Search

In response to the sustainability issues that agriculture faces in advanced economies, agroecology has gained increasing relevance in scientific, political, and social debates. This has promoted discussion about transitions to agroecology, which represents a significant advancement. Accordingly, it has become a growing field of research. We reviewed the literature on and in support of farm transitions to agroecology in advanced economies in order to identify key research challenges and suggest innovative research paths. Our findings can be summarized as follows: (1) Research that supports exploration and definition of desired futures, whether based on future-oriented modeling or expert-based foresight approaches, should more explicitly include the farm level. It should stimulate the creativity and design ability of farmers and other stakeholders, and also address issues of representation and power among them. (2) Research that creates awareness and assesses farms before, during or after transition requires more holistic and dynamic assessment frameworks. These frameworks need to be more flexible to adapt to the diversity of global and local challenges. Their assessment should explicitly include uncertainty due to the feedback loops and emergent properties of transitions. (3) Research that analyzes and supports farms during transition should focus more on the dynamics of change processes by valuing what happens on the farms. Research should especially give more credence to on-farm experiments conducted by farmers and develop new tools and methods (e.g., for strategic monitoring) to support these transitions. This is the first review of scientific studies of farm transitions to agroecology. Overall, the review indicates that these transitions challenge the system boundaries, temporal horizons, and sustainability dimensions that agricultural researchers usually consider. In this context, farm transitions to agroecology require changes in the current organization and funding of research in order to encourage longer term and more adaptive configurations.

In the context of climate change and growing urbanization, farms are expected to face more frequent extreme events (e.g. heat peaks, droughts, frost, hail or floods). Floods being the natural hazard that generates the most damage in the world, we focus our research on the impact of floods on farms and their adaptation to climate change. Flood damage on farms is the result of complex phenomena involving biophysical processes on the one hand, and farmers' decisions on rehabilitation and adaptation on the other. Some impacts are directly visible; others may be delayed and persist over time (Bremond et al, 2013). When they persist over time, they may impact the development trajectory of farms. For the observation of impacts, there are in fact two challenges: to identify the diversity of impacts and their temporality of occurrence. Field surveys following flood events often focus on short-term damage and on impacts resulting from extreme events, leaving long-term damage and minor events aside. In this paper, we aim to propose a methodology that we are currently implementing to jointly and recursively improve the observation and modelling of flood impacts on agricultural systems. In particular, we wish to define more precisely the diversity of flood impacts and long-term damage on farms by taking into account the development trajectory of farms. To this end, we combine two complementary approaches: observation and modelling. Our work is implemented in the framework of the system of observations of the impacts of floods (so-ii, http://so-ii.org), in the Greater Montpellier area, coordinated by our team. Based on our experience, we have developed diverse methods for both short-term and long-term monitoring such as: surveys, participative workshops, drone pictures, predictive models. To cover on field-observations, quantitative and qualitative surveys have been carried out over the years on the so-ii territory. Interviews took place both after a flood, to gather quantitative data, and several years after a flood (Bremond et al, 2020), to collect quantitative and qualitative data. In parallel, we are exploring the use of drones to gather pictures of plots from post-flood to several years later in order to understand mechanisms behind the impacts of floods on soil, plant material and adaptations. To reinforce long-term monitoring, we have set up a network of impact observers, on the so-ii territory, with whom we agreed to work over the long-term (about fifteen years). On the other hand, we are working on the use of predictive models to estimate flood damage to farms. For now, 3 models have been developed: floodam.agri (Bremond et al, 2022), R-EVA (Bremond et al, 2012) and COOPER (Nortes Martinez et al, 2021). Each of these models have different application levels (respectively plot, farm and cooperative system) and time-scales which helps for the observation of farm trajectories as a whole. In conclusion, we will discuss the perspectives and limitations of this approach. We will open up on the perspectives to share our methodology so it can be adapted to other territories by stakeholders who are interested in setting up a similar system of observation.

International audience; In the global change context, an efficient management of the available resources has become one of the most important topics particularly for the sustainable crop development. Many questions concern the evolution of the rice farming systems in Camargue in Southeastern France, which play a crucial role in controlling the soil salinity. Their surface area significantly decreased from 20 000 ha in 2010 to 14 000 ha in 2014. The arrival of the new Sentinel satellites makes it possible to evaluate these crop evolutions. The objectives of this study were to propose operational methodologies to accurately assess the surface areas of the main crops, rice, wheat, and grassland, from classifications based on multispectral data; map agricultural practices (sowing and harvest residue burning); and elaborate a farm typology based on variables computed from remote sensing data to better understand the farming strategies. Dense time series of Sentinel images acquired at high spatial resolution (10 m) were analyzed for 2016 and 2017. A satisfactory accuracy was obtained for land use classification with 88% of correctly classified fields. The accuracy obtained for the estimation of the sowing date varied according to the studied year from 8 to 12 days, and burned areas were correctly identified (80%). The farm typology allowed to cluster farms at the territory level.

International audience; AbstractThe Companion Modeling approach aims to stimulate learning and sharing of perceptions among several actors through the co-construction and use of a model, while the Integrated Assessment of Agricultural Systems aims to assess a large set of performances of current and alternative systems. Our research question was: to what extent does the joint use of these two methods foster sustainability thinking? We show how the co-design of evolving agriculture scenarios and the use of Integrated Assessment of Agricultural Systems methods guided the implementation of a Companion Modeling process applied to the Valensole Plateau, southeastern France. An initial phase, based on a preliminary diagnosis, aimed to collectively formulate the issues to be addressed, and to engage the actors in a collective dynamic. The second phase consisted of leading actors in the co-construction of a simulation model to represent the agricultural system to be evaluated using economic, social, and environmental indicators at different scales. In the final phase, the current situation and the co-designed scenario were collectively analyzed. The scenario was conceived as a “worst case”, embedded with drivers such as input costs and climate change. Opposing adaptation strategies were envisioned by farmers and other actors. Results show how debates between participants adaptively shaped the model. Scenario simulations highlighted better environmental performance at the expense of economic performance. The differences between farms led to discussions on technical management. Overall, the process stimulated collective reflection among the actors and created a learning experience with the sharing of perceptions and knowledge about the agricultural system. The model was considered relevant by the participants, who expressed interest in continuing its exploration. Our approach, which used a unique combination of Integrated Assessment of Agricultural Systems and scenario development with Companion Modeling, could be adapted to other areas and issues, although tool transfer remains an open question.

