Agricultural research increasingly relies on data-driven approaches for crop yield prediction that complement more established crop growth models, including machine learning techniques. However, these approaches rely on large training datasets. Here, we present the Crop Yields, Climate, Soils, and Satellites (CYCleSS) dataset, a large-scale crop yield dataset derived from precision yield data for 934 fields across England on which a variety of crops are grown. In addition, the data also contains satellite-derived remote sensing data, weather data, and data on soil type, all aligned at a grid resolution of 10 km. Weather data is available at a daily temporal resolution, satellite data at 5-day resolution, while crop yield data is available at yearly resolution. This effort has been made possible through careful anonymisation of the yield data while preserving the alignment with remote sensing, weather, and soil data. This data will be useful both to train machine learning models of yield prediction as well as to parameterize mechanistic crop growth models. Furthermore, the anonymisation procedure itself will be of interest to the research community, as it represents a solution to a common problem on the interface of agricultural research and farming practice.

Plant diseases occurring across wild and crop plants present modelling and management challenges. Wild plant and crop pathosystems differ in ecological structure, evolutionary dynamics and responsiveness to human intervention. At the interface, pathogens may spill over, spill back, persist or evolve, shaped by host diversity, dispersal processes and landscape connectivity. The potential importance of factors including pathogen dispersal, host life history and spatial configuration are examined through a qualitative comparison of case studies: Puccinia graminis, Phakopsora pachyrhizi, Xylella fastidiosa, Pyricularia oryzae Triticum lineage and Austropuccinia psidii. These examples illustrate how wild hosts may function as reservoirs, recombination partners or spillover targets, and how their role influences management efficacy and evolutionary risk. We explore the consequences of this wild-crop interface through two central questions: (i) how should plant diseases involving wild and cultivated pathosystems be managed, and (ii) what proportion of management effort should be allocated to each system? We show the principles underpinning answers to these questions via a conceptual framework based on a generic compartmental model incorporating asymmetric transmission and system-specific interventions, thereby accounting for key aspects of pathogen spread within and between wild host and crop populations. Finally, we identify critical data needs and modelling directions to better inform disease management on the wild plant-crop interface and argue for a more integrative approach bridging ecological and anthropogenic drivers of epidemics. This article is part of the theme issue 'Wild plant pathosystems'.

Societal Impact Statement The UK Plant Health Risk Register (PHRR) has so far identified 581 Plant Pests and Diseases (PPDs) that could invade the United Kingdom and affect 74 tree species. The combined effects of multiple invasions on trees are little understood and seldom considered. We estimate future invasion rates and tree losses from the PHRR risk scores using historical data. We project potentially severe losses to ecologically and economically important species like oak, apple, poplar and pine due to combined impacts of multiple PPDs. Our analysis provides a framework to better understand and address these pressing biosecurity challenges. Summary Plant pests and diseases (PPDs) pose a serious threat to trees and forests globally. The UK Plant Health Risk Register (PHRR) provides semiquantitative estimates of invasion probability and impact on host plants for PPDs thought to pose a risk to the United Kingdom to help prioritize biosecurity activities. The PHRR currently contains 636 PPDs potentially affecting 74 tree species found in the United Kingdom. We explore the implications of converting these ordinal likelihood and impact scores to quantitative estimates of invasion probabilities and losses to tree productivity, in terms of increased mortality or reduced growth, in the next 25 years. We use recent invasions to quantify likely losses, and generate plausible invasion probabilities from historical observations. Assuming that the expectation of loss in 25 years is the probability of invasion multiplied by the fractional loss, and that the impacts of multiple PPDs affecting the same host species are independent, we find that many tree species would suffer severe production declines under plausible invasion probability and impact estimates. Despite several important knowledge gaps, our analysis provides a framework for projecting how trees and forests might be impacted in future and helps to highlight the risk posed by PPDs to biodiversity and ecosystem services.

El Niño indirectly moderates rice yields in China through effects on migratory pests, highlighting the need to incorporate transboundary ecological processes into climate risk assessments and agricultural forecasting.

