You are on page 1of 9

ARTICLE IN PRESS

G Model
EURAGR-25216; No. of Pages 9

Europ. J. Agronomy xxx (2013) xxx–xxx

Contents lists available at ScienceDirect

European Journal of Agronomy


journal homepage: www.elsevier.com/locate/eja

Agroecological principles for the redesign of integrated crop–livestock


systems
Thierry Bonaudo a,∗ , Amaury Burlamaqui Bendahan b , Rodolphe Sabatier a ,
Julie Ryschawy a , Stéphane Bellon c , François Leger a , Danièle Magda d , Muriel Tichit a
a
AgroParisTech, umr 1048 SADAPT, Paris, France
b
EMBRAPA Roraima, Boa Vista, RR, Brazil
c
INRA, ur 767 Ecodéveloppement, Avignon, France
d
INRA, umr 1248 AGIR, Toulouse, France

a r t i c l e i n f o a b s t r a c t

Article history: Combining crops and livestock within integrated crop–livestock systems (ICLS) represents an opportunity
Received 2 April 2013 to improve the sustainability of farming systems. The objective of this paper is to analyse how agroeco-
Received in revised form 2 August 2013 logical principles can help farmers to redesign and improve the resilience, self-sufficiency, productivity,
Accepted 18 September 2013
and efficiency of ICLS. Relying on case studies from Brazil and France, we examine how the transforma-
tion of two conventional, specialised systems into more integrated-production systems illustrates the
Keywords:
different dynamics towards agroecological ICLS. The French case study, based on self-sufficient farming
Integrated crop–livestock systems
systems belonging to a sustainable agriculture network, highlights that cost-cutting management led to
Agroecology
Brazil
a win–win strategy comprising good economic and environmental performances. The farms decreased
France their dependence on external inputs and had only a limited loss of production. The past trajectories of the
Transition farms illustrate how increasing the interactions between subsystems improved the self-sufficiency and
efficiency of the farms. The Brazilian case study compares slash-and-burn agriculture in the Amazonian
region with the recovery of degraded grazing area by ICLS. A small increase in chemical inputs linked to a
diversification of productions led to a large increase in production and a large decrease in environmental
impacts (deforestation). The Brazilian case study also illustrates how the diversification of production
increased the resilience of the system to market shocks. Reconstructing the links among soil, crops, and
animals following agroecological principles could improve the different performances of ICLS. New agroe-
cological ICLS, benefiting from diversified productions and increased interactions between subsystems,
are likely to offset the trade-off between agricultural production and environmental impacts observed
in current ICLS.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction entails the use of nonfarming areas, promoting fertility transfer


across diverse and heterogeneous landscapes. In such systems,
Integrated crop–livestock systems (ICLS) are considered an effi- animals play a key role in recycling and increasing the efficiency
cient design for sustainable, ecologically based farming systems of resource usage (Schiere and Kater, 2001; Schiere et al., 2002).
(Gliessman, 2006; Russelle et al., 2007; Hendrickson et al., 2008a). Beyond the interactions between crop and livestock, the elabo-
The sustainability of ICLS relies on the complementarities between ration of diversified production is likely to favour economies of
crops and livestock and the connectedness of livestock to the land scope and minimise the need for external inputs (Ryschawy et al.,
(Naylor et al., 2005; Wilkins, 2008). The complementarities first 2012; Vermersch, 2007; Wilkins, 2008). Integrating complemen-
correspond to animal abilities to valorise natural and cultivated tary activities and promoting functional diversity (Tichit et al.,
vegetal resources, especially nonfood biomass (Schiere et al., 2002). 2011) are also acknowledged to increase the self-sufficiency of ICLS
Animals eat and convert products, by-products, and residues that compared with specialised farming systems (e.g. farms specialised
are not suitable for human consumption; and they maintain soil producing for livestock feed, human food, and/or economic bene-
fertility and crop production by recycling faeces. Animal mobility fits) (Vermersch, 2007; Wilkins, 2008).
The potential benefits of ICLS entail strong interactions between
the plant and animal components of the system; but under a unique
∗ Corresponding author. Tel.: +33 1 44 08 18 00; fax: +33 1 44 08 18 53. definition of ICLS, a wide variety of contrasting mixed-farming sys-
E-mail address: thierry.bonaudo@agroparistech.fr (T. Bonaudo). tems exist worldwide (Schiere and Kater, 2001). Ewing and Flugge