International audience; Highlights:• We used a participatory approach to design adaptation strategies by spatially combining adaptation measures.• Adaptation strategies were evaluated by numerical modeling and stakeholder opinions such as feasibility and desirability.• A few adaptation strategies maintained vineyard production volume at mid-century, but with spatial heterogeneity.• Low yield production sectors, which were less impacted by climate change, showed higher adaptive capacity.• Researchers' and stakeholders' joint participation helped the design of locally pertinent adaptations to climate change.Abstract: Context: Climate change threatens wine growing systems in varying ways because of their high diversity, even at a local scale. This diversity needs to be considered when designing and assessing adaptation strategies to coordinate better with these diverse local conditions.Objective: We developed a participatory modeling approach to (1) design adaptation strategies in a viticultural watershed in southern France, (2) numerically and spatially evaluate their effects under future climatic conditions, and (3) discuss the results with stakeholders.Methods: We organized two sets of collective workshops, before and after a simulation phase. During the workshops, we designed four adaptation strategies that correspond to different ways to combine adaptation measures proposed by stakeholders. A spatially explicit model was used to evaluate the effects of six adaptation measures (late varieties, irrigation, reducing canopy size, adjusting cover cropping, reducing density, and shading) at field scale and combinations of them at watershed scale. Simulations were realized under a high-emissions climate change scenario RCP 8.5. Model-based evaluations were followed by discussions with stakeholders. The cost-effectiveness of adaptation strategies was estimated at farm scale using an indicator designed by the stakeholders.Results and conclusions: The spatial combination of adaptation measures in a viticultural watershed provided options for adapting wine growing systems to climate change. A delayed harvest strategy with currently available late varieties provided only minimal relief from high temperature during ripening. A water stress limitation strategy would compensate for production losses if disruptive adaptations (e.g., reduced density) were adopted and if more vineyards were irrigated. A relocation strategy would encourage premium wine production in the constrained mountainous areas, where grapevines systems are historically adapted to limited water conditions. A soil improvement strategy was mentioned but not evaluated in the model due to scarce data and literature on the possible improvement of soil water holding capacity.Significance: The sharing of knowledge between researchers, technical experts, and wine growers enabled the construction of a common understanding of the local impacts and adaptation potential to climate change in the watershed. In further research, this knowledge could help decision makers to define pathways for adaptive actions at farm scale.

International audience; La diversité pédoclimatique et socio-économique des paysages viticoles constitue un défi pour évaluer les impacts du changement climatique. Les modèles sont des outils pertinents pour explorer quantitativement les futurs possibles, et les approches participatives aident à rendre compte de la diversité locale. Notre approche implique des acteurs de la viticulture et de la gestion de l'eau pour construire, tester et discuter les résultats d'un couplage de modèles : les impacts du changement climatique sur le bilan hydrique et la phénologie sont simulés avec des modèles existants, et un modèle original de rendement a été développé. Le couplage et l’exécution de ces modèles dans la plateforme OpenFLUID permet des simulations spatialement explicites sur un bassin versant viticole méditerranéen. La production de raisin diminuerait de 5 à 14% en 2100, avec une diminution plus importante dans les zones irriguées à haut rendement, malgré un doublement des apports d’eau. Une augmentation de température de 6°C pendant la maturation des baies menacerait la qualité du vin. L'approche participative a permis la construction d'un modèle ad hoc, et la discussion de ses résultats et limites avec les acteurs. Cet outil est prometteur pour identifier et tester les adaptations locales des vignobles au changement climatique.

International audience; In France, the agro-ecological systems are the basis for experimenting with alternative practices to industrial agriculture, renewing man to nature relations and considering ecological regulations to reduce the use of chemical inputs. Among these agro-ecological practices, the use of horses for mechanical weeding in viticulture has returned to the main French wine-growing regions. In addition to reducing the use of fossil fuels, the horse has a physical and emotional intelligence, which is useful to the winegrower. In fact, the communication between the winegrower and the horse requires developing the attention and observation skills necessary to an agro-ecological approach. It can take a long time to build this cross-comprehension between man and horse. That is why most winegrowers start this practice by using horse traction service providers. How does this observation and this understanding of the condition of the agro-ecosystem with the trio of horse, service provider and winegrower work? Can working with horse traction service providers be considered an agroecological practice? An exploratory study conducted in 2020 puts forward a description and understanding of the activities of the winegrower and the equine traction service provider by retracing the main stages of their cooperation in integrating the horse as an actor. This article presents the first results of this study. It focuses on the main characteristics of the work of the service provider with the horse on the vineyard estate, and on the influence of the climate context and the components of the agro-ecosystem on this work. The findings need to be developed further along two main axes: the role of the horse in the agro-ecosystem and the transformation of the activities of the service provider and the winegrower from year to year.

