INRA Prod. Anim., Managing animal diversity 2019, 32 (2), 263e-280e in livestock farming systems: types, methods and benefits

Marie-Angélina MAGNE1, Marie-Odile NOZIÈRES-PETIT2, Sylvie COURNUT3, Émilie OLLION4, Laurence PUILLET5, David RENAUDEAU6, Laurence FORTUN-LAMOTHE7 1AGIR, University of Toulouse, ENSFEA, INRA, INPT, INP- EI PURPAN, 31320, Castanet-Tolosan, France 2INRA, UMR SELMET, 2 place Viala, 34000, Montpellier, France 3University of Clermont Auvergne, AgroParisTech, Inra, Irstea, VetAgro Sup, UMR Territoires, F-63000, Clermont-Ferrand, France 4ISARA-Lyon, AGE team 23 Rue Jean Baldassini, 69007, Lyon, France 5UMR Modélisation Systémique Appliquée aux Ruminants, INRA, AgroParisTech, Université Paris-Saclay, 75005, Paris, France 6INRA, UMR PEGASE, Agrocampus Ouest, 35590, St Gilles, France 7GenPhySE, University of Toulouse, INRA, INPT, ENVT, 31320, Castanet-Tolosan, France *E-mail: [email protected]

„„Increasing biodiversity, and in particular animal diversity, is often described as a promising means of adapting livestock systems to potential hazards and facilitating the agroecological transition. However, the majority of farmers, consultants, teachers, and researchers still find it difficult to view biodiversity as an asset in livestock management. Here, we develop a conceptual framework and provide perspectives to support the analysis of animal diversity and its management in livestock farming systems, in order to realise the benefits of this approach.

Introduction et al., 2015; Duru and Therond, 2015). 2013), and fluctuations in feed intake First, it has contributed to a decline in (Delaby et al., 2009). biodiversity, including agrobiodiversity During the second half of the 20th — the diversity of cultivated plants To overcome these limitations, breed- century, the development of a produc- and domestic animals (Altieri, 1987) ing programmes in all livestock sectors tivist model of livestock farming led — which ultimately calls into ques- have been adjusted to better integrate to an increase in animal production tion the ability of agricultural systems functional traits, particularly in dairy through i) the selection of plants and to adapt to global change (Hoffmann, cattle (Phocas et al., 2014). However, animals based on production traits; 2013). Second, the resulting short-term the effectiveness of breeding adjust- ii) the use of inputs (e.g. concentrated increases in productivity are often neg- ments in overcoming these limitations feed, agrochemical fertilisers, and vet- atively correlated with long-term pro- has been questioned. For this reason, erinary products) to reduce environ- ductivity (Weiner, 2003). For example, alternative livestock farming systems mental heterogeneity and the effects the selection of dairy cows for high milk (LFSs) have been proposed, whose of limiting factors; and iii) the standardi- yields has led to a deterioration in their strength lies in their ability to meet the sation and modernisation of production fertility (Mackey et al., 2007; Sørensen needs of the agroecological transition. methods as well as the specialisation et al., 2008) and longevity (Knaus, 2009), Agroecology is based on the hypothesis of farms and regions (Mazoyer, 1982). and to an increased susceptibility to that biodiversity, and in particular agro- The limitations of this model have now health problems (Oltenacu and Broom, biodiversity, is an essential resource to been well documented (Brussaard 2010), to climatic variations (Gauly et al., ensure the sustainability of livestock

https://doi.org/10.20870/productions-animales.2019.32.2.2496 INRA Productions Animales, 2019, numéro 2 264e / Marie-angélina magne et al.

Figure 1. The sequential approach used to build, test, and apply the conceptual framework for an integrated analysis of animal diversity management in livestock integrated analysis of the management production. of animal diversity in LFSs (Section 3). Finally, we conducted a self-assessment­ Examining several research studies Construction of the conceptual Step 1 of animal diversity in LFSs using framework, its components, of our approach for developing and a preliminary analytical grid and their various attributes testing the conceptual framework in order to provide some proposals for its

Assessment of the function use in research, education, and consult- Selecting and describing and relevance of the conceptual ing for LFSs (Section 4). Step 2 four case studies using framework for an integrated analysis the conceptual framework of the management of animal diversity in LFSs 1. Presentation of the conceptual framework for Performing a cross-analysis Identification of gaps in the literature regarding the integrated analysing animal diversity Step 3 of the four research studies based on the conceptual framework analysis of the management and its management in of animal diversity in LFSs livestock farming systems

Conducting a self-assessment Description of some proposals The conceptual framework we Step 4 of the approach used to develop for the use of the conceptual and test the conceptual framework framework in research, education, developed has four main components and consulting on LFSs (Figure 2). farms and increase their adaptive animal diversity in LFSs, and charac- Two components (Figure 2, top) aim to capabilities in suboptimal and variable terised them using an analytical grid analyse animal diversity in the context environments (Darnhofer et al., 2010; we constructed based on our own of the different forms it takes (described Biggs et al., 2012; Dumont et al., 2013; expertise. This enabled us to develop in detail in § 1.1) and the organisational Duru et al., 2015). To date, the evidence and consolidate the different compo- and time scales at which it is created that has been collected to support this nents of our conceptual framework and benefits are seen (§ 1.2). The two has focused mostly on plant agrobio- and to identify the various attributes other components (Figure 2, bottom) diversity, particularly that of grasslands of each (Section 1). In the second step, aim to analyse its management in the (Damour et al., 2018). Studies on animal we selected four research studies and LFS (§ 1.3) and the benefit(s) that the agrobiodiversity (hereafter referred to described them using the concep- farmer derives or expects from it (§ 1.4). as animal diversity) are fewer, more tual framework we developed. These It should be noted that, while diversity recent, and scattered. We hypothe- studies were selected with the goals at the population scale is fundamental sise that this is primarily the result of of obtaining different perspectives on to the long-term management of diver- a lack of a shared conception among animal diversity in LFSs and examin- sity within species and breeds, we focus stakeholders in livestock production of ing the ways, and the degree to which, here on its management at the scale of what is meant by animal diversity and research on animal diversity can be bro- the LFS. Therefore, animal populations its management at the scale of the LFS. ken down into the different component are viewed as resources available to parts of the framework. We used these farmers in the acquisition of animal The aims of this article are thus to to test the functionality and relevance diversity to be used, maintained, and develop a conceptual framework for an of the conceptual framework in differ- enhanced within the LFS. integrated analysis of the management ent contexts for the integrated analysis of animal diversity at the scale of the LFS of the management of animal diversity „„1.1. The different forms of animal diversity and to propose areas for future research in LFSs (Section 2). In the third step, to explore the conditions under which it we performed a cross-analysis of the Diversity is characterised by varia- is beneficial in the long term. To do this, four research studies, with which we tion, heterogeneity, and differences, as we implemented a four-step sequential identified some commonalities and opposed to the concepts of uniformity, approach (Figure 1). differences in the application of the homogeneity, and similarity. conceptual framework. With this, we First, we examined several stud- were also able to identify some areas Genetic diversity refers to the degree ies from the literature that explored for future research with respect to the of variation (or polymorphism) of genes

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /265e

Figure 2. Conceptual framework for an integrated analysis of animal diversity management in livestock farming systems (LFSs). focused on associations between cat- tle and sheep (Cournut et al., 2012; - Genetic Meisser et al., 2013; D'Alexis et al., 2014) - Organisational scales - Phenotypic and cattle and horses (Martin-Rosset - Species - Time scales and Trillaud-Geyl, 2011). Few studies - Functional have focused on associations between Forms of animal Scales at which ruminant and monogastric species, diversity diversity is created and benefits are seen even though they may play potentially complementary roles in the food chain

Animal of an ecosystem (respectively, forage diversity intake and polymer degradation of in FSs plant cell-wall constituents vs. seed and tuber intake and degradation of other Management of Benefits delivered animal diversity to farmers from constituents). animal diversity

- Farmer (s)/manager (s) These three types of diversity may - Livestock farming practice (s) - Production (quantity all contribute to the final form exam- and quality) n a tion - Management indicators - Efficiency, autonomy b i ined here, that of functional diversity - Determinants - Resistance, resilience C o m (Figure 3; Tichit et al., 2011). This con- cept refers to the association of animals within a single species, breed, or line. groups). They are the joint result of with different biological or ecological It is the basis of genetic selection, the genetic diversity and the effects of the functions or aptitudes (e.g. meat breed basic principle of which is to choose physical, chemical, and social environ- vs. dairy breed), or with different bio- for reproduction, from the existing ment. Phenotypic diversity (e.g., age, logical and ecological responses to dis- variability in a given population, ani- weight, conformation) can have direct turbance (e.g. climate-stress tolerant mals that have favourable versions (or effects on the type and quality of ani- vs. sensitive phenotype). For example, alleles) of genes of interest, in order to mal products produced for the market. in aquaculture, ecosystem functioning pass them on from one generation to can be optimised by constructing an the next and thus enable the genetic Species diversity refers to the asso- assemblage of fish species (detritivores, improvement of the population (spe- ciations of different animal species vegetarians, carnivores) with comple- cies, breed, or line). The genes of inter- on farms. To date, most studies have mentary characteristics, functions, and est can be related to production traits Figure 3. The different forms of animal diversity and the organisational scales at or functional traits, such as resistance which it is created and is expressed. to pathogens or the ability to reproduce in a given environment (Phocas et al., Functional diversity 2017; see § 2.4). Genetic diversity is also the basis for actions designed for the

conservation and promotion of rare or diversity m a l limited-number breeds (Audiot, 1995). i Genetic Phenotypic Species of a n

s diversity diversity diversity e

Phenotypic diversity describes the y p T variability of observable and measur- able traits in an animal. These traits can ivestocSystème Animal Population be physical (size, coat colour, horn con- Geènneess OOrrggaanne erdTroupeau d’ésystemlevage figuration), physiological (age, parity, digestive efficiency, milk production, weight, conformation), behavioural Scales of organisation Dotted arrows: organisational scales at which animal diversity originates; Dashed arrows: organisational (activity on pasture, relationship scales at which the benefits of animal diversity are seen; Solid arrows: functional diversity supported by with humans), or biochemical (blood other forms of animal diversity.

