Assessing Biodiversity in Arable Farmland by Means of Indicators: an Overview

AGRONOMIE – ENVIRONNEMENT

Assessing biodiversity in arable farmland by means of indicators: an overview

Christian BOCKSTALLER Abstract: Maintaining biodiversity is one of the key issues of sustainable agriculture. It is ¸ Francoise LASSERRE-JOULIN now stated that innovation to enhance biodiversity in arable land requires operational Sophie SLEZACK-DESCHAUMES assessment tools like indicators. The goal of the article is to provide an overview of Severine PIUTTI available indicators. Besides measured indicators and simple indicators based on Jean VILLERD management data, we focus on predictive indicators derived from operational models Bernard AMIAUD and adapted to ex ante assessment in innovative cropping design. The possibility of use Sylvain PLANTUREUX for each indicator type is discussed. Key words: environmental assessment, indicator, model, validation, biodiversity, INRA, UMR 1121 ecosystemic services Nancy-Universite - INRA, IFR 110, Nancy-Colmar, BP 20507, 68021 Colmar France

Maintaining biodiversity is one of the intensification, among them extensifica- Indicators can be basic variables (e.g. key issues of sustainable development, tion and even suppression of chemical amount of input) or simple combination and agriculture is highly concerned in input like in organic farming (Hole et al., of these variables (balance, ratio) as well this perspective. The term was first 2005), reconsideration of field margin as field measurements, the former being suggested in 1985 at a conference on management to enhance semi-natural also called ‘‘indirect’’ and the latter biological diversity in the USA and was area of farmland (Marshall and Moonen, ‘‘direct’’ indicator regarding biodiver- popularized since the Rio Conference in 2002). It is now stated that this process of sity (Burel et al., 2008). Indicators can 1992 (Le Guyader, 2008). It is now innovation to enhance biodiversity also be derived from model outputs and commonly accepted that biodiversity in arable land requires operational assess- thus can be considered as ‘‘predictive can address the biological diversity at ment tools. These tools should evaluate indicators’’. By this way, an indicator different levels: i) the compositional, the current state at different scales, can be obtained from the average including the genetic, species, com- identify the causes of biodiversity impov- of model output, transformed into munity, habitat diversity, ii) the struc- erishment in a diagnosis phase, and scores or even expressed as the ratio tural, iii) the functional encompassing assess the effects of innovative solution of a model output and a reference processes within that level (Clergue et cropping systems (Bockstaller et al., value, as for pesticide risk indicators al., 2005). In the 2000s, the Millennium 2008b). This led many authors to plead (Bockstaller et al., 2009). This type of Ecosystems Assessment (2005) intro- for research on biodiversity indicators indicators expresses an explicit link duced the concept of ecosystemic (Carpenter et al., 2006) which have not between input variables addressing services provided by biodiversity, like to be confounded with bioindicators the causes, and an output reflecting pollination, and pest control. (Duelli and Obrist, 2003). The latter an effect on environment. Models can In arable area the change in land use, the use a component of biodiversity to assess be roughly separated in operational intensification and simplification of crop- something else, like the accumulation of models using a limited and available ping systems, as well as the drastic a pollutant. set of input variables and complex reduction of semi-natural elements From a general point of view, the term models which are too difficult to imple- (hedges, trees, wet zones, etc.) have ‘‘indicator’’ can refer to many defini- ment by non-scientists. If measured led to a significant decrease of biodiver- tions (Heink and Kowarik, 2010) as indicators are totally relevant for ex post sity in arable land (Le Roux et al., 2008). shown in figure 1 (Bockstaller et al., assessment of the state of biodiversity to Different options were developed to 2008b). Those authors set a typology evaluate e.g. the results of agri-environ- mitigate negative effect of agriculture based on the nature of the indicators. mental scheme (Kleijn et al., 2006), they

To cite this article: Bockstaller C, Lasserre-Joulin F, Slezack-Deschaumes S, Piutti S, Villerd J, Amiaud B, Plantureux S. Assessing biodiversity in arable farmland by means of indicators: an overview. OCL 2011;18(3):137-44. doi : 10.1684/ocl.2011.0381 doi: 10.1684/ocl.2011.0381

OCL VOL. 18 N8 3 mai-juin 2011 137

Article disponible sur le site http://www.ocl-journal.org ou http://dx.doi.org/10.1051/ocl.2011.0381 Management* Effects on Soil* Climate biodiversity

