Search

Abstract In the last two decades, an exponentially growing number of meta-analyses (MAs) synthesize thousands of peer-reviewed studies on the environmental impacts of farming practices (FPs). This paper describes the iMAP-FP evidence library, a comprehensive dataset on the effects of 34 categories of FPs (such as agronomic practices, cropping and livestock systems, land management options and mitigation techniques) on 34 impacts including climate mitigation, soil health, environmental pollution, water use, nutrients cycling, biodiversity, and agricultural productivity. Through systematic screening, 570 MAs published since 2000 were selected and categorized according to the type of FP. We assessed their impacts, the geographic regions covered, and their quality. We extracted 3,811 effects and their statistical significance associated with sustainable FPs (intervention) compared to a control (typically conventional agriculture) across 223 different intervention-control pairs. Our dataset is accompanied with an online free-access library, which includes a catalogue of synthetic reports summarizing the available evidence on each evaluated FP.

Abstract Many countries have implemented policies to reduce the use of chemical inputs in agriculture. However, these policies face many obstacles that limit their effectiveness. The purpose of this paper is to review the main challenges associated with reducing chemical inputs in agriculture and to propose potential solutions. Our analysis, based on a literature review linking agronomy and economics, shows that several agronomic options have proven effective in reducing chemical inputs or mitigating their negative impacts. We argue that the organization of the agri-food system itself is a major barrier to their implementation. Involving all stakeholders, from the chemical input industry to consumers, and designing appropriate policy frameworks are key to address this issue. We recommend combining different policy instruments, such as standards, taxes and subsidies, in a simplified and coherent way to increase effectiveness and ensure better coordination in the adoption of sustainable practices.

Abstract During extensive periods without rain, known as dry-downs, decreasing soil moisture (SM) induces plant water stress at the point when it limits evapotranspiration, defining a critical SM threshold (θcrit). Better quantification of θcrit is needed for improving future projections of climate and water resources, food production, and ecosystem vulnerability. Here, we combine systematic satellite observations of the diurnal amplitude of land surface temperature (dLST) and SM during dry-downs, corroborated by in-situ data from flux towers, to generate the observation-based global map of θcrit. We find an average global θcrit of 0.19 m3/m3, varying from 0.12 m3/m3 in arid ecosystems to 0.26 m3/m3 in humid ecosystems. θcrit simulated by Earth System Models is overestimated in dry areas and underestimated in wet areas. The global observed pattern of θcrit reflects plant adaptation to soil available water and atmospheric demand. Using explainable machine learning, we show that aridity index, leaf area and soil texture are the most influential drivers. Moreover, we show that the annual fraction of days with water stress, when SM stays below θcrit, has increased in the past four decades. Our results have important implications for understanding the inception of water stress in models and identifying SM tipping points.

Climate change has a significant impact on rice grain appearance quality; in particular, high temperatures during the grain filling period increase the rate of chalky immature grains, reducing the marketability of rice. Heat-tolerant cultivars have been bred and released to reduce the rate of chalky grain and improve rice quality under high temperatures, but the ability of these cultivars to actually reduce chalky grain content has never been demonstrated due to the lack of integrated datasets. Here, we present a dataset collected through a systematic literature search from publicly available data sources, for the quantitative analysis of the impact of meteorological factors on grain appearance quality of various rice cultivars with contrasted heat tolerance levels. The dataset contains 1302 field observations of chalky grain rates (%) - a critical trait affecting grain appearance sensitive to temperature shocks - for 48 cultivars covering five different heat-tolerant ranks (HTRs) collected at 44 sites across Japan. The dataset also includes the values of key meteorological variables during the grain filling period, such as the cumulative mean air temperature above the threshold temperature (TaHD), mean solar radiation, and mean relative humidity over 20 days after heading, obtained from a gridded daily meteorological dataset with a 1-km resolution developed by the National Agriculture and Food Research Organization. The dataset covers major commercial rice cultivars cultivated in Japan in different environmental conditions. It is a useful resource for analyzing the climate change impact on crop quality and assess the effectiveness of genetic improvements in heat tolerance. Its value has been illustrated in the research article entitled “Effectiveness of heat tolerance rice cultivars in preserving grain appearance quality under high temperatures - A meta-analysis”, where the dataset was used to develop a statistical model quantifying the effects of high temperature on grain quality as a function of cultivar heat tolerance.

Abstract Terrestrial vegetation emits vast amounts of monoterpenes into the atmosphere, influencing ecological interactions and atmospheric chemistry. Global emissions are simulated as a function of temperature with a fixed exponential relationship (β coefficient) across forest ecosystems and environmental conditions. We applied meta-analysis algorithms on 40 years of published monoterpene emission data and show that relationship between emissions and temperature is more sensitive and intricate than previously thought. Considering the entire dataset, a higher temperature sensitivity (β = 0.13 ± 0.01 °C−1) is derived but with a linear increase with the reported coefficients of determination (R2), indicating that co-occurring environmental factors modify the temperature sensitivity of the emissions that is primarily related to the specific plant functional type (PFT). Implementing a PFT-dependent β in a biogenic emission model, coupled with a chemistry – climate model, demonstrated that atmospheric processes are exceptionally dependent on monoterpene emissions which are subject to amplified variations under rising temperatures.

Abstract Following a request from the European Commission, the EFSA Panel on Plant Health performed a quantitative risk assessment for the EU of Retithrips syriacus (Mayet) (Thysanoptera: Thripidae), a polyphagous thrips, regarded as a tropical/subtropical pest occurring in several countries of Africa, South America, Asia and in the EU in Cyprus. The current risk assessment focused on potential pathways for entry, the climatic conditions allowing establishment, the expected spread capacity and the impact considering a time horizon of 10 years (2023–2032). The Panel identified the import of cut roses, persimmons, table grapes, as well as plants for planting of the genera Acalypha and Terminalia from third countries and those of Persea americana (avocado) from Israel as the most relevant entry pathways to consider. Over the next 10 years, an annual median estimate of 95 (90% Certainty Range, CR, ranging from 13 to 1832) potential R. syriacus founder populations per year are expected to successfully transfer to a suitable host in the EU NUTS2 regions where the climatic conditions are predicted as suitable for establishment; this value drops to a median of 4.6 founder populations per year (90% CR: 1 every 1.9 years – 85.6 per year) after considering the actual probability of establishment of a potential founder population. The estimated number of founder population per year is mostly driven by the import of cut roses and plants for planting. If such founder populations were to establish, R. syriacus is estimated to spread at a median rate of 0.05 km/year (90% CR 0.02–2.30 km/year) after a median lag phase of 1.1 years (90% CR 0.3–3.3 years). The overall impact on yield (expressed as % of the total agricultural production) directly attributable to R. syriacus when considering: (i) the main R. syriacus hosts in the EU, (ii) the areas of the EU where establishment is possible, (iii) the current agricultural practices and (iv) the evidence of impact from the countries where the pest is established for a long time, was estimated at 0.065% as the median value of the uncertainty distribution (90% CR 0.001%–0.571%). Options for risk reduction are discussed, but the effectiveness was not quantified.