1161-0301/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.eja.2013.09.010

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

2 T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx

(2004) link the wide variety of ICLS to contrasting social and pae- rely on external input supplies to increase the production function
doclimatic contexts. ICLS are among the major and most diversified to systems that rely on the management of metabolic and immune
types of farming systems in the world (Herrero et al., 2010). There is functions to boost the production function with minimised exter-
also a gradient of ICLS ranging from low external input agriculture nal inputs (Fig. 1). A significant share of the production can be
(LEIA) to high external input agriculture (HEIA). LEIA corresponds exported while maintaining the functional integrity of the agroe-
to a high level of integration between crops and livestock, which cosystem (Thompson and Nardone, 1999). Thus, self-sufficiency
allows a high rate of recycling of natural resources and requires becomes the fourth property determining agroecosystem
a low amount of input; but LEIA provides little economic profit sustainability.
and low production volumes (Schiere et al., 2002). On the other Based on this conceptual framework and the ecological prin-
hand, HEIA corresponds to high production levels achieved with ciples proposed by Altieri (2002a), Altieri and Nicholls (2005),
high inputs. In HEIA, both components crops and livestock coex- Dumont et al. (2013), and Pretty (2013), we consider six princi-
ist independently. The non-ecological diversification of production ples for the analysis and design of sustainable ICLS (Fig. 2). Two of
units is achieved with low integration, overuse of natural resources, the principles refer to the structure of the farming system:
and high rates of consumption of external inputs. LEIA and HEIA
correspond to the two extremes in a wide range of the produc- (i) Diversity, i.e. heterogeneity, in land-use patterns and biotic and
tive and environmental efficiencies of ICLS (quantity of external abiotic components.
inputs/quantity of agricultural production). (ii) Maximisation of ecological (e.g. predator–prey) or production-
The development of sustainable ICLS comprises a wide port- based (e.g. complementarities between production cycles)
folio of designs (Schiere and Kater, 2001; Darnhofer et al., 2010). interactions.
We study a variety of existing mixed systems to better understand
the links between integration and ecoefficiency. Such studies are
Diversifying the components of the system and increas-
a first step towards understanding the process of transition from
ing the interactions among them are crucial to maintain the
conventional to ecological systems (Mc Rae et al., 1990). In this
metabolic, immune, and production functions. As in ecosystems,
paper, we use the conceptual framework of agroecology to ana-
the functional diversity, rather than the absolute diversity, of the
lyse ICLS and reveal combinations of practices that reinforce the
components sustains the three key functions of the farming system
integration of crops and livestock. First, we present the conceptual
(Tichit et al., 2011).
framework, based on six agroecological principles, for analysing
The four other principles refer to the management of the farming
and designing more sustainable ICLS. We apply the framework to
system:
two contrasting case studies from Brazil and France. Finally, we
discuss the main drivers that either favour or work against the
development of ICLS. We conclude with some questions for the (iii) Consider the production, immune, and metabolic functions
development of a research agenda. simultaneously to preserve the functional integrity of the
agroecosystem (Thompson and Nardone, 1999).
(iv) Close the energy and material cycles; i.e. minimise losses and
2. Conceptual framework external inputs, and substitute chemical inputs with natural
inputs.
To analyse the variety of ICLS, we utilise the conceptual frame- (v) Optimise the nutrient availability for crops and animals.
work of agroecology. Agroecology was defined, in a broad sense, Nutrient availability is more often a question of temporal
as ‘the ecology of sustainable food systems’ (Francis et al., 2003). settlement between supply and demand than a question
In this paper we consider a more restrictive definition: ‘the appli- of absolute availability. The temporal dimension of man-
cation of ecological concepts and principles to the design and agement must therefore be at the center of the redesign
management of sustainable agroecosystems’ (Gliessman, 1998). In process.
this definition, the key concept is the ecosystem: ‘. . . a functional (vi) Develop the collective management at the landscape level,
system of complementary relations, between living organisms and including seminatural elements. Ecological processes like pest
their environment. Natural ecosystems reflect a long period of evo- control or pollination depend on the landscape scale. ICLS man-
lution in the use of local resources and adaptation to local ecological agement has to extend beyond farm boundaries, which leads
condition’ (Gliessman, 2005). We consider three key functions of to collective landscape management among farmers and other
an ecosystem: production, metabolic function, and immune func- users, including both farmed and seminatural elements. The
tion. The organic analogy of the metabolic function is the process of questions behind landscape scale management are as much
degradation and synthesis of materials and energy in ecosystems. related to biotechnical aspect as they are to the issue of collec-
The metabolic function refers to the nutrient cycling and energy tive management.
flow in ecosystems that occur through a complex set of biotic and
abiotic interactions (e.g. hydrogeochemical processes and trophic 3. Case studies
interactions). The immune function refers to the health mainte-
nance and the resistance to disturbance of the ecosystem (Holling, 3.1. Case studies in context
1973). The immune function refers to the mechanisms of popula-
tion regulation (Pretty, 2013) and nurturing proper functioning of The combination of the six principles for the analysis and
natural pest control (Altieri, 2002a; Dumont et al., 2013). The pro- design of sustainable ICLS supports the productivity, efficiency,
duction function refers to the quantities of primary and secondary self-sufficiency, and resilience of the farming system. In the case
production. In the ecological literature, ‘complex natural ecosys- studies, we analyse the links between practices, principles, and
tems’ are considered to be simultaneously productive, efficient, and properties and illustrate multiple redesign paths and agroeco-
resilient (Tilman et al., 2001). logical trajectories. We compare two contrasting ICLS with their
The design of local, sustainable agricultural systems with simi- relative conventional systems. We focus the analysis on the
lar properties should, therefore, mimic the structure and functions structure of the system components (livestock, crops, etc.) and the
of natural ecosystems (Pretty, 1994; Gliessman, 1998, 2005; Altieri, links among the components through different flows and inter-
2002a,b). Designing such systems entails shifting from systems that actions. The various technical–economic performance indicators

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx 3

Agroecosystem
METABOLIC
FUNCTION
management
External METABOLIC
inputs FUNCTION

IMMUNE
FUNCTION
IMMUNE
FUNCTION
PRODUCTION PRODUCTION
FUNCTION FUNCTION

a b
Fig. 1. Production, metabolic and immune functions of (a) a conventional agroecosystem, with the production function based on external inputs (fertiliser and pesticide), in
which immune and metabolic functions are not supported, or are even externalised from the agroecosystem; (b) a sustainable agroecosystem with beneficial effects among
the production, metabolic, and immune functions.

are presented in Figs. 3 and 4. For both systems, we calculate the We present two contrasting case studies. The first case study
level of intensity, as the quantity of inputs per hectare of utilised is representative of the specialised and intensive systems of a
agricultural area (inputs in US $/ha), and the efficiency of input temperate area (Brittany, France). The second case study is rep-
use (inputs/agricultural production). We obtained the data from resentative of the extensive systems of a tropical area (Amazonia,
published literature on conventional systems and monitoring Brazil). In both areas, the farms are mixed in the sense of Schiere
schemes of ICLS farms. et al. (2002): most of the feed for livestock is produced on the farms,

a
Feed cost: $660/ha Index of pesticide application: 0.66
Mechanization expenses: $429/ha Impacts Total N load: 161 kg N/ha
Total inputs: $1080/ha Total energy consumption: 65 Eqf/1000 l milk

Inputs Commercial
Crops Animals products
Pasture
Crops
Prim-Holstein
Grasslands High yielding
Milk production:
Fodder Manure 4403 l/ha
Ecosystems Income: $775/ha
services

b Feed cost: $376/ha


Index of pesticide application: 0.21
Total N load: 121 kg N/ha
Mechanization cost: $344/ha Impacts
Total energy consumption: 43 Eqf /1000 l milk
Total inputs: $720/ha

Inputs Crops Pasture, corn, … Animals Commercial


Crops products
Prim-Holstein
Fodder Moderate yielding
Milk production:
Grasslands 4255 l/ha
Manure
& Legume
Income: $1087/ha
Ecosystems
services