Climate change will alter the geographical locations most suited for crop production, but adaptation to these new conditions may be constrained by edaphic and socio-economic factors. Here we investigate climate change adaptation constraints in banana, a major export crop of Latin America and the Caribbean. We derived optimal climatic, edaphic and socio-economic conditions from the distribution of intensive banana production across Latin America and the Caribbean, identified using remote sensing imagery. We found that intensive banana production is constrained to low-lying, warm aseasonal regions with slightly acidic soils, but is less constrained by precipitation, as irrigation facilitates production in drier regions. Production is limited to areas close to shipping ports and with high human population density. Rising temperatures, coupled with requirements for labour and export infrastructure, will result in a 60% reduction in the area suitable for export banana production, along with yield declines in most current banana producing areas.

Southern and central China harbor some of the most diverse and ecologically significant forest ecosystems in China, making them priority regions for biodiversity conservation and afforestation planning. Within this broader context, the Qinling Mountains serve as a representative case study, as they are recognized as a global biodiversity hotspot and one of China’s most important temperate broad-leaved forest regions. Understanding the spatial distribution of dominant tree communities in this area is essential for biodiversity conservation and ecosystem management in the face of increasing anthropogenic pressures and climate change. This study aimed to (1) identify the potential distribution and richness patterns of dominant tree communities in the Qinling Mountains and adjacent regions, and (2) assess their vegetation associations and key environmental drivers to inform afforestation and conservation planning; and (3) Explore the potential species richness and endemicity to identify the most suitable areas for afforestation of target communities or group of the dominant tree communities. We applied ensemble stacked species distribution models (S-SDMs) using ecological niche modeling to predict the potential distribution of target dominant tree communities across China, capturing diverse ecological regions including the Qinling region. Key environmental variables included climatic, edaphic, and anthropogenic factors. Vegetation associations were explored using non-metric multidimensional scaling (NMDS) analysis based on environmental gradients and field data collected from the Qinling Mountains and surrounding areas. The S-SDMs showed high predictive performance (AUC > 0.8, TSS > 0.5). Key predictors of species distribution included mean diurnal temperature range (Bio2), human influence, precipitation seasonality (Bio15), and organic soil carbon. Richness and endemicity hotspots were concentrated in southeast and south-central China, particularly in Hainan and Taiwan Islands. NMDS analysis revealed three distinct tree associations structured along environmental gradients, notably human disturbance, precipitation, and soil characteristics. Spatial prioritization highlighted provinces such as Henan, Anhui, Shandong, and Jiangsu as key areas for afforestation by dominant tree communities. The findings along with the developed interactive map offer valuable insights for landscape-scale biodiversity conservation, afforestation efforts, and ecosystem restoration, supporting the achievement of global Sustainable Development Goals (SDGs). Incorporating climate change modeling into future analyses will be critical to ensuring the long-term resilience and success of restoration initiatives.

The prevalence of crop insect pests, which damage crops and reduce their yield, is increasing globally owing to changes in climate and land use, posing a threat to food security. In this Review, we synthesize evidence on how tropical, temperate, migratory and soil crop pests respond to changes in climate, land use and agricultural practices. In general, crop pests are responding to warming with expanded geographic ranges, advanced phenological events and increased number of reproductive generations per year. Increased pest damage under warming is projected to exacerbate yield losses of 46%, 19% and 31% under 2 °C warming for wheat, rice and maize, respectively. Pests at mid–high latitudes respond more positively to warming than those in the tropics. Moderate drought can increase pest damage to crops owing to enhanced feeding on plants as a water source and decreased resilience of plants and natural enemies of pests. Increased precipitation reduces small pests through washing them away, but favours pests in general through buffering thermal-hydro stresses. Land use change, such as deforestation and conversion to cropland, enhances warming and reduces biodiversity, leading to enhanced crop damage. Agricultural intensification, particularly fertilization and irrigation, increases the quality and quantity of host plants and buffers pests from environmental extremes, favouring proliferation. Globalization of trade networks increases pest invasions, with associated damage exceeding US $423 billion in 2019. Future research should examine the mechanisms underlying changes in pest status and develop monitoring and prediction systems to inform management approaches. Crop yield losses to insect pests pose a risk to food security. This Review assesses global trends of crop pest prevalence associated with global environmental change, identifies the underlying ecological mechanisms and proposes strategies for effective, sustainable management of pests to support future food security.

