Frédérique Angevin

Ex ante assessment of the sustainability of alternative cropping systems: implications for using multi-criteria decision-aid methods. A review

Sustainability is a holistic and complex multi-dimensional concept encompassing economic, social and environmental issues, and its assessment is a key step in the implementation of sustainable agricultural systems. Realistic assessments of sustainability require: (1) the integration of diverse information concerning economic, social and environmental objectives; and (2) the handling of conflicting aspects of these objectives as a function of the views and opinions of the individuals involved in the assessment process. The assessment of sustainability is therefore increasingly regarded as a typical decision-making problem that could be handled by multi-criteria decision-aid (MCDA) methods. However, the number and variability of MCDA methods are continually increasing, and these methods are not all equally relevant for sustainability assessment. The demands for such approaches are also rapidly changing, and faster ex ante assessment approaches are required, to address scales currently insufficiently dealt with, such as cropping system level. Researchers regularly carry out comparative analyses of MCDA methods and propose guidelines for the selection of a priori relevant methods for the assessment problem considered. However, many of the selection criteria used are based on technical/operational assumptions that have little to do with the specificities of ex ante sustainability assessment of alternative cropping systems. We attempt here to provide a reasoned comparative review of the main groups of MCDA methods, based on considerations related to those specificities. The following main guidelines emerge from our discussion of these methods: (1) decision rule-based and outranking qualitative MCDA methods should be preferred; (2) different MCDA tools should be used simultaneously, making it possible to evaluate and compare the results obtained; and (3) a relevantly structured group of decision-makers should be established for the selection of tool variants of the choosen MCDA methods, the design/choice of sustainability criteria, and the analysis and interpretation of the evaluation results.

MASC, a qualitative multi-attribute decision model for ex ante assessment of the sustainability of cropping systems

Realistic assessments of sustainability are often viewed as typical decision-making problems requiring multi-criteria decision-aid (MCDA) methods taking into account the conflicting objectives underlying the economic, social and environmental dimensions of sustainability, and the different sources of knowledge representing them. Some MCDA-based studies have resulted in the development of sustainable agricultural systems, but the new challenges facing agriculture and the increasing unpredictability of their driving forces highlight the need for faster ex ante (‘Before-the-event’) assessment frameworks. These frameworks should also (i) provide a more realistic assessment of sustainability, by integrating a wider range of informal knowledge, via the use of qualitative information; (ii) address alternative scales, such as cropping system level, improving granularity for the handling of sustainability issues and (iii) target a larger panel of decision-makers and contexts. We describe here the MASC model, which is at the center of a framework addressing these objectives. The MASC model has at its core a decision tree that breaks the sustainability assessment decisional problem down into simpler units as a function of sustainability dimensional structure (economic, social and environmental), generating a vector of 32 holistic ‘mixed’ (quantitative and qualitative) elementary criteria rating cropping systems. The assessment process involves the calculation of these criteria, their homogenization into qualitative information for input into the model and their aggregation throughout the decision tree based on ‘If-Then’ decision rules, entered by the user. We present the model and describe its first implementation for the evaluation of four cropping systems generated from expert knowledge, and discuss its relevance to the objectives cited above. The MASC model has several advantages over existing methods, due to its ability to handle qualitative information, its transparency, flexibility and feasibility.

Describing and locating cropping systems on a regional scale. A review.

At regional scale issues such as diffuse pollution, water scarcity and pollen transfer are closely related to the diversity and location of cropping systems because agriculture interacts with many other activities. Although sustainable land use solutions for territorial development and natural resource management are needed, very few agro-environmental studies account for both the coherence and the spatial variability of cropping systems. The originality of this article is to review methods that describe and locate cropping systems within large areas. We mainly based our analysis on four case studies using the concept of cropping systems on a regional scale, but differing in their objectives and extents. We found that describing and locating cropping systems in space meets not only decision-making stakes but also a scientific stake that allows multi-simulations over large areas when models require cropping system information. Simulation models are indeed necessary when the study aims at estimating cropping system externalities. Then, the involved process determines the extent, and the model determines the support unit, unless socio-economic considerations prevail. In this case, as well as when no model is involved, it is often considerations related to stakeholders that determine extent and support unit choices. On a regional scale, the cropping system must be described by only a few variables whose selection depends on the study objective and the involved processes. Collecting cropping system information for all support units is often simplified by identifying determining factors of cropping systems. However, obtaining deterministic relations between easily accessible factors and cropping system variables is not always possible, and sometime accessing modalities of determining factors for all support units is also difficult. We found that describing and locating cropping systems relied very much on expertise and detailed survey data. The development of land management practice monitoring would facilitate this description work.

Spatial sensitivity of maize gene-flow to landscape pattern: a simulation approach

Pollen dispersal is a critical process defining connectivity among plant populations. In the context of genetically modified (GM) crops in conventional agricultural systems, strategies based on spatial separation are promoted to reduce functional connectivity between GM and non-GM crop fields. Field experiments as well as simulation studies have stressed the dependence of maize gene flow on distances between source and receptor fields and on their spatial configuration. However, the influence of whole landscape patterns is still poorly understood. Spatially explicit models, such as MAPOD-maize, are thus useful tools to address this question. In this paper we developed a methodological approach to investigate the sensitivity of cross-pollination rates among GM and non-GM maize in a landscape simulated with MAPOD-maize. The influence of landscape pattern on model output was studied at the landscape and field scales, including interactions with other model inputs such as cultivar characteristics and wind conditions. At the landscape scale, maize configuration (proportion of and spatial arrangement in a given field pattern) was shown to be an important factor influencing cross-pollination rate between GM and non-GM maize whereas the effect of the field pattern itself was lower. At the field scale, distance to the nearest GM maize field was confirmed as a predominant factor explaining cross-pollination rate. The metrics describing the pattern of GM maize in the area surrounding selected non-GM maize fields appeared as pertinent complementary variables. In contrast, field geometry and field pattern resulted in little additional information at this scale.