Flow (materials and energy / year) Agroecosystem


Interactions (space /time) Ecosystem

Fig. 2. (a) French conventional ICLS: Farm size is 71 ha, with 22% arable and 78% main fodder area. Main fodder area comprises 72% grasslands and 28% maize silage. The herd
is made up of 48 Prim-Holstein cows with high milk potential (6300 l/cow/year). Reproductive management relies on artificial insemination only. First calving occurs at 28
months; 60% of calving occurs in autumn and winter and there is no calf suckling. The farmer aims at maximising the expression of the genetic potential of the cows; the feed
ration is made of 60% maize silage, 30% grazed herbage, and 10% grass hay with 800 kg of concentrate feed/cow/year, 60% of which are produced on the farm. Average milk
production is 4403 l/ha and average stocking rate is 1.36 livestock unit/ha. Eqf (equivalent liter fuel) is a unit of measurement corresponding to 35 MJ; (b) French sustainable
agriculture farm: Farm size is 64 ha, with 18% arable and 82% main fodder area. Main fodder area comprises 85% grasslands and 15% maize silage. The herd is made up of
49 Prim-Holstein with moderate milk potential (5700 l/cow/year). Reproductive management combines artificial insemination and natural mating. First calving occurs at 28
months, 70% of calving is grouped in spring, and there is calf suckling during weeks 3 to 4. The farmer limits his expectations of the expression of the genetic potential of the
cows; the feed ration is made of 60% grazed herbage, 30% grass hay, and 10% maize silage with 450 kg of concentrate feed/cow/year, 80% of which are produced on the farm.
Average milk production is 4255 l/ha and average stocking rate is 1.09 livestock unit/ha. Eqf (equivalent liter fuel) is a unit of measurement corresponding to 35 MJ.

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

4 T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx

a
Free environmental resources Impacts
Loss biodiversity, GHG, …
Slash and burn, for new agricultural land
and fertility
Forest
Crops Animal
Inputs Commercial
Corn Rice Pasture products
Mineral salt, Zebu Nelore
Veterinary products, … Beans
beef
Herbicide
Total inputs: $54-240/ha Cattle meat,
B. brizantha Manure Income: $12-80/ha
Ecosystems
Services
Climate and low pest pressure

b
Free environmental resources Impacts Use of pesticide
wood
Crops Commercial
Inputs Pasture, corn, Animals products
Orchard Cassava, … Poultry Ovine
Mineral salt, …
veterinary products, Fish Manure
Pigs
pesticide, seed, feed, Corn Cassava Milk by- Fish, cattle, cheese,
Total inputs: $178/ha Manure product milk, chicken, eggs,
B. brizantha pigs, ovine, manure
Manure Zebu beef
Ecosystems B. humidicula orange
and milk
Services Income: $153/ha

Climate and low pest pressure

Flow (materials and energy/year) Agroecosystem


Interactions (space/time) Ecosystem

Fig. 3. (a) An Amazonian conventional model of slash-and-burn cattle ranching: Farm size is between 100 and 400 ha, with 17% arable and 83% grasslands. The farmer aims
to maximise the area of grasslands. The small farmers burn 3 to 6 ha of forest per year, to plant annual food crops (rice, bean, cassava, corn) for human consumption (Bonaudo
et al., 2007). With this practice, he benefits from two free environmental resources: land and the associated forest rent. After 30 or 40 days, the pasture is sowed with the annual
crops, to be well developed and cover all the soil after harvesting. The large holders can deforest over 100 ha and plant pasture directly. The B. brizantha produces between 8
and 20 tons DM/ha/year (Heuzé et al., 2013). The Nelore breed has medium or low production potential. The average stoking rate is 1 head/ha of grassland; the reproductive
management is natural mating, and calving occurs all year round. The first calving occurs at more than 28 months, and the calving interval is around 20 months. Cattle gain
between 400 and 550 kg carcass weight (CW), are slaughtered between 36 and 48 months (Poccard-chapuis et al., 2011; Cederberg et al., 2011). The meat production is
between 25 and 80 kg CW/ha/year (Poccard-chapuis et al., 2011; Veiga and Tourrand, 2001). Cederberg et al. (2011) have estimated the average production in Amazonia at
42 kg CW/ha/year in 2006. Barros et al. (2002) calculated that mineral supplementation corresponds to 77% of the production cost. We calculated an input cost from 54 to 240
US$/ha/year. Barros et al. (2002) calculated production comparable costs. Barros et al. (2002) calculate a profitability of $19 to 41/ha/year. With an economic model, Bowman
et al. (2012) estimated the annual profit of cattle ranching to be $12 to 80/ha. Siegmund-Schultze et al. (2010) estimated the annual return at $8.7/ha for small holders; (b) the
Amazonian ICLS model: Farm size is 332 ha, composed of 189 ha of primary forest and 15 ha of permanent preservation area. The agricultural area of 143 ha was established
in 20 years. Each year a small plot of forest was cleared for implanting annual crops followed by pasture. There are currently 89 ha of B. brizantha and 34 ha of B. humidicula,
2.5 ha of annual crops, 1.5 ha for the orchard and some farm buildings and housing, and 1 ha dedicated to fish farming. The animal subsystems are composed of poultry
production (380 head corresponding to 650-700 kg Live weight/year and 700 eggs/year), sheep (225 kg Live weight/year), pigs (620 kg Live weight/year), fish production (21
tons Live weight/year), and a herd of mixed cattle (130 head) producing calves, cows (13 tons Live weight/year), and milk to make cheese (1.12 tons/year). Integration of crop
and pasture requires herbicide (6 l/ha glyphoosate, 2 l/ha 2,4-dichlorophenoxyacetic acid, 1.5 l/ha nicosulforon, 0.2 l/ha tricolpir), 98 kg/ha/P2O5, 126 kg/ha/K2O, and mineral
fertilisation (N 140 kg/ha). The production is 5.5 t/ha of grain (versus less than 1 t/ha with deforestation). Before this technique, the work was only manual; we measured
more than 27 h labor/ha of annual crop production. Today the manual work is estimated at 8 h/ha with 6 additional hours of tractor per hectare. We estimate the production
cost at $178/ha/year and the income of all activities at $331/ha/year. The profit is $153/ha (before tax, debt repayment, and payment of family labor).

and a small share of the gross margin comes from the crop com- 3.2. Shifting from high external input dairy systems to
ponent. The environmental impact in the first case study is the self-sufficient ICLS in Brittany
nitrogen excess resulting from the concentration of zero-grazing
livestock systems. The environmental impact in the second case The general goal of the farms belonging to the Réseau Agricul-
study is the deforestation, biodiversity loss, and climate change ture Durable1 (RAD) is to sustain the level of added value derived
resulting from agricultural and cattle expansion (Laurance et al., from dairy production without necessarily maximising outputs per
2001; Nepstad et al., 2009; Bowman et al., 2012). The case studies animal or per unit area (Alard et al., 2002). Farm management
reflect two agriculture-farming integration strategies, one based on thus gives priority to gross margin, rather than gross product,
self-sufficiency of the herd and the other based on diversifying the and to the internal resources of the farm, rather than external
system components and maximising the interactions among the resources. The overall production strategy combines two priorities:
components. For the Brazilian case, we explore how system diver- self-sufficiency and parsimony in external input use. Both priorities
sification can increase the production per hectare and potentially apply to the chemical and feed inputs used in both the plant and
reduce deforestation and land competition between crop and live- the animal production processes.
stock. For the French case, we analyse how the reconnection of the To use inputs sparingly in the production processes, priority is
livestock to the land-base can close the nutrient cycles. Although given to maximising the use of grazed herbage, at the expense
the systems seem far from one another, international trade fluxes of maize silage, and to reducing the use of concentrate feeds as
make them interdependent. The Brittany concentrate feed uses
the soybean cropped in Brazil. Such global interdependency is
one of the drivers of deforestation in the Amazon and Cerrado 1
Sustainable agriculture network: This organization gathers farmers from western
biome. France, mostly middle sized dairy farms, with a mean of 50 cows in 2010.