International audience; Highlights:• Spontaneous service plants strategies present a high spatial and temporal diversity.• Strategies can be measured by calculating the Coverage Index (CI) for each vineyard.• Higher CI was linked to quality labels, organic crop protection and low target yields.• No link was found between CI and water and soil resources in vineyards.• Target services by winegrowers are consistent with the CI and period coverage.Abstract: 'Service plants' include spontaneous vegetation or sown species of cover crops associated with perennial crops in the rows or inter-rows with a high potential to provide ecosystem functions and services. In vineyards, service plants target specific services depending on the management strategy implemented by the winegrower, including the plant species, the surface covered, the plant growth control and destruction date. Understanding the management strategies linked to their associated target services at the regional scale is necessary to better help winegrowers, advisers and policy makers regarding an adapted use of service plants. To do this, we conducted a survey in 2016 among 334 winegrowers in Languedoc-Roussillon region in France, enquiring about their service plant management practices during the season 2014-2015. Given the diversity of the strategies of service plant management, we proposed a typology analyzing their spatial and temporal dimensions. Further, we present a Coverage Index (CI), which combines both temporal and spatial dimensions of the service plant management strategies. We conducted a multiple components analysis and clustering to create a vineyard typology and applied linear models to find correlations between the CI and specific vineyard characteristics. Three quarters of interviewed winegrowers sowed or maintained service plants in their vineyards; 41 % used a winter service plant strategy; 8.4 % a semi-permanent and 27.3 % a permanent service plant strategy. The preferred surface coverage strategy was full surface during grapevine dormancy and its reduction to half of the inter-rows after grapevine budburst. However, the diversity of surface coverage strategies during the grapevine vegetative period was remarkable. Lower water resources and specific soil characteristics were not linked to the service plant management strategies. Higher CI was associated with vineyards presenting quality labels (PDO and Organic), independent wine-making and lower target yields, showing that the added value of producing high quality wine plays an important role when implementing service plants in vineyards. Overall, our study showed: i) the popularity of spontaneous service plant strategies; ii) the spatial and temporal diversity of service plant management strategies and iii) the utility of the CI to study the implementation of service plants and to understand the motivations and constraints of their use.

International audience; Highlights:• A participatory approach was used to study farm resilience after a shock.• Shocks differed according to farm type, but were all related to water.• Mental models were used with local actors to design future farming systems.• Performance was assessed according to eight complementary indicators.• According to local actors, all farm types would adapt, suggesting a good resilience.Abstract:The Mediterranean region is expected to be a hotspot for climate change, making the resilience of farming systems a major challenge. Some studies have used quantitative models at the farm scale to analyze the resilience of farming systems but with little involvement of stakeholders. We used a participatory approach with local actors on the Saïs plain of Morocco to design possible future states and qualitatively assess the resilience of typical farm types (FTs) experiencing major change. Our approach combined individual interviews of farmers and local actors, mainly public, with participatory collective meetings to identify representative FTs and their performance, project their evolutions and future performance in the face of change, and evaluate their resilience. Performance, defined according to literature, interviews and meetings, included different types of capital, income, yields, markets, support of public policies, and water access. Four FTs were considered: highly irrigated horticulture (FT1), rainfed cereals (FT2), partly irrigated cereal-legumes (FT3), and mostly irrigated fruit-tree/horticulture (FT4). The primary driver for FT2 and FT3 was climate change; for FT1 and FT4, it was limiting access to water resources. Participants designed more diversified systems for all FTs in relation to those changes. Rankings of FT performance did not change between current and future states. Performance did not evolve significantly, but FT4 was seen as the most resilient and FT2 the least. These qualitative results differ somewhat from other studies mobilizing quantitative approaches, but they highlight the potential of local adaptation to limit the impacts of global change on vulnerable agriculture.

CONTEXT During crises, adaptation or recovery measures or plans at local or national scales may not necessarily address longer-term or structural problems such as climate change mitigation. OBJECTIVE This article describes farmers and policymakers' responses to mitigate the adverse effects of Covid-19 on the agricultural sector. We then assess the responses' possible effects on greenhouse gas (GHG) emissions. METHODS The study is based on surveys conducted with farmers, traders, and extension staff in Burkina Faso, Colombia, and France, and literature. We used the Cool Farm Tool calculator to assess GHG emissions associated with fertilizer production, crop production and produce transportation to international markets for the three main cash crops in the three countries. RESULTS AND CONCLUSIONS We identified contrasting responses by the agricultural sector mostly driven by changes in the consumption patterns at local or international levels. We also identified contrasting state responses to mitigate Covid-19. These responses at farm and policy scales led to similar trends in decreasing carbon dioxide (CO2) emissions across the studied countries. However, none of the studied countries linked Covid-19 response measures to long-term climate change mitigation actions. Therefore, an opportunity to sustain Covid-19 induced short-term decreases in GHG emissions was overlooked. SIGNIFICANCE Analyzing the impacts that Covid-19 had on agricultural systems and the decision taken by policymakers to handle its direct and indirect effects can help society draw lessons on how to improve climate action.

International audience; The assessment of rice yield at territory level is important for strategic economic decisions. Assessing spatial and temporal yield variability at regional scale is difficult because of the numerous factors involved, including agricultural practices, phenological calendars, and environmental contexts. New remote sensing data acquired at decametric resolution (Sentinel missions) can provide information on this spatial variability. The study objective was thus to evaluate the potential of Sentinel-2 images for monitoring rice cropping systems and yield from farm to region scales. The approach considered both observations and modeling. In-depth farmers surveys were carried out in the Camargue region, Southeastern France. The novelty was to use operational tools (BVNET and PHENOTB) to compute leaf area index, to daily interpolate this biophysical variable from 44 images acquired in 2016 and 2017 for each rice field, and to derive key phenological parameters from the analysis of the temporal profiles. The STICS crop model was spatially used, considering the biophysical variables derived from remote sensing. We tested four simulation strategies, differing in the integration intensity of remote sensing information into the model. Results have shown that (1) Sentinel-2 data allowed distinguishing early and late rice varieties. (2) The phenological stages mapped at the regional level allowed to better understand the agricultural practices of farmers. (3) The assimilation of remote sensing data to the STICS crop model significantly improved yield estimation and provided useful information on the spatial variability observed at regional scale. It was the first time that Sentinel-2 data are used with STICS crop model to assess rice yield at both farm and regional scale in the Camargue area. The proposed method is based on free open data and free access model, easily reproducible in other environmental contexts.

