ABSTRACT. Rural areas are of particular importance with respect to the agro-food sector and should be specifically addressed within this scope. Open datasets related to agriculture and rural regions, as well as data publication and data linking of external data sources contributed by different public and private stakeholders, allow to provide specific and high-value applications and services for the support in the planning and decision-making processes of different stakeholders groups, related to the agricultural, forestry, rural development and environmental domains. The presentation will focus on the definition, discovery, and utilization of Open Data in different rural applications.

ABSTRACT. Nowadays we have large amounts of information at our hands for the management of an agriculture business. The so-called data driven agriculture, with the use of Big Data and Data Science technologies, can lead to huge improvements in productivity. It is also a big opportunity for local IT companies and startups. But there are many challenges for the development of a technological product or service in this field: the data is diverse, exposed by different institutions and in different formats (many of them proprietary). Also the amount of data involved is very large, requiring specific tools for processing and analysis.

Palenque is a Big Data platform for Agriculture. With Palenque we intend to attract national IT companies and startups to the sector of digital farming and data-driven agriculture and to boost the adoption of this technology among farmers. The main goals are: 1) offer a common technical framework, with tools and computing infrastructure, for the development of data-intensive applications. 2) generate an ecosystem of technology suppliers to facilitate the adoption of technology and the development of innovative services. 3) facilitate the information sharing among different actors (farmers, researchers, suppliers, public agencies) through data standards and protocols.

ABSTRACT. In this paper, we describe an online software platform, which combines geophysical information from various diverse sources, together with big data analysis, in order to estimate the impact of the agricultural sector on the environment, considering land, water, biodiversity and natural areas such as forests and wetlands. Based on the P-Sphere project, this platform promotes a more sustainable agricultural sector by designing and developing an information and knowledge-based software tool, using a big-data approach for managing and analyzing a wide range of geospatial and mainstream information from the context of Catalunya (Spain), which can be accessible by standard communication technologies such as the internet/web and mobile apps.

ABSTRACT. In this paper, an algorithm for the generation of interpretable maps yielding management zones in agriculture is presented. In the literature, there are two main types of methods aiming at generating zones from spatial data: i) data classification based on a priori classes ii) data segmentation inspired from image analysis. Both have limitations: difficulty to take into account the spatial contiguity in the first case, or sensitivity to seed location in the second one. The proposed approach avoids these drawbacks, by defining a global zoning quality criterion that is used to guide an optimization procedure based on quantile values. It avoids the definition of a priori classes, while assigning interpretable labels to the generated zones. The algorithm provides effective visualization to help users analyze and reason about data. The algorithm is applied to real field data from precision agriculture.

ABSTRACT. Informed decisions on soil fertility management should be based on up-to-date soil analytical results, but this is not always the case. Reasons can include: i) soil analytical services become expensive when applied regularly to cover the full farm variability, ii) the analytical results can take time to be delivered, and iii) the link between analytical results and management action can be complex. In addition, the information needs to be presented in an appropriate format (e.g. a farm management system can differ within and between countries), which hinders development of general-purpose information and communications technology (ICT) solutions. This submission describes how the two services Lab-in-a-box and Soil Scanner tackle the described obstacles. In addition the flexibility of the fully automated crop and region specific soil advisory service system is outlined.

ABSTRACT. In epidemiological monitoring networks, many data are routinely collected in order to establish a phytosanitary diagnostics and report on its evolution. In France, This system was restructured in 2009 and these data are valued in the weekly editions and in the final campaign reports. Nevertheless, the actors of the system lack time with less than 16 hours between the collection of data in the fields and the release of the bulletin. They also lack tools to automate data analysis, including comparisons with prior years or data-based predictions. These data seem largely under-valued to establish the risk analysis and its evolution in the short term. The first objective of this work is to develop an operational statistical approach to analyze epidemiological dynamics based on the field observations collected. The second objective is to make this analysis procedure included in a dashboard available in an online tool that mobilizes all available data.