INRA Productions Animales, 2019, numéro 2 266e / Marie-angélina magne et al.

habitats (with respect to, e.g., reproduc- at which constituent elements inter- assessed at the organisational scale of tion, thermophilia, or feeding; Néori act through biological processes; and the herd and a time scale of several pro- et al., 2004). (ii) the organisational and time scales duction campaigns. Simulations based at which animal diversity is expressed, on this model revealed that increased „„ 1.2. Organisational i.e. the broader scale at which animal phenotypic diversity of goats in the and time scales at which diversity delivers its potential bene- herd, i.e. the combination of “specialist” animal diversity is created and expressed fits to the farmer. The scales at which and “generalist” goats, was beneficial in animal diversity are created are thus reducing inter-annual variability in the Because an LFS is, by definition, intrinsically linked to the forms of diver- herd's milk production in response to composed of two sub-systems — a sity present (§ 1.1), while the scales at environmental variations. management system and a biological which it is expressed are closely linked system (Dedieu et al., 2008) — it inte- to methods of management (§ 1.3) and „„1.3. The management of animal diversity grates different coexisting and nested the benefits (potential or realised) the in livestock farming systems forms of animal diversity across the farmer derives from such management various organisational scales inherent in (§ 1.4). Characterising and understanding all living organisms. Its overall function- the management of animal diversity ing and dynamics emerge from those Thus, for models of LFSs to accurately first requires the identification of the of the underlying organisational scales assess the benefits of different manage- person(s) managing the diversity, the and their interactions, which are in turn ment approaches, the organisational management practices that influence it, driven by the practices of the farm man- and time scales in question must be the management indicators in use, and ager. Specifically, the functional dynam- clearly identified. For example, to assess the determinants of these practices. ics of a livestock production enterprise the benefits of animal diversity in goat arise from those of the constituent herds in terms of its effect on resilience, Within LFSs, the farmer and his/her groups of animals, which are in turn the a model was developed that combined working partners on the farm are the product of the functional dynamics of analyses at the scale of the individual main managers of animal diversity. animals' constituent organ systems (e.g. animals and that of the herd (Tichit However, depending on how animal organs of the digestive, reproductive, et al., 2012). The individual animal and breeding and selection are organized, and respiratory systems). Furthermore, the production campaign were defined it may be necessary to identify the type in addition to the nesting within organ- as the relevant organisational and time of farmers who play a relevant role. This isational scales, there is also nesting scales, respectively, at which the pheno- is particularly true in the pig or poultry that occurs within different time scales. typic diversity of goats is created; here, sectors, in which animal diversity can be Indeed, the dynamics of LFSs are based the functional diversity supported by managed by nucleus breeders, multipli- on processes specific to each organisa- this phenotypic diversity was quantified ers, and/or producers (§ 2.4). tional scale, which take place according as differences in the allocation of feed to different rhythms and at different resources to the biological functions of Livestock farmers manage animal time scales. For instance, cellular met- reproduction and lactation. “Generalist” diversity through a variety of practices: abolic processes take place on a scale goats were defined as having good herd management (e.g. feeding and of minutes, while herd processes such reproductive performance (kidding pasture management, reproduction, as female replacement take place on a rate = 90%) and average milk perfor- and health management), herd config- scale of several years. mance (700 L/lactation) regardless of uration (e.g. plans for replacement, cull- environmental conditions. “Specialist” ing, and mating), and processing, sale, To characterize, and realise the ben- goats were instead defined as having and marketing (sale of a range of animal efits of animal diversity in LFSs, it can average reproductive performance products: animals sold for reproduction be helpful to understand the interre- (kidding rate = 70%) and high milk or fattening, milk, eggs, meat, fish, and lationships of these different scales production in favourable environments their derivatives). These three kinds of and the ways in which they nest. This (1,000 L/lactation) and low milk pro- management practices make it possible involves distinguishing (Figure 3): (i) the duction in unfavourable environments to create or acquire animal diversity, to organisational and time scales at which (600 L/lactation). Then, the expression manipulate it (e.g. to allocate and seg- animal diversity is created, i.e. the scales of this diversity and its benefits were ment it, to exploit it), and to maintain

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /267e

it or not (to reduce or increase). Thus, a diversity naturally results from farm- downstream marketing ­operators and farmer who sells all of his lambs under ers’ livestock management practices animal-product industries create a a quality and origin certification may be without being explicitly desired or demand for homogeneous or hetero- compelled to reduce the phenotypic managed. In the former case, animal geneous animals depending on the diversity within his sheep enterprise by diversity is intentionally managed by marketing chains and periods in ques- grouping births and/or increasing sup- farmers based on objectives and indi- tion, sometimes even providing farmers plementation of twin lambs, in order cators, whether explicitly stated or not. technical advice or financial incentives to create batches of lambs that are as to meet their expectations. The man- homogeneous as possible in terms of To implement animal diversity man- agement of the diversity of animal weight, conformation, and fat content. agement practices, farmers seek out, products, which operates on its own Conversely, a farmer raising lambs for gather, and use information both on scales, complements that carried out by direct sale may seek to increase phe- and off the farm (Magne et al., 2011). farmers. The goal of this type of organ- notypic diversity by organising lamb- For example, they may examine data isation is to ensure that the supply of ing and rearing, particularly through on the biotechnical system as a whole, animal products, which is unevenly breeding and feeding practices, in such i.e. the indicators they use to charac- distributed, fluctuating, and by nature a way as to have heterogeneous lambs terise on-farm animal diversity and to subject to uncertainty across both time available throughout the year and thus assess the benefits they derive from and space, meets the constant demand meet the expectations of individual it (and that their partners derive indi- (consumption) that is concentrated in consumers (Nozières, 2014). Among rectly from it; § 1.2). However, they urban centres (Malassis, 1979). livestock management practices, argu- may also analyse data on the physical, ably the three most important for the social, and economic environment of „„1.4. The benefits to livestock farmers management and exploitation of ani- the farm, and on how animal diversity of managing animal mal diversity in LFSs are the choice of may be affected by both internal and diversity animals for breeding and replacement, external changes. All or a part of this the differentiation of feeding practices information can be formalised through The fourth component of our concep- according to the end use of products, the use of indicators, including perfor- tual framework addresses the potential and batching practices. These practices, mance indicators such as milk yield and or real benefits derived or expected and the land use that is associated with inter-annual variations on a given farm from the management of animal diver- them, can all be organised at different or between farms, as well as selection sity in LFSs. Here, we only considered time scales (within a year, over many indexes built for an entire population. farm manager(s) as beneficiaries, years) and organisational scales (§ 1.2). although there are other possibilities Thus, in LFSs that mix cattle and sheep The choice of practices for the man- (e.g. breeding organisations, citizens, species, grassland can be grazed sepa- agement of animal diversity depends consumers, etc.). The benefits are the rately, alternately, or simultaneously by on two types of determinants: those advantages that the farmer derives the different species, depending on the that are internal to the farm or the from the biological and ecological pro- farmer's objectives for each species in individual farmers (e.g. his/her values, cesses supported by the animal diver- terms of production and the use of for- standards, and objectives) and those sity under his/her management. age resources (§ 2.3). In some situations, that are external (e.g. orientation of instead, farmers do not implement dif- supply chains, resources available in The first potential benefit con- ferent practices to manage on-farm the environment). These determinants cerns improvement in the production animal diversity; this is the case, for are factors that hinder or facilitate of goods (i.e. animal products and example, with farmers who feed all the implementation of animal diver- by-products) that contribute to the dairy cows in their herd in the same sity management practices in LFSs. gross production on which the farmer's way (homogeneity in practices), even For example, by strongly restricting income relies. Proper management of if the animals are of different breeds the genetic variability available in the animal diversity can promote comple- and/or milk potential (§ 2.2). Finally, it is breeding stock placed on the market, mentarity in animal production cycles worth identifying if the animal diversity animal breeding societies and enter- and thus enable the expansion of the is managed by the farmer consciously prises can hinder the development range of products available at a given or unconsciously. In the latter case, the of animal diversity in LFSs. Similarly, time and over a production campaign.

INRA Productions Animales, 2019, numéro 2 268e / Marie-angélina magne et al.