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx 5

Fig. 4. Agroecosystems: properties, principles, and practices for design. Ecological and management principles support the four emergent properties and depend on a diversity
of agro-ecological practices.

much as possible (Fig. 3a and b). Several management practices in the feeding strategy helps to cut both harvesting costs and dis-
enable the minimisation of inputs in the production processes. tribution costs. Relying on long-term grasslands decreases the cost
First, the farmer does not seek the full expression of the genetic of grassland renovation (seeds and establishment); it also limits
potential of the animals. The herd is made up of moderate-yield ploughing and the need for mineral fertiliser in the subsequent
cows (5700 l/cow/year), with heifers first mated at 19 months of crop. Finally, the use of long-term grasslands globally impairs the
age. The lower milk production per cow is compensated for by development of weeds in the crop rotation and improves the poten-
reduced metabolic stress, fewer health problems and reproduc- tial for organic carbon sequestration in the soil.
tive failures, and higher lifetime production. The animal production The use of complex grass-legume mixtures brings several
cycle is organised according to herbage seasonal dynamics. Sev- advantages. It allows a drastic reduction in the use of min-
enty percent of calving occurs in the spring. The animals thus have eral fertilisers and herbicides, and the increased complementarity
high nutritional requirements during the period of high grass avail- between the nitrogen and water uptake of the different species
ability. The main fodder area comprises a very large share of the reduces the risks of nitrate leaching. The mixtures improve the
grasslands (>80%), and a limited area is dedicated to maize silage. feeding value and stability of the pastures over time and reduce
The cows receive low levels of concentrate feed (c. 450 kg of con- the need for concentrates. Because grass–legume mixtures are usu-
centrate/cow/year), 80% of which is produced on the farms. Such ally integrated in crop rotations, the legumes allow a reduction in
a configuration of the feeding system implies a moderate stocking fertiliser inputs.
rate (1.09 livestock units/ha) and a moderate rate of milk produc- Grazing management is the key to the efficiency of the system.
tion (4255 l/ha). The grassland is divided in two parts. The base area, located close
The grassland plots comprise a high proportion of long-duration to the milking parlour, is dedicated only to grazing. The base area
(5 to 15 years) grass–legume mixtures. The plots are primarily is divided into six to eight plots, the sizes of which are calculated
exploited by grazing. The grazing of the cows on the functionally to fit with the herd size. The complementary area is mowed in the
diverse mixtures brings several advantages for the optimisation of spring and used as security pasture in other seasons when grass
the agroecosystem functioning. First, the efficient photosynthetic production on the base area is less predictable. The system guaran-
capacities of the grasses optimise the conversion of carbon diox- tees adequate fodder availability throughout the year, enables the
ide, water, and minerals into biomass. Second, as an alternative grazing season to be extended in the summer, autumn, and winter,
to mineral fertilisation, the legume species fix nitrogen. Synergies and minimises animal movements.
between the grasses and the legumes benefit biomass production. RAD farms involved in the ‘low-input fodder system’ agrien-
Third, the grazing animals fulfil three functions. They harvest (and vironmental scheme demonstrate significant environmental
distribute) their own feed; return nutrients to the soil; and help improvements (Le Rohellec et al., 2009); the total nitrogen load
to control weeds, thus minimising mechanical interventions (har- (i.e. excreta, mineral and organic fertilisers, and manure) is 33%
vesting, ploughing) and fuel costs. The grazing on long duration, lower than in conventional farms. The frequency index of pes-
multispecies grasslands is an essential key for the design of self- ticide applications fell by two-thirds 3 years after the scheme
sufficient and sparing systems. Maximising the share of grassland was adopted. The improvement is mainly due to the increase in