ABSTRACT. Nitrogen (N) is indispensable to maintain agricultural productivity, and at the same time, it is necessary to reduce its potential environmental impacts. Therefore, it is important to apply N fertilizer efficiently. Site-specific (SS) N application can increase N use efficiency (Raun et al 2002). Heterogeneity of soil fertility on a field determines to what extend that efficiency can be improved. Besides agro-environmental indicators, economic efficiency also needs to be considered. Profitability of variable rate application (VRA) for SS N fertilization is questionable (Bullock et al 2002). When yield response to N does not strongly vary within a field, investment in VRA technologies may not be paid off (Anselin et al 2004; Liu et al 2006). SS N application does not necessarily show advantages in terms of risk aversion over uniform management, when spatial and temporal variability is taken into account (Whelan & McBratney 2000). However, this situation may change, if an overall restriction for N supply is introduced. This paper investigates how the economic potential of SS application for N fertilizer changes when such a restriction is applied. This may help to determine, if profitability can be improved by VRA technologies in an N restriction situation.

ABSTRACT. The decline in industrial Sulfur (S)-containing emissions (protocols of Helsinki, 1985; Oslo, 1994; Kyoto, 1997) has led to an S depletion in soil resulting in yield losses and crop quality degradation. This highlights the importance to monitor S fertilization, especially in rapseed (Brassica napus L.) a high S-demanding crop. SuMoToRI (Sulfur Model Towards Rapeseed Improvement, Brunel-Muguet et al. 2015a) is a process-based model that predicts rapeseed growth, S allocation (between plant parts) and S partitioning (organic vs. mineral) until the onset of pod formation in relation to S availability, temperature and radiation. Recent work showed evidence for interactions effects between S availability and increased temperature during the grain filling period in rapeseed (Brunel-Muguet et al. 2015b). In the context of climate change, we aim to identify the effects of warmer temperatures on rapeseed growth. We performed simulations with climatic datasets from the last 20 years and from projections with the four RCP (Representative Concentration Pathway) scenarios (Inman, 2011), in different locations and under S limiting and non-limiting condition. First simulations (two contrasting locations, projected data from RCP4.5 scenario) showed that increased temperatures impacted the timing of phase initiation and duration and as a result earlier S demand from the leaves..

ABSTRACT. With the increasing scarcity of natural ressources, the unsustainability of the conventional agriculture and the need of food security, agronomy engineering is facing a serious challenge. Different projects aim to design innovative cropping systems and simulation is used to test the robustness of field experiments results. While there are already crop models that simulate correctly crop development, there are few models that simulate the way farmer conduct their cropping systems. Yet, it is a key challenge because innovative cropping systems are often based on practices which take into account the state of the environment. Using “fixed calendar dates” for the activities is no more relevant. The modeling and simulation platform RECORD allows creating such models. A decision modeling framework is proposed, a graphical plugin has been developed to help modelers to sketch and implement their decision models and to link them with biophysical models. The sequence of technical operations is formally modeled as a graph p. An example will be presented to illustrate how farmers’ decision-making is formalized , how the coupling is done between the crop model and the decision model , how crop rotations are simulated, and how the results are used for bio-economic evaluation.

ABSTRACT. In each modeling process the question of the available data to build a model is crucial. Classically, when a researcher wants to model a biological process, it is necessary to go through a data acquisition step through experimentation which often require several years. It is now possible to access new sources of data acquired on a purpose far away from the perspective of modeling matter. This is the case of the territory bio-aggressors monitoring data of cereals crops used in the weekly edition of the crop health bulletins. Therefore, a large volume of observation data is generated each week and its valorization is currently limited to the edition of the bulletin. We then reflected upon the value of those data in the risk assessment of straw cereals during the cropping season. What is the probability of disease occurrence this year? When will it appear? Or how will it evolve? The development of software such as R greatly facilitates the manipulation of massive data and associated statistical analyzes. Since 2014, the ARVALIS modeling and statistics team has been developing methods adapted to those data with the aim of offering more reliant indicators to assist farmers on the necessity to spray.

ABSTRACT. Globalisation, environmental and climate changes multiply the risks of emergence and re-emergence of diseases or animal pests on crops. To control and prevent epidemics, the role of epidemiological surveillance systems is becoming more and more important. In France, the Ecophyto national action plan to reduce pesticide use enhanced the national Epidemio Surveillance Network (ESN). How to design and optimise such a network is still a research subject. The study presented here uses a qualitative and simple model based on Dynamic Bayesian Networks to explore the influence of spatial size (number of observed sites) and temporal size (number of years of past observations) of the network on the number of pesticide treatments applied, the severity of epidemics and the expected net margin.