„„2.1. Study 1: Similarly, by supporting functional maintain itself, but also to adapt to Within-herd animal complementarity, it is possible to changes in its environment (Dedieu diversity and its benefits improve the quality and/or quantity of and Ingrand, 2010). Animal diversity from the perspectives animal products. For example, mixing can buffer a hazard, whether climatic, of dairy cattle farmers specialist and generalist dairy cattle economic, sanitary, or zootechnical, and breeds within a herd has been shown offer options for adaptation and trans- The work of Ollion (2015) described to generate worthwhile performance formation of the LFS based on current or how dairy farmers characterise the trade-offs between milk yield, milk predicted conditions (Darnhofer et al., forms of animal diversity in their herds, content, and meat yield, thanks to com- 2010; Nozières et al., 2011). In such a how this diversity changes over time, the plementarity in breed features (§ 2.2). way, associations between meat sheep driving factors behind these changes, Similarly, co-grazing of cattle heifers and dairy cattle permit readjustments and the benefits that motivate change. and dairy goats was shown to improve in the use of grassland resources and This research thus analysed intraspe- goat weight gain and overall animal stock according to the respective feed- cific animal diversity from the point of production while improving pasture ing requirements and purpose of each view of the farmers themselves. It was management (D'Alexis et al., 2014). species (Cournut et al., 2012), which can based on interviews carried out in a be used to adapt to climatic hazards; sample of 39 farmers. Twenty-five farm- The second potential benefit is likewise, management of phenotypic ers had purebred herds: 11 Holstein, improvement in the efficiency of the diversity in sheep (breeding, feed- 7 Montbeliarde, 6 Normande, and LFS, defined here as the ratio between ing, and batching) allows marketing 1 Ferrandaise. Nine farmers had multi- the value of the products obtained and practices to adapt to market changes breed herds that contained Holstein the resources expended for their pro- (Nozières and Moulin, 2016). cows along with those of one or two duction. When the focus is on enhanc- other breeds, which comprised 10 to ing the value of internal farm resources, While each of these three types of 75% of the herd; the breeds consid- then efficiency includes self-sufficiency. benefits can be assessed independently, ered were Montbeliarde (5 herds), Thus, the management of animal diver- they are more often considered in com- Normande (2 herds), Simmental sity in LFSs can be a strategy for better bination in discussions of the manage- (1 herd), and both Montbeliarde and use of the available on-farm resources ment of animal diversity on farms (§ 3 Abondance (1 herd). Finally, five farm- (feed resources, labour, land, equip- and 4). ers had crossbred herds, with crossed ment) and a means by which to develop cows representing 50 to 100% of the the complementarity of the constituent 2. Four studies herd. These herds were based on three- elements of the system to “better use the of animal diversity to five-breed crossbreeding schemes, whole range of resources and conditions in livestock farming involving Holstein, Normande, Jersey, offered” (Reboud and Malezieux, 2015). systems characterised Montbeliarde, Scandinavian Red, New As an example, mixing meat sheep and using the conceptual Zealand Holstein, and Brown Swiss. dairy cows on the same farm enables framework Almost all of the farmers interviewed better use of the diversity of on-farm (n = 37/39) stated that they had animal grassland resources (both in space and Understanding what types of animal diversity in their herd (Figure 4). time), as the two species have different diversity are beneficial in LFSs and the feeding behaviours, feed requirements, conditions under which these benefits In their narratives, the farmers and susceptibilities to gastrointestinal arise involves consideration of all four addressed two forms of animal diver- strongyles. The association of livestock components of the proposed concep- sity: genetic and phenotypic. For a species with different production cycles tual framework. To demonstrate this, majority of the farmers (24/39), with- can also increase working efficiency by we applied the conceptual framework in-herd diversity was based on the spreading labour demands more evenly to four research studies carried out at genealogy of each animal: 19 farm- over the production campaign. the National Institute for Agricultural ers distinguished their cows accord- Research (INRA) in order to analyse ing to the genetic origin of the dam A third potential benefit concerns their approaches with respect to animal (“Olympus, well, all her daughters have improvement in the resilience of the diversity and its management in LFSs the same character: they are going to LFS, defined here as its capacity to (Table 1). be docile”), while 5 farmers based their

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /269e

statements on the genetic origin of the are more vigorous and spiteful than or physiological traits (biological trade- sire (“some bulls disappointed us while Holstein; they have more character”). offs such as milk yield vs. milk content, others we wouldn’t have bet on provided The farmers also characterised the milk vs. longevity or fertility; n = 9/39). us four or five daughters about which phenotypic diversity of their animals we have nothing bad to say”). For the according to physical traits (morphol- Of the 37 farmers who identified ani- other sampled farmers (15/39), animal ogy, coat colour, horn type; n = 31/39), mal diversity in their herd, 28 found diversity referred instead to the genetic behavioural traits (character, feeding that it had declined over time, whether variation among breeds or crossbreeds behaviour; n = 18/39), ability to adapt intentionally or not. An analysis of their (“Montbeliarde and Abondance breeds milk production to hazards (n = 34/39), practices showed that this decrease

Table 1. Presentation of the four research studies within the conceptual framework developed to describe and understand animal diversity in livestock farming systems.

Scales at which animal diversity is created Type of Forms and expressed Potential or real benefits Animal diversity management: animal of animal of animal diversity modality (M), indicator (I), production diversity to the farm managers or determinant (D) Organisational Time scale scale

Production of goods (through heterosis, M: unintentional management From the complementarity of milk/ vs. intentional management I: Genotype cows’ careers Study 1. meat, milk yield/milk behavioural, genetic, physical, and Animal  Herd to the ongoing Dairy cattle content) ; Efficiency and performance D: Avoiding Phenotype campaign of (economic, labour); inbreeding and routine at work; production Resilience (to climate increasing heterosis variability)

M: three management strategies differing in their approach to animal diversity (reduction, Production of goods (milk/ segregation, amplification) meat; milk yield/milk and the ways of using and Genotype Year content); Efficiency (of Study 2. valuing it (complementarity and Animal  Herd (campaign of concentrates); Resilience Dairy cattle of production and efficiency) I: Functional production) (cows’ ability to reproduce); functional diversity supported Trade-offs between the by breed diversity D: degree above-mentioned benefits of standardisation and intensification of the livestock system

M: four management strategies differing in the degree of interaction between livestock Production of goods species I: functional diversity (cow's milk/lamb meat; Study 3. supported by species diversity Year fodder production); Dairy cattle (dairy cattle and meat sheep) Species Herd  Farm (campaign of Efficiency (feeding, labour, and meat D: work organisation, degree production) economic); Resilience sheep of standardisation and (to climate and economic intensification of the livestock variability) system, product certification under protected designation of origin

M: Sorting and directing the sale of breeding stock according to Economic efficiency for their heat-resistance phenotype I: Genotype nucleus breeding farmers; Batch of predictors of pigs' robustness to Study 4. Pigs and Animal  Herd Resistance to heat stress animals heat D: degree of standardisation phenotype and economic efficiency for and intensification of producers the livestock systems at every stage of the chain

INRA Productions Animales, 2019, numéro 2 270e / Marie-angélina magne et al.

Figure 4. Responses from 39 dairy cattle farmers regarding the dynamics of change and the benefits of animal diversity in their herds. Montbeliarde (Mb), Normande (No), Simmental (Si), and Brown Swiss (Br), Farmers were labelled generalist breeds. A herd (39) was said to be multi-breed (MB) if it

Is there animal diversity in your herd was composed of cows of both types of breeds and neither type of breed repre- Yes sented more than 80% of the total num- (37) o (2) ber of cows in the herd. An analysis of the individual performance of cows of How has this animal diversity evolved over time both breed types within 22 MB herds (of which 13, 3, 3, and 3 combined Ho with Heterogenisation Stable Homogenisation (3) (6) (28) Mb, Si, Br, and No respectively) con- firmed the functional complementarity hat benefit(s) do you see from the diversity in your herd of the two breed types with respect to - one(12) certain traits. Thus, within a herd, spe- - Heterosis (2) - o avoid inbreeding (2) - o avoid inbreeding (8) cialist-breed cows produced more milk - Heterosis (3) - o avoid routine (6) - conomic (1) (on average +1,137 kg/year) for a lon- - o be able to adapt (3) - conomic (3) - o be able to adapt (3) ger lactation length (on average +38 d) than generalist-breed cows. However, the latter group produced milk with resulted from using the same herd and adapting to the changing market- more protein (average +2.1 g/kg) and management practices (selection for place, e.g. steering production towards fat (average +2.2 g/kg) contents and replacement, culling, mating manage- milk yield or milk content according to had a higher lactation rank (+5 months). ment) for all the animals in the herd to milk prices, and buffering the impacts produce an “optimal” cow. For the nine of various hazards: “What I am saying is The benefits of managing the diver- other farmers, animal diversity was that if you have climate variability over sity of breed types within MB herds maintained or increased (n = 3) over the years, with some feeding situations were first assessed by comparing the time thanks to the use of dairy cross- that are not optimal, in a mixed herd, performance of all milk-recording MB breeding practices. some animals will resist better than oth- herds in Aveyron (n = 83) to that of ers, and if the situation is reversed, some milk-recording single-breed herds, Finally, 25 farmers indicated that animals will be able to optimise their per- whether composed of a specialist breed they derived various benefits from formance. So it should buffer the various (SB) (405 SB herds) or a generalist breed within-herd animal diversity (Figure 4). fluctuations.” (GB) (117 GB herds) (Table 2). For them, behavioural and physical diversity improved labour efficiency „„2.2. Study 2: Compared to SB herds, MB herds Management practices by facilitating the management of were shown to have a better trade- and benefits of multi-breed inbreeding, which requires the main- dairy cattle herds off among milk production (both milk tenance of a broad range of genetic yield and milk contents), reproductive variability for female replacement and Magne et al (2016) studied the per- performance (which translates into diversification of the bulls used for mat- formance and management of multi- herd resilience through the ability to ing. Breed diversity (4/37) was said to breed dairy cattle herds in Aveyron regenerate), and feed concentrate improve the production of goods in (Southern France), which represented efficiency. Specifically, while MB herds an LFS when breeds have contrasting 16% of the milk-recording farms in this produced less milk than SB herds, strengths: milk yield (Holstein) vs. milk region in 2010. The type of diversity they were more efficient in converting content (Montbéliarde, Simmental), or studied was functional diversity sup- feed-concentrates, had better repro- iconic cultural status (Abondance) and ported by genetic diversity (specialist ductive performance, and their milk high milk-yield performance (Holstein). vs. generalist breed). The Holstein breed had a higher protein content. Instead, Animal diversity was also cited as an (Ho) was defined as a milk-specialist MB herds demonstrated poorer repro- aid in managing uncertainties (6/37) breed, while other breeds, including ductive performance than GB herds, but