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

6 T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx

grassland area and the thresholds imposed for maize and wheat support any ecological or management principle (Fig. 4a). Produc-
management. RAD farms report 33% lower energy consumption tion is the unique objective, based on a very poor agrobiodiversity.
per 1000 l of milk than conventional farms because of less fertilisa- The only interaction between crops and animal subsystems is ani-
tion, less fuel use for mechanisation, and savings on concentrate mal defecation during grazing. The system is based on free natural
feeds. The moderate decrease in productivity (−200 l of milk/ha resources (forest rent). The forest is perceived only as a stock of
in RAD vs. conventional farms) is largely compensated for by an space and fertility to exploit. The externalities of deforestation, low
overall reduction in input costs (feed costs are nearly halved and production levels and high impacts on the environment (degraded
mechanisation expenses are lower), which leads to a higher net land, loss of biodiversity, climate change), remain ignored or
income per ha. denied.
The general approach of the RAD farms is to find an optimal rela- The goal of the Amazonian ICLS is to intensify agricultural land
tionship between the animal component and the on-farm resource to generate higher income per ha and stop deforestation. The farm-
component. There is no attempt to optimise animal production ers combine a wide variety of species in the crop and livestock
independently of the resources available on the farm. The optimised components, leading to many interactions within and between
component is thus the relationship between the animal demand the components. The property area of 332 ha comprises 60% pri-
and the available on-farm feed resources. Several management mary forest and ‘permanent preservation area’ on river banks and
practices help to close nutrient cycles and optimise nutrient avail- steep slopes, corresponding to ecological infrastructures. Grass-
ability. The high share of grazed grassland in the feeding strategy lands (37% of the total farm) are composed of B. brizantha (73%)
contributes to input minimisation and thus consolidates farm self- and Brachiaria humidicula (27%), exploited through rotational graz-
sufficiency. It also improves farm efficiency, because it is combined ing. Less than 3% of the area is dedicated to field crops (e.g. corn
with a limited expectation for milk production (e.g. not seeking and cassava for the family and farmyard animals) and tree crops.
the full expression of the genetic potential of the cows). The limit Given the impossibility of continued deforestation, the low yields
on the farmers’ expectations is important, because it maintains the of annual crops and the degradation of the old pastures (plots 15 to
level of production within the limits at which marginal costs remain 20 years old), ICLS farmers implement long-term rotations between
economically viable. Another important advantage of RAD farms is pastures and annual crops. As a first step, they make a mulch of the
their lower sensitivity to input-price fluctuations; the systems are pasture to improve the soil. In the second step, the farmers directly
expected to be more resilient to market uncertainties, because they sow annual crops with mineral fertilisation. After 30 or 40 days, the
have a lower dependency on external inputs. The lower depend- farmers sow grass along with the annual crops. After harvesting the
ency on nonrenewable energy offers a potential advantage in light annual crops, cattle graze at high density for one or two days and
of expected higher energy prices in the future. The grassland-based harvest the crop residues and grass biomass. The grazing promotes
feeding strategy does not increase the sensitivity of the systems tillering and pasture regrowth. After a rest period of 20 to 30 days,
to climate change, because the lower stocking rate of the farms livestock graze on the plot for 6 to 8 months before an annual crop
amounts to a buffering strategy to compensate for interannual vari- is replanted via the same procedure. The strategy is to practice 2
ability in grass growth. These self-sufficient ICLS systems open up to 3 years of direct seeding (with input) and 5 to 10 years of pas-
new options for existing farming systems, because they display ture (with no input). The system produces approximately five times
a win-win relationship between environmental performance and more grain than the slash-and-burn system and may increase the
economic performance. average stocking rate of 1 head of livestock per hectare of grassland
to more than 2 heads per hectare of grassland.
3.3. Shifting from cattle ranching on one-species grasslands to Production diversification and the design of complementarities
ICLS in Amazonia are the founding characteristics of the system. The animal sub-
system comprises poultry, sheep, pigs, fish (tambaqui, Colossoma
There is a huge agricultural variety of designs and practices macropomum), and a herd of mixed cattle producing calves and milk
in Amazonia between small, medium, and large holders. They all, to make cheese. The family on the farm consumes a small share of
however, share the same general goal: occupy and valorise land chickens, pigs, sheep, and cow’s milk and sells the other products.
while developing low-risk and low-cost agricultural productions. There are multiple synergies within and between the subsystems.
Farm management gives priority to cattle ranching (Bonaudo et al., The manure produced by the pigs, poultry, and cattle is used to fer-
2007). The grazing system is based on Brachiaria brizantha, which tilise the fish pond, the pasture, and the orchard (Fig. 4b). Maize
represents 90% of the sown grass in Amazonia (Poccard-chapuis and cassava are used to feed poultries and pigs. The pigs also bene-
et al., 2011). This species has a good ability to cover the soil, which fit from the by-products of cheese making. Cattle feeding depends
prevents the development of invasive weeds. It has good pro- on grazing with only a mineral salt complement.
ductivity without fertiliser, even in relatively poor soil, and good There is a relatively good efficiency in the use of inputs. Even
resistance to drought (Veiga and Tourrand, 2001; Heuzé et al., with higher input use, profits per ha of agricultural land are 2 to 20
2013). To valorise pasture with low feed quality, farmers use the times higher than in regional studies (Barros et al., 2002; Siegmund-
Nelore zebu cattle breed, which is well adapted to tropical regions. Schultze et al., 2010; Bowman et al., 2012). Another advantage is
The Amazonian ranching system is adapted to marginal regions diversifying and securing sources of family income. The poultry,
with little infrastructure. The only inputs are basic equipment like milk, and fruits provide regular income, while fish farming and
corral and fences made with local wood, mineral salt, and some cattle ranching provide a more important income once per year.
veterinary products. Pasture maintenance is achieved through good The farmers also have a commercial business in town to sell their
management of grazing pressure. Weeds are removed either by products directly, aggregate value, and thus increase their self-
hand or by regular fires. Even with low production costs, the income sufficiency.
per hectare is very low (Arima and Uhl, 1996) The solution to the The system responds partly to the ecological principles of bio-
low income levels is farmland expansion by forest slash and burn logical diversification and the maximisation of interactions mainly
(Arima and Uhl, 1996; Bommel et al., 2010). through the number of animal species (ruminant, monogastric,
Even if some individual practices (e.g. no tillage, overlapping and fish). A diversity of animals in interaction within the farm-
crop and grass cycles, permanent grassland, choice of locally- ing system partially addresses some management principles, such
adapted breeds and varieties, low use of pesticides and veterinary as closing nutrient cycles, via the production of feed, recycling
products) have potentially low impact, the whole system does not of by-products, and organic fertilisation of cultures. The system