In Europe, the heterogeneous features of crop systems with majority of small to medium sized agricultural holdings, and diversity of crop rotations, require high-resolution information to estimate cropland Net Ecosystem Exchange (NEE) and its two main components of Gross Ecosystem Exchange (GEE) and the Ecosystem Respiration (RECO). In this context, this paper presents an assimilation of high-resolution Sentinel-2 indices with eddy covariance measurements at selected European cropland flux sites in a new modified version of Vegetation Photosynthesis Respiration Model (VPRM). VRPM is a data-driven model simulating CO2 fluxes previously applied using satellite-derived vegetation indices from the Moderate Resolution Imaging Spectroradiometer (MODIS). This study proposes a modification of the VPRM by including an explicit soil moisture stress function to the GEE and changing the equation of RECO. It also compares the model results driven by S2 indices instead of MODIS. The parameters of the VPRM model are calibrated using eddy-covariance data. All possible parameters optimization scenarios include the use of the initial version vs. the proposed modified VPRM, S2, or MODIS vegetation indices, and finally the choice of calibrating a single set of parameters against observations from all crop types, a set of parameters per crop type, or one set of parameters per site. Then, we focus the analysis on the improvement of the model with distinct parameters for different crop types vs. parameters optimized without distinction of crop types. Our main findings are: (1) the superiority of S2 vegetation indices over MODIS for cropland CO2 fluxes simulations, leading to a root mean squared error (RMSE) for NEE of less than 3.5 μmolm-2s-1 with S2 compared to 5 μmolm-2s-1 with MODIS (2) better performances of the modified VPRM version leading to a significant improvement of RECO, and (3) better performances when the parameters are optimized per crop-type instead of for all crop types lumped together, with lower RMSE and Akaike information criterion (AIC), despite a larger number of parameters. Associated with the availability of crop-type land cover maps, the use of S2 data and crop-type modified VPRM parameterization presented in this study, provide a step forward for upscaling cropland carbon fluxes at European scale.

Abstract Future European agriculture should achieve high productivity while limiting its impact on the environment. Legume-supported crop rotations could contribute to these goals, as they request less nitrogen (N) fertilizer inputs, show high resource use efficiency and support biodiversity. However, legumes grown for their grain (pulses) are not widely cultivated in Europe. To further expand their cultivation, it remains crucial to better understand how different cropping and environmental features affect pulses production in Europe. To address this gap, we collected the grain yields of the most cultivated legumes across European countries, from both published scientific papers and unpublished experiments of the European projects LegValue and Legato. Data were integrated into an open-source, easily updatable dataset, including 5229 yield observations for five major pulses: chickpea (Cicer arietinum L.), faba bean (Vicia faba L.), field pea (Pisum sativum L.), lentil (Lens culinaris Medik.), and soybean (Glycine max (L.) Merr.). These data were collected in 177 field experiments across 21 countries, from 37° N (southern Italy) to 63° N (Finland) of latitude, and from ca. 8° W (western Spain) to 47° E (Turkey), between 1980 and 2020. Our dataset can be used to quantify the effects of the soil, climate, and agronomic factors affecting pulses yields in Europe and could contribute to identifying the most suitable cropping areas in Europe to grow pulses.

Abstract Statistical synthesis of data sets (meta-analysis, MA) has become a popular approach for providing scientific evidence to inform environmental and agricultural policy. As the number of published MAs is increasing exponentially, multiple MAs are now often available on a specific topic, delivering sometimes conflicting conclusions. To synthesise several MAs, a first approach is to extract the primary data of all the MAs and make a new MA of all data. However, this approach is not always compatible with the short period of time available to respond to a specific policy request. An alternative, and faster, approach is to synthesise the results of the MAs directly, without going back to the primary data. However, the reliability of this approach is not well known. In this paper, we evaluate three fast-track methods for synthesising the results of MAs without using the primary data. The performances of these methods are then compared to a global MA of primary data. Results show that two of the methods tested can yield similar conclusions when compared to global MA of primary data, especially when the level of redundancy between MAs is low. We show that the use of biased MAs can reduce the reliability of the conclusions derived from these methods.

Management practices that increase the surface albedo of cultivated land could mitigate climate change, with similar effectiveness to practices that reduce greenhouse gas emissions or favor natural CO _2 sequestration. Yet, the efficiency of such practices is barely quantified. In this study, we quantified the impacts of seven different management practices on the surface albedo of winter wheat fields (nitrogen fertilizer, herbicide, fungicide, sowing, harvest, tillage, and crop residues) by analyzing observed daily albedo dynamics from eight European flux-tower sites with interpretable machine learning. We found that management practices have significant influences on surface albedo dynamics compared with climate and soil conditions. The nitrogen fertilizer application has the largest effect among the seven practices as it increases surface albedo by 0.015 ± 0.004 during the first two months after application, corresponding to a radiative forcing of −4.39 ± 1.22 W m ^−2 . Herbicide induces a modest albedo decrease of 0.005 ± 0.002 over 150 d after application by killing weeds in the fallow period only, resulting in a magnitude of radiative forcing of 1.33 ± 1.06 W m ^−2 which is higher than radiative forcing of other practices in the same period. The substantial temporal evolution of the albedo impacts of management practices increases uncertainties in the estimated albedo-mediated climate impacts of management practices. Although these albedo effects are smaller than published estimates of the greenhouse gas-mediated biogeochemical practices, they are nevertheless significant and should thus be accounted for in climate impact assessments.

High-dimensional climate data collected on a daily, monthly, or seasonal time step are now commonly used to predict crop yields worldwide with standard statistical models or machine learning models. Since the use of all available individual climate variables generally leads to calculation problems, over-fitting, and over-parameterization, it is necessary to aggregate the climate data used as predictors. However, there is no consensus on the best way to perform this task, and little is known about the impacts of the type of aggregation method used and of the temporal resolution of weather data on model performances. Based on historical data from 1981 to 2016 of soybean yield and climate on 3447 sites worldwide, this study compares different temporal resolutions (daily, monthly, or seasonal) and dimension reduction techniques (principal component analysis (PCA), partial least square regression, and their functional counterparts) to aggregate climate data used as inputs of machine learning and linear regression (LR) models predicting yields. Results showed that random forest models outperformed and were less sensitive to climate aggregation methods than LRs when predicting soybean yields. With our models, the use of daily climate data did not improve predictive performance compared to monthly data. Models based on PCA or averages of monthly data showed better predictive performance compared to those relying on more sophisticated dimension reduction techniques. By highlighting the high sensitivity of projected impact of climate on crop yields to the temporal resolution and aggregation of climate input data, this study reveals that model performances can be improved by choosing the most appropriate time resolution and aggregation techniques. Practical recommendations are formulated in this article based on our results.