International audience; The diversity of both pedoclimatic and socio-economic contexts in agricultural landscapes makes evaluating the impacts of climate change a challenge. Models are pertinent tools to quantitatively explore possible futures, and participatory approaches help account for local diversity. We developed a model through an approach that involves stakeholders in the model's construction and testing and in the discussion of results. We simulated spatially explicit impacts of climate change on water balance, grape phenology, and yield in a Mediterranean vineyard watershed for two climate scenarios. Results show a decrease in grape production of 7–14% by 2100. Yield decrease would be higher in irrigated high-yield areas despite a doubling of irrigation water supply. A forecasted 6 °C increase in temperature during berry ripening would threaten wine quality. The approach allows the communication of model results and limitations to stakeholders. It is a promising means of identifying potential local adaptations to climate change.

The frequent, widespread use of pesticides in agriculture adversely affects biodiversity, human health, and water quality. In 2008, the French government adopted an environmental policy plan, “Ecophyto 2018”, to halve pesticide use within 10 years. Trends in synthetic pesticide sales and use in France were described, through three different indicators: the number of unit doses (NUD), the quantity of active ingredient (QAI), and the treatment frequency index (TFI). Changes in pesticide use on seven of the principal arable crops in France since the implementation of this policy plan were analyzed, together with the impact of changes in pesticide use on water quality. No evidence was found for a decrease in pesticide sales at national level between 2008 and 2013. In terms of the TFI values for individual crops, the only decrease in pesticide use observed since 2001 was for soft wheat. This decrease was very slight, and pesticide use did not decline more rapidly after 2006 than before. Changes in pesticide use differed between French regions and crops. Water pollution did not decrease during the period studied. Possible explanations for the lack of effectiveness of the French environmental plan are considered in the context of European legislation.

International audience; The assessment of yield variability at the territory scale is often difficult due to the lack of knowledge on various factors involved, e.g., the agricultural practices of farmers and phenological calendars. Remote sensing can help to provide precise and timely information on crops particularly with the unprecedented amount of free Sentinel data within the Copernicus programme. This study focuses on the evaluation of the contribution of the new Sentinel 2 data to provide phenological information on rice cropping systems in the Camargue region in the South-Eastern France. Dense time series of data acquired at high spatial resolution (10m) were analyzed for 2016 and 2017 and were used to map rice, compute Leaf Area Index. Various methods combining remote sensing information were compared with two different crop models (STICS and SAFY) to assess the yields. The performances are discussed according to surveys made with farmers.

International audience; In the global change context, an efficient management of the available resources has become one of the most important topic particularly for sustainable crop development. Many questions concern the evolution of the rice farming systems in Camargue, in the Southeastern France, which play a crucial role to control the soil salinity and whose the surfaces significantly decreased these last years from 20 000 ha in 2010 to around 14000 ha in 2014. The arrival of the new satellites Sentinel makes it possible to evaluate these crop evolutions. The objectives of this study were to propose operational methodologies to: 1) accurately assess the surfaces of the main crops, rice, wheat and grassland in 2016 and 2017 from classifications based on multispectral data, (2) map some agricultural practices (sowing and harvest residue burning), and (3) elaborate a farm typology for the Camargue region based on variables computed from remote sensing data to better understand the farmer strategies.

International audience; Understanding the range of approaches available for assessing the impacts of agricultural changes at landscape scale is important when addressing local to global issues. Using a topic modelling approach, we reviewed the literature on impact modelling of agricultural landscapes. A search in Web of Science using the keywords model, agricultural systems and landscape yielded 1,975 hits, of which 514 papers met our selection criteria. The most salient terms fell within six groups: change, scale, pollution, biodiversity, practices and terms on biophysical/regulatory conditions. We identified four main topics: water quality, water quantity/energy crops, biodiversity and Integrated Assessment. Water management issues were more likely to be covered in North American researches, while issues related to Integrated Assessment were mainly covered in European studies. We found no relationship between topic and model type. We conclude that future integrated studies should consider the diversity of agricultural systems in governance of water and biodiversity issues.

International audience; This chapter is written by researchers at UMR Innovation (Montpellier) who propose an external critical analysis of the TATA-BOX project. Firstly, we highlight the main contributions of both the project and the book, which cover different stages of a participatory research project, by crossing several disciplinary viewpoints. We note multiple outputs that can strengthen the capacities of farmers and researchers. We then develop three questions that echo our own works on innovations and agroecological transition: (i) How to associate agroecological issues with the diversity of practices and projects in a given area which is not necessary in a transition towards "strong ecological modernization"?; (ii) What are the conditions for disseminating/outscaling the TATA-BOX approach? In particular, can the project be replicated without the support of researchers? (iii) How to best integrate the political dimensions of ecological transition at different scales? We conclude that TATA-BOX could become a "political object" that promotes agroecological transition to local authorities, national policy makers and media, and international networks.

International audience; The aim of this paper is to provide a better understanding of potentialities of the new Sentinel-1 radar images for mapping the different crops in the Camargue region in the South France. The originality relies on deep learning techniques. The analysis is carried out on multitemporal Sentinel-1 data over an area in Camargue, France. 50 Sentinel-1 images processed in order to produce an intensity radar data stack from May 2017 to September 2017. We revealed that even with classical machine learning approaches (K nearest neighbors, random forest, and support vector machine), good performance classification could be achieved with F-measure/Accuracy greater than 86 % and Kappa coefficient better than 0.82. We found that the results of the two deep recurrent neural network (RNN)-based classifiers clearly outperformed the classical approaches. Finally, our analyses of Camargue area results show that the same performance was obtained with two different RNN-based classifiers on the Rice class, which is the most dominant crop of this region, with a F-measure metric of 96 %. These results thus highlight that in the near future, these RNN-based techniques will play an important role in the analysis of remote sensing time series.

