and XRF elemental analysis were performed, and the data was used to estimate ruminant carrying capacity of plots based on metabolisable energy and crude protein requirements for production. Results: Fertiliser applications increased carrying capacity due to yield improvements but reduced crude protein while increasing cellulose and hemicellulose. Increased growth appeared to have a dilution effect on some essential minerals, particularly Ca, Mg, Mn, and P. Sodium nitrate produced higher carrying capacities per unit of nitrogen compared to ammonium sulphate or organic manure. Conclusions: The findings highlight trade-offs in improved grasslands between forage quality, quantity, biodiversity, and management inputs. Results show that fertiliser applications enhance carrying capacity by increasing forage yield but potentially at the cost of reduced nutritional quality and species diversity. This study also provides the first comprehensive nutritional analysis of the Park Grass plots, revealing how historical fertiliser treatments influenced forage quality and ruminant carrying capacity over 160 years. Significance: Studying the trade-offs and gradients within grassland systems is essential for understanding the balance between productivity and biodiversity. This study also contributes to the rich dataset available on the Park Grass Experiment, providing future opportunities and insight, whilst also highlighting the importance of long-term experimental studies in the agricultural and environmental sciences

Agricultural organic waste can enhance aggregate organic carbon stability, which is crucial for soil carbon sequestration in croplands. However, it is unclear how aggregate organic carbon stability changes with different nature-based nutrient management practices, especially with partial organic substitution. This study aimed to elucidate how different organic wastes (chicken manure, biochar, straw, and carbon-based materials from kitchen waste) influence aggregate organic carbon stability, including aggregate stability, the content of physically protected organic carbon, and the decomposability of aggregate carbon. The improvement of aggregate organic carbon stability was trialed in a 4-year field experiment with equivalent nitrogen and organic carbon input under nature-based nutrient management. The results showed that all nature-based nutrient management practices improved aggregate organic carbon stability compared to no nutrient addition. Biochar application dramatically improved aggregate organic carbon stability by 5.8-11.4 % in aggregate stability, 83.9-152.4 % in aggregate organic carbon, and 36.6-75.0 % in aggregate recalcitrant carbon content. By comparison, straw returning showed the lowest improvement in aggregate organic carbon stability, owing to substantial increases of microbial respiration and enzyme activities involved in carbon degradation. Organic carbon merely increased by 32.3 %, 33.6 %, and 29.5 % in large macroaggregates, small macroaggregates, and microaggregates, respectively. This study dissected the different efficiencies of nature-based nutrient management in improving aggregate organic carbon stability in vegetable fields. The findings highlight that appropriate nature-based nutrient management with organic waste could better implement the carbon neutrality in agroecosystems from the perspective of aggregate organic carbon stability.

Intensive arable agriculture uses agrochemicals to replace ecosystem services (e.g. pest control and soil health) while simultaneously degrading others (e.g. pollination). Agroecological farming aims to reduce this reliance on agrochemicals. Whether these practices maintain yields at a scale relevant to farm business viability is unclear. In a 4‐year replicated study across 17 English farms we assessed the ability of farmer co‐designed agroecological systems to support regulating services, beneficial invertebrates, crop yield and profitability. We test three management systems: (1) ‘business‐as‐usual (BAU)’ control; (2) ‘enhancing‐ES’ supporting beneficial invertebrates with wildflower field margins and protecting soils with cover crops; (3) ‘maximising ES’ with the further addition of soil organic matter and in‐field strips to bring beneficial invertebrates into the crop. Soil carbon stocks were highest in the maximising‐ES system. Predation and pollination ecosystem services were higher in the enhancing‐ES and maximising‐ES systems, as were earthworms and other populations of beneficial predatory and pollinating invertebrates. Pest snail biomass was also lowest in the enhancing‐ES and maximising‐ES systems, although aphid numbers were higher. The enhancing‐ES and maximising‐ES systems increase yields of cereals and oilseed rape. However, the loss of productive agricultural land and establishment costs exceeded the value of increased yields. Only enhancing‐ES breaks even with agri‐environmental subsidies. Synthesis and applications . These results highlight that while evidence for the role of ecosystem services in supporting crop yield can be found, overcoming economic constraints within conventional farming systems is likely to be a key barrier to widespread uptake. Agri‐environmental subsidy payments can offset these costs, but only for moderate interventions. Transition to more sustainable farming systems needs to overcome these economic constraints with new policy interventions.