Abstract Earthworms play a key role in soil carbon mineralization, but their effect is highly uncertain and suspected to vary as a function of several factors, particularly the earthworm density and time from earthworm inoculation. We conducted a meta-analysis considering these factors based on 42 experiments comparing carbon mineralization in the absence and presence of earthworms at different times. The results reveal an average carbon mineralization increase of 24% (sd 41%) in the presence of earthworms with an initial median earthworm density of 1.95 mg/g soil DM (Dry Mass) (sd 48%). We show that carbon mineralization due to earthworms was related to their density and time from inoculation. From a simple regression model using these two variables, the estimated impact of earthworms on carbon mineralization was 20% increase from 0 to 60 days and 14% decrease at day 350 for a density of worms commonly found in soils (0.5 mg/g soil DM). Finally, we proposed a simple equation that could be used in organic matter decomposition models that do not take macrofauna into account.

Measurement(s) biomass • N concentration Technology Type(s) field data collection • field data collection and lab analysis Factor Type(s) nitrogen nutrition index Sample Characteristic - Organism field crops Sample Characteristic - Environment agricultural systems Sample Characteristic - Location global dataset

Abstract Following a request from the European Commission, the EFSA Panel on Plant Health performed a risk assessment of the citrus fruit midge Resseliella citrifrugis (Diptera: Cecidomyiidae), an oligophagous species, which feeds on fruits of Citrus spp., and is reported from China. The pest was temporarily regulated in October 2022 (Regulation (EU) 2022/1941, under Art. 30 (2016/2031)). The entry risk assessment focused on the citrus fruit pathway. Three scenarios were considered: A0 (current practice, i.e. regulated pest for the EU), A1 (deregulation) and A2 (A0 with additional stand‐alone post‐harvest cold treatment). Based on the outputs of the entry model, under scenario A0, slightly less than 40 potential founder populations per year are expected (median; 90%‐uncertainty interval between about one per 30 years and about 3,000 per year). Under scenario A1, the risk of entry increases by about three times and reaches about 120 potential founder populations per year (median; 90%‐uncertainty interval between about one per 10 years and about 9,000 per year). Compared to scenario A0, the risk of entry is orders of magnitude lower for scenario A2 (median = about one potential founder population per 120 years; 90%‐uncertainty interval between one per about 600 million years and about two per year). The main uncertainties in the entry assessment are the probability of transfer, the RRO effectiveness (for scenario A2) and the disaggregation of consignments (transport of citrus fruit in boxes or lots to different locations). For all scenarios, the number of established populations is only slightly lower than the number of potential founder populations. Establishment is thus not expected to be a major constraint for this pest to then spread and cause impacts, despite the uncertainty about the pest thermal requirements. The median lag period between establishment and spread is estimated to be about 18 months (90%‐uncertainty interval between about 7 and 54 months). After the lag period, the median rate of spread by flying and due to transport of harvested citrus fruit from orchards to packinghouses is estimated at about 100 km/year (90%‐range between about 40 and 500 km/year). The main uncertainties in the spread assessment include the level of susceptibility of cultivars of different citrus species in the EU, the spread rate in China and the climate suitability of the initial spread focus in the EU. The median impact of R. citrifrugis in the EU citrus‐growing area (proportion of infested citrus fruit out of harvested citrus fruit) is estimated at about 10% (90%‐uncertainty interval between about 2% and 25%). Uncertainties affecting the impact assessment include the susceptibility of different citrus cultivars and the effect of the citrus fruit‐harvesting season in the EU (mainly winter, the less suitable season for the pest).

Abstract The internal validity of conclusions about effectiveness or impact in systematic reviews, and of decisions based on them, depends on risk of bias assessments being conducted appropriately. However, a random sample of 50 recently-published articles claiming to be quantitative environmental systematic reviews found 64% did not include any risk of bias assessment, whilst nearly all that did omitted key sources of bias. Other limitations included lack of transparency, conflation of quality constructs, and incomplete application of risk of bias assessments to the data synthesis. This paper addresses deficiencies in risk of bias assessments by highlighting core principles that are required for risk of bias assessments to be fit-for-purpose, and presenting a framework based on these principles to guide review teams on conducting risk of bias assessments appropriately and consistently. The core principles require that risk of bias assessments be Focused, Extensive, Applied and Transparent (FEAT). These principles support risk of bias assessments, appraisal of risk of bias tools, and the development of new tools. The framework follows a Plan-Conduct-Apply-Report approach covering all stages of risk of bias assessment. The scope of this paper is comparative quantitative environmental systematic reviews which address PICO or PECO-type questions including, but not limited to, topic areas such as environmental management, conservation, ecosystem restoration, and analyses of environmental interventions, exposures, impacts and risks.

Measurement(s) relative yield change Technology Type(s) crop simulation model Factor Type(s) geographic location • current average temperature • current annual precipitation • future mid-point • climate scenario • temperature change • annual precipitation change • CO2 ppm Sample Characteristic - Organism Zea mays • Oryza sativa • Glycine max • Triticum aestivum Sample Characteristic - Environment climate change Sample Characteristic - Location global Machine-accessible metadata file describing the reported data: https://doi.org/10.6084/m9.figshare.17427674

Using global flux tower observations, the authors show that atmospheric dryness always reduces photosynthesis, whereas soil dryness can increase photosynthesis if soil water stores are sufficient.

Abstract Following a request from the European Commission, the EFSA Panel on Plant Health performed a risk assessment of Citripestis sagittiferella (Lepidoptera: Pyralidae), the citrus pulp borer, an oligophagous pest reported from South‐East Asia and restricted to Citrus spp. The entry risk assessment focused on the citrus fruit pathway. Two scenarios were considered: scenario A0 (current practice) and A2 (additional post‐harvest cold treatment). Based on the outputs of the entry model obtained in scenario A0, the median number of founder populations in the EU citrus‐growing area is estimated to be slightly less than 10 per year (90%‐uncertainty interval between about one entry per 180 years and 1,300 entries per year). The risk of entry and the simulated numbers of founder populations are orders of magnitude lower for scenario A2 compared to scenario A0. The key uncertainties in the entry model include transfer, the cold treatment effectiveness, the disaggregation factor and sorting. The simulated numbers of established populations are only slightly lower than the numbers of founder populations. As the probability of establishment has little impact on the number of established populations, it is not a major source of uncertainty, despite the lack of data on the thermal biology of the pest. The median lag period between establishment and spread is estimated to be slightly more than 1 year (90%‐uncertainty interval between about 2 months and 33 months). After the lag period, the median spread rate by natural means (flying) and due to transport of harvested citrus fruit from orchards to packinghouses is estimated at about 100 km/year (90%‐uncertainty interval between about 40 and 500 km/year). The main sources of uncertainties affecting the spread rate include the extent to which environmental factors could hamper the build‐up of the populations and the lack of data on the spread rate at the origin. The median impact of C. sagittiferella in the EU citrus‐growing area is estimated at about 10% of infested fruits among the harvested citrus fruits (90%‐uncertainty interval between about 2% and 25%). Uncertainties affecting the impact assessment include the susceptibility of different Citrus species and cultivars.