Abiltiy to trace cause-effect relationship

Predictive indicator based on Predictive indicator complex model based on M(x1, …, xn, p1, pk) operational model (e.g. Vegpop2) Measured f(x , …, x ) 1 p indicator (e.g. Flora-predict) y , y Simple indicators 1 2 (e.g. number of x , x , x /x , x -x 1 2 1 2 1 2 birds species) (e.g. % semi-natural area)

Integration of process Feasibility

Figure 1. Typology of indicators base on the construction method and evaluation of their quality (inspired from Bockstaller et al., 2008b) do not allow trace the cause. Simple allows assess the biodiversity, whereas indicators used at different scale from indicators can complete the information the operational model Flora-Predic pro- field to national level. An exhaustive on the causes or ‘‘pressure’’. Predictive vides a probability of presence. This review of proposals for different taxa can indicators offer a compromise between output only indicates the occurrence of be found in Burel et al. (2008). Species simple indicators and measured indica- species and can be considered as an of almost all taxonomic groups have tors regarding feasibility and degree of indicator. This last group of predictive been proposed (Lindenmayer and integration of process. They can be used biodiversity indicators remains poorly Likens, 2011). The indicator may for ex ante assessment to predict effect of covered by scientific literature. In this address all the species of a given taxon simulated system. Such indicators are article, we aimed at providing an over- or a given category like the number necessary for agronomist working on view of the available biodiversity indi- threatened species given by the Red innovative cropping design (Sadok et al., cators covering the three types of Lists for a given region (see list given by 2008). indicators, with a focus on this last Delbaere (2003)). It can also focus on Since the 90s, scientific publications on group of predictive indicators. We the diversity of keystone species, i.e. a biodiversity indicators have increased to illustrate it with recent initiatives con- species supporting the functioning of a reach 100 articles per year in the last cerning predictive indicators addressing ecosystem and the survival of many years (Burel et al., 2008). In the last biodiversity for different taxa mainly other species as well as umbrella species, decade, several reviews were published plants, invertebrate and soil microbial i.e. species which needs a large area to communities, this for spatial scale rang- survive and offers possibility of existence but their scope was beyond agriculture addressing natural land (e.g. Levrel, ing from field to agricultural landscape. to many other species (Clergue et al., 2007). Others focused on agriculture 2005). However examples of such but covered only measured indicators specific species are scarce in agriculture. and secondary simple indicators derived Measured (direct) In field experiments testing new from management data (Braband et al., indicators designed cropping systems, agronomist 2003; Buchs€ et al., 2003; Delbaere, assessed biodiversity by some measured 2003; Burel et al., 2008). Clergue et al. Since the 80s, a vast number of indicators like plant diversity (Vereijken, (2005) gave two examples of models measured indictors were proposed in 1997; Pacini et al., 2003). Among which can be used to derive indicators: the literature as previously reviewed by invertebrates, indicator based on the Vegepop2 a complex model predicting Noss (1990) and more recently by diversity of carabid beetle were pro- the effect on field boundary flora and Lindenmayer and Likens (2011). Indi- posed by many authors because they Flora-predict (Amiaud et al., 2005), an cators based on species diversity and/or are relatively easy to assess by simple operational model for meadow flora. abundance among a given taxon or pitfall although they were criticized as The model Vegpop2 predicts a dynamic several taxa (e.g. birds, plants, carabid indicator of biodiversity (Duelli, 1997). of biomass for different species that beetles, etc.) are the most commonly Doring€ and Kromp, 2003 analysed the