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /271e

Table 2. Average performance of 83 multi-breed (MB) herds, 405 single-breed herds composed of specialist-breed (SB) cows, or 117 single-breed herds composed of generalist-breed (GB) cows from milk-recording farms in Aveyron (France) in 2010.

Performance, expressed as average Single specialist Multi-breed (MB) Single generalist (standard deviation) breed (SB) herds herds * breed (GB) herds **

Milk yield (kg/cow/year) 7,497a(1,091) 6,457b(1,059) 6,028c(879)

Protein content (g/kg) 33.1c(0.9) 33.6b(1.2) 34.9a(1.1)

Fat content (g/kg) 41.8b(1.6) 42ab(1.8) 42.5a(2.1)

Somatic cell count (1,000 cells/mL) 266.9a(100.2) 265.8a(105.2) 205.8b(86.3)

Calving interval (days) 430a(30) 414b(26) 399c(27)

Concentrate distributed (kg/cow/year) 1,747a(416) 1,537b(397) 1,581b(342)

Concentrate efficiency (g/kg milk) 234b(52) 239b(50) 266a(64) a, b, c: averages with a different letter differ at the threshold of P < 0.05. *Of the 83 MB herds, 53 had Holstein as the dominant breed, 19 had Montbeliarde, 5 had Brown Swiss, and 6 had Simmental. **Of the 117 GB herds, 50 were composed of purebred Montbeliarde, 34 Simmental, 32 Brown Swiss, and 1 Normande. they produced more milk, with equiva- Ho in terms of better reproductive per- their practices according to breed type: lent protein and fat contents, and had formance and the ability to produce i) they used beef-breed crossbreeding higher feed-concentrate efficiency and milk with higher protein and fat con- on a mean of 30% of their generalist self-sufficiency. However, MB herds tents using local fodder resources and cows to optimise their growth poten- did not present any advantages with fewer feed-concentrates. Three of these tial and increase the slaughter value of respect to udder health compared to farmers even sought to combine these their calves, but they did not use this single-breed herds. functional complementarities into practice on Ho cows unless insemina- a new genotype by using rotational tion failed three times; ii) four of them Based on an analysis of farmers’ dairy crossbreeding. In accordance practiced early calving for Ho cows and management practices in the 22 MB with their desire for feed-concentrate late calving for generalist cows; and iii) herds, three groups of farmers were efficiency, farmers in group 1 prac- they adapted the amount of concen- identified based on differences in their ticed late calving (so that heifers could trates distributed to cows based on milk approaches to using, replacing, and use low-feed-value fodder resources) production. Group 3 was composed exploiting the diversity of breed types. and distributed the same amount of of eight herds: seven combined Ho Regardless of the group, farmers always feed-concentrates to all cows based and Mb cows and one used Ho and Br used the Ho breed to increase herd milk on the feed requirements of general- cows. Four farms had predominantly Ho yield. Instead, the choice of generalist ist-breed cows selected for their milk cows, three had equivalent proportions breed to pair with Ho depended on yield traits. Group 2 was composed of of both breeds, and one had predom- the exact type of functional comple- eight farms; the generalist breeds were inantly generalist-breed cows. In this mentarity that was desired, which dif- Mb (four farms), Br (two), Si (two), and case, the functional complementarity fered among the three groups. Group No (one). Five farms were predomi- desired was between the milk yield of 1 had six farms, in which the generalist nantly composed of Ho cows and three Ho cows and the disease resistance of breeds were Si (2 farms), No (2 farms), had equivalent proportions of Ho and generalist-breed cows. However, all or Br (2 farms). Three of these farms generalist-breed cows. These farmers practices were centred around the man- were predominantly based on Ho cows managed MB herds to optimise two pro- agement of Ho cows, to try to reduce (accounting for 60% to 80% of the herd) duction goals, milk and meat, and used the degree of dissimilarity between the and three were predominantly based the generalist breed for its functional two types of breeds to optimise milk on generalist-breed cows. Farmers in complementarity with Ho in terms of yield and streamline work processes. this group used the generalist breed producing less milk but more meat. For this reason, one farmer even prac- for its functional complementarity with These farmers differentiated some of ticed absorption crossbreeding. The

INRA Productions Animales, 2019, numéro 2 272e / Marie-angélina magne et al.

age at first calving was 24 months for the three groups did not significantly between the two species and between all cows, in order to minimise the length differ in terms of performance except the animals and grassland and crop of unproductive periods and prevent for milk yield/cow/year (in decreasing resources, in the context of the farms' the fattening of generalist cows. The order: group 3, 2, and 1) and feed-con- structural characteristics and the farm- amount of concentrates was adapted to centrate efficiency (equal for group 1 ers' production objectives. They also an animal’s level of production regard- and 3 and less for group 2). Meat pro- identified some options these farmers less of breed. Although their goal was duction benefits could not be assessed chose to deal with climate variability. to maintain the functional complemen- due to a lack of data. Four strategies were described for the tarities of the two breeds, the farmers management of mixed-species systems; also took care to maintain the genetic „„2.3 Study 3: Management these were characterised by differences strategies of livestock composition of the breeds because in i) the regime(s) of land use and grazing systems that combine dairy they were invested in it. cattle and meat sheep for the two species, and ii) the degree of productivity expected by the farmer for Finally, a comparison of herd perfor- Cournut et al (2012) analysed the the grassland and the herds (Figure 5). mance in these three groups provided management of 18 farms that com- Because these management strategies evidence that MB herds were beneficial bined dairy cattle and meat sheep in were based on different degrees of inte- for their demonstrated complemen- Auvergne (Central France). They char- gration and overlap between species, tarity between milk yield and feed-­ acterised, spatially and temporally, the tactics used by farmers to adapt to concentrate efficiency. Indeed, herds in how farmers managed the interactions climate hazards also differed.

Figure 5. Four different livestock management strategies for the coexistence of dairy cattle and meat sheep in the region of Auvergne, France.

Dairy Dairy Dairy Dairy cattle cattle cattle cattle

Meat Meat Meat sheep Meat sheep sheep sheep

Segregated Side-by-side Adjusted Integrated management management management management

R 2 ha of AA 14 ha of AA 13 ha of AA ha of AA

T dairy cows 6 dairy cows 35 dairy cows 3 dairy cows 5 ewes 33 ewes 33 ewes 26 ewes

R 3-4 partners 3 partners 2-3 partners ouples or individual T

S Altitude of 12 m Altitude of m Altitude of m Altitude of 6 m

AD S Permanent grassland empory grassland Maie silage Cereals T I I

T 6 litres/cow/year litres/cow/year 5 5 litres/cow/year 4 5 litres/cow/year 1,1 lambs/ewe/year 1,4 lambs/ewe/year , lambs/ewe/year , lambs/ewe/year O D R P