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx 7

could still be greatly improved, especially by reducing the purchase Farm structure is a second factor in the development of ICLS.
of feed for the fish. A number of agricultural practices also con- The general tendency for farm enlargement has limited the sur-
tribute to the principle of closing cycles: direct seeding, mulching, vival or the redevelopment of ICLS. Scale economies were enabled
no tilling, and especially the introduction of long crop rotations only through specialisation to achieve work productivity gains on
to avoid deforestation. The principle of optimising nutrient avail- enlarged farms (Hendrickson et al., 2008b). Furthermore, farm-
ability is practiced by overlapping crop cycles and by high-density ers need specific skills to understand not only crop or livestock
sowing. These practices contribute to the control of erosion and management but also potential interactions between crops and
weeds, the objective being to ensure a high soil cover year round. livestock. As observed in the RAD case study, farmers can learn and
The soil covering helps to reduce the use of herbicides, which test new ways of producing while limiting input use. Adaptive man-
remains the main negative output. Finally, farmers manage the agement, for example, is a learning-by-doing process that involves
seminatural elements of the landscape by keeping areas of per- an increased dedication to system observation. The complexity of
manent preservation, maintaining landscape heterogeneity, and integration can limit the adoption of integrated crop–livestock sys-
avoiding diffuse pollution of rivers. In this way, the farmers man- tems, as agricultural education is currently focused on single-crop
aging such ICLS consider the three functions of the agroecosystem or livestock production. As observed in the Brazilian case study,
simultaneously. smart management can combine diversified production with the
This ICLS, through the spatial and temporal integration of several limited use of inputs. Nevertheless, these drivers of change condi-
crops and livestock, achieves high yield and income, stops defor- tion farmers’ choices and could thus allow the development of ICLS
estation, and reduces greenhouse gas emissions. The ICLS produces when a specific tradition or dynamic of ICLS is involved (Ryschawy
more and is more efficient, self-sufficient, and resilient than spe- et al., 2013). As we observed through our case studies, some farm-
cialised livestock systems. The ICLS is an alternative to land-sparing ers can develop ICLS even if major drivers promote conventional
strategies (Green et al., 2005) and promotes long-term thinking on specialised farming systems.
the roles of forests and farming in the provision of essential ecosys- The third driver of ICLS development concerns the regional loca-
tem services (e.g. soil fertility, low pest pressure, stable climate). tion, determining market, and soil–climatic contexts. The survival
of ICLS is mostly observed in intermediate areas, where cash-
crop production is possible but does not enable high-production
4. Discussion yields. In the most favourable regions where higher yields are
achieved, farming systems are specialised for cash crop produc-
4.1. Positive and negative drivers for the development of tion. Agricultural sectors have thus been reorganised: downstream
sustainable ICLS and upstream industries enable the low-price provisioning of min-
eral fertiliser and pesticides, as in the Brazilian Mato Grosso. Such
Although mixed crop–livestock systems could have important a concentration of production leads to specific farmer networks
contributions to a number of key ecological processes, they have developed according to geographical proximity. In particular, some
been highly marginalised worldwide. Three major groups of fac- livestock production sectors are concentrated in specific regions
tors favour the specialisation of farming systems and could thus of the world (Gaigné, 2012). Geographical location thus leads to
explain the decline in ICLS over time (Ryschawy et al., 2013; Veysset lower prices for a specific type of production within a specific area.
et al., 2005). The general economic and political environment For instance, in Brittany, the concentration of pork, poultry, and
explains a part of farm specialisation. Since the 1960s, economies of dairy-milk production and industries enable low prices for live-
scale, favoured by specialisation and farm enlargement to respond stock feed and low transportation prices for products, reinforcing
to mass-production objectives, have caused a strong decrease in the agglomeration process and farm specialisation.
the number of ICLS worldwide (Wilkins, 2008; Vermersch, 2007;
Ryschawy et al., 2013). Economies of scope, achieved through prod- 4.2. The transformation of both systems led to more integration
uct diversification, were limited relative to the economies of scale between the plant and animal components
(Gaigné, 2012). Increases in work productivity and the decline of
the relative energy prices (e.g. feed and pesticides) have favoured Our case studies illustrate that there are several paths to greater
such economies of scale. Thus, worldwide economic tendencies sustainability. Starting from LEIA and HEIA systems, increased
limited the development of ICLS. Nevertheless, increasing mar- crop–livestock integration caused two opposed trajectories to con-
ket uncertainties could contribute to the redevelopment of ICLS verge on an agroecological ICLS (Fig. 5). The French case illustrates
farming systems, which are more resilient to market instabilities how the combination of moderate deintensification with increased
(Ryschawy et al., 2012). crop–livestock integration can drastically reduce external inputs
In Europe, incentives of the Common Agricultural Policy (fertilisers, pesticides, feed, etc.) via better recycling of crops and
contributed to specialisation through output-based subsidies livestock by-products (straw, manure, etc.). The strategy mitigates
(Chatellier and Guyomard, 2008). The externalisation of environ- environmental externalities and causes a limited production loss.
mental costs also contribute to the highly-specialised reductionist On-farm integration, as practiced by RAD farmers, enables excreta
agriculture. The dominance of a reductionist approach in research to be recycled and limits the use of fuel and fossil fuel, while feed-
and development led to policies focused on single-criterion per- ing animals in a way that allows them to produce efficiently. Such
formances. Such approaches limited the potential to promote ICLS systems, which give strong attention to the conservation of the
and led mostly to high external-input agricultural systems, as resource base via trade-offs with production levels, square with
shown in both case studies. New political incentives favouring agroecological thinking. The Amazonian case illustrates how the
a systemic approach could enable the new development of ICLS. combination of moderate intensification and increased crop live-
For example, in Brazil, researchers and political stakeholders are stock integration can substitute for a high level of natural inputs
considering the intensification of deforested areas as an option (forest rent) with a moderate level of chemical inputs (herbicide
enabling limits to deforestation without compromising the fur- and fertilisers), leading to an increase in production while mitigat-
ther growth of agricultural production. Public policies, including ing environmental impacts.
payments for ecosystem services and the creation of a market for Both trajectories took advantage of the agroecological princi-
alternative agriculture, are also needed to fund advisory systems ples favoured by crop–livestock integration and maintained the
that promote the development of ICLS. systems on more sustainable space of the set of possible outcomes

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

8 T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx

Fig. 5. Trajectories of different ICLS in the set of feasible systems. HEIA and LEIA stand for High External Input Agriculture and Low External Input Agriculture, the two
extremes of a gradient of old forms of ICLS. Arrows 1 and 2 illustrate the dynamics on the continuum of systems between the two extremes: arrow 1 corresponds to
conventional intensification and arrow 2 corresponds to ecologisation. Red arrows illustrate the amazonian and the RAD (Réseau Agriculture Durable; a network of sustainably
managed French farms) case studies. Arrow 3 corresponds to orthogonal dynamics: agroecological transition that do not necessarily optimizes production but decreases the
input/production ratio.