138 OCL VOL. 18 N8 3 mai-juin 2011 ability of different carabid beetle species well know, and present a power of land (Le Roux et al., 2008) we classify to indicate the impact of change of communication to the society (Levrel, them in two groups: i) indicators related cropping systems from intensive to 2007). to management of farmland at different organic. Since diversity cannot be only be scales ii) indicators addressing cropping practices, which can be expressed in At smaller level, soil is one of the major reduced to the number of species but amount of inputs per area unit or in reservoirs of microbial diversity, one also to their abundance and evenness in percentage of area disturbed by fertil- gram of soil containing between 3,000 distribution, several composite indices izer, pesticides, irrigation, tillage, both and 11,000 genomes (Torsvik and were proposed, to go further than the being expressed at different scales. Ovreas, 2002). Nevertheless, assessing species number. Table 1 shows some Table 2 gives some examples used in the diversity (in terms of species number) examples but more can be found in different assessment methods (Bockstal- remained a challenge for microbial Magurran (2004) who supplied infor- ler et al., 2008a; Bockstaller et al., 2009) ecologists, traditional techniques based mation about their statistical relevance. and information about their validation. on isolation and culturing being too Among them, the Shannon index is one By validation we mean here correlation selective (Gardi et al., 2009). Molecular the most popular one although some studies between indicator output and techniques like fingerprinting based on authors criticized its statistical relevance measurement of diversity for different 16S rDNA sequences were commonly and do not consider it as a true diversity taxa. We used data from the work of used to reveal patterns considered as ‘‘a index but as an entropy evaluation (Jost, Billeter et al. (2008). picture’’ of the microbial communities. 2006). The reciprocal Simpson’s indica- Nevertheless, the evaluation of the tor is a real diversity indicator and its microbial specific diversity is poorly outputs can be more easily interpreted: It is equal to the species number in case Predictive indicators informative in relation to soil functioning based on model (Maron et al., 2011). Therefore molec- of even distribution of species (all species have the same percentage in ular techniques targeting key microbial Like for other environmental issues (e.g. the sample) and decreases with increase genes implicated in ecosystem functions nitrate leaching), researchers developed of unevenness. It highlights the number were valuable tools to assess functional mechanistic models to predict the of dominant species in a certain way. In diversity (genetic structure, abundance dynamic of population for a given any case, no single index can provide all vs level of expression of functional micro- species, its survival probability, and the information directly. Our proposal is bial genes). The development of multi- ecological process like predator-prey to implement simultaneity a composite parametric indices integrating these data interaction. They worked mainly in index like the reciprocal Simpon’s is now needed to better understand the ecology for natural area (Guisan and indicator, and at least two other indices relationships between soil microbial Zimmermann, 2000) with mechanistic to explain the former, like the number of diversity and function. models and also operational static species and an the evenness index. At higher scale, only one indicator is models (Gontier et al., 2006) but few currently available in France at present concerned farmland. Besides the Vege- Simple (indirect) pop2 model for field boundary flora (see time: the diversity and abundance of common birds (about 120 species) indicators based on above Clergue et al., 2005), another which are divided in generalists, farm- management data model covering a species of farmland, land, forest and urban areas. Results of the Corn Bunting deserves attention but the assessment over more than 10 years Many proposals also exist for this type of remains relatively complex, needing show a clear decline in farmland special- indicators. Among the 91 indicators spatial data (Meyer et al., 2007). ists. The strength of this indicator is that listed for agriculture by Delbaere However in recent years, some opera- is can be relatively easy to obtain, (2003) more than the half belongs to tional approaches were developed with species being easy to determine, easy this type. Considering the general the specific objective to assess the effect to interpret, the cause of variation being model explaining biodiversity in farm- of farmland and crop management on biodiversity as shown in table 3. Most of Table 1. Example of composite indices of biodiversity the methods were developed in frame of a multicriteria assessment (excepted Indicator Methode de calcul Sanderson et al., 1995; Keichinger, 2001; Butler et al., 2009). SALCAbd Species number S = S s i (Jeanneret et al., 2006) was developed avec s :ith species i to complete the SALCA method based S (a) Shannon Index H’ = - pi .Lnpi on life cycle analysis. Actually this with pi = proportion of species i (entre 0 et 1) biodiversity component does not tackle Evenness E = H’/Ln S the whole production cycle but only the 2 farm level like the other methods Simpson’s Index D = S pi presented in table 3. Indicators of this Reciprocal Simpson’s index 1/D group provide output in form of a Buckland arithmetic Occurrence index: BuckArith-OI= probability of presence for one or a S 100/S Oi /Ri reduced number of species, or in form of with Oi = site number where species i is observed risk or impact scores. Whereas some Ri = site number where species i was observed models tackle in a explicit way a broad

OCL VOL. 18 N8 3 mai-juin 2011 139 Table 2. Example of simple indicators used in environmental methods (Bockstaller et al., 2008a) or proposed in initiative (e.g. IRENA at EU level (EEA, 2005))