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /273e

In the segregated management strat- out to make use of forage left behind in dairy cattle and meat sheep farming egy (three farms), the two species were by cows. Farmers promoted their eco- systems differ according to farmers’ raised separately on different sites, nomic security in several ways: expand- management strategies, which are tied essentially operating as two distinct ing the range of animal products offered in particular to the desired compromise farms. Here, farmers prioritised the pro- (milk and cheese processed on-farm, between production, efficiency, and duction of high-quality milk for process- beef-crossbred calves, and small lambs resilience. ing into PDO cheeses (Saint-nectaire, produced at low cost), improving prod- Fourme d'Ambert) and the production uct values (through organic farming „„2.4. Study 4: Managing phenotypic of milk and sheep meat based on high- certification and/or direct sales), diver- diversity for heat tolerance er-altitude grassland. These extensive sifying on-farm activities (e.g. agritour- in pig farming systems LFSs, based on permanent grassland, ism), and increasing feed self-sufficiency had a low animal stocking rate (0.8 live- (through crop rotation including cereals In commercial pig farms, environ- stock units per ha of principal forage and temporary meadows and a low ani- mental and especially climate variabil- area (LU/PFA)). Farmers used species mal stocking rate: 0.7 LU/ha of PFA). On ity prevents farmers from realising the diversity to maximise the potential of these farms, increased species diversity full genetic potential of their animals, upland meadows for producing cow’s enabled the expansion of the range of which results in decreases in the farm's milk and sheep meat. In summer, ewes products offered, as well an increase in economic efficiency. Improvements and heifers grazed in open mountain the value of milk cows (from the use of in animal robustness, defined here as pastures, increasing the resilience of on-farm forage resources and cheese the ability of an animal to express its these LFSs to climate variability. In the processing) and low-cost lambs. The genetic potential in a wide range of side-by-side management strategy (five functional complementarity of the dif- environments, have therefore become farms), the two species were also man- ferent species was used to extract fur- a priority for the whole pig sector, aged on separate grassland areas, but ther value, through sequential grazing most particularly for nucleus breeding these pastures were all grouped around and potentially co-grazing. When con- farmers and multipliers. However, even a single farm site. Farmers prioritised the fronted with climate hazards, farmers within the same breeding line, heat sen- productivity of each species (an aver- preferentially adapted their manage- sitivity varies from one pig to another: age of 7,000 litres of milk/cow/year and ment of the sheep herd using a wide some animals have a lower and more 3 lambs/ewe/2 years) and, to support range of changes, from adjusting the variable growth rate in tropical environ- this goal, also sought to maximise crop fodder resources allocated to them to ments than in temperate environments productivity through a crop rotational decreasing their numbers. Finally, in (Figure 6). Study 4 focused on charac- system based on productive temporary the integrated management strategy terising and using this phenotypical meadows, cereals, and maize. The ani- (four farms), all grassland areas were variability to discriminate among ani- mal stocking rate here was the highest available for grazing by the two spe- mals according to their robustness (Dou of the four strategies (1.1 LU/ha of PFA) cies, with a primary focus on sequential et al., 2017). The expected benefit was and fodder was frequently purchased grazing. Crop rotation included some improvement in the resilience and eco- to overcome limitations arising from cereals and temporary grasslands, and nomic efficiency of pig systems (at the climate variability. In this case, spe- the animal stocking rate was low (0.7 multiplier or producer stages) by mak- cies diversity was used to diversify the LU/ha of PFA). Farmers’ priorities were ing animal performance less sensitive farms’ products (cow’s milk and sheep simplifying the organisation of work to climate variability. meat) and distribute them over the year. (labour efficiency) and enhancing the Fodder resources were managed inten- quality of products by using on-farm The purpose of this study was to pro- sively to meet the requirements of all fodder resources (efficiency). On these vide tools to identify the least heat-sen- animals; farmers made little use of the farms, the low production requirements sitive (i.e. most robust) individuals in an species’ functional complementarity expected for all animal species rendered animal population using biomarkers to better exploit pasture resources. In the entire farm resistant to climate and from easily accessible sources (such as the adjusted management strategy (six economic variability. blood, saliva, or faeces). It showed that farms), both species grazed their own robust or heat-sensitive animals can be dedicated pastures, but sequential graz- Overall, this study revealed that the identified with good precision from a ing (cows followed by ewes) was carried benefits derived from mixed-species blood sample and plasma metabolome

INRA Productions Animales, 2019, numéro 2 274e / Marie-angélina magne et al.

Figure 6. Live weight distribution at 23 weeks of age in two populations of pigs that were actually used by dairy cattle of the same line (backcross LWxCR) reared in temperate (blue, n = 634 pigs) or tropical conditions (orange, n = 664 pigs; adapted from Rosé et al. 2017). farmers to characterise, manage, and 4.5 assess the benefits of animal diversity in their herds; however, it did not test their 4.0 objectivity or validity. In studies 2 and 3, characterisation of animal diversity was 3.5 defined a priori based on the functional 3.0 complementary of breed types (study 2) or theoretical species (study 3); the valid- 2.5 ity of these descriptions were then eval- uated a posteriori through assessments 2.0 of performance between breed types 1.5 within MB herds and through farmers’

Freuency () narratives, respectively. Finally, studies 1.0 1 and 3 identified the benefits derived from the management of animal diver- 0.5 sity on farms based on farmers’ narra- 0.0 tives. Instead, study 2 assessed benefits 40 60 80 100 120 140 by (i) comparing the performance of MB Pigs live weight at 23 wees of age (g) herds with that of single-breed herds ropical emperate and (ii) comparing farmers’ management ­strategies among different MB herds. analysis. Although these markers have its four components. Study 4 focuses not yet been tested in farm manage- mainly on the first two components (the The analysis also enables identifica- ment, they could be used by multipli- forms of animal diversity and the organ- tion of the partial contributions of each ers to guide the phenotypic diversity of isational scale at which it originates), research study to three main research breeding stock and avoid (for example) because the challenge in that study was areas: characterising animal diversity the export of heat-sensitive animals to to objectively characterise the functional and understanding the underlying hot regions. These predictive biomark- diversity of pigs (thermotolerance) at the biological and ecological mechanisms; ers could also be used to identify the scales at which it is created, based on the characterising management options genetic markers influencing the com- definition of predictors intended to pro- for animal diversity to take advantage plex trait of heat sensitivity for applica- vide tools for management in LFSs. This of it in LFSs; and assessing the benefits tions in breeding programmes, which study therefore represents processes derived from the management of ani- could construct a robustness index of that are upstream of the characterisation mal diversity in LFSs, including their breeding animals. In commercial farms, of management practices and the assess- trade-offs (Table 3). Based on a brief this phenotypic diversity could be man- ment of benefits in LFSs. Conversely, review of the literature, in addition to a aged by selecting animals for reproduc- the three other studies reviewed here thorough analysis of these four research tion who are adapted to the targeted focused mainly on the other two com- studies, we identified the need for a pre- climatic context. ponents of the conceptual framework, cise description of these research areas related to animal diversity manage- and we formulated a set of concrete 3. Cross-sectional analysis ment in LFSs and its benefits. However, research directions for each. of the four studies each study took a different approach and research perspectives to these two components (Table 3). „„3.1. Better characterising and understanding animal Studies 1 and 3 were based exclusively diversity for management on farmers’ narratives (opinions and/ in livestock farming systems The analysis of these four studies or practices), while study 2 combined based on the conceptual framework farmers’ narratives with quantitative We propose two areas for future reveals that animal diversity is often analyses of within-herd performance research to better characterising and incompletely studied with respect to data. Study 1 identified key indicators understanding animal diversity in LFSs.

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /275e

Table 3. Types of research approaches applied in each of the four studies reviewed using the conceptual framework, and the contribution of each to three research areas identified for the integrated analysis of animal diversity in livestock farming systems.

Study 2 Study 3 Study 1 Study 4 Multibreed Dairy cattle Dairy cattle Pigs dairy cattle and meat sheep

Farmers’ narratives X X X Research (opinions and/or practices) approaches used Data on herd/animal X X

Characterising animal diversity and understanding the underlying ** ** * *** mechanisms

Research Characterising options for animal areas * *** *** diversity management

Assessing the benefits derived *** *** * from the use of animal diversity

The number of stars reflects the degree of contribution of each of the four studies to the three research areas identified as key for the understanding of the management of animal diversity in livestock farming systems : * marginal; ** important; *** major

The first is to better describe and parameters/biomarkers) that are used to ­diversity. Here, the frameworks devel- understand functional diversity and its categorise functional groups of animals oped in ecology by Biggs et al. (2012) and links with other forms of animal diver- (or describe functional diversity) within Vigglizo (1994) to describe biodiversity sity (Figure 2). Indeed, the four studies LFSs and to understand the biological could help to refine the differentiation examined here reveal that a key role in and ecological mechanisms behind and characterisation of animal entities LFSs is played by functional diversity them. Here, in reviewing the approaches on a farm. These authors proposed the supported by other forms of animal used in the literature for assigning ani- identification of: the variety of the differ- diversity. That is, increased species diver- mals to functional groups within an LFS, ent entities, i.e. their nature and number; sity, such as an association between we identified several limitations. The first their relative abundance; their degree of two livestock species for co-grazing, is is the lack of available and easily acces- dissimilarity; and the nature and inten- only beneficial if the species have com- sible variables to carry out this type of sity of interactions between them. The plementary grazing abilities (e.g. diet characterisation. Indeed, this was the concept of ‘variety’ corresponds to the preferences, grass heights in pastures) main challenge in study 4, the aim of forms of animal diversity in our concep- that make different contributions to the which was to identify biomarkers that tual framework and was used in each functioning of the LFS as a whole (e.g. can be easily measured in bodily fluids of the four reviewed research studies. regulating the dynamics of vegetation such as blood and urine for the descrip- By contrast, relative abundance was in time and space, regulating gastro- tion of pigs according to their heat sen- only addressed in study 2, in analyses intestinal strongyles). As has already sitivity. In studies 2 and 3, individuals of the proportions of generalist and been discussed in the fields of ecology, were grouped based on the theoretical specialist cows within MB herds. The livestock production, and agronomy functional complementarity of animals; degree of dissimilarity refers to the (Petchey et al., 2004; Tichit et al., 2011; however, it was not always possible to variables used to differentiate animals Duru et al., 2013, respectively), it is func- test these assumptions due to the lack of into functional groups (e.g. heat stress tional diversity that must be studied available data on animal performance. response of pigs in study 4, differences rather than species richness per se. To overcome this limitation, it would in the biological abilities of generalist be helpful to have routine monitoring and specialist cows in study 2). Finally, The second research direction is of useful phenotypic or genetic param- the nature and intensity of interactions closely related to the first: to refine eters in livestock farming. The second between entities refers to the functional the variables (e.g. morphological and limitation regards the type of variables complementarity effects we identified behavioural traits; animal performance; that can be used to characterise animal in the reviewed studies, such as the