Abstract Following a request from the European Commission, the EFSA Panel on Plant Health performed a risk assessment of Xanthomonas citri pv. viticola (Xcv). This pest causes bacterial canker of grapevine and is reported from Brazil and India. Two scenarios were considered: scenario A0 (current practice) and A2 (additional control measures). For the fresh grape import pathway, scenario A0 results in an order of magnitude of about one entry per 10 years (median; 90% uncertainty interval between ca. one entry per 18,000 years and ca. five entries per year). For the Vitis spp. plants for planting for research/breeding purposes import pathway, the risk of entry is several orders of magnitude smaller than the risk due to fresh grape import. This outcome is also obtained under scenario A2. The key entry uncertainties include import volume and transfer (for plants for planting), transfer and the disaggregation factor (for fresh grapes) and the limited availability of epidemiological data. The extent of the area favourable for Xcv establishment in the EU is uncertain, illustrating the limitations of climate suitability assessments when based on few data points and little epidemiological information. Nevertheless, the risk of Xcv establishment is only slightly lower than the risk of Xcv entry, i.e. no major establishment constraints are expected for most entries. Similarly, the risk of Xcv establishment is assessed as only slightly lower under current climate compared to the climate of 2041–2060. For grapevine growing areas in the EU with average yearly temperature above 17°C, the lag phase between establishment and spread is expected to be about 3 years (median; 90% range between ca. 6 months and ca. 6 years). Under the same scenario, the rate of spread by natural means is assessed to be ca. 300 m/year (median; 90% range between ca. 35 and ca. 800 m/year). The spread rate would be considerably higher considering movements of plants and cutting tools or machinery. The percentage of grapevine plants infected by Xcv in production sites as yearly average over a 30‐year production cycle is estimated to be ca. 17% (median; 90% range between ca. 1.5% and ca. 46%) in table grapes and ca. 12% (median; 90% range between ca. 0.7% and ca. 37%) in wine grapes. Impacts have been reported to be severe in Brazil and India, but the estimates provided here show that there is considerable uncertainty about expected impacts in the EU.

Using decision support systems to schedule fungicide application based on disease risk provides similar protection to calendar-based strategies but uses 50% less fungicide, according to a global meta-analysis.

Abstract Following a request from the European Commission, the EFSA Panel on Plant Health performed a pest risk assessment of Amyelois transitella (Lepidoptera: Pyralidae), the navel orangeworm, for the EU. The quantitative assessment considered two scenarios: (i) current practices and (ii) a requirement for chilled transport. The assessment focused on pathways of introduction, climatic conditions and cultivation of hosts allowing establishment, spread and impact. A. transitella is a common pest of almonds, pistachios and walnuts in California, which is the main source for these nuts imported into the EU. Based on size of the trade and infestation at origin, importation of walnuts and almonds from the USA was identified as the most important pathways for entry of A. transitella. Using expert knowledge elicitation (EKE) and pathway modelling, a median estimate of 2,630 infested nuts is expected to enter the EU each year over the next 5 years (90% certainty range (CR) from 338 to 26,000 infested nuts per year). However, due to estimated small likelihoods of transfer to a host, mating upon transfer and survival of founder populations, the number of populations that establish was estimated to be 0.000698 year−1 (median, 90% CR: 0.0000126–0.0364 year−1). Accordingly, the expected period between founding events is 1,430 years (median, 90% CR: 27.5–79,400 year). The likelihood of entry resulting in establishment is therefore considered very small. However, this estimate has high uncertainty, mainly concerning the processes of transfer of the insect to hosts and the establishment of founder populations by those that successfully transfer. Climate matching and CLIMEX modelling indicate that conditions are most suitable for establishment in the southern EU, especially around the Mediterranean basin. The median rate of natural spread was estimated to be 5.6 km/year (median, 90% CR 0.8–19.3 km/year), after an initial lag period of 3.1 year (mean, 90% CR 1.7–6.2 year) following the establishment of a founder population. If A. transitella did establish, estimated median yield losses in nuts were estimated to be in the order of 1–2% depending on the nut species and production system. A scenario requiring imports of nuts to be transported under chilled conditions was shown to provide potential to further reduce the likelihood of entry.

Abstract To identify newborns at risk of developing ASD and to detect ASD biomarkers early after birth, we compared retrospectively ultrasound and biological measurements of babies diagnosed later with ASD or neurotypical (NT) that are collected routinely during pregnancy and birth. We used a supervised machine learning algorithm with a cross-validation technique to classify NT and ASD babies and performed various statistical tests. With a minimization of the false positive rate, 96% of NT and 41% of ASD babies were identified with a positive predictive value of 77%. We identified the following biomarkers related to ASD: sex, maternal familial history of auto-immune diseases, maternal immunization to CMV, IgG CMV level, timing of fetal rotation on head, femur length in the 3rd trimester, white blood cell count in the 3rd trimester, fetal heart rate during labor, newborn feeding and temperature difference between birth and one day after. Furthermore, statistical models revealed that a subpopulation of 38% of babies at risk of ASD had significantly larger fetal head circumference than age-matched NT ones, suggesting an in utero origin of the reported bigger brains of toddlers with ASD. Our results suggest that pregnancy follow-up measurements might provide an early prognosis of ASD enabling pre-symptomatic behavioral interventions to attenuate efficiently ASD developmental sequels.

There are big uncertainties in the contribution of irrigation to crop yields. Here, the authors use Bayesian model averaging to combine statistical and process-based models and quantify the contribution of irrigation for wheat and maize yields, finding that irrigation alone cannot close yield gaps for a large fraction of global rainfed agriculture.

Abstract Conservation agriculture (CA) has been promoted to mitigate climate change, reduce soil erosion, and provide a variety of ecosystem services. Yet, its impacts on crop yields remains controversial. To gain further insight, we mapped the probability of yield gain when switching from conventional tillage systems (CT) to CA worldwide. Relative yield changes were estimated with machine learning algorithms trained by 4403 paired yield observations on 8 crop species extracted from 413 publications. CA has better productive performance than no-till system (NT), and it stands a more than 50% chance to outperform CT in dryer regions of the world, especially with proper agricultural management practices. Residue retention has the largest positive impact on CA productivity comparing to other management practices. The variations in the productivity of CA and NT across geographical and climatical regions were illustrated on global maps. CA appears as a sustainable agricultural practice if targeted at specific climatic regions and crop species.