Indicator Unit Example of method Scale Validation (correlation using the indicator, with given taxon)a or list containing it Percentage of area cropped % IRENA Region, country nsb in organic farming Percentage of area with % IRENA Region, country nsb agri-environmental scheme Percentage of % DIALECTE, IDEA Farm, landscape, Herbs, birds, bees, bugs, semi-natural area region, country hoverflies, carabids, spiders, Habitat diversity None Farm, landscape, Bees region, country Percentage of area with % IRENA, KULc, REPROc farm-landscape Not studied high value nature Hedgerow length in farm m RAD Farm, landscape Not studied Percentage of well-managed % Projet OTPA Farm, landscape nsb hedgerows Median size of field ha KULc, REPROc Farm, landscape nsb Average number of none Farm, landscape Bees, bugs, carabids crops per farm Crop diversity None (Shannon KULc, REPROc Farm, landscape nsb index Percentage of area % Farm, landscape nsb cropped intensively Pesticide use Number of treatments/ha, g of active ingredient/ha Fields, farm, landscape No correlation Pesticide use Treatment DIALECTE, IDEA, KULc, Fields, farm, landscape nsb frequency index REPROc Percentage of non % DIALECTE, Farm, landscape nsb sprayed area Nitroen fertilizer use kg N/ha Fields, farm, landscape Birds Percentage of intensively % DIALECTE, IDEA Fields, farm, landscape Herbs fertilized area Percentage of % DIALECTE, IDEA Fields, farm, landscape nsb irrigated area a see Billeter et al. (2008). Correlations studies included herbs, woody plants, birds, bees, bugs, hoverflies, carabids, spiders b not studied in Billeter et al. (2008) c German methods (see Bockstaller et al. (2009)) number of species, for plants based on equations (Sanderson et al., approach, using a dataset and deriving (Sanderson et al., 1995) or severa taxa 1995; Jeanneret et al., 2006; Butler et a posteriori regressions between a var- (Butler et al., 2009), most them focus on al., 2009) or functions (Meyer- iable assessing the occurrence of bird a few number of species or few taxa Aurich et al., 2003), decision tree using and different sets of input variables without explicit information on species. fuzzy subsets allowing cope with uncer- (table 3). The IBEA method (Anonymous., 2011) tainty and avoid effect of knife-edge does not address the species level but limit of classes (Keichinger, 2001; Sattler only biodiversity in general through an et al., 2010). More recently several Discussion ‘‘environment quality’’ and a ‘‘genetic authors (Sadok et al., 2009; Messean diversity’’ components. et al., 2010; Anonymous., 2011)) devel- This article aimed at providing an oped a qualitative approach based on overview of the biodiversity indicator Behind the calculation of such predic- decision tree using the DEXi tool available to agronomists working on tive indicators, different aggregation (Bohanec et al., 2008). Tichit et al. cropping systems at field and to other approaches are used: scoring systems (2010) worked on a totally different stakeholders working on higher scales.

140 OCL VOL. 18 N8 3 mai-juin 2011 Table 3. Examples of predictive indicator with their main characteristics

Name Taxonomic gro2up Species Expresion of Aggregation function Input variable or Scale Reference (species) number result component VEM Plants 534 Probability of presence Calculation of suitability index from the British e.g. grazing (yes/no) Fields, farm, Sanderson between 0 and 1 National Vegetation Classification Slurry application (yes/no) landscape et al., (1995) A probability of presence is derived from this index Mineral fertilization (kg N/ha) Keichinger, Pheasant 1 Score between 0 Decision tree with fuzzy subsets for global Soil cover Farm, landscape (2001) Partridge 1 (maximum impact) indicators and components Crop diversity Field hare 1 and 10 (no impact) Machine use Wild rabbit 1 Pesticide risk Irrigation Semi-natural area Meyer-Aurich Amphibian Not explicit Disturbance impact Continuous function e.g. Field, farm, et al., (2003) Partridge 1 scored between 0 number of tillage perturbation landscape (none) and 1 (high) amount of nitrogen number of herbicides SALCA bd Plants Not explicit Score between 1 Each cropping practices scored and calculation Scale Field, farm Jeanneret Mamamals (negative impact) and of an average value et al. (2006) Birds 5 (positive impact) Amphibians Snails and slugs Spides Carabid beetle Orthoptera Bees and bumblebees Butterflies Butler et al. Birds 63 Risk score between 0 Scoring system: assessing impact on species e.g. Spring to autumn sowing Regional (2009) Bumblebees 14 (none) and 3 to 6 (high) needs (e.g. diet, forage habitats) Increased agrochemical inputs Butterflies 23 Loss of non-cropped habitat Mamals 44 Land drainage Broadleaf plants 190 MASC None No 4 qualitative classes Decisison tree based on decision tree Crop diversity Cropping system Sadok et al. (DExi software) Non sprayer area (2009) Treatment frequency indexa DEXiPM Flying natural enemies Not explicit 5 qualitative classes Decisison tree based on decision tree e.g Cropping system Messean et al. Polllinators (DExi software) Deep tillage, and neighboring (2010) Soil natural enemies Treatment frequency indexa, habitat Weeds Habitat management for Flora of semi-natural area soil natural ennemies