The research and improvement of methods to be used for crop monitoring are currently major challenges, especially for radar images due to their speckle noise nature. The European Space Agency’s (ESA) Sentinel-1 constellation provides synthetic aperture radar (SAR) images coverage with a 6-day revisit period at a high spatial resolution of pixel spacing of 20 m. Sentinel-1 data are considerably useful, as they provide valuable information of the vegetation cover. The objective of this work is to study the capabilities of multitemporal radar images for rice height and dry biomass retrievals using Sentinel-1 data. To do this, we train Sentinel-1 data against ground measurements with classical machine learning techniques (Multiple Linear Regression (MLR), Support Vector Regression (SVR) and Random Forest (RF)) to estimate rice height and dry biomass. The study is carried out on a multitemporal Sentinel-1 dataset acquired from May 2017 to September 2017 over the Camargue region, southern France. The ground in-situ measurements were made in the same period to collect rice height and dry biomass over 11 rice fields. The images were processed in order to produce a radar stack in C-band including dual-polarization VV (Vertical receive and Vertical transmit) and VH (Vertical receive and Horizontal transmit) data. We found that non-parametric methods (SVR and RF) had a better performance over the parametric MLR method for rice biophysical parameter retrievals. The accuracy of rice height estimation showed that rice height retrieval was strongly correlated to the in-situ rice height from dual-polarization, in which Random Forest yielded the best performance with correlation coefficient R 2 = 0.92 and the root mean square error (RMSE) 16% (7.9 cm). In addition, we demonstrated that the correlation of Sentinel-1 signal to the biomass was also very high in VH polarization with R 2 = 0.9 and RMSE = 18% (162 g·m − 2 ) (with Random Forest method). Such results indicate that the highly qualified Sentinel-1 radar data could be well exploited for rice biomass and height retrieval and they could be used for operational tasks.

National audience

Chapitre 10; International audience; The transformations currently underway in agriculture are throwing the spotlight on the studies and methods currently used to design innovative farming systems. This chapter analyses the specificity of five approaches for co-designing technical systems that have been tested in France and in some African and Latin American countries. They are based on a high level of interaction between actors involved in these approaches, and are facilitated by a range of intermediary objects such as modelling tools and agricultural experimentation in a rural environment. They have led to the production of operational and scientific knowledge on technical changes and the conditions under which they can be implemented at the farm level, as well as the institutional conditions conducive to the emergence of new systems. These approaches rely on skills that go beyond those pertaining to agronomy alone; the inclusion of social science researchers may become necessary, in particular to analyse how to hybridize various types of knowledge in order to accompany innovation on farms and in territories.

International audience; The development and improvement of methods to map agricultural land cover are currently major challenges, especially for radar images. This is due to the speckle noise nature of radar, leading to a less intensive use of radar rather than optical images. The European Space Agency Sentinel-1 constellation, which recently became operational, is a satellite system providing global coverage of Synthetic Aperture Radar (SAR) with a 6-days revisit period at a high spatial resolution of about 20 m. These data are valuable, as they provide spatial information on agricultural crops. The aim of this paper is to provide a better understanding of the capabilities of Sentinel-1 radar images for agricultural land cover mapping through the use of deep learning techniques. The analysis is carried out on multitemporal Sentinel-1 data over an area in Camargue, France. The data set was processed in order to produce an intensity radar data stack from May 2017 to September 2017. We improved this radar time series dataset by exploiting temporal filtering to reduce noise, while retaining as much as possible the fine structures present in the images. We revealed that even with classical machine learning approaches (K nearest neighbors, random forest, and support vector machines), good performance classification could be achieved with F-measure/Accuracy greater than 86% and Kappa coefficient better than 0.82. We found that the results of the two deep recurrent neural network (RNN)-based classifiers clearly outperformed the classical approaches. Finally, our analyses of the Camargue area results show that the same performance was obtained with two different RNN-based classifiers on the Rice class, which is the most dominant crop of this region, with a F-measure metric of 96%. These results thus highlight that in the near future these RNN-based techniques will play an important role in the analysis of remote sensing time series.This article belongs to the Special Issue High Resolution Image Time Series for Novel Agricultural Applications

Cixiid planthoppers are considered of major economic importance, as they can transmit phytoplasmas responsible for many plant diseases. While thorougly studied in vineyards, the epidemiology of stolbur phytoplasma, transmitted by Hyalesthes obsoletus Signoret, was rarely investigated on minor crops as lavender, where it leads to ‘yellow decline’ disease and large economic losses. The objective of this paper is to understand the effect of the local landscape characteristics on the presence and density of H. obsoletus in the ‘Plateau de Valensole’, southern France. Potential host plants of H. obsoletus were surveyed in three contrasted zones (in terms of crops and disease intensity), by uprooting plants and capturing adults in emergence traps. The localization and potential movements of H. obsoletus from the host plants towards lavandin (infertile hybrid of lavender) were determined using yellow sticky traps. Clary sage plants were found as major hosts of H. obsoletus. Flying insects were also caught in fields of lavandin, although emergence traps and plant uprooting did not confirm this crop as a winter host, i.e., as a reservoir for the insect. Based on one zone, we showed that attractiveness may depend on crop (clary sage or lavandin) and on its age, as well as on the distance to the supposed source field. These results suggest that clary sage could be an important host of H. obsoletus, whose density largely varies between zones. Genetic studies would be required to confirm the role of clary sage in the dissemination of yellow decline of lavandin.