Peatland restoration has been suggested as a key method for the UK to meet national, legally binding climate targets. This can involve blocking up drainage ditches or erosion features, as well as encouraging regeneration of peatland vegetation through Sphagnum reintroduction or removal of scrub or trees. It is unclear, however, how suitable future conditions will be for both peat accumulation and Sphagnum survival. We applied three bioclimatic envelope models for blanket bogs in Britain to assess how future climate is likely to deviate from current conditions, focussing on four national parks with significant peatland area (Dartmoor, the Flow Country, the Peak District and Snowdonia). We also assessed the likelihood of thresholds being passed at which irreversible desiccation of Sphagnum moss may occur. Our bioclimatic envelope models use updated climate projections (bias-corrected UKCP18 projections under Representative Concentration Pathways (RCP) 2.6, 4.5 and 8.5) that are more accurate in the upland regions in which blanket bogs can occur, and use thresholds of blanket bog occurrence which are tailored to Britain. This gives us higher confidence in the results as compared to previous models. Our results show substantial losses in areas suitable for peatland by 2061-2080 under all RCPs. Under RCP8.5 there is virtually no peatland within its current bioclimatic envelope in our case study areas and only limited areas in Snowdonia under RCP4.5, suggesting these regions will be outside the ideal conditions that lead to peat accumulation. Only western Scotland retains substantial areas suitable for peat. The frequency of Sphagnum desiccation events is projected to increase by between 44% and 82% which will likely result in decreased success of hummock forming species, particularly at easterly sites where rainfall is lower, though wetter microsites will likely allow more drought-tolerant species to persist. Policy implications. Action should be taken to raise water tables at degraded sites to limit the impact of future drought conditions. However, climatic conditions being outside the current bioclimatic envelope may make full restoration challenging. Sphagnum reintroduction programmes may have greater success utilising drought-tolerant species as hummock forming species are at greater risk of die off during desiccation events.

Crop rotation and intercropping with Allium spp. (e.g. garlic, onion) are recognized as biological controls of Fusarium wilt in several crops. However, the non-target effects of this strategy on the soil microbiome are largely unknown. We evaluated the effect of cultivating Musa basjoo, Allium tuberosum (Chinese leek), and their co-cultivation on the rhizosphere and bulk soil microbiome under glasshouse conditions. We do not report the impact of Chinese leek on Foc abundances, as this has been previously reported. The bacterial and fungal communities in the rhizosphere and bulk soil varied depending on the plant type. The microbiomes of allium and bananas were consistently dissimilar, and the co-cultivation treatment contained constituent taxa from both. However, the effect of allium on soil community composition was outweighed by that of bananas in the co-cultivation scenario. Discriminative taxa for allium included members of the genera Sphingomonas, Microbacterium, Flavihumibacter, Brevundimonas, Pseudolabrys, Ramlibacter, Trichoderma, Mortierella and Fusarium. In the case of bananas, predominant biomarkers encompassed members of Rhizobacter, Noviherbaspirillum, Pseudarthrobacter, Aquabacterium, Pseudomonas, Tausonia and Humicola Biomarkers and predictions of functional gene abundances suggest that shifts in the soil microbiome induced by allium are correlated with increases in microorganisms exhibiting potential disease suppression and antibiotic or antifungal traits, whereas those in bananas are associated with plant-growth-promoting microorganisms. The biocontrol efficacy of allium co-cultivation may there involve shifts in the soil microbiome as well as direct impacts of root exudate chemistry on Fusarium plant pathogens. ### Competing Interest Statement The authors have declared no competing interest.