OCL Sattler et al. Birds 1 Between 0 (low) and 1 Decision tree with fuzzy subsets e.g. cropping practices, Field (2010) Amphibian 1 Herbicides treatment a O.1 N 18 VOL. Mamal 1 frequency index , Hoverﬂy Not explicit fertilization (kg N/ha) Field ﬂora Not explicit 8

a-un2011 mai-juin 3 Tichit et al. Birds 2 Occurrence of bird Logistic regression function 11 management and Field (2010) (yes/no) 11 habitats variables IBEA Anonymous, None None 5 qualitative classes Decisison tree based on e.g. Farm (2011) decision tree (DExi software) Mineral fertilization (kg N/ha) Treatment frequency indexa Area with tillage 141 a Sum of ratio actual pesticide rate/recommended rate. Here the index is calculated separately for fungicides, herbicides, insecticides. We organized the presentation in three in the Netherlands, which assesses bio- the frame of multi-criteria assessment groups of indicators according to their diversity at farm level, each species is method addressing also other environ- type of construction. As it was pointed weighted by a factor addressing its rarity. mental issues or even other dimensions out by several review articles, a very Although there is a growing agreement of sustainability. Their advantage is that large number of measured indicators to assess not only biodiversity but also they do not need taxonomic identifica- are available using many taxonomic the linked ecosystemic services like tion while providing information of groups. The implementation of such pollination (Le Roux et al., 2008), those impacts of crop management factors indicators requires taxonomic knowl- are assessed in very indirect way by on one or several species, or a given edge what is the main hindrance to their assessing the diversity of species taxon. Most of them can be used by implementation for non-specialists like involved in the service. This is the case agronomists working on design of many agronomists. This led some for biological control. But possibility of innovative cropping systems. authors to propose parataxonomic predation between predator guilds lim- Like for the other types of indicators, their approach which assess the diversity by its the interest of species diversity, which predictive quality remains a question. sorting living organisms into morpho- is a very complex concept (Straub and Correlations were found by some authors logic groups (Duelli and Obrist, 2003). Snyder, 2008). Measurement of amount between indicator output and measured This approach remains very controver- of captured prey by a predators, or diversity of the taxon (Butler et al., 2009, sial (Abadie et al., 2008). Lindenmayer determination of stomach content of excepted for broadleaf species), at and Likens (2011) analysed in a very predator with help of molecular biology the level of occurrence of species exhaustive way the use of measured (King et al., 2008) are useful to assess the (Sanderson et al., 1995), or with expert biodiversity indicator. They pointed out service in experimental conditions. The judgments (Keichinger, 2001). The the lack of justification for the choice of implementation of such techniques in development of qualitative approach the species and the poor predictive routine to derive an indicator remains based on the DEXi tool (Bohanec et al., quality of many measured indicators. questionable for the moment. 2008) should be noticed. The construc- Some authors tried to validate meas- tion of such qualitative models is rela- ured indicators. At European level no The second group of simple indicators based on management variable are tively easy but the methodology requires group species could be correlated to the more methodological investment, espe- diversity of all taxa (Billeter et al., 2008), more easy to assess than measured ones and are therefore used in several multi- cially on the sensitivity (Bergez et al., although such test across different 2010). No validation was undertaken for ecosystems is not relevant according thematic environmental assessment method (Braband et al., 2003). Their these methods based on DEXi. Actullay Lindenmayer and Likens (2011). At field such indicators constructed with expert- level, a comparison of individual taxa of predictive quality remains in general poor although Billeter et al. (2008) based methods may be limited by invertebrates with the whole diversity of uncertainty and subjectivity. While invertebrates accross taxa yielded the found some correlations with the diversity of some taxa (table 2). But the expert-based indicators are built in a best correlation for heteroptera and ‘‘top-down’’ manner, a new trend con- aculate hymnoptera (bees, wasp and correlations were observed for a very broad range of conditions regarding sists in adopting a ‘‘bottom-up’’ point of ants), (Duelli and Obrist, 2003). Arena, view by building indicators from obser- coleotera (among them carabid bee- landscape structures, farming systems. Such indicators do not take into account vation data. The approach of Tichit et al. tles), syrphidae or red list species (2010) is one option. Machine-learning showed a poor correlation. interactions between management and pedo-climatic conditions and do not techniques (Shan et al., 2006) such as decision trees is another. They allow Such considerations led several authors refer in an explicit way to species. In any case, simple indicators based on man- produce indicators that are more objec- to state that assessment of biodiversity tive, scientifically sound and still easy to should rest on several taxa (e.g. Carignan agement data are useful to communi- cate with farmers as they focus on interpret. Uncertainty is still present but and Villard, 2002; Duelli and Obrist, can be represented and handled using 2003). In any case, biodiversity indicators information which farmers know and understand (Tichit et al., 2010). They accurate formalisms such as fuzzy logic. A are often expressed in composite indices work on the development of a biological like those presented in table 1.The can be used to analyze results obtained with measured indicators. The Nature control indicator is ongoing within the Shannon indicator is the most popular Casdar entomophage project (2009- but does not seem to be the most Balance is an example of assessment 1 method of biodiversity at farm level 2011) . The availability of adequate relevant (Jost, 2006). A major drawback dataset is a major requirement for the is that such indicators are only quantita- combining both types of indicators (Oppermann, 2003). development of indicators following this tive and do not take into account the methodology. nature of species. An increase of bio- In the difference with previous reviews diversity can also be due to invasive on biodiversity indicators in farmland (e. species (Lamb et al., 2009). Those g. Braband et al., 2003; Buchs€ et al., Conclusion authors tested the ability of several 2003), our article does not only focus on indices to show changes of biodiversity. measured indicators or simple indicators A vast number of biodiversity indicators Traditional indices like the Simpson and using management data, but also on are currently available for the agrono- Shannon indices (see table 1) yielded predictive indicators. In the last decade poorly in comparison with the Buckland several initiatives were proposed for arithmetic Occurrence index. In the assessment of agricultural systems, a 1 http://78.155.145.122/rmtbiodiv/moo- approach of van Wenum et al. (1999) majority of them being developed in dle/course/category.php?id=21.