Rice blast disease is a threat for European rice growers, who apply chemical treatments each year to limit its impact on rice yield and milling quality. Good agronomic practices such as varietal choice and reduced nitrogen fertilization can also be effective in limiting the impact of the disease, which largely varies across sites and growing seasons. Here we present a three-year experiment (2013–2015), in which blast disease severity was dynamically sampled on four varieties grown with two nitrogen doses (standard and double farmer fertilization) in three sites located in Northern Italy (i.e., the largest European rice district). No chemical treatments were applied on these experimental plots, which were compared to blast-treated controls. Field yield and yield after milling (t ha −1 ) were measured to assess the impact of rice blast. Disease progress curves of leaf and panicle blast were analyzed via F-test for site, nitrogen dose, rice variety, and year. The areas under disease progress curves were correlated with yield losses via linear regression. Finally, a 4-way analysis of variance was performed using field yield losses and head rice yield as dependent variables. Results Blast epidemics were significantly affected by all the factors considered, with rice variety and year as the most important sources of variability. Areas under disease progress curves were significantly correlated with losses in field yield and even more in yield after milling, with panicle blast proving to be the most impactful symptomatology. Year and variety ranked first and second among the factors explaining yield losses, both in field and after milling. These results confirm the effectiveness of varietal choice to reduce blast impact, indicating that fungicide applications should be conditional to the conduciveness of weather conditions.

Crop simulation models are commonly used to forecast the performance of cropping systems under different hypotheses of change. Their use on a regional scale is generally constrained, however, by a lack of information on the spatial and temporal variability of environment-related input variables (e.g., soil) and agricultural practices (e.g., sowing dates) that influence crop yields. Satellite remote sensing data can shed light on such variability by providing timely information on crop dynamics and conditions over large areas. This paper proposes a method for analyzing time series of MODIS satellite data in order to estimate the inter-annual variability of winter wheat sowing dates. A rule-based method was developed to automatically identify a reliable sample of winter wheat field time series, and to infer the corresponding sowing dates. The method was designed for a case study in the Camargue region (France), where winter wheat is characterized by vernalization, as in other temperate regions. The detection criteria were chosen on the grounds of agronomic expertise and by analyzing high-confidence time-series vegetation index profiles for winter wheat. This automatic method identified the target crop on more than 56% (four-year average) of the cultivated areas, with low commission errors (11%). It also captured the seasonal variability in sowing dates with errors of ±8 and ±16 days in 46% and 66% of cases, respectively. Extending the analysis to the years 2002–2012 showed that sowing in the Camargue was usually done on or around November 1st (±4 days). Comparing inter-annual sowing date variability with the main local agro-climatic drivers showed that the type of preceding crop and the weather conditions during the summer season before the wheat sowing had a prominent role in influencing winter wheat sowing dates.

Reducing pesticide use is a major challenge in agriculture, and farmers are encouraged to develop integrated practices. The aim of this study was to describe, analyse and assess the current contribution of various options to pest, disease and weed management on the reduction of fungicides, herbicides and insecticides by quantifying both their use and their effect on pesticide use intensity, within the context of different wine-growing regions. We conducted a statistical analysis of pesticide use and crop management over two seasons (2006 and 2010) for 11 French wine-growing regions (10,000 data). We used the Treatment Frequency Index (TFI) indicator to characterise the intensity of pesticide use. We analysed three crop management options: timing of first fungicide spray, type of soil cover cropping, and type of weed control. For each management option, we compared alternative or low-input practices to normal chemical-based practices, e.g., mechanical weed control (alternative option) vs. herbicide (normal chemical option). To strengthen the results, two quantitative variables completed the analysis: fungicide mean sprayed dose in the field and number of fungicides sprayed in the field. Results showed that each alternative or low input practice had an impact on TFI, but that this impact differed between wine-growing regions and between the management options under consideration. Regarding fungicides, our results showed that late timing of first fungicide spray (compared to regional reference) contributed to a reduction up to 50% of the part of the TFI due to fungicides, as compared to fields sprayed at median or early timing. Regarding herbicides, results of low-input/alternative practices were more homogeneous in terms of their relative impact on TFI reduction. The differences in use and impact of crop protection practices were more important between regions than between the two studied years. Our results may help policy makers to target methods for decreasing pesticide use, based on particularities of wine-growing regions and locally realistic practices.

International audience; European farmers are facing challenges that call for important transformations on their agricultural production systems, including an increasing number of regulations aimed at reducing environmental impacts from farming practices. Climate change is also expected to affect agricultural production in most European regions, and in Southern Europe this effect is expected to negatively impact yields. In this study, we present the application of an innovative participatory approach to assess the potential of innovative agricultural systems to reconcile environmental sustainability with economic viability while contributing to local and global food security. Our approach consisted of combining (1) the participation of local stakeholders in the design of narrative scenarios, and (2) an integrated assessment of scenarios through the calculation of indicators at different scale with a bio-economic model. We tested our approach with a case study situated in the Camargue region of Southern France. Rice is currently the main crop in this region, but farmers there face adverse economic conditions linked to the recent reform of European Common Agriculture Policy. After identifying the main drivers of change, local stakeholders developed narrative scenarios and described how farmers would adapt within the context of those changes. These elements were then translated into model inputs. At the regional level, the four scenarios led to variations in farmland acreage (28,000–33,000 ha), as well as the proportion of rice crops (19–75%) and areas cultivated under organic farming standards (8–43%). The four scenarios also led to different values for indicators of agricultural economic welfare, food production, and environmental impacts. Trade-offs between these indicators and the associated objectives assigned to agriculture were identified and discussed with the stakeholders. We end with a discussion of the limitations and advantages of our approach to the participatory development and assessment of locally developed narrative scenarios.