This is a critical moment for land use policy globally, with many countries (e.g. the UK and the European Union) currently undertaking significant green reforms of their agricultural policies. Despite their importance for maintaining agricultural outputs and plant diversity, the effects of artificial soil enrichment on pollinators remain poorly understood. Our two-year study at the world’s longest-running ecological experiment, Park Grass, Rothamsted, examines the relationship between soil fertilisation, grassland yield and biodiversity. Our data show a large and significant negative effect of the major plant nutrients (NPK) on the abundance, species richness and functional diversity of both pollinators and flowering plants. The results also indicate a large and significant trade-off between productivity and biodiversity. Our findings are a salutary reminder of the challenge in reconciling conflicting aims in farmland management and strongly suggest that financial incentives are necessary to offset yield reductions to improve biodiversity outcomes in agricultural grasslands.

Faced with the biodiversity extinction crisis and climate change, alternative approaches to food production are urgently needed. Decades of chemical-based weed control have resulted in a dramatic decline in weed diversity, with negative repercussions for agroecosystem biodiversity. The simplification of cropping systems and the evolution of herbicide resistance have led to the dominance of a small number of competitive weed species, calling for a more sustainable approach that considers not only weed abundance but also community diversity and composition. Agroecological weed management involves harnessing ecological processes to minimize the negative impacts of weeds on productivity and maximize biodiversity. However, the current research effort on agroecological weed management is largely rooted in agronomy and field-scale farming practices. In contrast, the contributions of landscape-scale interventions on agroecological weed management are largely unexplored (e.g., interventions to promote pollinators and natural enemies or carbon sequestration). Here, we review current knowledge of landscape effects on weed community properties (abundance, diversity, and composition) and seed predation (a key factor in agroecological weed management). Furthermore, we discuss the ecological processes underlying landscape effects, their interaction with in-field approaches, and the implications of landscape-scale change for agroecological weed management. Notably, we found that (1) landscape context rarely affects total weed abundance; (2) configurational more than compositional heterogeneity of landscapes is associated with higher alpha, beta, and gamma weed diversity; (3) evidence for landscape effects on weed seed predation is currently limited; and (4) plant spillover from neighboring habitats is the most common interpretation of landscape effects on weed community properties, whereas many other ecological processes are overlooked. Strikingly, the drivers of weed community properties and biological regulation at the landscape scale remain poorly understood. We recommend addressing these issues to better integrate agroecological weed management into landscape-scale management, which could inform the movement towards managing farms at wider spatiotemporal scales than single fields in a single season.

Glyphosate, the most widely used herbicide, is linked with environmental harm and there is a drive to replace it in agricultural systems. We model the impacts of discontinuing glyphosate use and replacing it with cultural control methods. We simulate winter wheat arable systems reliant on glyphosate and typical in northwest Europe. Removing glyphosate was projected to increase weed abundance, herbicide risk to the environment, and arable plant diversity and decrease food production. Weed communities with evolved resistance to non-glyphosate herbicides were not projected to be disproportionately affected by removing glyphosate, despite the lack of alternative herbicidal control options. Crop rotations with more spring cereals or grass leys for weed control increased arable plant diversity. Stale seedbed techniques such as delayed drilling and choosing ploughing instead of minimum tillage had varying effects on weed abundance, food production, and profitability. Ploughing was the most effective alternative to glyphosate for long-term weed control while maintaining production and profit. Our findings emphasize the need for careful consideration of trade-offs arising in scenarios where glyphosate is removed. Integrated Weed Management (IWM) with more use of cultural control methods offers the potential to reduce chemical use but is sensitive to seasonal variability and can incur negative environmental and economic impacts.

The achievements of the Green Revolution in meeting the nutritional needs of a growing global population have been won at the expense of unintended consequences for the environment. Some of these negative impacts are now threatening the sustainability of food production through the loss of pollinators and natural enemies of crop pests, the evolution of pesticide resistance, declining soil health and vulnerability to climate change. In the search for farming systems that are sustainable both agronomically and environmentally, alternative approaches have been proposed variously called 'agroecological', 'conservation agriculture', 'regenerative' and 'sustainable intensification'. While the widespread recognition of the need for more sustainable farming is to be welcomed, this has created etymological confusion that has the potential to become a barrier to transformation. There is a need, therefore, for objective criteria to evaluate alternative farming systems and to quantify farm sustainability against multiple outcomes. To help meet this challenge, we reviewed the ecological theories that explain variance in regulating and supporting ecosystem services delivered by biological communities in farmland to identify guiding principles for management change. For each theory, we identified associated system metrics that could be used as proxies for agroecosystem function. We identified five principles derived from ecological theory: (i) provide key habitats for ecosystem service providers; (ii) increase crop and non-crop habitat diversity; (iii) increase edge density: (iv) increase nutrient-use efficiency; and (v) avoid extremes of disturbance. By making published knowledge the foundation of the choice of associated metrics, our aim was to establish a broad consensus for their use in sustainability assessment frameworks. Further analysis of their association with farm-scale data on biological communities and/or ecosystem service delivery would provide additional validation for their selection and support for the underpinning theories.