Measurement(s) yield trait • crop Technology Type(s) digital curation Factor Type(s) precipitation balance • temperature • crop species • soil texture • management of crop rotation • management of soil cover • management of field fertilization • management of weed and pest control • management of crop irrigation • geographic location Sample Characteristic - Organism Viridiplantae Sample Characteristic - Environment cultivated environment Sample Characteristic - Location global Machine-accessible metadata file describing the reported data: https://doi.org/10.6084/m9.figshare.13303118

This study analyses the quality of six regression algorithms in forecasting the monthly price of maize in its primary international trading market, using publicly available data of agricultural production at a regional scale. The forecasting process is done between one and twelve months ahead, using six different forecasting techniques. Three (CART, RF, and GBM) are tree-based machine learning techniques that capture the relative influence of maize-producing regions on global maize price variations. Additionally, we consider two types of linear models—standard multiple linear regression and vector autoregressive (VAR) model. Finally, TBATS serves as an advanced time-series model that holds the advantages of several commonly used time-series algorithms. The predictive capabilities of these six methods are compared by cross-validation. We find RF and GBM have superior forecasting abilities relative to the linear models. At the same time, TBATS is more accurate for short time forecasts when the time horizon is shorter than three months. On top of that, all models are trained to assess the marginal contribution of each producing region to the most extreme price shocks that occurred through the past 60 years of data in both positive and negative directions, using Shapley decompositions. Our results reveal a strong influence of North-American yield variation on the global price, except for the last months preceding the new-crop season.

Abstract Background Antibiotic resistance is a major concern for public and environmental health. The role played by the environment in disseminating resistance is increasingly considered, as well as its capacity for mitigation. We reviewed the literature on strategies to control dissemination of antibiotic-resistant bacteria (ARB), antibiotic resistance genes (ARG) and mobile genetic elements (MGE) in the environment. Methods This systematic review focused on three main strategies: (i) restriction of antibiotic use (S1), (ii) treatments of liquid/solid matrices (S2) and (iii) management of natural environment (S3). Articles were collected from seven scientific databases until July 2017 and from Web of Science until June 2018. Only studies reporting measurements of ARB, ARG or MGE in environmental samples were included. An evidence map was drawn from metadata extracted from all studies eligible for S1, S2 and S3. Subsets of studies were assessed for internal and external validity to perform narrative and quantitative syntheses. A meta-analysis was carried out to assess the effects of organic waste treatments (random-effect models). Review findings Nine hundred and thirty-one articles representing 1316 individual studies (n) were eligible for S1 (n = 59), S2 (n = 781) and S3 (n = 476) strategies, respectively. Effects of interventions to control the dissemination of antibiotic resistance in the environment were primarily studied in strategy S2. A partial efficiency of wastewater treatment plants (WWTPs) to reduce antibiotic resistance in treated effluent was reported in 118 high validity studies. In spite of the heterogeneity in published results, the meta-analysis showed that composting and drying were efficient treatments to reduce the relative abundance of ARG and MGE in organic waste, by 84% [65%; 93%] and 98% [80%; 100%], respectively. The effect of anaerobic digestion was not statistically significant (51% reduction [− 2%; 77%]) when organic waste treatments were compared together in the same model. Studies in strategies S1 and S3 mainly assessed the effects of exposure to sources of contamination. For instance, 28 medium/high validity studies showed an increase of antibiotic resistance in aquatic environments at the WWTP discharge point. Some of these studies also showed a decrease of resistance as the distance from the WWTP increases, related to a natural resilience capacity of aquatic environments. Concerning wildlife, nine medium/high validity studies showed that animals exposed to anthropogenic activities carried more ARB. Conclusions and implications Knowledge gaps were identified for the relationship between restriction of antibiotic use and variation of antibiotic resistance in the environment, as well as on possible interventions in situ in natural environment. Organic waste treatments with thermophilic phase (> 50 °C) should be implemented before the use/release of organic waste in the environment. More investigation should be conducted with the datasets available in this review to determine the treatment efficiency on ARG carried by specific bacterial communities.

Assessing the productive performance of conservation agriculture (CA) has become a major issue due to growing concerns about global food security and sustainability. Numerous experiments have been conducted to assess the performance of CA under various local conditions, and meta-analysis has become a standard approach in agricultural sector for analysing and summarizing the experimental data. Meta-analysis provides valuable synthetic information based on mean effect size estimation. However, summarizing large amounts of information by way of a single mean effect value is not always satisfactory, especially when considering agricultural practices. Indeed, their impacts on crop yields are often non-linear, and vary widely depending on a number of factors, including soil properties and local climate conditions. To address this issue, here we present a machine learning approach to produce data-driven global maps describing the spatial distribution of the productivity of CA versus conventional tillage (CT). Our objective is to evaluate and compare several machine-learning models for their ability in estimating the productivity of CA systems, and to analyse uncertainty in the model outputs. We consider different usages, including classification, point regression and quantile regression. Our approach covers the comparison of 12 different machine learning algorithms, model training, tuning with cross-validation, testing, and global projection of results. The performances of these algorithms are compared based on a recent global dataset including more than 4,000 pairs of crop yield data for CA vs. CT. We show that random forest has the best performance in classification and regression, while quantile regression forest performs better than quantile neural networks in quantile regression. The best algorithms are used to map crop productivity of CA vs. CT at the global scale, and results reveal that the performance of CA vs. CT is characterized by a strong spatial variability, and that the probability of yield gain with CA is highly dependent on geographical locations. This result demonstrates that our approach is much more informative than simply presenting average effect sizes produced by standard meta-analyses, and paves the way for such probabilistic, spatially-explicit approaches in many other fields of research.

In this study, we investigate and forecast the impact of crop production shocks on the global prices of three major international agricultural commodities: maize, soybean, and cocoa. We perform a thorough assessment of the forecasting performances of five econometric and machine learning models using 60 years of data. First, we train the models on production and price data to forecast the monthly price variations for each crop separately considering different time horizons. Next, we implement a cross-validation procedure to identify the models with the most accurate forecasting ability for each crop. After choosing the best forecaster, we identify the most influential producing areas using several local and global model-agnostic interpretation tools. Our findings indicate significant differences among commodities in terms of prediction accuracy, with cocoa exhibiting a higher level of prediction error compared to less volatile markets like maize and soybean. Our results reveal a significant influence of Northern America’s maize and soybean production on the global prices of these commodities. The effects of production on prices are asymmetrical: small decreases in US production lead to substantial price increases, while small increases in production do not systematically decrease prices. In contrast, cocoa price variations are influenced by production coming from several regions, not from a single one.