The future of global rice productions in top producing countries is undermined by the impact of climate change threatening food security in the near future. In those European Mediterranean areas where rice is cultivated, this peculiar cropping system plays a crucial role in terms of sociocultural and ecological issues, and the climate change impact is still scarcely investigated. In this study, we explored the future trends of potential rice yields in the region considering the multiple sources of uncertainty associated with climate and yield predictions. Two rice crop models (STICS and WARM) were calibrated using 20 field experiments carried out in two main European rice producing areas − i.e., the Italian Lomellina and the French Camargue. These models were then applied under a range of climate change scenarios in 2030 and 2070 time frames, considering projections from the combination of four General Circulation Models and two extreme Representative CO2 Concentration Pathways (RCP 2.6 and 8.5). We compared the simulated yield levels with no adaptation, and designed adaptation strategies based on the anticipation of sowing date and the adoption of varieties with longer crop cycle. Our results showed that with no adaptation yields would decrease on average by 8% in 2030 and 12% in 2070 in Camargue and Lomellina. Future simulated yields in the two areas were lower than in the baseline in 67% (Camargue) and 84% (Lomellina) of the cases. The implementation of both adaptation strategies proved to be effective in reversing the situation, leading to an average yield increase of 28% and 25% in 2030 and 2070, respectively. The associated probability of lower yields than in current conditions was 24% in the two sites. Despite the uncertainty in predictions, mainly related to site, GCM and RCP, our findings indicate that the European rice sector has the potential to enhance current production levels in a changing climate, if longer cycle varieties will be grown in Mediterranean rice areas.

International audience; This paper investigates the causes of imprecision of the observations and uncertainty of the authors who create Volunteer Geographic Information (VGI), i.e., georeferenced contents enerated by volunteers when participating in some citizen science project. Specifically, various aspects of imprecision and uncertainty of VGI are outlined and, to cope with them, a nowledge-based approach is suggested based on the creation and management of “contextualized VGI”. A case study example in agriculture is reported where contextualized VGI can be created bout in situ crops observations by the use of a smart app that supports volunteers by means of both an ontology and the representation of the context of the geo-localization. Furthermore, n approach to cope with both ill-defined knowledge and volunteer’s uncertainty or imprecise observations is defined based on a fuzzy ontology with uncertainty level-based approximate easoning. By representing uncertainty and imprecision of VGI, users, i.e., consumers, can exploit quality checking mechanisms to filter VGI based on their needs.

Organic and low-input systems are proposed as ways to reduce the environmental impacts of agriculture. Previous studies have shown that yields of organic systems can be ∼19 to 25% lower than conventional systems. An intermediary, low-input system could be less damaging for the environment than conventional systems, while reducing yield losses in comparison with organic systems. In this study, we performed a meta-analysis to compare low-input systems to conventional and organic systems. Our analysis is based on data of cropping system experiments conducted in Europe and North America, and focuses on two important crops, maize (Zea mays L.) and soft winter wheat (Triticum aestivum L.). Pesticide use was greatly reduced for low-input systems as compared with conventional for the two crops (50% for maize, 70% for wheat on average). Mean mineral N use was also reduced by 36% for maize and 28% for wheat in low-input relative to conventional. Maize yields in low-input systems were not different from those in conventional systems, and were higher than yields in organic systems (yield ratio of low-input vs. organic = 1.24). Wheat yields in low-input systems were lower than yields in conventional systems (yield ratio of low-input vs. conventional = 0.88), but were substantially higher than yields in organic systems (yield ratio of low input vs. organic = 1.43). This is one of the first meta-analyses to assess performance in terms of pesticide use intensity, and yields, with clear evidence emerging that low-input systems can markedly reduce pesticide application, without strongly reducing crop yields, relative to conventional systems.

Cultivar resistance plays a major role in current disease management strategies, but its efficacy is usually short-lived unless resistance deployment strategies to ensure resistance durability can be designed. Using a spatially explicit model, we evaluated cropping system mosaics that were designed by stakeholders involved in field agronomic practices to manage phoma stem canker of winter oilseed rape. We simulated pathogen population adaptation to a newly introduced major resistance gene (RlmX) to estimate the durability of the resistance under various scenarios of cropping system mosaics within a small region. Our objective was first to find descriptors of agricultural landscape that are relevant for resistance management and then to study the relationship between cropping practices applied in nearby fields and the genetic structure of the pathogen population in fields cropped with RlmX-cultivars. Key cropping practices were characterized with different metrics for several buffer sizes (100–2000 m) around target fields; and these indicators were used in linear models to predict pathogen evolution. Indicators describing local cultivar composition were very informative; adding information on tillage, but not nitrogen fertilization or fungicide treatment, could marginally increase the goodness of fit. The effects of cropping practices on resistance durability could be shown when the landscape was characterized within 500 m around RlmX-fields. We conclude that, in order to study and ultimately design landscapes promoting resistance durability against phoma stem canker, it is sufficient to take into account a relatively small portion of the landscape around RlmX-cultivars, focusing on cultivar choice and tillage practices of RlmX cultivated fields.

In agricultural landscapes, pest and disease control mainly rely on cropping system characteristics and location. We have combined a participatory approach and a spatially-explicit model, to design and assess cropping system scenarios for future contextual changes in the small region of Picardie, France. Phoma stem canker of winter oilseed rape (WOSR), a widespread disease responsible for major economic losses, was used as a case study to investigate the effects of modifying cropping systems, at a regional scale, on disease controlling and on the sustainability of a newly introduced specific resistance gene (RlmX). Meta-models were fitted to assess the effects of cropping practices and weather conditions included in scenarios on three complementary model outputs: size of the pathogen population, yield loss, and fraction of the virulent population on RlmX-cultivars. We ran three replicates of each cropping practice scenario, varying the location of WOSR and associated cropping characteristics within the region. Outputs differed slightly between replicates, but there were no significant differences between replicates (alpha = 0.001) for each output. The size of the pathogen population was well explained by winter oilseed rape acreages, cultivar landscape composition and tillage, and the yield loss was explained by weather conditions and cultivars. The fraction of the virulent population on RlmX-cultivars was explained primarily by cultivar landscape composition. These differences in explanatory variables for the different outputs highlighted their complementarity. Cropping practices and weather impacts on the three variables remained consistent among the explored possible future contexts. Highlighting the most efficient cropping practices to be applied in the event of different future changes could help local decision-makers to design cropping practices in the face of contextual change.