With a growing body of research associating livestock agriculture with faster global warming, higher health costs and greater land requirements, a drastic shift towards plant-based diets is often suggested as an effective all-round solution. Implicitly, this argument is predicated on the assumption that the reallocation of resources currently assigned to animal production systems will automatically result in the efficient cultivation of human-edible crops without negative environmental, health or socioeconomic consequences. In reality, however, the validity of this assumption warrants careful examination, as a farm’s capability to adopt a new agricultural system is multifaceted and context-specific. Through a transdisciplinary review of literature, here we discuss examples of unintended consequences that could arise from the conversion of grasslands into arable production, including potentially adverse impacts on yield stability, biodiversity, soil fertility and beyond. We contend that few of these issues are being methodically considered as part of the current food security debate and call for a closer examination of supply-side constraints.

Unprecedented climate and land use changes are having major impacts on water-based ecosystem services (ES). It is crucial, therefore, to get an in-depth understanding of current levels of such ES provision, and how they could be impacted by changed environments or management. However, applying existing models for water-related ES pose substantial challenges, which include the need for in-depth specialist hydrological knowledge, the requirement for numerous datasets and parameters that may not be consistently available, and high computational costs. Additionally, there is often a mismatch between the resolution of the model output and parcel-level land management, at which ES information is often most valuable for supporting decision-making. Here we detail a rapid method for assessing water retention potential by estimating an area's 'sponginess' - its capacity to absorb and retain precipitation. Our approach builds upon a topography-adjusted Curve Number methodology, a widely recognised and straightforward tool for estimating water run-off. We calculated run-off across 1 km grid cells (representing field- to farm-scale land parcels) in Great Britain based on the land's sponginess under storm conditions. The primary objective was to investigate the applicability of the method in estimating point-values - i.e., the independent contribution of each grid square - within the context of known limitations. The results enable the identification of areas with higher potential for the ES of water retention. Our model output illustrates the spectrum of sponginess across Great Britain, ranging from less than 30% of precipitation retained in city regions to as high as 99% in some rural, agricultural areas. Importantly, we demonstrate that the model is easy to run, can be used with freely-available data, and produces outputs compatible with grid-based models for other ES. Overall, the model provides an accessible approach to estimating the ES of water retention to researchers worldwide, even in data-scarce areas. ### Competing Interest Statement The authors have declared no competing interest.

Archetypes of land- and socio-ecological systems, generated using unsupervised classification methods, enable the assimilation of complex environmental and socio-economic information. Such simplification has considerable potential to feed into decision support systems for sustainability planning. But, the usefulness of archetypes depends on how well they relate to sustainability criteria, such as ecosystem service (ES) delivery, that are external to the input datasets employed for archetype generation. Sensitivities in such post-hoc association analyses, and the associated utility of the archetype framework in a decision support context, remain unexplored. Here we emulated post-hoc association analysis procedures using simulated socio-ecological datasets and ES response variables. Our simulations revealed a substantial influence on analysis performance from (1) the number of variables used as inputs in archetype generation, (2) the correlation structure of input datasets, (3) the type and distribution of input variables, and (4) the functional form (linear or non-linear) characterising the relationship between ES variables and their predictors. We observed near-identical performance when archetypes were generated using K-means clustering and Self-Organising Maps (SOMs) - two commonly used archetype classification methods. Further, better archetype classifier performance did not guarantee better discrimination of ES value distributions between archetypes. Our results suggest that designing a framework to generate archetypes for sustainability planning, and the selected methodological choices therein, should place greater emphasis on what the archetypes will be used for in downstream analyses, and not focus solely on archetype classifier performance. This would better ensure the identification of archetypes adaptable to a diverse array of sustainability indicators and sufficiently robust for monitoring decision outcomes over time. ### Competing Interest Statement The authors have declared no competing interest.