142 OCL VOL. 18 N8 3 mai-juin 2011 mist and stakeholders who needs to 2010 the XIth ESA Congress. Montpellier Doring€ TF, Kromp B. Which carabid species assess the impact of agriculture on (France) : Agropolis International Editions, benefit from organic agriculture? - a review of biodiversity. A great diversity of indica- 2010. comparative studies in winter cereals from tors goes together with this abundance. Billeter R, Liira J, Bailey D, et al. Indicators for Germany and Switzerland. Agric Ecosyst The selection of indicators among the biodiversity in agricultural landscapes: a pan- Environ 2003; 98: 153-61. three types of indicators given by the European study. J Appl Ecol 2008; 45: 141-50. Duelli P. Biodiversity evaluation in agricultural typology proposed in this article, will Bockstaller C, Galan MB, Capitaine M, et al. landscapes: An approach at two different depend on the goal of the assessment Comment evaluer la durabilite des systemes scales. Agric Ecosyst Environ 1997; 62: 81-91. work. We recommend to use measured en production veg etale ? In: Reau, R, Dore, T Duelli P, Obrist MK. Biodiversity indicators: indicators when users aim at assessing (Eds.), Systemes de culture innovants et the choice of values and measures. Agric efficacy of a policy, of an agri-environ- durables: quelles methodes pour les mettre Ecosyst Environ 2003; 98: 87-98. mental scheme. But this requires several au point et les evaluer. Educagri, Dijon EEA. Agriculture and environment in EU-15; the species to give a ‘‘whole picture’’ and a (France), 2008a. IRENA indicator report. European Environ- justification of the species choice. Sim- Bockstaller C, Guichard L, Keichinger O, et al. mental Agency (EEA), Copenhagen (Dane- ple indicators can be used by farmers’ Comparison of methods to assess the sustain- mark), 2005. advicers in a first stage to work with ability of agricultural systems. A review. Agron Gardi C, Montanarella L, Arrouays D, et al. farmers to make them aware of the Sustain Dev 2009; 29: 223-35. Soil biodiversity monitoring in Europe: problem. They can also complete the Bockstaller C, Guichard L, Makowski D, et al. ongoing activities and challenges. Eur J Soil information given by measured indica- Agri-environmental indicators to assess crop- Sci 2009; 60: 807-19. tors. But it remains a challenge to prove ping and farming systems. A review. Agron Gontier M, Balfors B, Mortberg U. Biodiver- the link between both type of indica- Sustain Dev 2008b; 28: 139-49. sity in environmental assessment - current tors, to highlight the causes of the Bohanec M, Messean A, Scatasta S, et al. A practice and tools for prediction. Environ observed impacts on biodiversity. The qualitative multi-attribute model for eco- Impact Assess Rev 2006; 26: 268-86. third option is to use predictive indica- nomic and ecological assessment of genet- tors resulting from an operational model Guisan A, Zimmermann NE. Predictive hab- ically modified crops. Ecol Model 2008; 215: itat distribution models in ecology. Ecol or assessment function linking causes 247-61. Model 2000; 135: 147-86. and effect. They are useful for agrono- Braband D, Geier U, Kopke U. Bio-resource Heink U, Kowarik I. What are indicators? On mists working on ex ante assessment of evaluation within agri-environmental assess- newly designed cropping systems or to the definition of indicators in ecology and ment tools in different European countries. environmental planning. Ecol Indic 2010; 10: test policy options. Finally, more Agric Ecosyst