Extreme weather events have severe impacts on food systems, especially for smallholders in global food value chains (GFVCs). There is an urgent need to understand (a) how climate shocks manifest in food systems, and (b) what strategies can enhance food system resilience. Integrating satellite, household and trade data, we investigate the cascading impacts after two consecutive hurricanes on smallholder banana farmers in Dominican Republic, and determinants of their recovery. We found that farmers experienced an ‘all-or-nothing’ pattern of damage, where 75% of flooded farmers lost > 90% of production. Recovery of regional production indicators took ca. 450 days. However, farm-level recovery times were highly variable, with both topographic and human capital factors determining recovery. Utilising this case study, we show that engaging in a GFVC impeded recovery via ‘double exposure’ of production loss and losing market access. Our results suggest that strategies to enhance resilience, with a particular focus on recovery, in GFVCs should promote trader loyalty, facilitate basin-scale collaboration and expand risk-targeted training.

Ten years ago, (black) stem rust - the most damaging of wheat (Triticum aestivum) rusts - re-emerged in western Europe. Disease incidences have since increased in scale and frequency. Here, we investigated the likely underlying causes and used those to propose urgently needed mitigating actions. We report that the first large-scale UK outbreak of the wheat stem rust fungus, Puccinia graminis f. sp. tritici (Pgt), in 2022 may have been caused by timely arrival of airborne urediniospores from southwest Europe. The drive towards later-maturing wheat varieties in the UK may be exacerbating Pgt incidences, which could have disastrous consequences. Indeed, infection assays showed that two UK Pgt isolates from 2022 could infect over 96% of current UK wheat varieties. We determined that the temperature response data in current disease risk simulation models are outdated. Analysis of germination rates for three current UK Pgt isolates showed substantial variation in temperature response functions, suggesting that the accuracy of disease risk simulations would be substantially enhanced by incorporating data from prevailing Pgt isolates. As Pgt incidences continue to accelerate in western Europe, we advocate for urgent action to curtail Pgt losses and help safeguard future wheat production across the region.

Societal Impact Statement Bananas are the world's most popular dessert fruit and a staple starch crop for millions in low‐ and middle‐income countries. The banana export trade that supplies North America, Europe, and other wealthy nations has a history fraught with exploitation and conflict. The price of cheap bananas has been environmental degradation, violence, and poverty. Only recently have efforts to address the power imbalances in this trade been made. Voluntary certification schemes aim to address multiple sustainability issues, while research into biological control, accelerated plant breeding, and efficient irrigation will help prepare the industry for emerging threats from pests, diseases, and climate change. Summary Bananas are the world's favorite dessert fruit, a staple starch crop for millions, and an important source of income for producers across the tropics and subtropics. Bananas evolved and diversified as giant perennial herbs of open habitats within the humid forests of Southeast Asia and West Oceania and were domesticated around 7000 years BP through a series of hybridization events. This review considers the journey from rainforest riversides to intensively managed monoculture plantations, focussing on the Cavendish banana that comprises nearly the entire global export trade. Climate change increasingly threatens economic sustainability in several major producer regions, requiring responses such as efficient irrigation systems to maintain productivity and water security. Pests and diseases are spreading globally and have severe direct impacts on production as well as indirect impacts via harm to ecological and human health caused by pesticides. New pest and disease management methods employing biological controls and enhancing soil health and new plant breeding techniques must be developed and implemented. The banana production and trade system has been characterized by power imbalances between international firms that own plantations and supply the market and the local agricultural workers who cultivate and harvest the fruit. Voluntary certification schemes have been developed to address the numerous environmental, social, and economic sustainability issues faced by the industry. There are indications, from research on biological disease control to new deals on wages and benefits for banana workers, that change is slowly coming to the global banana trade.

Climate change is increasing the risk of extreme weather events, potentially threatening crop yields and global food security. A key benefit of international free trade is risk sharing, because global aggregate production is more stable than that of individual countries which may be adversely affected by extremes. Here we test the hypothesis that diverse sourcing of crops from multiple trading partners reduces exposure to extreme weather, using a detailed trade matrix and range of extreme weather indices. We find that countries with high source diversity have moderate exposure, but that there is wide variation in the degree of exposure in countries reliant on domestic crop production. Global aggregate production and export volatility is stable or declining for most crops, suggesting that source diversification will increase resilience to both climatic and non-climatic supply shocks.