ABSTRACT. CHN is a mechanistic crop model developed by ARVALIS - Institut du végétal which first purpose is to be used during the agricultural season as a decision-making tool. Crop growth and its limiting factors can be then estimated with CHN, in real time and in projection until the end of crop season thanks to frequency calculations. Model outputs then allows farmers to manage in the same time nitrogen fertilization and irrigation. CHN crop model has been for now implemented on bread wheat, durum wheat and maize. Once parameter optimization of the different modules of CHN has been made, the model is able to predict dry matter production and nitrogen quantity absorbed by the plant. CHN is considered enough precise and accurate to provide first agronomic valorizations of model outputs.

By using CHN, one can envision supporting farmers by giving them decision support thanks to estimations of cereal growth in the field and into the soil and climate conditions of the cultivation season, taking into account carbon, water and nitrogen fluxes and also the effect of hydric and nitrogenous stresses.

ABSTRACT. Improving agricultural markets in DC providing a better access to information is the main objective of Market Information Systems. Most MIS adopted mobile phones but their effective use by farmers remains marginal. What are the constraints of adoption? Based on surveys to recipients, we analyze the adequacy of the use of SMS to disseminate information to smallholder farmers. In Madagascar, two main MIS recently introduced the use of mobile phone. After a few months, feedbacks from the targeted producers were collected. Results show that mobile phone penetration is uneven (46 to 75% of households, depending on the areas) and only 10% farmers use it to market their products. Radio is used for entertainment. The level of understanding of the SMS ranks from 39% in the remote area to 86% around the capital. A large majority of farmers see these SMS as an improvement of their general knowledge; fewer declare that they will use them for them for marketing. More than half of them are willing to pay to receive these SMS. However, the main constraints are: (i) rapid “loss” of the recipients (change of number, loss of phone…), (ii) technical constraints (difficulties to recharge the battery, phone network coverage…).

ABSTRACT. The RESOURCE project was developed to improve the knowledge on migratory waterbirds and their utilisations for their better management in sahelian countries. Thus, it also focus on the socio-economic impact of the utilisation of waterbirds through tourism and hunting: What species? How many? How? When? Why (benefits)? And also, what are the actors perceptions and knowledge of wetlands, waterbirds, migration, legislation and hunting? After a two steps co-building questionnaire, we designed a socio-economic, socio-cultural and cognitive survey protocol and use the KoBoCollect as a data co-production tool with the actors of the Lakes Chad and Fitri and the Nile Delta. The first Chadian data is currently on the KoBoToolbox site and Egyptian data collecting will begin in april 2017. So, we will present the work carried out during this first data collection campaign from 15 February to 15 May 2017. If the use of KoBoCollect already shows benefits in terms of access and digital creation, remote access and data sharing, speed of pre-analysis and conduct of investigation, it also question the relationship between time saved on direct digitization of data and time lost in the collection of qualitative and literal data with a digital interface.

ABSTRACT. The current wave of agricultural development is raising many questions about the direction of this development process. The authors argue that in order to understand the real driving forces behind the transformation of agriculture and make valid statements of the future direction of it, one must look at the facts through the lens of history and social macro-evolution. Agriculture (as other sectors like health and education) integrated industrial and information elements from the very start of the industrial revolution, and these elements shaped the sector together. Agriculture can be seen as an underside of industrial and information society. Current trends are indicating the coming of a new control structure based on networks, connectivity and mutual partnership, which manifests itself in new forms of resource distribution, decision making and collaboration. To name this new control structure, the most appropriate choice is ‘isocratic control’. The study aims to demonstrate the evolution of different control mechanism in agriculture and calls for a wider framework to understand agricultural development in the information society.