Pesticides pose serious threats to both human health and the environment. In Europe, farmers are encouraged to reduce their use, and in France a recent environmental policy fixed a target of halving the pesticide use by 2018. Organic and integrated cropping systems have been proposed as possible solutions for reducing pesticide use, but the effect of reducing pesticide use on crop yield remains unclear. Here we use a set of cropping system experiments to quantify the yield losses resulting from a reduction of pesticide use for winter wheat in France. Our estimated yield losses resulting from a 50% reduction in pesticide use ranged from 5 to 13% of the yield obtained with the current pesticide use. At the scale of the whole country, these losses would decrease the French wheat production by about 2 to 3 millions of tons, which represent about 15% of the French wheat export.

International audience; Farmers increasingly need to adjust their management practices to accommodate unexpected events such as drought and preserve the longevity of their production project. Applied to grassland use e.g. grazing or cutting, such an exploitation of flexibility requires some background knowledge about where and when freedom of choice can be exercised. To this end, we design an analysis framework for grassland-based farming systems in mountainous or less-favored areas. An expert-based discrimination tree characterizes organizational flexibility by determining the range of possible types of grassland use under various topographic and farming constraints such as suitability for mechanization and ease of access to a field. A set of time windows evaluates the timing flexibility in grassland use, each associated with a combination of a grassland community type and a type of grassland use. Compared with field data obtained in the French Aubrac region, the outputs of the discrimination tree match for 139 of 165 grassland fields. For a particular type of grassland use, the set of time windows proves that timing flexibility in grassland use between grassland community types can increase by 15 days over a 37 days time range. When applying the two components of the analysis framework to a farm case, it shows that 24% of the farm area offers organizational flexibility, with several possibilities for grassland use. Timing flexibility for bringing forward or delaying the use of the grassland fields is unused in the farm. Most of the dates of grassland first use are similar irrespective of the diversity of grassland communities. The application of the analysis framework offers a sound evaluation of the potential flexibility to establish where and when it is possible to adjust management practices to cope with unexpected events. It can also be helpful in setting up coherent alternatives to the observed management strategies that can then be expanded in dynamic simulation models enabling deeper analysis.

Chapitre 10; National audience; Les mutations en cours au sein de l’agriculture interrogent les travaux et les méthodes relatifs à la conception de systèmes agricoles innovants. Ce chapitre analyse la spécificité de cinq démarches de co-conception de systèmes techniques testées en France et dans différents pays d’Afrique et d’Amérique latine. Elles se basent sur des interactions fortes entre les acteurs impliqués dans ces démarches, facilitées par une diversité d’objets intermédiaires tels que la modélisation ou l’expérimentation agronomique en milieu paysan. Elles ont permis de produire des connaissances opérationnelles et scientifiques sur des changements techniques et leurs conditions de mise en oeuvre à l’échelle de l’exploitation ainsi que sur les conditions institutionnelles favorables à l’émergence de nouveaux systèmes. Ces démarches mobilisent des compétences ne relevant pas seulement de l’agronomie. L’intégration de chercheurs relevant des sciences humaines s’avère centrale, en particulier pour analyser comment hybrider des connaissances multiples en vue d’accompagner l’innovation au sein des exploitations et des territoires.

[Departement_IRSTEA]Territoires [TR1_IRSTEA]SEDYVIN [ADD1_IRSTEA]Adaptation des territoires au changement globalCahier thématique; National audience; Ce deuxième cahier diffuse des connaissances scientifiques (non exhaustives) sur les effets du changement climatique sur l’agriculture et la forêt en Provence-Alpes-Côte d’Azur. Les cultures (vergers, céréales, riz, maraîchage…) et les forêts (feuillus, conifères) bénéficient des bienfaits du climat méditerranéen qui offre des conditions favorables au développement des plantes sous certaines conditions, mais souffrent aussi des événements extrêmes qui le ponctuent à intervalles irréguliers. Le climat méditerranéen qui sévit en région PACA pourrait se résumer par Toulourenc (« tout ou rien » en provençal), du nom du cours d’eau à caractère torrentiel qui coule dans la vallée étroite située au pied du versant nord du mont Ventoux. Avec le changement climatique actuel, les aspects négatifs de notre climat sont appelés à se renforcer et font déjà peser sur les terroirs agricoles et les forêts emblématiques de la région de nouvelles contraintes auxquelles il est nécessaire de faire face pour éviter des conséquences trop néfastes. Mais la pérennité et le développement des systèmes agricoles et forestiers ne dépendent pas seulement de l’évolution du climat. L’urbanisation, l’occupation des sols, les pollutions locales (sol, air, eau), les incendies, mais aussi les pratiques culturales et la gestion forestière, jouent un rôle fondamental. Il convient donc de privilégier une approche transversale. Cette publication souligne les conséquences du changement climatique sur l’agriculture et la forêt en prenant soin d’identifier les enjeux environnementaux, économiques et sociaux, les risques à l’échelle régionale et locale, mais aussi les solutions susceptibles de réduire les impacts (atténuation, adaptation) et les éventuelles opportunités à saisir. Comme dans le précédent cahier, la contribution des chercheurs et experts, exerçant leur métier en région PACA et dans les territoires limitrophes4, sous forme d’articles et de zooms, apporte des éléments de compréhension afin de mieux cerner les problématiques liées au changement climatique.