The identification of sustainable fertilisation practices is essential to reduce agriculture’s impact on the environment while insuring sufficient crop production. The use of enhanced efficiency fertilisers (EEFs) is thought to improve nitrogen (N)-fertiliser uptake by crops while reducing nutrient losses to the environment. EEFs’ performance has been assessed in several meta-analyses and systematic reviews, which are heterogeneous in content and quality of reporting. This provides fragmented information and makes it difficult to conclude about their ability to provide more sustainable fertilisation. Here we synthetise evidence from 26 meta-analyses and reviews selected by a systematic literature search to describe the separate effects of four commonly used EEFs—nitrification inhibitors, urease inhibitors, double inhibitors, or controlled-release fertilisers—on the environment, nutrient use efficiency, soil fertility, and crop production. A unique contribution of this review is the assessment of the quality of the selected papers and the synthesis of their results through a systematic framework. Results showed that compared to conventional fertilisers, EEFs generally increased soil nutrients, crop yield, and N use efficiency, and reduced N leaching, emissions of greenhouse gases and air pollutants. Some differences were found between the different EEFs; while urease inhibitors, double inhibitors, and controlled-released fertilisers decreased ammonia emission compared to conventional fertilisers, nitrification inhibitors increased these emissions or did not affect them. The results were consistent when excluding low-quality studies from the analyses. Overall, this global synthesis indicates that EEFs could maintain crop yields while reducing some of the negative environmental impacts of conventional N-fertilisers. Attention should be paid to the potential increase of ammonia emissions by nitrification inhibitors and additional evidence is needed on the potential side effects on soil health, biodiversity, and water quality.

Agricultural price shocks strongly affect farmers' income and food security. It is therefore important to understand and anticipate their origins and occurrence, particularly for the world's main agricultural commodities. In this study, we assess the impacts of yearly variations in regional maize productions and yields on global maize prices using several statistical and machine-learning (ML) methods. Our results show that, of all regions considered, Northern America is by far the most influential. More specifically, our models reveal that a yearly yield gain of +8% in Northern America negatively impacts the global maize price by about –7%, while a decrease of –0.1% is expected to increase global maize price by more than +7%. Our classification models show that a small decrease in the maize yield in Northern America can inflate the probability of maize price increase on the global scale. The maize productions in the other regions have a much lower influence on the global price. Among the tested methods, random forest and gradient boosting perform better than linear models. Our results highlight the interest of ML in analyzing global prices of major commodities and reveal the strong sensitivity of maize prices to small variations of maize production in Northern America.

Abstract Recently, yield shocks due to extreme weather events and their consequences for food security have become a major concern. Although long yield time series are available in Europe, few studies have been conducted to analyze them in order to investigate the impact of adverse climate events on yield shocks under current and future climate conditions. Here we designated the lowest 10th percentile of the relative yield anomaly as yield shock and analyzed subnational wheat yield shocks across Europe during the last four decades. We applied a data‐driven attribution framework to quantify primary climate drivers of wheat yield shock probability based on machine learning and game theory, and used this framework to infer the most critical climate variables that will contribute to yield shocks in the future, under two climate change scenarios. During the period 1980–2018, our attribution analysis showed that 32% of the observed wheat yield shocks were primarily driven by water limitation, making it the leading climate driver. Projection to future climate scenarios RCP4.5 and RCP8.5 suggested an increased risk of yield shock and a paradigm shift from water limitation dominated yield shock to extreme warming induced shocks over 2070–2099: 46% and 54% of areas were primarily driven by extreme warming under RCP4.5 and RCP8.5, respectively. A similar analysis conducted on yields simulated by an ensemble of crop models showed that models can capture the negative impact of low water supply but missed the impact of excess water. These discrepancies between observed and simulated yield data call for improvement in crop models.

Botrytis cinerea, a fungal pathogen that causes gray mold on grapes, can decrease yield, substantially reduce wine quality, and therefore cause significant economic losses. In a context of increasing awareness of environmental and human health, biopesticides are a potential alternative to synthetic chemical treatments to produce grapes and wine in compliance with high food standards. However, the effectiveness of biopesticides is not well known and more research is needed to help winegrowers assess their ability to control wine diseases. Our study aims to assess the efficacy of two commercial biopesticides, based on potassium bicarbonate and Aureobasidium pullulans, in reducing the incidence of gray mold (i.e., the proportion of grape bunches that are diseased). We use data from an on-farm trial network managed over 3 years (from 2014 to 2016) in a major wine producing region located in Southwestern France, and fit Bayesian generalized linear multilevel models able to take the variability of treatment effect across trials into account. The fitted models were then used to estimate the efficacy on incidence as a function of the severity (i.e., the proportion of diseased grape berries in a bunch) in an untreated plot in order to determine if the effectiveness of the treatments depends on the disease pressure. At average disease severity (i.e., 3%), the efficacy on disease incidence at the network level was equal to 20% [95% CI = (−0.1; 37.3)] and 13% [95% CI = (0.2; 24.7)] for potassium bicarbonate and A. pullulans, respectively. For both biopesticides, the efficacy on incidence for a new site-year is highly uncertain, but potassium bicarbonate had a lower uncertainty and a lower application cost compared to A. pullulans. Our results confirm that potassium bicarbonate is an interesting biopesticide under farming conditions in organic vineyards in southwestern France, but the amount of uncertainty points to the need for further research.

Grape downy mildew (GDM) is a major disease of grapevine that has an impact on both the yields of the vines and the quality of the harvested fruits. The disease is currently controlled by repetitive fungicide treatments throughout the season, especially in the Bordeaux vineyards where the average number of fungicide treatments against GDM was equal to 10.1 in 2013. Reducing the number of treatments is a major issue from both an environmental and a public health point of view. One solution would be to identify vineyards that are likely to be heavily attacked in spring and then apply fungicidal treatments only to these situations. In this perspective, we use here a dataset including 9 years of GDM observations to develop and compare several generalized linear models and machine learning algorithms predicting the probability of high incidence and severity in the Bordeaux region. The algorithms tested use the date of disease onset and/or average monthly temperatures and precipitation as input variables. The accuracy of the tested models and algorithms is assessed by year-by-year cross validation. LASSO, random forest and gradient boosting algorithms show better performance than generalized linear models. The date of onset of the disease has a greater influence on the accuracy of forecasts than weather inputs and, among weather inputs, precipitation has a greater influence than temperature. The best performing algorithm was selected to evaluate the impact of contrasted climate scenarios on GDM risk levels. Results show that risk of GDM at bunch closure decreases with reduced rainfall and increased temperatures in April-May. Our results also show that the use of fungicide treatment decision rules that take into account local characteristics would reduce the number of treatments against GDM in the Bordeaux vineyards compared to current practices by at least 50%.