The benchmarking of farm environmental sustainability and the monitoring of progress towards more sustainable farming systems is made difficult by the need to aggregate multiple indicators at the relevant spatial scales. We present a novel framework for identifying alternative pathways to improve environmental sustainability in farming systems that addresses this challenge by analysing the co-variance of indicators within a landscape context. A set of sustainability indicators was analysed within the framework of a published set of Farm Management Archetypes (FMAs) that maps the distribution of farming systems in England based on combinations of environmental and management variables. The archetype approach acknowledges that sustainability indicators do not vary independently and that there are regional constraints to potential trajectories of change. Using Pareto Optimisation, we identified optimal combinations of sustainability indicators (“Pareto nodes”) for each FMA independently, and across all FMAs. The relative sustainability of the archetypes with respect to one another was compared based on the proportion of Pareto nodes in each FMA. Potential for improvement in sustainability was derived from distances to the nearest Pareto node (either within or across FMAs), incorporating the cost of transitioning to another archetype based on the similarity of its environmental variables. The indicators with the greatest potential to improve sustainability within archetypes (and, therefore, should have a greater emphasis in guiding management decisions) varied between FMAs. Relatively unsustainable FMAs were identified that also had limited potential to increase within archetype sustainability, indicating regions where more fundamental system changes may be required. The FMA representing the most intensive system of arable production, although relatively unsustainable when compared to all other archetypes, had the greatest internal potential for improvement without transitioning to a different farming system. In contrast, the intensive horticulture FMA had limited internal potential to improve sustainability. The FMAs with the greatest potential for system change as a viable pathway to improved sustainability were dairy, beef and sheep, and rough grazing, moving towards more mixed systems incorporating arable. Geographically, these transitions were concentrated in the west of England, introducing diversity into otherwise homogenous landscapes. Our method allows for an assessment of the potential to improve sustainability across spatial scales, is flexible relative to the choice of sustainability indicators, and—being data-driven—avoids the subjectivity of indicator weightings. The results allow decision makers to explore the opportunity space for beneficial change in a target landscape based on the indicators with most potential to improve sustainability.

Carabid beetles are major predators in agro-ecosystems. The composition of their communities within crop environments governs the pest control services they provide. Field margins and landscape features are known to affect carabid community composition, yet evidence is currently lacking that can be used to support land management decisions targeted at optimising predation services at the farm scale. We used experimental margins across a farm site to test carabid communities in crop areas, margins, and adjacent habitats sampled in the summer. We used novel subterranean trapping with standard pitfall trapping, to distinguish above ground and below ground activity of adults and larvae in different farm habitats. Crop type was the major influence on carabid communities in crop areas. This was followed by landscape influences in terms of adjacent habitat and boundary features, and whilst significant, margin type explained relatively little variance in summer carabid communities in-field. Trap type revealed differential activity by species. Responses to crop type, landscape factors, and margin type also varied by species. Overall, abundances were less in association with margins than control of no margin. Particularly, abundances were lower in the spillover zone adjacent to grass margins, and in the wildflower margins themselves. Carabid larvae showed notably higher abundances in association with an absence of field margins. Measures to boost key carabid species in crop areas should be considered at a farm scale, taking into account potential barrier effects, and potential buffer effects.

Rumex obtusifolius (broad-leaved dock) is a problematic weed that reduces yield and nutritional value of forage in grasslands of temperate regions worldwide. We conducted an on-farm study to identify management practices and environmental factors that influence the risk of the occurrence of R. obtusifolius in high densities in permanent, productive grasslands used for forage production. Following a common protocol, a paired case-control design was implemented in Switzerland (CH), Slovenia (SI), and United Kingdom (UK) to compare parcels with high densities of R. obtusifolius (cases, >= 1 plant m(-2)) with nearby parcels free of or with very low densities of the species (controls, <= 4 plants 100 m(-2)). A total of 40, 20, and 18 pairs were recorded in CH, SI, and UK respectively. Parameters measured included data about management practices and history, vegetation cover and composition, and soil nutrients and texture. Across countries, increased vegetation cover reduced the relative risk of R. obtusifolius occurrence. By contrast, increased soil phosphorus and potassium and high soil bulk density raised the relative risk. These effects were consistent across countries, as no interactions between country and any of the factors were observed. The two indicator species for case parcels, Plantago major and Poa annua, were typical species of disturbed areas and fertile soils, while indicators for control parcels were characteristic of grasslands under medium to high management intensity (e.g., Festuca rubra, Cynosorus cristatus, Anthoxantum odoratum). We conclude that the risk for grassland infestation with R. obtusifolius can be significantly affected by management practices. Prevention measures should target phosphorus and potassium fertilisation to the forage plants' requirements, minimise soil compaction, and maintain dense swards.