Environ 2003; 98: 423-34. research is necessary on the validation 584-93. Buchs€ W, Harenberg A, Zimmermann J, et al. of all those reviewed biodiversity indi- Hole DG, Perkins AJ, Wilson JD, et al. Does Biodiversity, the ultimate agri-environmental cators. There is also a need of new organic farming benefit biodiversity? Biol indicator? Potential and limits for the appli- Conserv 2005; 122: 113-30. indicators assessing not only diversity cation of faunistic elements as gradual per se but also the impact of crop and Jeanneret P, baumgartner D, Freiermuth R, indicators in agroecosystems. Agric Ecosyst landscape management on ecosystemic Environ 2003; 98: 99-123. et al. Methode d’evaluation de l’impact des services. activites agricoles sur la biodiversite dans les Burel F, Garnier E, Amiaud B, et al. Chapitre1. bilans ecologiques. Salca bd. Agroscope FAL Les effets de l’agriculture sur la biodiversite. Reckenholz, Zurich, 2006. In: Le Roux X, Barbault R, Baudry J, Burel F, Jost L. Entropy and diversity. Oikos 2006; 113: REFERENCES Doussan I, Garnier E, Herzog F, Lavorel S, Lifran R, Roger-Estrade J, Sarthou J-P, M. T 363-75. Abadie JC, Andrade C, Machon N, et al. On (Eds.), Agriculture et biodiversite. Valoriser les Keichinger O. Evaluation de l’impact des the use of parataxonomy in biodiversity synergies. Expertise scientifique collective, pratiques agricoles d’exploitations de monitoring: a case study on wild flora. synthese du rappport, INRA (France), 2008. grandes cultures sur la valeur cyneg etique a Biodivers Conserv 2008; 17: 3485-500. Butler SJ, Brooks D, Feber RE, et al. A cross- l’aide d’indicateurs agro-ecologiques. Doc- Amiaud B, Pervanchon F, Plantureux S. An taxonomic index for quantifying the health of torat INPL-ENSAIA, Nancy, 2001. expert model for predicting species richness farmland biodiversity. J Appl Ecol 2009; 46: King RA, Read DS, Traugott M, et al. in grasslands: Flora-predict. In: Lillak, R, 1154-62. Molecular analysis of predation: a review of Viiralt, R, Linke, A, Geherman, V (Eds.), Carignan V, Villard MA. Selecting indicator best practice for DNA-based approaches. Mol 13th International Occasional Symposium of Ecol 2008; 17: 947-63. the European-Grassland-Federation, Tartu species to monitor ecological integrity: A (Estonia), 2005. review. Environ Monit Assess 2002; 78: 45-61. Kleijn D, Baquero RA, Clough Y, et al. Mixed biodiversity benefits of agri-environment Anonymous. Un outil de diagnostic de Carpenter SR, DeFries R, Dietz T, et al. schemes in five European countries. Ecol Lett l’impact des pratiques agricoles sur la bio- Millennium Ecosystem Assessment: Research 2006; 9: 243-54. diversite : Outil « IBEA ».In:Impact des needs. Science 2006; 314: 257-8. pratiques agricoles sur la Biodiversite des Clergue B, Amiaud B, Pervanchon F, et al. Lamb EG, Bayne E, Holloway G, et al. Indices Exploitations Agricoles. FNE, INRA, MNHN, Biodiversity: function and assessment in for monitoring biodiversity change: Are some Paris, 2011. agricultural areas. A review. Agron Sustain more effective than others? Ecol Indic 2009; 9: 432-44. Bergez JE, Carpani M, Monod H, et al. Dev 2005; 25: 1-15. Sensitivity analysis of DEXi type models Delbaere B. An inventory of biodiversity Le Guyader H. La biodiversite : un concept applied to design cropping systems. In : indicators in Europe, 2002. European Environ- flou ou une realit e scientifique ? Cour Environ Wery J, Shili-Touzy I, Perrin A, (Eds). Agro ment Agency (EEA), Copenhagen, 2003. 2008; n855: 7-26.