Classical land rent theories imply that the best land is cultivated first. This principle forms the basis of many land-use studies, but empirical evidence remains limited, especially on a global scale. In this paper, we estimate the effects of agricultural suitability and market accessibility on the spatial allocation of cultivated areas at a 30 arc-min resolution in 15 world regions. Our results show that both determinants often have a significant positive effect on the cropland fraction, but with large variations in strength across regions. Based on a quantile analysis, we find that agricultural suitability is the dominant driver of cropland allocation in North America, Middle East and North Africa and Eastern Europe, whereas market accessibility shows a stronger effect in other regions, such as Western Africa. In some regions, such as South and Central America, both determinants have a limited effect on cropland fraction. Comparison of high versus low quantile regression coefficients shows that, in most regions, densely cropped areas are more sensitive to agricultural suitability and market accessibility than sparsely cropped areas.

The production of anthropogenic reactive nitrogen (N) has grown so much in the last century that quantifying the effect of N enrichment on plant growth has become a central question for carbon (C) cycle research. Numerous field experiments generally found that N enrichment increased site-scale plant biomass, although the magnitude of the response and sign varied across experiments. We quantified the response of terrestrial natural vegetation biomass to N enrichment in the Northern Hemisphere (>30° N) by scaling up data from 773 field observations (142 sites) of the response of biomass to N enrichment using machine-learning algorithms. N enrichment had a significant and nonlinear effect on aboveground biomass (AGB), but a marginal effect on belowground biomass. The most influential variables on the AGB response were the amount of N applied, mean biomass before the experiment, the treatment duration and soil phosphorus availability. From the machine learning models, we found that N enrichment due to increased atmospheric N deposition during 1993–2010 has enhanced total biomass by 1.1 ± 0.3 Pg C, in absence of losses from harvest and disturbances. The largest effect of N enrichment on plant growth occurred in northeastern Asia, where N deposition markedly increased. These estimates were similar to the range of values provided by state-of-the-art C–N ecosystem process models. This work provides data-driven insights into hemisphere-scale N enrichment effect on plant biomass growth, which allows to constrain the terrestrial ecosystem process model used to predict future terrestrial C storage.

In France, the 2016 winter wheat harvest was at its lowest since over 50 years. Here, Ben-Ari et al. show the role of seasonal temperature and precipitation extremes in this loss, and accounting for both of these variables explains large, historical yield loss events.

Numerous meta-analyses have been conducted in the last three decades to assess the productive and environmental benefits resulting from a diversification of cropping systems. These meta-analyses assessed one or several diversification strategies (e.g., rotations, cover crops, agroforestry) according to various outcomes (e.g., productivity, profitability, biodiversity). To date, no dataset has provided a comprehensive synthesis of existing experimental data on crop diversification. We present here a dataset containing 2382 effect sizes published in 99 meta-analyses covering 3736 experimental studies worldwide (https://figshare.com/s/c15a93e96c95f89ddd89). We also provide an extensive appraisal of the quality of each meta-analysis and a quantification of the redundancy of primary studies between meta-analyses. Our database hence provides (i) a quantification of the impacts of a variety of diversification strategies on crop production, the environment and economic profitability at the global scale and, (ii) a quality and redundancy assessment that may be used as a reference for future studies. Keywords: Agroforestry, Cover crop, Intercropping, Rotation, Variety mixture, Systematic review, Meta-synthesis

Abstract This opinion updates the information included in the previous EFSA Scientific Opinion concerning the in planta control measures for Xylella fastidiosa, with a systematic review and critical analysis of the potential treatment solutions that have been published against this pest so far. The output of this opinion focuses on the application of chemical or biological treatments on living plants. In vitro studies, hot water treatments, use of resistant varieties and vector control are excluded from the review. The use of antibiotics is not considered due to the risk of antimicrobial resistance development. The use of weakly virulent or avirulent strains of X. fastidiosa is covered in this review, although this organism is an EU quarantine plant pest and its introduction in the EU territory is banned. Experiments were recently conducted to assess the effect of application of zinc, copper, and citric acid biocomplex, of N‐acetylcysteine, and of ‘diffusible signal factor’ (and of its homologs). Their results showed that these control measures were sometimes able to reduce symptoms caused by X. fastidiosa. Recent experiments also showed that several species of endophytic microorganisms, some bacteriophages and inoculation of weakly virulent/avirulent strains of X. fastidiosa could offer some protection against the Pierce's disease. However, based on the reviewed results, the Panel concludes that, although several published experiments show some effects in reducing symptoms development, the tested control measures are not able to completely eliminate X. fastidiosa from diseased plants. The Panel confirms as previously stated that there is currently no control measure available to eliminate the bacteria from a diseased plant in open field conditions.

An amendment to this paper has been published and can be accessed via the original article.

West Africa is one of the regions the most concerned with structural food and nutrition security. Consequently, agricultural development pathways and scenarios are under high scientific and political scrutiny in this region. Rice, as a rapidly growing staple plays a key role in the West African diet representing close to 40% of the total volume of cereal consumed in the region. In the context of the 2008 food price crisis several West African countries have since proclaimed rice self-sufficiency as a target. Here, we show that rice yields tend to be, on average over the entire region, less stable (by a range of 15%–30%) than that of alternative crops, possibly substitutable in diets. The regions where yields of alternative crops are more stable than those of rice correspond to the main climatic regions where these crops are grown: sorghum, millet in the Sahelian and Sudanian regions and tubers in the Guinean region. Rice yields are significantly less stable for 33% of the comparisons considered and are significantly more stable than any alternative crop for 15% of the comparisons in few areas without clear longitudinal patterns. Models accounting for climate variability explain up to 17% of the variance of the data and reveal that yield variability differences between rice and alternative crops tends to widen in the areas where the monsoon precipitation is more variable between-years. The highest levels of variability of rice yields compared to alternative crops are recorded in regions where the monsoon varies strongly between years. Our analysis advocates for an explicit account of yield stability in West African rice expansion scenarios and supply strategies.

Plant pest invasions cost billions of Euros each year in Europe. Prediction of likely places of pest introduction could greatly help focus efforts on prevention and control and thus reduce societal costs of pest invasions. Here, we test whether generic data-driven risk maps of pest introduction, valid for multiple species and produced by machine learning methods, could supplement the costly species-specific risk analyses currently conducted by governmental agencies. An elastic-net algorithm was trained on a dataset covering 243 invasive species to map risk of new introductions in Europe as a function of climate, soils, water, and anthropogenic factors. Results revealed that the BeNeLux states, Northern Italy, the Northern Balkans, and the United Kingdom, and areas around container ports such as Antwerp, London, Rijeka, and Saint Petersburg were at higher risk of introductions. Our analysis shows that machine learning can produce hotspot maps for pest introductions with a high predictive accuracy, but that systematically collected data on species’ presences and absences are required to further validate and improve these maps.