The global banana industry is threatened by one of the most devastating diseases: Fusarium wilt of banana. Fusarium wilt of banana is caused by the soilborne fungus Fusarium oxysporum f. sp. cubense (Foc), which almost annihilated the banana production in the late 1950s. A new strain of Foc, known as tropical race 4 (TR4), attacks a wide range of banana varieties, including Cavendish clones, which are the source of 99% of banana exports. In 2019, Foc TR4 was reported in Colombia, and more recently (2021) in Peru. In this study, we sequenced three fungal isolates identified as Foc TR4 from La Guajira (Colombia) and compared them against 19 whole-genome sequences of Foc TR4 publicly available, including four genome sequences recently released from Peru. To understand the genetic relatedness of the Colombian Foc TR4 isolates and those from Peru, we conducted a phylogenetic analysis based on a genome-wide set of single nucleotide polymorphisms (SNPs). Additionally, we compared the genomes of the 22 available Foc TR4 isolates, looking for the presence-absence of gene polymorphisms and genomic regions. Our results reveal that (i) the Colombian and Peruvian isolates are genetically distant, which could be better explained by independent incursions of the pathogen to the continent, and (ii) there is a high correspondence between the genetic relatedness and geographic origin of Foc TR4. The profile of present/absent genes and the distribution of missing genomic regions showed a high correspondence to the clades recovered in the phylogenetic analysis, supporting the results obtained by SNP-based phylogeny.

The global banana industry is threatened by one of the most devastating diseases: Fusarium wilt (FWB). FWB is caused by the soil-borne fungus Fusarium oxysporum f. sp. cubense ( Foc ), which almost annihilated the banana production in the late 1950s. A new strain of Foc , known as tropical race 4 (TR4), attacks a wide range of banana varieties including Cavendish clones which are the source of 99% of banana exports. In 2019, Foc TR4 was reported in Colombia, and more recently (2021) in Peru. In this study, we sequenced three fungal isolates identified as Foc TR4 from La Guajira (Colombia) and compared them against 19 whole-genome sequences of Foc TR4 publicly available, including four genome sequences recently released from Peru. To understand the genetic relatedness of the Colombian Foc TR4 isolates and those from Peru, we conducted a phylogenetic analysis based on a genome-wide set of single nucleotide polymorphisms (SNPs). Additionally, we compared the genomes of the 22 available Foc TR4 isolates looking for the presence-absence of gene polymorphisms and genomic regions. Our results reveal that (i) the Colombian and Peruvian isolates are genetically distant, which could be better explained by independent incursions of the pathogen to the continent, and (ii) there is a high correspondence between the genetic relatedness and geographic origin of Foc TR4. The profile of present/absent genes and the distribution of missing genomic regions showed a high correspondence to the clades recovered in the phylogenetic analysis, supporting the results obtained by SNP-based phylogeny.

Organic agriculture, employing manures or composts, has been proposed as a way of mitigating undesirable impacts of mineral fertilizer use. Of particular interest is the effect of fertilizer regime on soil microbes, which are key to nutrient cycling, plant health and soil structure. However, the effect of fertilizers on soil microbial diversity remains poorly understood. Since biological diversity is an important determinant of ecosystem function and a fundamental metric in community ecology, the effects of fertilizer regimes on soil microbial diversity are of theoretical and applied interest. Here, we conduct a meta-analysis of 37 studies reporting microbial diversity metrics in mineral fertilized (NPK), organically fertilized (ORG) and unfertilized control (CON) soils. Of these studies, 32 reported taxonomic diversity derived from sequencing, gradient gel electrophoresis, or RFLP. Functional diversity, derived from Biolog Ecoplate™ measures of carbon substrate metabolism, was reported in 8 studies, with 3 studies reporting both diversity metrics. Bacterial and archaeal diversity was reported in 28 taxonomic studies, and fungal diversity in 8 taxonomic studies. We found that functional diversity was 2.8 % greater in NPK compared with CON, 7.0 % greater in ORG vs CON, and 3.8 % greater in ORG vs NPK. Bacterial and archaeal taxonomic diversity was not significantly different between NPK and CON, but on average 2.9% greater in ORG vs CON, and 2.4 % greater in ORG vs. NPK. Fungal taxonomic diversity was not significantly different between any treatment pairs. There was very high residual heterogeneity in all meta-analyses of soil diversity, suggesting that a large amount of further research is required to fully understand the influence of fertilizer regimes on microbial diversity and ecosystem function.