INRA Productions Animales, 2019, numéro 2 276e / Marie-angélina magne et al.

functional complementarity­ in grazing an ­example, in study 4, biomarkers for The second research direction is to between meat ewes and dairy cows in heat sensitivity in pigs were developed analyse the changes in farmers' work mixed-species herds or the complemen- for use in sorting breeding stock before (physical, interpersonal, and cognitive tarity between milk yield and milk con- they were sold to producers. Depending work) that are linked with the manage- tent in MB herds. However, other types on the geographical origin (temperate ment of animal diversity in LFSs. In stud- of interactions, such as those commonly or tropical) or season in question, mul- ies 1 and 3, work requirements were studied in ecology (e.g. neutralism, coop- tipliers can use this tool to enhance shown to determine certain choices eration, competition, learning) remain to the value of animal diversity on their regarding animal diversity manage- be described and documented in LFSs. farms and provide breeding animals or ment. In the literature, managing agro- The frameworks proposed by Biggs semen that are adapted to the climatic biodiversity in agricultural systems is et al. (2012) and Vigglizo (1994)can thus conditions of different pig producers. often mentioned as being accompanied be used to better characterise animal This strategy would lead to a reduction by a necessary reorganisation of work diversity, regardless of the organisa- in the diversity of traits associated with (Duru et al., 2015; Darnhofer et al., 2010). tional scale under consideration (e.g. heat-stress robustness within a given However, no study has yet provided herd, farm, territory, country). producer's farm, but would maintain empirical evidence for this, particularly trait diversity at the population scale when the agrobiodiversity under study „„3.2. Characterising animal (regardless of the stage considered). is animal diversity in LFSs. Specifically, diversity management In the future, such biomarkers could farmwork should be studied in both its options for long-term benefits be used as phenotypes to identify the (i) organisational dimensions (How are genetic markers related to heat toler- the tasks to manage animal diversity, To better characterise different ance. The use of these genetic markers and the workforce available for them, options for the management of animal in breeding programmes would then organised in space and time?) and (ii) diversity and maximise the long-term make it possible to select animals that cognitive dimensions (What informa- benefits that can be derived from them, are less sensitive to climate variability. tion do farmers use to manage animal we propose three directions for future However, such a selection process could diversity? How do they construct new research. eventually lead to a reduction in animal indicators to manage it and adapt their diversity not only on individual farms, objectives according to the benefits The first is to more thoroughly explore which may prefer pigs that are robust they achieve?). the effects of basic and combined live- in any climatic environment, but also stock management practices on animal in the animal population as a whole. The final research direction is to diversity, and vice versa, at different Finally, as shown in studies of crop-live- identify factors that help or hinder the time and organisational scales. Indeed, stock integration (Moraine et al., 2017), ­management of animal diversity in LFSs. our analysis consistently identified a a key issue in the management of ani- In the studies reviewed here, some of lack of formal knowledge of the prac- mal diversity is the spatial scale under these factors became apparent, such as tices (and combinations thereof) that consideration (within-farm, groups of market organisation and liberalisation, make it possible to create animal diver- farms, territory, etc.). The development the organisation of animal genetic selec- sity, to use/enhance it, and to regener- of management programmes whose tion, and farmers' standards. However, ate/replace it over time. An example of scope extends beyond the scale of few studies have explicitly analysed this type of knowledge was provided by an individual farm would aid in over­ the factors that support or impede the study 2, which identified and described coming barriers related to specialisa- management of animal diversity in LFSs three management strategies in MB tion in livestock sectors, farms, and (Magne et al., 2017). The challenge is herds. Another critical question regards farmers' skills. However, the organisa- therefore to identify and categorise them the effects of such management prac- tional barriers present at such a scale according to whether they arise from tices over the long term and at differ- are not insignificant; addressing these human (e.g. values, standards, knowl- ent organisational scales, depending would require a collective restructuring edge), socio-economic (e.g. market on the type of animal sector under of animal configurations, management, ­organisation), technological (e.g. pheno- ­consideration: from producers to and marketing practices at the farm typing techniques), or biophysical (e.g. multipliers and nucleus breeders and scale to enhance animal diversity at the level of exposure to climatic, economic, from LFSs to animal populations. As ­territorial scale. and health hazards) influences. It will be

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /277e

critical to analyse these elements not similar to those raised by multi-criteria asked regarding the management of only from the point of view of farmers assessments (Lairez et al., 2017). In par- animal diversity in LFSs. Its application to but also from the point of view of other ticular, systems and methods must be case studies could facilitate discussion stakeholders in the socio-technical envi- ­developed to acquire and process herd between experts from different back- ronment of LFSs. performance data of different types (e.g. grounds around the steps to be taken health, reproduction, production of dif- toward this goal. Indeed, its use here on „„3.3. Assessing the benefits ferent products, product quality, work different research studies has initiated of managing animal efficiency and quality) and at different these types of discussions between diversity in livestock farming systems time and organisational scales. livestock production researchers­ with different research objectives and work- Our comparative analysis of the four The second research direction is an ing on different kinds of livestock sys- studies revealed that research efforts improved integration of the coevolu- tems. The discussion space provided should focus on assessing the benefits­ tion of management practices, animal by the framework can be opened up to derived from the management of animal diversity, and LFS performance into other researchers and, more broadly, to diversity in LFSs in order to broaden the the assessment of the benefits derived. other ­stakeholders in livestock-related base of evidence for its use. For exam- Assessments of benefits must take into sectors. ple, in study 4, the expected ­benefit of account two key challenges: i) the influ- using heat-tolerant animals is increased ence of the techniques used by ­farmers Specifically, it would be useful to resilience on the producers’ farms. to manage animal diversity; and ii) their share and discuss the framework with This outcome, though, remains to be evolution over the medium to long researchers in: ­experimentally verified; the effective term, to account for the dynamics of resilience of such farms must be com- the biological processes underlying (i) livestock production, working on pared to those composed of pigs with animal diversity and their interactions species or production systems that different degrees of robustness. Similarly, with farming practices. This last point were not examined here (e.g. poultry in studies 1 and 3, the benefits that is particularly important when seek- or fish). This could confirm the general ­farmers said they derived from ­managing ing to assess the resilience derived applicability of the conceptual frame- animal diversity on their farms need to be from animal diversity management. work and highlight differences in the assessed using quantitative data analy- However, current research methods management of some under-­explored sis. Based on these observations, we pro- are not always suitable (or available) types of animal diversity, such as pose two research directions to increase for long-term assessment. For this, two monogastric-ruminant associations or knowledge on the benefits of managing ­promising tools are livestock mon- intra- or interspecific functional com- animal diversity in LFSs. itoring and modelling/simulation. plementarities that were not described Modelling also has the advantage of in the studies reviewed here; The first is to assess the benefits being able to separate the manage- derived from the management of ani- ment component of livestock farming (ii) genetics, to merge and communi- mal diversity in LFSs in combination, systems from the biological component cate research on animal diversity at the not in isolation. Indeed, studies have (Tichit et al., 2011; Blanc et al., 2013). scales of LFSs and animal populations shown that what is expected or gained (Phocas et al., 2017). Indeed, it appears from the management of animal diver- 4. Scientific and practical that at these two scales of creation, the sity in LFSs is most often a combination implications organisation and use of animal diversity of ­several benefits (e.g. production are intimately linked, even though they of goods, resilience, efficiency, self-­ „„4.1. An integrative are often considered independently; sufficiency in terms of inputs, quality of framework to support life at work) rather than just one. These discussion between (iii) microbiology and veterinary medi- combinations can also involve different stakeholders cine and agronomy, to better formalise trade-offs that livestock farmers must the relationships between animal be willing to make. Before such assess- The conceptual framework we pro- diversity in LFSs and the associated ments can be conducted, though, meth- pose here is a mental map of elements biodiversity. Such biodiversity can be odological questions must be addressed, to be integrated and questions to be microbial diversity hosted by farm ani-

INRA Productions Animales, 2019, numéro 2 278e / Marie-angélina magne et al.