ABSTRACT. ICT supports farmers by facilitating access to markets through real-time data on market prices, weather forecasts, and information on planting techniques. A major impediment for smallholder farmers to fully exploit the benefits of ICT is the lack of sufficient infrastructure in rural areas. On one hand, carefully designed government policies and regulations can promote greater ICT penetration and the provision of ICT services by private sector. On the other hand, cumbersome regulation of the ICT sector can hinder competition and inhibit the creation of innovative solutions that are responsive to users’ needs. This paper employs a new data set on ICT regulations across 62 countries around the world to explore the impact of sectoral policies on rural and agricultural development. The results suggest that the higher quality of ICT regulatory framework is associated with higher mobile internet market penetration and better developed ICT infrastructure. More specifically, the type of licensing regime and efficiency of spectrum allocation can play important parts in encouraging the private sector to invest and rollout mobile networks in rural areas. As evidenced by the European Union countries greater liberalization of the ICT sector supports ubiquitous connectivity.

ABSTRACT. Oilseed rape is the third oilseed crop (FAO, 2016) mainly grown for edible oil production, meals used for cattle feeding and more recently Biodiesel®. Like most of the Brassicaceae species, rapeseed is a high S-demanding crop which raises the importance to tightly monitor S fertilization, especially in a context of soil S oligotrophy mainly due to reductions in industrial SO2 emissions (protocols of Helsinki, 1985; Oslo, 1994; Kyoto, 1997). This can be supported through a modelling approach. SuMoToRI (Sulfur Model Towards Rapeseed Improvement, Brunel-Muguet et al. 2015) is a process-based model that predicts rapeseed growth, S allocation (between plant parts) and S partitioning (organic vs. mineral) until the onset of pod formation in relation to S availability, temperature and radiation. In this study, we perform global sensitivity analyses (SA) under contrasting S supply conditions (i.e. amount, date of inputs and fractioning) by focusing on three main parameters i.e. RUE (Radiation Use Efficiency), the SLA (Surface Leaves Area) and Beta (Carbon leaf allocation coefficient). The final objectives to rank the parameters according to their impacts on plant performances under contrasting S supply strategies and to identify the most suitable plant parameter value combinations under these contrasting S supply strategies.

ABSTRACT. The dynamic crop model AZODYN-rapeseed simulates crop growth yields and seed quality under the main soil and climate conditions in France. AZODYN-rapessed estimates the above-ground biomass production from climate data and soil characteristics. It works on a daily time step and describes plant phenology, leaf expansion and biomass production and its allocation to the grain. We tested the model in various situations of soil (deep, median or superficial), and of climate (Western or Eastern region). The model outputs was compared with measurements for biomass, plant nitrogen content and soil nitrogen content in different regions of France from 2004 to 2014. The first evaluation assessed model robustness. The model will help to better understand the link between crop development and fertilization practices and provide new scientific knowledge on rapeseed development.

We plan to quantify the variability of crop performance in response to the main abiotic stresses occurring in the field (light, temperature, water, nitrogen). We compared the abiotic stress level in experiments located in the differents parts of France Several cases will be presented in this communication. The model can be used in order to characterize the nitrogen stress level (NNI) of a field instead of measurements.

ABSTRACT. Grass growth is difficult to model as it depends on several factors which are grass management, vegetation and soil and weather conditions. As the development of emerging technologies facilitates the continuous acquisition of information about soil, weather and vegetation, it offers new perspectives for modelling. The objective of the study is to explore possible correlations between growth measurements and data from sensors, that could be used to build more accurate and user-friendly grass growth models. A randomized complete block design was used in 2016 in a French experimental farm to follow the growth of a ryegrass sward from every two weeks measurements during 3 regrowth cycles. The effects of fertilization, date of first grass cut and age of regrowth were accounted for. Sensors developed by the Parrot industry and located in the experimental design allow recording data about soil and climate conditions. Data-linkage between date-adjusted means of grass growth and sensor’s information is promising. Sensors data reflect the suitability of weather and soil conditions for grass growth. However, the pore water electrical conductivity recorded by sensors does not seem to detect fertilisation events. Further analysis of data will allow proposing growth models directly fed with sensors’ information.