We identify major knowledge gaps in the primary impacts of extreme weather and climate change across the UK’s food system, its functioning and their interactions to provide information to support adaptation and resilience planning. Future shocks and stresses due to changes in weather and climate extremes will have significant impacts on the UK food system. Key knowledge gaps remain in our understanding of their impacts on non-cereal crops, livestock and fisheries production, on the food chain beyond primary production, on the longer-term impacts, and in an integrated, full system view of impacts that accounts for cumulative impacts, interactions, feedbacks and the interplay between domestic and overseas elements of the UK food system. These knowledge gaps need to be urgently addressed to ensure future climate resilience of the UK food system. There are several areas where research could better support decision-making towards increased resilience to weather and climate shocks in both food policy and business sectors. We note the need for a step change in the collection, quality, synthesis and application of a broad range of weather and food chain data and information across time and space. There is a need to develop tools to support the inclusion of the “missing middle” of food chain and policy discussions that incorporate weather and climate impacts: processing/packaging, transport, storage, wholesale, retail and disposing/reusing. Greater integration of climate, biophysical, social, political and economic research is required to characterise geo-political influences on food system climate resilience. Further work is needed to assess adaptation actions needed in response, and their knock-on trade-offs and consequences across sectors, and their interactions. The challenges identified here suggest the need for challenge-led, connective, interdisciplinary approaches to future funding initiatives in support of achieving food system resilience to weather and climate shocks.

Plant pathology has developed a wide range of concepts and tools for improving plant disease management, including models for understanding and responding to new risks from climate change. Most of these tools can be improved using new advances in artificial intelligence (AI), such as machine learning to integrate massive data sets in predictive models. There is the potential to develop automated analyses of risk that alert decision-makers, from farm managers to national plant protection organizations, to the likely need for action and provide decision support for targeting responses. We review machine-learning applications in plant pathology and synthesize ideas for the next steps to make the most of these tools in digital agriculture. Global projects, such as the proposed global surveillance system for plant disease, will be strengthened by the integration of the wide range of new data, including data from tools like remote sensors, that are used to evaluate the risk ofplant disease. There is exciting potential for the use of AI to strengthen global capacity building as well, from image analysis for disease diagnostics and associated management recommendations on farmers' phones to future training methodologies for plant pathologists that are customized in real-time for management needs in response to the current risks. International cooperation in integrating data and models will help develop the most effective responses to new challenges from climate change.

Food system resilience has multiple dimensions. We draw on food system and resilience concepts and review resilience framings of different communities. We present four questions to frame food system resilience (Resilience of what? Resilience to what? Resilience from whose perspective? Resilience for how long?) and three approaches to enhancing resilience (robustness, recovery, and reorientation-the three "Rs"). We focus on enhancing resilience of food system outcomes and argue this will require food system actors adapting their activities, noting that activities do not change spontaneously but in response to a change in drivers: an opportunity or a threat. However, operationalizing resilience enhancement involves normative choices and will result in decisions having to be negotiated about trade-offs among food system outcomes for different stakeholders. New approaches to including different food system actors' perceptions and goals are needed to build food systems that are better positioned to address challenges of the future.

Global food security is threatened by climate change, both directly through responses of crop physiology and productivity, and indirectly through responses of plant-associated microbiota, including plant pathogens. While the interactions between host plants, pathogens and environmental drivers can be complex, recent research is beginning to indicate certain overall patterns in how plant diseases will affect crop production in future. Here, we review the results of three methodological approaches: large-scale observational studies, process-based disease models and experimental comparisons of pathosystems under current and future conditions. We find that observational studies have tended to identify rising temperatures as the primary driver of disease impact. Process -based models suggest that rising temperatures will lead to latitudinal shifts in disease pressure, but drying conditions could mitigate disease risk. Experimental studies suggest that rising atmospheric CO2 will exacerbate disease impacts. Plant diseases may therefore counteract any crop yield increases due to climate change.