mals (see the review of Calenge et al., potential, for three reasons. First, the the ­often-­negative view that these 2014 on the links between the microbi- framework structures the analysis of stakeholders have of animal diversity ota and animal phenotypes), microbial animal diversity­ management in LFSs in LFSs, which tends to be unwanted diversity hosted by the environment around four unambiguous components and minimised because it is thought to ­surrounding the animal (building, pas- (Figure 2). Second, the four studies be inefficient in the current socioeco- ture), or even cultivated or uncultivated examined with respect to these four nomic context. For example, consul- plant diversity; components constitute concrete cases tants often consider multi-breed dairy that help to establish the framework. cattle herds to be in a phase of transi- (iv) human and social sciences, parti­ Finally, the cross-analysis of these four tion to single-breed herds, or that they cularly economics, to enable more accu- studies enables the identification of are a consequence of the purchase of a rate assessments of the benefits derived commonalities and differences in the second herd, whereas for many farm- from animal diversity in LFSs and to management of animal diversity in ers, multi-breed herds are beneficial better understand the socio-technical LFSs. These three kinds of elements and should be maintained. In addition, and sociological determinants of dif- have been shown to facilitate learning the conceptual framework can be used ferent ways of managing it (or the lack and understanding of complex sub- by and with consultants for a collective thereof) in LFSs. jects (Barth, 2013), such as the man- reimagining of the data that should be agement of animal diversity in LFSs. collected in LFSs to produce bench- The framework can also be used The framework could therefore be marks on the benefits (or lack thereof) by researchers and stakeholders in used by teachers of livestock produc- of managing animal diversity. It would the livestock sector (e.g. farmers, tion in universities or technical agricul- then be useful for advisory organisa- consultants, breeding organisations, tural schools to provide students with tions to fund, build, and monitor, over downstream operators) to express the ability to apply biodiversity, one the long term, networks for livestock and promote different points of view of the key agroecological concepts, to farmers centred around the use of (convergences and divergences) and the management of LFSs. Indeed, the animal­ diversity. to integrate these into the design of French curriculum for technical agricul- different options for managing animal tural schools specifically mentions the Conclusion diversity in LFSs. goal of training future livestock farm- ers and livestock consultants to con- „„4.2. Contribution sider animal diversity as a resource for In this article, using a review of and implications for training managing LFSs from an agroecological research studies conducted at INRA and advice in livestock farming systems perspective. Finally, the framework can and a brief overview of the literature also serve as a tool for the exchange of on LFS research, we developed an orig- This article is not an exhaustive and information on the current barriers and inal conceptual framework for an inte- systematic review of the literature on support systems for the management grated analysis of animal diversity and animal diversity and its management of animal diversity in LFSs. In doing so, its management at the scale of LFSs. in LFSs, and is not intended to define it provides an opportunity to imagine We then proposed some directions “good management practices of ani- technical, organisational, cognitive, and for future research to identify the con- mal diversity”, as these must be con- economic means for overcoming cur- ditions under which animal diversity sidered in context. However, we think rent barriers. could be beneficial in the long term. In that the conceptual framework we so doing, we promote a formal concept developed is generic enough to be For livestock consultants, the appli- of the management of animal diversity applied to different livestock farming cation of the framework to the four in LFSs, a key concept in agroecology situations and used by different live- studies presented here serves as a but one that livestock stakeholders, stock stakeholders, particularly teach- concrete example of how animal including researchers, find difficult to ers and consultants. diversity can be managed in LFSs and grasp and use. By applying the frame- provides evidence that this diversity work to four research studies, we illus- The framework, and more gener- can be successfully managed by farm- trate the forms that animal diversity ally the approach underlying its use ers and be beneficial to them. This can take at different organisational and in this article, has strong educational makes it possible to counterbalance time scales, the way it can be managed

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /279e

in LFSs, and the benefits that can be diversity that are beneficial to them in in livestock production science. Her derived from it for farmers. This con- their individual contexts and over the work on the role and services provided ceptual framework is a tool for inte- long term. by biodiversity in agriculture was pio- grating the knowledge produced, or neering and opened up innumerable to be produced, on the management Acknowledgements research perspectives. Muriel, thank of animal diversity in LFSs and is an you for the scientific legacy you have aid for discussions and plans for future left us on this theme. research projects, both theoretical and We dedicate this article to Muriel applied. It is also a resource that can Tichit, who died in April 2019. She This article was translated with be used in training and advising farm- served as research director at INRA www.DeepL.com/Translator. The ers on the design and development of within the Department of Action and authors warmly thank Lindsay Higgins options for the management of animal Development Sciences and held a PhD for English language editing.

References Altieri M.A., 1987. Agroecology: the scientific basis of Dedieu B., Faverdin P., Dourmad J.Y., Gibon A., 2008. lenges for dairy cow production systems arising from alternative agriculture. Westview Press, Boulder CO, Système d’élevage, un concept pour raisonner les climate change in Central Europe – a review. Animal, USA, 227p. transformations de l’élevage. In : Numéro spécial, 20 7, 843-859. ans de recherches en productions animales à l’INRA. Audiot A., 1995. Races d’hier pour l’élevage de demain Hoffmann I., 2013. Adaptation to climate change--ex- re Charley B., Herpin P., Perez J.M. (Éds). INRA Prod. 1995, 1 Édition, INRA Éditions, Paris, France, 229p. Anim. 21, 45-58. ploring the potential of locally adapted breeds. Animal, 7, 346-362. Biggs R., Schlüter M., Biggs D., Bohensky E.L., Delaby L., Faverdin P., Michel G., Disenhaus C., Peyraud BurnSilver S., Cundill G., Dakos V., Daw T.M., Evans J.L., 2009. Effect of different feeding strategies on lac- Knaus W., 2009. Dairy cows trapped between per- L.S., Kotschy K., Leitch A.M., Meek C., Quinlan A., tation performance of Holstein and Normande dairy formance demands and adaptability. J. Sci. Food Raudsepp-Hearne C., Robards M.D., Schoon M.L., cows. Animal, 3, 891-905. Agricult., 89, 1107-1114. Schultz L., West P.C., 2012. Toward principles for enhancing the resilience of ecosystem services. Annu. Dou S., Villa-Vialaneix N., Liaubet L., Billon Y., Giorgi Lairez J., Feschet P., Botreau R., Bockstaller C., Fortun- Rev. Environ. Resour., 37, 421-448. M., Gilbert H., GourdineJ.L., Riquet J., Renaudeau D., Lamothe L., Bouvarel I., Aubin J., 2017. L’évaluation 2017. 1HNMR-Based metabolomic profiling method multicritère des systèmes d’élevage pour accompa- Blanc F., Ollion E., Puillet L., Delaby L., Ingrand S., to develop plasma biomarkers for sensitivity to chro- gner leurs évolutions : démarches, enjeux et questions Tichit M., Friggens N.C., 2013. Évaluation quantitative nic heat stress in growing pigs. PLoS ONE 12(11): soulevées. INRA Prod. Anim., 30, 255-268. de la robustesse des animaux et du troupeau : quels e0188469. principes retenir ? Renc. Rech. Rum., 20, 265-272. Mackey D.R., Gordon A.W., McCoy M.A., Verner M., Mayne C.S., 2007. Associations between genetic Dumont B., Fortun-Lamothe L., Jouven M., Thomas Brussaard L., Caron P., Campbell B., Lipper L., Mainka merit for milk production and animal parameters M., Tichit M., 2013. Prospects from agroecology and S., Rabbinge R., Babin D., Pulleman M., 2015. and the fertility performance of dairy cows. Animal, industrial ecology for animal production in the 21st Reconciling biodiversity conservation and food secu- 1, 29-43. century. Animal, 7, 1028-1043. rity : scientific challenges for a new agriculture. Curr. Opin. Environ. Sustain., 2, 34-42. Magne, M.A., Cerf, M., Ingrand, S., 2011. Comment Duru M., Therond O., 2015. Livestock system sustain- les éleveurs choisissent-ils et utilisent-ils des informa- Cournut S., Conrard A., Bertrand J., Ingrand S., 2012. ability and resilience in intensive production zones : tions pour conduire leur exploitation ? Cah. Agricult., Intérêt de la mixité d’espèces pour accroître la flexibilité which form of ecological modernization ? Reg. 20, 421-427. des élevages : l’exemple des élevages bovin lait + ovin Environ. Change, 15, 1651-1665. viande en Auvergne. Renc. Rech. Rum., 19, 273-276. Magne M.A., Thénard V., Mihout S., 2016. Initial Duru M., Jouany C., Roux X., Navas M.L., Cruz P., 2013. insights on the performances and management D’Alexis S., Sauvant D., Boval M., 2014. Mixed grazing From a conceptual framework to an operational of dairy cattle herds combining two breeds with systems of sheep and cattle to improve liveweight gain: approach for managing grassland functional diversity ­contrasting features. Animal, 10, 892-901. a quantitative review. J. Agricult. Sci., 152, 655-666. to obtain targeted ecosystem services: Case studies from French mountains. Renew. Agr. Food Syst., 29, Magne M.A., Ollion E., Cournut S., Mugnier S., Sabatier Damour G., Navas M.L., Garnier, E., 2018. A revised 239-254. R., 2017. Some key research questions about the trait-based framework for agroecosystems including interest of animal diversity for the agroecological decision rules. J. Appl. Ecol., 55, 12-24. Duru M., Therond O., Martin G., Martin-Clouaire R., transition of livestock farming systems. In: First Magne M.A., Justes E., Journet E.P., Aubertot J.N., Agroecol. Europe Forum Fostering synergies between Darnhofer I., Bellon S., Dedieu B., Milestad R., 2010. Savary S., Bergez J.E., Sarthou J.P., 2015. How to movement, Science and Practice, 25-27 October 2017, Adaptiveness to enhance the sustainability of farming implement biodiversity-based agriculture to enhance Lyon, France, p70. systems. A review. Agron. Sustain. Dev., 30, 545-555. ecosystem services: a review. Agron. Sustain. Dev., 35, 1259-1281. Malassis L., 1979. Economie agro-alimentaire : tome I : Dedieu B., Ingrand S., 2010. Incertitude et adapta- économie de la consommation et de la production tion : cadres théoriques et application à l’analyse de Gauly M., Bollwein H., Breves G., Brügemann K., agroalimentaire. Édition Cujas, Paris, France, 437p. la dynamique des systèmes d’élevage. In : Robustesse, Dänicke S., Daş G., Demeler J., Hansen H., Isselstein rusticité, flexibilité, plasticité, résilience… les nou- J., König S., Lohölter M., Martinsohn M., Meyer Martin-Rosset W., Trillaud-Geyl C., 2011. Pâturage veaux critères de qualité des animaux et des systèmes U., Potthoff M., Sanker C., Schröder B., Wrage N., associé des chevaux et des bovins sur des prairies d’élevage. Sauvant D., Perez J.M. (Éds). Dossier, INRA Meibaum B., von Samson-Himmelstjerna G., Stinshoff permanentes : premiers résultats expérimentaux, Prod. Anim., 23, 81-90. H., Wrenzycki C., 2013. Future consequences and chal- Fourrages, 207, 211-214.