ABSTRACT. The production systems are nowadays dependent on sanitary coverage based on the use of pesticides. With the society's evolution towards sustainable protection for both environment and health, the development of an innovative cropping system becomes an important issue. Many programs have been developed to try to reduce the use of these phytosanitary inputs. Among the most decisive, the Grenelle Environment initiated a turning in policies by adopting commitment 129, within 10 years. However, the constraints linked to the production requirement and disease pressure, make it difficult to put in place a reasoned pest management without accompaniment. Promété has an agro-ecological approach, with development of Decision Support System in agriculture. The aim of decision support is to guide farmers by giving answers to questions that arise during the decision-making process: “Do I need to treat?”. We have developed an approach based on integration of climatic, pathological and phenological data, at plot level. In 2016, trials carried out in French vineyards, have shown a reduction up to 46% in the Treatment Frequency Index compared to a conventional or organic crop production. These results show the importance of a modeling strategy at the plot level, for the rational use of phytosanitary products.

ABSTRACT. Dealing with Big data is a great opportunity for agriculture. This presentation will show that this is also the opportunity for the development of interdisciplinarity. There is a need for semantics to structure Big Data and make it meaningful for larger communities. The more able to group and integrate data we are, the more we give them value. This requires the use of standardized metadata that are machine readable and the use of formalized semantics. Structuring Big data will be relevant provided that we develop Big Data analytics.

ABSTRACT. Strawberries are one of the most valuable crops in Florida. The state produces about 16 million flats of strawberries every year, which represents 15% of nation’s berries and virtually all the berries grown during the winter. Anthracnose fruit rot and Botrytis fruit rot are the most important diseases for production of annual strawberries in Central Florida and worldwide. The Strawberry Advisory System (SAS) is a disease alert system available on AgroClimate.org that warns growers about the risk of disease incidence and recommend best options for control based on risk level and past applications. SAS utilizes 6 weather stations belonging to the Florida Automated Weather Network (FAWN) to calculate infection risk levels based on air temperature and leaf wetness duration observations. SAS has been quite successful and a mobile phone app has also been developed to facilitate growers’ access to the information provided. More recently we decided to increase the system usefulness by developing and integrating customized monitoring stations to the system using wireless sensor networks. Our main goal is to understand the spatial variability of the disease risk in a field using Wireless Sensor Networks (WSN) and investigate the potential for precision application of fungicides in Florida strawberry farms.

ABSTRACT. The distance between research findings and the Danish farmers is short. Among other things is due to efficient communication of the findings via the portal “landmand.dk” (farmer.dk). More than 20,000 farmers use this portal, corresponding to 83 % of all Danish farmers! The reason for this high adoption rate is the heavy individualisation of the contents which focus on each individual farm.

This article describes both the basic vision and the actual setup – in such a way that it is transferable to similar solutions in other countries.

ABSTRACT. Digital technologies applied at the farm level will certainly change the ways in which agricultural and food commodities are produced and marketed. They allow, among other things, to fulfill the requirement of triple economic, productive and environmental performance and to bring producers closer to consumers. Combined with this, new forms of business are emerging. The agricultural sector is starting to see a large number of start-ups with the same ambition and approach than NATUs companies. The API-AGRO project, initially a collaborative research project financed by the French Ministry of Agriculture, was transformed into a start-up in early 2017. The value proposition of the young company is to make data meet with uses by widening its initial ecosystem with a large number of actors of French agriculture. The start-up has chosen to segment its business model as much on the uses as on the transactions of data flows while remaining in a dynamics of openness. Its API-centric model allows, thanks to the technical platform, to propose a support to create ecosystems of partnerships between companies. The platform connect and cross data to create innovative applications because powered directly by multi-sided information systems.

ABSTRACT. Targeted Information and Communication Technology (ICT) tools can support the application of ecological principles in agriculture. However, many ICT tools released in the last decades for the agricultural sector have not been taken up in practice. Here we present the co-design of a soil health monitoring tool with two organic farmers communities. This work aims (i) to study the value of participatory tool development in avoiding implementation lock-in; (ii) to develop a tool that integrates modelling of experimental data together with self-assessments by the farmers. Each stage of the co-design is characterized by an interaction among farmers, agronomists and programmers. The participatory methodology used in in this work includes a series of planned interaction steps and feedback loops. The tool includes two functions: (1) soil health self-assessment for harvesting and sharing data of on-field spade test and (2) soil organic matter dynamics for forecasting the effect of fertilization practices on soil organic matter dynamics. The application is composed of four main components: farm data, soil health self-assessment data, soil model (based on Rothamsted carbon model), open data on soil and weather conditions. We are currently at the tool development phase, testing each part of the tool with the target communities.