Diseases of fruit and foliage caused by fungi and oomycetes are generally controlled by the application of fungicides. The use of decision support systems (DSSs) may assist to optimize fungicide programs to enhance application on the basis of risk associated with disease outbreak. Case-by-case evaluations demonstrated the performance of DSSs for disease control, but an overall assessment of the efficacy of DSSs is lacking. A literature review was conducted to synthesize the results of 67 experiments assessing DSSs. Disease incidence data were obtained from published peer-reviewed field trials comparing untreated controls, calendar-based and DSS-based fungicide programs. Two meta-analysis generic models, a “fixed-effects” vs. a “random-effects” model within the framework of generalized linear models were evaluated to assess the efficacy of DSSs in reducing incidence. All models were fit using both frequentist and Bayesian estimation procedures and the results compared. Model including random effects showed better performance in terms of AIC or DIC and goodness of fit. In general, the frequentist and Bayesian approaches produced similar results. Odds ratio and incidence ratio values showed that calendar-based and DSS-based fungicide programs considerably reduced disease incidence compared to the untreated control. Moreover, calendar-based and DSS-based programs provided similar reductions in disease incidence, further supporting the efficacy of DSSs.

The diversification of cropping systems encompasses different strategies that may help maintain or enhance the sustainability of agriculture. Thousands of experiments have been carried out around the world since almost five decades to evaluate and compare the performances of various diversification strategies in a wide array of agroecosystems and climates. Although these analyses have been synthesized in a growing number of meta-analyses, the information remains somewhat fragmented. A multicriteria systematic synthesis of worldwide agricultural diversification is still lacking. Here, we review all meta-analyses conducted on crop diversification strategies and produce a detailed overview of their results and of their quality. We identified and analyzed 99 meta-analyses summarizing the results of more than 3700 agronomic experiments on seven crop diversification strategies. Among these strategies, rotation and associated plant species are dominant in the literature followed by intercropping, agroforestry and landscape heterogeneity. Our analysis reveals that rotation and intercropping are associated with yield increases. Agroforestry systematically induces an improvement of biodiversity and soil quality—in particular soil organic carbon. We show that, regardless of the context, a combination of several diversification strategies outperforms any individual strategy. Our review reveals that a significant knowledge gap remains, in particular regarding water use, farmers’ profitability, product quality and production stability. Few meta-analyses investigate the performance of landscape heterogeneity and of systems with species other than cereals and pulses. Additionally, we show that most of the meta-analyses studied cannot be considered fully transparent and reproducible. Their conclusions should therefore be interpreted with caution. Our systematic mapping provides a benchmark to guide and improve the relevance and reliability of future meta-analyses in agronomy.

Quantitative models have several advantages compared to qualitative methods for pest risk assessments (PRA). Quantitative models do not require the definition of categorical ratings and can be used to compute numerical probabilities of entry and establishment, and to quantify spread and impact. These models are powerful tools, but they include several sources of uncertainty that need to be taken into account by risk assessors and communicated to decision makers. Uncertainty analysis (UA) and sensitivity analysis (SA) are useful for analyzing uncertainty in models used in PRA, and are becoming more popular. However, these techniques should be applied with caution because several factors may influence their results. In this paper, a brief overview of methods of UA and SA are given. As well, a series of practical rules are defined that can be followed by risk assessors to improve the reliability of UA and SA results. These rules are illustrated in a case study based on the infection model of Magarey et al. (2005) where the results of UA and SA are shown to be highly dependent on the assumptions made on the probability distribution of the model inputs.

The world's population is predicted to exceed nine billion by 2050 and there is increasing concern about the capability of agriculture to feed such a large population. Foresight studies on food security are frequently based on crop yield trends estimated from yield time series provided by national and regional statistical agencies. Various types of statistical models have been proposed for the analysis of yield time series, but the predictive performances of these models have not yet been evaluated in detail. In this study, we present eight statistical models for analyzing yield time series and compare their ability to predict wheat yield at the national and regional scales, using data provided by the Food and Agriculture Organization of the United Nations and by the French Ministry of Agriculture. The Holt-Winters and dynamic linear models performed equally well, giving the most accurate predictions of wheat yield. However, dynamic linear models have two advantages over Holt-Winters models: they can be used to reconstruct past yield trends retrospectively and to analyze uncertainty. The results obtained with dynamic linear models indicated a stagnation of wheat yields in many countries, but the estimated rate of increase of wheat yield remained above 0.06 t ha⁻¹ year⁻¹ in several countries in Europe, Asia, Africa and America, and the estimated values were highly uncertain for several major wheat producing countries. The rate of yield increase differed considerably between French regions, suggesting that efforts to identify the main causes of yield stagnation should focus on a subnational scale.

Yield dynamics of major crops species vary remarkably among continents. Worldwide distribution of cropland influences both the expected levels and the interannual variability of global yields. An expansion of cultivated land in the most productive areas could theoretically increase global production, but also increase global yield instability if the most productive regions are characterized by high interannual yield variability. In this letter, we use portfolio analysis to quantify the tradeoff between the expected values and the interannual variance of global yield. We compute optimal frontiers for four crop species i.e., maize, rice, soybean and wheat and show how the distribution of cropland among large world regions can be optimized to either increase expected global crop production or decrease its interannual variability. We also show that a preferential allocation of cropland in the most productive regions can increase global expected yield at the expense of yield stability. Theoretically, optimizing the distribution of a small fraction of total cultivated areas can help find a good compromise between low instability and high crop yields at the global scale.

Nitrous oxide (N(2)O) is a greenhouse gas with a global warming potential approximately 298 times greater than that of CO(2). In 2006, the Intergovernmental Panel on Climate Change (IPCC) estimated N(2)O emission due to synthetic and organic nitrogen (N) fertilization at 1% of applied N. We investigated the uncertainty on this estimated value, by fitting 13 different models to a published dataset including 985 N(2)O measurements. These models were characterized by (i) the presence or absence of the explanatory variable "applied N", (ii) the function relating N(2)O emission to applied N (exponential or linear function), (iii) fixed or random background (i.e. in the absence of N application) N(2)O emission and (iv) fixed or random applied N effect. We calculated ranges of uncertainty on N(2)O emissions from a subset of these models, and compared them with the uncertainty ranges currently used in the IPCC-Tier 1 method. The exponential models outperformed the linear models, and models including one or two random effects outperformed those including fixed effects only. The use of an exponential function rather than a linear function has an important practical consequence: the emission factor is not constant and increases as a function of applied N. Emission factors estimated using the exponential function were lower than 1% when the amount of N applied was below 160 kg N ha(-1). Our uncertainty analysis shows that the uncertainty range currently used by the IPCC-Tier 1 method could be reduced.