INRA Productions Animales, 2019, numéro 2 280e / Marie-angélina magne et al.

Mazoyer M.L., 1982. Origines et mécanismes de repro- Séminaires Méditerranéens (115). CIHEAM Eds. FAO- G., Frappat B., González-García E., Hazard D., Larzul duction des inégalités régionales de développement CIHEAM Network for Research and Development in C., Lubac S., Mignon-Grasteau S., Moreno-Romeux agricole en Europe. Écon. Rurale., 150, 25-33. Sheep and Goats, Montpellier, France, 633-636. C., Tixier-Boichard M., Brochard M., 2017. Quels pro- grammes d’amélioration génétique des animaux pour Meisser M., Frey C.F., Deléglise C., Mosimann E., Nozières M.O., Moulin C.H., Dedieu B., 2011. The herd, des systèmes d’élevage agro-écologiques ? INRA Prod. 2013. Pâturage mixte bovins – ovins en moyenne a source of flexibility for livestock farming systems Anim., 30, 31-46. montagne : trois années d’essai dans le Jura suisse. faced with uncertainties? Animal, 5, 1442-1457. Fourrages, 216, 305-311. Reboud X., Malezieux E., 2015. Vers une agroécologie Ollion E., 2015. Evaluation de la robustesse des vaches biodiverse : enjeux et principaux concepts mobilisés. Moraine M., Duru M., Therond O., 2017. A social-eco- laitières : entre aptitudes biologiques des animaux et Innov. Agron., 43, 1-6. logical framework for analyzing and designing inte- stratégies de conduite des éleveurs. Thèse de doctorat. grated crop-livestock systems from farm to territory Université Blaise Pascal-Clermont-Ferrand II, France. Sørensen M.K., Norberg E., Pedersen J., Christensen levels. Renew. Agric. Food Syst., 32, 43-56. L.G., 2008. Invited review: crossbreeding in dairy cat- Oltenacu P.A., Broom D.M., 2010. The impact of tle: a Danish perspective. J. Dairy Sci., 91, 4116-4128. Néori A., Chopin T., Troell M., Buschmann A.H., genetic selection for increased milk yield on the wel- Kraemer G.P., Halling C., Shpigel M., Yarish C., 2004. fare of dairy cows. Anim. Welf., 19, 39-49. Tichit M., Puillet L., Sabatier R., Teillard F., 2011. Integrated aquaculture: rationale, evolution and state Multicriteria performance and sustainability in live- of the art emphasizing seaweed biofiltration in mod- Petchey O.L., Hector A., Gaston K.J., 2004. How do stock farming systems: Functional diversity matters. ern mariculture. Aquaculture, 231, 361-391. different measures of functional diversity perform? Livest. Sci., 139, 161-171. Ecology, 85, 847-857. Nozières M.O., 2014. La commercialisation des pro- Tichit M., Puillet L., Martin O., Douhard F., Friggens duits, source de flexibilité pour les éleveurs ? le cas Phocas F., Bobe J., Bodin L., Charley B., Dourmad J.Y., N.C., Sauvant D., 2012. Livestock farming and uncer- de l’élevage ovin allaitant en Languedoc-Roussillon., Friggens N.C., Hocquette J.F., Le Bail PY., Le Bihan- tainties: exploring resilience with viability tools. In: Thèse de doctorat. Montpellier SupAgro, France. Duval E., Mormède P., Quéré P., Schelcher F., 2014. 63rd Ann. Meet. Europ. Assoc. Anim. Prod., Bratislava, Des animaux plus robustes : un enjeu majeur pour Slovakia, 338. Nozières M.O., Moulin C.H., 2016. L’hétérogénéité le développement durable des productions animales biologique des agneaux : Une contrainte à gérer nécessitant l’essor du phénotypage fin et à haut débit. Vigglizo E., 1994. The Response of Low-Input ou un atout à valoriser ? In: The value chains of In : Phénotypage des animaux d’élevage. Phocas F. Agricultural Systems to Environmental Variability. A Mediterranean sheep and goat products. Organisation (Ed). Dossier, INRA Prod. Anim., 2, 181-194. Theoretical Approach. Agricult. Syst., 44, 1-17. of the industry, marketing strategies, feeding and pro- duction systems. Options Méditerranéennes. Série A: Phocas F., Belloc C., Bidanel J., Delaby L., Dourmad Weiner J., 2003. Ecology — the science of agriculture J.Y., Dumont B., Ezanno P., Fortun-Lamothe L., Foucras in the 21st century. J. Agricult. Sci., 141, 371-377. Abstract Biodiversity, and particularly animal diversity managed by farmers, is often viewed as a valuable resource for effecting the agroecological transition and improving the resilience of livestock farming systems (LFSs). However, this topic has not been thoroughly studied; what little research is available tends to be recent and scattered. This paper aims to develop a conceptual framework for an integrated view of the management of animal diversity in LFSs and identify some avenues for future research on this topic. The framework addresses four interrelated components: the form(s) of animal diversity under consideration, the organisational and temporal levels at which it is created and has relevance, the forms of its management in LFSs, and the benefits for farmers that arise from this management. Here, we apply this framework to four research studies that took contrasting approaches to the topic of animal diversity. Through the evaluation of these four components, a cross-analysis of these case-studies revealed that this approach enables the management of animal diversity in LFSs in an integrated way and ensures its benefits in the long term. Our work also highlights three novel research directions aimed at strengthening the characterisation of animal diversity, the understanding of its management in LFSs, and the assessment of the various benefits derived from its management. Finally, some applications for the use of the framework in research, teaching, and consulting on LFSs are suggested.

Résumé Gérer la diversité animxxale dans les systèmes d’élevage : laquelle, comment et pour quels bénéfices ? La biodiversité, et en particulier la diversité animale, est présentée comme un levier prometteur pour la transition agroécologique et la résilience des systèmes d’élevage. Or, les travaux traitant de cette question sont peu nombreux, récents et épars. Cet article vise à développer un cadre conceptuel pour analyser de manière intégrée la diversité animale et ses modalités de gestion à l’échelle des systèmes d’élevage et à proposer des pistes de recherche pour y contribuer. Ce cadre est structuré en quatre composantes : i) les formes que recouvre la diversité animale, ii) les niveaux organisationnel et temporel auxquels elle se construit et s’exprime, iii) ses modes de gestion et iv) les bénéfices retirés par l’éleveur. Quatre études de recherche contrastées en termes de diversité animale analysée ont été revisitées au travers du cadre conceptuel proposé. Leur relecture et leur analyse transversale montrent l’intérêt d’articuler les quatre composantes du cadre pour raisonner de manière intégrée la gestion de la diversité animale en élevage afin d’en tirer parti sur le long terme. Elles permettent aussi d’identifier trois fronts de recherche à investiguer conjointement : affiner la caractérisation de la diversité animale en élevage, mieux caractériser ses modes de gestion y compris ses déterminants et approfondir l’évaluation des différents bénéfices retirés de sa gestion. Des pistes d’utilisation du cadre en recherche, en enseignement et dans conseil en élevage sont enfin proposées.

MAGNE M.-A., NOZIÈRES-PETIT M.-O., COURNUT S., OLLION É., PUILLET L., RENAUDEAU D., FORTUN-LAMOTHE L., 2019. Managing animal diversity in livestock farming systems: which diversity? Which forms of management ­practices? For which benefits? INRA Prod. Anim., 32, 263e-280e. https://doi.org/10.20870/productions-animales.2019.32.2.2496

INRA Productions Animales, 2019, numéro 2 Managing animal diversity in livestock farming systems: types, methods and benefits /281e

INRA Productions Animales, 2019, numéro 2