ABSTRACT. In a first part we describe the ICT equipment of French farmers with special emphasis on mobile tools since we observed ten years ago that nearly farmers having a significant economic impact were already equipped with PCs. In a second part we try to describe the use of social networks by farmers. We observe the deep interest of farmers for the use of video to present their farm or the operations that they are performing on a daily basis. These videos distributed through the social networks may help a better recognition of role and works of farmers better than long talks and texts.

ABSTRACT. attached

ABSTRACT. Whiteflies are major pests on a wide variety of crops all around the world. Although it has been reported that whiteflies communicate acoustically, the relationship between acoustic signals and behavior of whiteflies has not been clarified. In this paper, a simultaneous video and sound acquisition system for detecting and monitoring whiteflies is constructed, and its performance is evaluated by experiments. Whiteflies of Bemisia tabaci biotype B on a cucumber leaf in a small clear case are lighted using LED to obtain video images by camera. High-sensitive microphone is located close to cucumber leaf and sound of whiteflies are recorded. From the experiments, we found two types of sound, and by comparing sound and image, one of them is found to be flying sound. Hence, this proposed system can be a useful tool for detecting and monitoring whiteflies.

ABSTRACT. Monitoring changes in behaviours could provide insight into the reproduction status, health, and overall well-being of dairy cows. Traditional methods based on direct observation of the herd, either live or from video recordings, are becoming increasingly labour-intensive and time-consuming as herd size increases. Thus, automatic behaviour recognition systems using accelerometers in combination with machine learning algorithms become more important to continuously and accurately quantify cows’ behaviours. The aim of this study is to propose methods for classifying three behaviours (lying, standing, and feeding) of indoor dairy cows using leg-mounted accelerometers. Lying, standing, and feeding behaviours of 16 lactating dairy cows were logged for 6 hours with 3D-accelerometers attached to the right hind leg of the cows. The behaviours were simultaneously recorded using visual observation (live and backed-up with video-recordings) as reference. Different features were extracted from the logged raw data and classification algorithms (K-nearest neighbours, naïve Bayes, and support vector machine) were used to classify the cows’ behaviours. The models allow excellent classification of the lying behaviour (precision 99%, sensitivity 98%), followed by feeding (precision 81%, sensitivity 86%). Standing was the most difficult behaviour to classify with a maximum precision and sensitivity of 68% and 76%, respectively.

ABSTRACT. In this paper we address the problem of food resources allocation in pastoral based dairy systems. The food resource allocation for a dairy herd consists in determining how to distribute the available resources considering different objectives. In Uruguay, the resource allocation is done by dividing the herd into feeding groups considering the characteristics of the cows, and then distributing those groups in different feeding areas. This type of allocation can be addressed by doing a combinatorial optimization model. Considering the difficulty of large-scale optimization problem, where traditional exact approaches cannot be applied, a very good alternative are metaheuristics. In this context, metaheuristics have been used to obtain good quality approximate solutions in a reasonable execution time. In this work, the main goal was to determine how to allocate the available resources in order to satisfy multiple objectives, considering a multi-period approach, and integrating operational decisions using more detailed models. We developed an optimization multi-objective evolutionary algorithm to generate an approximation of the Pareto front, taking as decision variables a sub-set of input parameters, and using different objective functions. The model developed considers the different time scales and allows users to graphically explore the properties of the solutions.

ABSTRACT. Pigs are easily affected by the surrounding environment. So when managing a swine farm, it is important to provide an appropriate environment for pigs. Current swine farms in Korea have difficulty affording an appropriate environment for pigs because of Korea is a country with a wide range of temperature and many swine farms have an aging infrastructure. This study introduces the new Internet of Things(IoT)-based swine farm environment management system that addresses the problems in existing swine farm environment management system. The new IoT-based swine farm environment management system collects swine farm’s real-time data from sensors and makes a decision making algorithm. Then, swine farm’s facilities are controlled automatically by the system for making proper environment depend on farm’s circumstances. From this system, farmers can empirically evaluate real time information and adjust each facet of the swine farm accordingly. The new IoT-based swine farm environment management system is expected to maximize efficiency of managing swine farm.