represented by Fig. 5. The integrations that we described are par- paragons in the future. Further research should compare ICLS struc-
tial. We hypothesise that better crop–livestock integration and the ture and function, degree of integration, and performance, using a
further application of agroecological principles can further improve set of technical, economical, environmental, and social indicators.
the performances of both systems, which are already on their way Finally, crop–livestock systems can help the redesign of both indus-
to what agroecological ICLS could be. trial and specialised systems by demonstrating empirical ways of
dealing with agroecological complexity and diversity. We believe
5. Conclusion that the empirical study of ICLS diversity could serve as a source
of data, knowledge, and inspiration for the necessary regreening of
Sustainability challenges require simultaneous progress in pro- conventional systems.
duction and environmental performance. ICLS offer good potential
for the design of systems that base their productivity on ecosystem References
function and services. Hence, six questions arise:
Alard, V., Béranger, C., Journet, M., 2002. A la recherche d’une agriculture durable.
Etude de systèmes herbagers économes en Bretagne. INRA Editions, Paris.
– How is diversity related to ecosystem function services in land- Altieri, M.A., 2002a. Agroecological principles and strategies for sustainable agri-
scapes dominated by ICLS? culture. In: Uphoff, N.T. (Ed.), Agroecological Innovations: Increasing Food
– What are the synergies and trade-offs among ecosystem ser- Production with Participatory Development. Earthscan publication Ltd., London,
pp. 40–46.
vices? Altieri, M.A., 2002b. Agroecology: the science of natural resource management for
– Which indicators are best for monitoring temporal and spatial poor farmers in marginal environments, Agriculture. Ecosystems and Environ-
synergies, synchronisms, or offsets among crop and animal com- ment 1971, 1–24.
Altieri, M.A., Nicholls, C.I., 2005. Agroecology and the Search for a Truly Sustain-
ponents?
able Agriculture. In: United Nations Environment Programme, first ed. United
– What are the action levers for optimising these synergies? Nations, Mexico D.F., pp. 291.
– Do the different ecosystem services show trade-offs or synergies? Arima, E., Uhl, C., 1996. Pecuária na Amazônia Oriental: desempenho atual e per-
spectivas. Imazon, Belém, pp. 96.
Barros, G.S.C., Zen, S., Bacchi, M.R.P., Ichibara, S.M., Osaki, M., Ponchio, L.A., 2002.
These questions invite collaborative research comprising agron- Economia da pecuária de corte na região norte do Brasil. Centro de Estudos
omy, ecology, economics, and the social sciences. Matching Avançados em Economica Aplicada (CEPEA)-ESALQ/USP, Piracicaba, pp. 76.
Bommel, P., Bonaudo, T., Barbosa, T., Veiga, J.B. da, Pak, M.V., Tourrand, J.F., 2010.
different kinds of knowledge, scientific and empirical, is also La relation complexe entre l’élevage et la forêt en Amazonie bresilienne:
required, as agroecology attempts to adapt solutions, e.g. ICLS con- une approche par la modelisation multi-agents. Cahiers Agricultures 19 (2),
figuration, to the local ecology of contexts. The codesign of models 104–111.
Bonaudo, T., Tourrand, J.F., Lossouarn, J., Poccard-chapuis, R., les 5 et 6 Decembre
by scientists and stakeholders (Lamanda et al., 2012; Voinov and 2007. Elevage bovin et espace forestier sur les fronts pionniers amazoniens:
Bousquet, 2010) can facilitate knowledge integration, formalisa- D’une opposition historique vers une nécessaire complémentarité. In: 14emes
tion, transparency, and a shared view. Modelling could be the link Recontres autour des recherches sur les ruminants, Paris.
Bowman, M.S., Soares-Filho, B.S., Merry, F.D., Nepstad, D.C., Rodrigues, H., Almeida,
between local situations and generic thinking (Sterk et al., 2007). O.T., 2012. Persistence of cattle ranching in the Brazilian Amazon: a spatial
The usefulness of multiple-objective optimisation was recently analysis of the rationale for beef production. Land Use Policy 29, 558–568.
emphasised for the design of mixed farming systems (Groot et al., Cederberg, C., Persson, U.M., Neovius, K., Molander, S., Clift, R., 2011. Including carbon
emissions from deforestation in the carbon footprint of Brazilian beef. Environ-
2012; Groot and Rossing, 2011). Models can explore the synergies
mental Science and Technology 45, 1773–1779.
and trade-offs among different ecosystem services (Teillard et al., Chatellier, V., Guyomard, H., 2008. Le bilan de santé de la PAC, le découplage et
2012; Sabatier et al., 2013). l’élevage en zones difficiles. INRA Sciences Sociales Recherches en Economie et
In the debate surrounding sustainable production, we show the Sociologie Rurales 6, 1–8.
Darnhofer, I., Bellon, S., Dedieu, B., Milestad, R., 2010. Adaptativeness to enhance the
importance of reconstructing links among soils, crops, and animals. sustainability of farming systems: a review. Agronomy for Sustainable Develop-
We conclude that crop–livestock systems could be agroecological ment 30, 545–555.

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010
ARTICLE IN PRESS
G Model
EURAGR-25216; No. of Pages 9