Describing and Locating Cropping Systems on a Regional Scale

At regional scale issues such as diffuse pollution, water scarcity and pollen transfer are closely related to the diversity and location of cropping systems because agriculture interacts with many other activities. Although sustainable land use solutions for territorial development and natural resource management are needed, very few agro-environmental studies account for both the coherence and the spatial variability of cropping systems. The originality of this article is to review methods that describe and locate cropping systems within large areas. We mainly based our analysis on four case studies using the concept of cropping systems on a regional scale, but differing in their objectives and extents. We found that describing and locating cropping systems in space meets not only decision- making stakes but also a scientific stake that allows multi-simulations over large areas when models require cropping system information. Simulation models are indeed necessary when the study aims at estimating cropping system externalities. Then, the involved process determines the extent, and the model determines the support unit, unless socio-economic considerations prevail. In this case, as well as when no model is involved, it is often considerations related to stakeholders that determine extent and support unit choices. On a regional scale, the cropping system must be described by only a few variables whose selection depends on the study objective and the involved processes. Collecting cropping system information for all support units is often simplified by identifying determining factors of cropping systems. However, obtaining deterministic relations between easily accessible factors and cropping system variables is not always possible, and sometime accessing modalities of determining factors for all support units is also difficult. We found that describing and locating cropping systems relied very much on expertise and detailed survey data. The development of land management practice monitoring would facilitate this description work.

Guidelines to design models assessing agricultural sustainability, based upon feedbacks from the DEXi decision support system

New agricultural systems are required to satisfy societal expectations such as higher quantity and quality of agricultural products, reducing environmental impacts, and more jobs. However, identifying and implementing more suitable agricultural systems is difficult due to conflicting objectives and to the wide diversity of scientific disciplines required to solve agricultural issues. Therefore, designing models to assess the sustainability of agricultural systems requires multi-criteria decision aid methods. The French agronomist community has recently developed 11 hierarchical and qualitative models to assess sustainability using the DEXi decision aid software. Here, we give guidelines to help designers to build their own specific models. First, we present the principles and applications of the DEXi software. Then, we provide guidance on the following steps of model designing: (1) initial analysis and planning of the design process, (2) selection and hierarchy of sustainability criteria, (3) indicator selection and building, (4) parameterization, (5) evaluation, and (6) model dissemination and uses. We then discuss advantages and drawbacks of this kind of modeling formalism, the role of a participatory approach, and the main properties to consider during the design process.

Impact of gene stacking on gene flow: the case of maize

To respect the European labelling threshold for the adventitious presence of genetically modified organisms (GMOs) in food and feed, stakeholders mainly rely on real-time PCR analysis, which provides a measurement expressed as a percentage of GM-DNA. However, this measurement veils the complexity of gene flow, especially in the case of gene stacking. We have investigated the impact of gene stacking on adventitious GM presence due to pollen flow and seed admixture as well as its translation in terms of the percentage of GM-DNA in a non-GM maize harvest. In the case of varieties bearing one to four stacked events, we established a set of relationships between the percentage of GM kernels and the percentage of GM-DNA in a non-GM harvest as well as a set of relationships between the rate of seed admixture and the percentages of GM material in a non-GM harvest. Thanks to these relationships, and based on simulations with a gene flow model, we have been able to demonstrate that the number of events and the stacking structure of the emitting fields impact the ability of a non-GM maize producer to comply with given GM kernel or GM-DNA thresholds. We also show that a great variability in the rates of GM kernels, embryos and DNA results from seed admixture. Finally, the choice of a unit of measurement for a GM threshold in seed lots can have opposite effects on the ability of farmers to comply with a given threshold depending on whether they are crop or seed producers.

Issues and challenges in landscape models for agriculture: from the representation of agroecosystems to the design of management strategies

ContextAgroecosystems produce food and many other services that are crucial for human well-being. Given the scales at which the processes underlying these services take place, agricultural landscapes appear as appropriate spatial units for their evaluation and management. The design of sustainable agricultural landscapes that value these services has thus become a pressing issue but faces major challenges stemming from the diversity of processes, their interactions and the number of scales at stake. Agricultural landscape modelling can provide a key contribution to this design but must still overcome several difficulties to offer reliable tools for decision makers.ObjectivesOur study aimed at shedding light on the main scientific and technical difficulties that make the building of landscape models that may efficiently inform decision-makers a complex task, as well as translating them in terms of challenges that can be further investigated and discussed.MethodsWe examine current issues and challenges and indicate future research needs to overcome the scientific and technical obstacles in the development of useful agricultural landscape models.ResultsWe highlight research perspectives to better couple landscape patterns and process models and account for feedbacks, integrate the decisions of multiple stakeholders, consider the spatial and temporal heterogeneity of data and processes, explore alternative landscape organisations and assess multiobjective performance.ConclusionCoping with the issues and challenges discussed in this paper should improve our understanding of agroecosystems and give rise to new hypotheses, thereby informing future research.

Management of herbicide-resistant weed beet: a simulation study.

The rapid evolution of economie and environmental constraints on farming systems require cropping systems design or assessment to be conducted in the context of market instability as well as economie (c.g. CAP in EU) and environmental policy changes. The Nitrate Directive (91/676/EC) is one of the oIdest EU environmental policy, designed to reduce water pollution by nitrate from agricultural sources, through a set of measures, defined at regional level, and mandatory for farmers of vulnerable zones.