OCL VOL. 18 N8 3 mai-juin 2011 143 Le Roux X, Barbault R, Baudry J, et al. (Eds.) Meyer BC, Mammen K, Grabaum R. A Sattler C, Nagel UJ, Werner A, et al. Inte- Agriculture et biodiversite. Valoriser les syner- spatially explicit model for integrating spe- grated assessment of agricultural production gies. Expertise scientifique collective, syn- cies assessments into landscape planning as practices to enhance sustainable develop- these du rapport, INRA (France), 2008. exemplified by the Corn Bunting (Emberiza ment in agricultural landscapes. Ecol Indic Levrel H. Quels indicateurs pour la gestion de calandra). J Nat Conserv 2007; 15: 94-108. 2010; 10: 49-61. la biodiversite. Paris: Institut français de la Millennium Ecosystems Assessment. Ecosys- Shan Y, Paull D, McKay RI. Machine learning biodiversite, 2007. tems and Human Well-Being: Synthesis. Wash- of poorly predictable ecological data. Ecol Lindenmayer DB, Likens GE. Direct measure- ington D.C. (USA): Island Press, 2005. Model 2006; 195: 129-38. ment versus surrogate indicator species for Noss RF. Indicators for monitoring biodiver- Straub CS, Snyder WE. Increasing enemy evaluating environmental change and bio- sity: a hierarchical approach. Conserv Biol biodiversity strengthens herbivore suppres- 1990; 4: 355-65. diversity loss. Ecosystems 2011; 14: 47-59. sion on two plant species. Ecology 2008; 89: Magurran AE. Measuring biological diversity. Pacini C, Wossink A, Giesen G, et al. Evalua- 1605-15. Oxford (UK): Blackwell Science, 2004. tion of sustainability of organic, integrated Tichit M, Barbottin A, Makowski D. A and conventional farming systems: a farm Maron P-A, Mougel C, Ranjard L. Soil methodological approach to identify cheap and field-scale analysis. Agric Ecosyst Environ microbial diversity: Methodological strategy, and accurate indicators for biodiversity 2003; 95: 273-88. spatial overview and functional interest. CR assessment: application to grazing manage- Biol 2011; in press. Sadok W, Angevin F, Bergez J-E, et al. Ex ante ment and two grassland bird species. Animal Marshall EJR, Moonen AC. Field margins in assessment of the sustainability of alternative 2010; 4: 819-26. northern Europe: their functions and inter- cropping systems: guidelines for identifying Torsvik V, Ovreas L. Microbial diversity and actions with agriculture. Agric Ecosyst Environ relevant multi-criteria decision aid methods. function in soil: from genes to ecosystems. 2002; 89: 5-21. Agron Sustain Dev 2008; 28: 163-74. Curr Opin Microbiol 2002; 5: 240-5. Messean A, Lo-Pelzer^ E, Bockstaller C, et al. Sadok W, Angevin F, Bergez J-E, et al. MASC: Outils d’evaluation et d’aide a la conception a qualitative multi-attribute decision model van Wenum J, Buys J, Wossink A. Nature de strategies innovantes de protection des for ex ante assessment of the sustainability of quality indicators in agriculture. In: Brouwer, grandes cultures. Innov Agron 2010; 8: 69-81. cropping systems. Agron Sustain Dev 2009; FM, Crabtree, JR (Eds.), Environmental indicators and agricultural policy. CAB Interna- Meyer-Aurich A, Zander P, Hermann M. 29: 447-61. tional, Wallingford (UK), 1999. Consideration of biotic nature conservation Sanderson RA, Rushton SP, Pickering AT, et al. targets in agricultural land use - a case study A preliminary method of predicting plant Vereijken P. A methodical way of prototyping from the Biosphere Reserve Schorfheide- species distributions using the British integrated and ecological arable farming Chorin. Agric Ecosyst Environ 2003; 98: National Vegetation Classification. J Environ systems (I/EAFS) in interaction with pilot 529-39. Manage 1995; 43: 265-88. farms. Eur J Agron 1997; 7: 235-50.

144 OCL VOL. 18 N8 3 mai-juin 2011