T. Bonaudo et al. / Europ. J. Agronomy xxx (2013) xxx–xxx 9

Dumont, B., Fortun-Lamothe, L., Jouven, M., Thomas, M., Tichit, M., 2013. Prospects Naylor, R., Steinfeld, H., Falcon, W., Galloway, J., Smil, V., Bradford, E., Alder, J.,
from agroecology and industrial ecology for animalproduction in the 21st cen- Mooney, H., 2005. Losing the links between livestock and land. Science 310,
tury. Animal 7 (6), 1028–1043. 1621–1622.
Ewing, M.A., Flugge, F., 26 Sept. to 1 Oct 2004. The benefits and challenges Nepstad, D., Soares-Filho, B., Merry, F., Lima, A., Moutinho, P., Carter, J., Bowman, M.,
of crop–livestock integration in Australian agriculture. In: Fisher, T., et al. Cattaneo, A., Rodrigues, H., Schwartzman, S., 2009. The end of deforestation in
(Eds.), New Directions for a Diverse Planet. Proc. fourth Int. Crop Sci. the Brazilian Amazon. Science 326, 1350.
Congr. The Regional In-stitute, Gosford, NSW, Available at www.cropscience. Poccard-chapuis, R., Bonaudo, T., Tourrand, J.F., Lossouarn, J., 2011. Élevage, filières
org.au/icsc2004/symposia/6/3/2005 ewingma.htm. et territoires en régions chaudes. INRA Productions Animales 24 (1), 129–144.
Francis, C., Lieblein, G., Gliessman, S., Breland, T.A., Creamer, N., Harwood, R., Pretty, J.N., 2013. Agricultural sustainability: concepts, principles and evidence.
Salomonsson, L., Helenius, J., Rickerl, D., Salvador, R., Wiedenhoeft, M., Simmons, Philosophical Transactions of the Royal Society of London, Series B 363, 447–465.
S., Allen, P., Altieri, M., Flora, C., Poincelot, R., 2003. Agroecology: the ecology of Pretty, J.N., 1994. Regenerating Agriculture. Earthscan Publications Ltd., London.
food systems. Journal of Sustainable Agriculture 22 (3), 99–118. Russelle, M.P., Entz, M.H., Franzluebbers, A.J., 2007. Reconsidering integrated
Gaigné, C., 2012. Organisation des filières animales et environnement. Vingt ans crop–livestock systems in North America. Agronomy Journal 99, 325–334.
après la directive nitrates. INRA Productions Animales 25, 375–388. Ryschawy, J., Choisis, N., Choisis, J.P., Gibon, A., 2013. Paths to last in mixed
Gliessman, S.R., 1998. Agroecology: Ecological Processes in Sustainable Agriculture. crop–livestock farming: lessons from an assessment of farm trajectories of
Ann Arbor Press, Chelsea, MI., pp. 384. change. Animal 7 (4), 673–681.
Gliessman, S.R., 2005. Agroecology and agroecosystems. In: Pretty, J. (Ed.), The earth- Ryschawy, J., Choisis, N., Choisis, J.P., Joannon, A., Gibon, A., 2012. Mixed
scan reader in sustainable agriculture. Earthscan, London, UK. crop–livestock systems: an economic and environmental-friendly way of farm-
Gliessman, S.R., 2006. Animals in agroecosystems. In: Agroecology: The Ecology ing? Animal 6, 1722–1730.
of Sustainable Food systems, second ed. CRC Press, Boca Raton, Florida, pp. Sabatier, R., Doyen, L., Tichit, M., 2013. Heterogeneity and the tradeoff between
269–285. ecological and productive functions of agrolandscapes: a model of cattle/bird
Green, R.E., Cornell, S.J., Scharlemann, J.P.W., Balmford, A., 2005. Farming and the interactions in a grassland landscape. Agricultural Systems (Available online 25
fate of wild nature. Science 307, 550–555. March 2013).
Groot, J.C.J., Oomen, G.J.M., Rossing, W.A.H., 2012. Multi-objective optimization and Schiere, J.B., Ibrahim, M.N.M., van Keulen, H., 2002. The role of livestock for sustaina-
design of farming systems. Agricultural Systems 110, 63–77. bility in mixed farming: criteria and scenario studies under varying resource
Groot, J.C.J., Rossing, W.A.H., 2011. Model-aided learning for adaptive management allocation. Agriculture, Ecosystems and Environment 90, 139–153.
of natural resources—an evolutionary design perspective. Methods in Ecology Schiere, J.B., Kater, L., 2001. Mixed Crop-Livestock Farming: A Review of Traditional
and Evolution 2, 643–650. Technologies Based on Literature and Field Experiences. In: FAO Animal Produc-
Hendrickson, J.R., Hanson, J.D., Tanaka, D.L., Sassenrath, G.F., 2008a. Principles tion and Health, Papers 152. FAO, Rome, Italy.
of integrated agricultural systems: introduction to processes and definition. Siegmund-Schultze, M., Rischkowsky, B., da Veiga, J.B., KingValuing, J.M., 2010. Cat-
Renewable Agriculture and Food Systems 23 (4), 265–271. tle on mixed smallholdings in the eastern Amazon. Ecological Economics 69,
Hendrickson, J.R., Liebig, M.A., Sassenrath, G.F., 2008b. Environment and integrated 857–867.
agricultural systems. Renewable Agriculture and Food Systems 23 (4), 304–313. Sterk, B., van Ittersum, M.K., Leeuwis, C., Wijnands, F.G., 2007. Prototyping and
Herrero, M., Thornton, P.K., Notenbaert, A.M., Wood, S., Msangi, S., Freeman, H.A., farm system modeling—partners on the road towards more sustainable farm
Bossio, D., Dixon, J., Peters, M., van de Steeg, J., Lynam, J., Parthasarathy Rao, P., systems? European Journal of Agronomy 26, 401–409.
Macmillan, S., Gerard, B., McDermott, J., Seré, C., Rosegrant, M., 2010. Smart Teillard, F., Allaire, G., Cahuzac, E., Léger, F., Maigné, E., Tichit, M., 2012. A novel
investments in sustainable food production: revisiting mixed crop–livestock method for mapping agricultural intensity reveals its spatial aggregation: impli-
systems. Science 327, 822–825. cations for conservation policies. Agriculture, Ecosystems and Environment 149,
Heuzé, V., Sauvant, D., Tran, G., 2013. Bread grass (Brachiaria brizan- 135–143.
tha). In: Feedipedia.org.: A Programme by INRA, CIRAD, AFZ and FAO, Thompson, P.B., Nardone, A., 1999. Sustainable livestock production: methodologi-
http://www.feedipedia.org/node/490, Last updated on March 19, 2013. cal and ethical challenges. Livestock Production Science 61, 111–119.
Holling, C.S., 1973. Resilience and stability of ecological systems. Annual Review of Tichit, M., Puillet, L., Sabatier, R., Teillard, F., 2011. Multicriteria performance and sus-
Ecology and Systematics 4, 1–23. tainability in livestock farming systems: functional diversity matters. Livestock
Lamanda, N., Roux, S., Delmottec, S., Merot, A., Rapidela, B., Adamd, M., Werye, J., Science 139, 161–171.
2012. A protocol for the conceptualisation of an agro-ecosystem to guide data Tilman, D., Reich, P.B., Knops, J., Wedin, D., Mielke, T., Lehman, C., 2001. Diversity
acquisition and analysis and expert knowledge integration. European Journal of and productivity in a long-term grassland experiment. Science 294, 843–845.
Agronomy 38, 104–116. Veiga, J.B. da, Tourrand, J.F., 2001. Pastagens cultivadas na Amazonia Brasileira:
Laurance, F.W., Cochrane, M.A., Bergen, S., Fearnside, P.M., Delamônica, P., Barber, situacao atual e perspectivas. Documentos Embrapa Amazonia Oriental 83, 36.
C., D’Angelo, S., Fernandes, T., 2001. The future of the Brazilian Amazon. Science Vermersch, D., 2007. L’éthique en friche, QUAE INRA ed. QUAE, Paris, pp. 117, Col-
291, 438–439. lection update sciences and technologies.
Le Rohellec, C., Falaise, D., Mouchet, C., Boutin, M., Thiebot, J., 2009. Ana- Veysset, P., Bebin, D., Lherm, M., 2005. Adaptation to agenda 2000 (CAP reform)
lyse de l’efficacité environnementale et énergétique de la mesure agri- and optimisation of the farming system of French suckler cattle farms in the
environnementale Système fourrager économe en intrants (SFEI) à partir de Charolais area: a model-based study. Agricultural Systems 83, 179–202.
l’analyse de pratiques de quarante-quatre signataires. Rencontres Recherches Voinov, A., Bousquet, F., 2010. Modelling with stakeholders. Environmental Mod-
Ruminants 16, 109–112. elling and Software 25, 1268–1281.
Mc Rae, R.J., Hill, S.B., Mehuys, F.R., Henning, J., 1990. Farm scale agronomic and Wilkins, R.J., 2008. Eco-efficient approaches to land management: a case for
economic conversion from conventional to sustainable agriculture. Advances in increased integration of crop and animal production systems. Philosophical
Agronomy 43, 155–198. Transactions of the Royal Society: Biological Sciences 363, 517–525.

Please cite this article in press as: Bonaudo, T., et al., Agroecological principles for the redesign of integrated crop–livestock systems. Eur. J. Agron.
(2013), http://dx.doi.org/10.1016/j.eja.2013.09.010

You might also like