Archivos de Zootecnia ISSN: 0004-0592 [email protected] Universidad de Córdoba España

García, A.; Perea, J.; Acero, R.; Angón, E.; Toro, P.; Rodríguez, V.; Gómez Castro, A.G. Structural characterization of extensive farms in andalusian dehesas Archivos de Zootecnia, vol. 59, núm. 228, diciembre, 2010, pp. 577-588 Universidad de Córdoba Córdoba, España

Available in: http://www.redalyc.org/articulo.oa?id=49518785011

How to cite Complete issue Scientific Information System More information about this article Network of Scientific Journals from Latin America, the Caribbean, and Portugal Journal's homepage in redalyc.org Non-profit academic project, developed under the open access initiative STRUCTURAL CHARACTERIZATION OF EXTENSIVE FARMS IN ANDALUSIAN DEHESAS*

CARACTERIZACIÓN ESTRUCTURAL DE LOS SISTEMAS GANADEROS DE LAS DEHESAS ANDALUZAS

García, A.1, Perea, J.1*, Acero, R.1, Angón, E.1, Toro, P.2, Rodríguez, V.1 and Gómez Castro, A.G.1

1Animal Production Department. University of Cordoba. Animal Production Building. Campus Rabanales. 14014 Cordoba. Spain. *[email protected] 2Pontificia Universidad Católica de Chile. Facultad de Agronomía e Ingeniería Forestal. Av. Vicuña Mackenna 4860. Santiago. Chile. *[email protected]

ADDITIONAL KEYWORDS PALABRAS CLAVE ADICIONALES Factorial analysis. Farm typology. Livestock farming Análisis factorial. Tipología de explotaciones. Sis- systems. temas ganaderos extensivos.

SUMMARY RESUMEN Three types of livestock farming systems Mediante análisis multivariante se establecie- are identified in Andalusian dehesas using ron tres tipos de sistemas ganaderos en las multivariate analysis. One of two conservationist dehesas andaluzas. En 49% de las explotaciones systems, both applying sustainable management se detectó un sistema ganadero denominado criteria, was detected in most of farms: dehesa conservacionista de dehesa que corresponde a farming system (49% of farms): small extensive pequeñas explotaciones de bovino y ovino con cattle and sheep farms, which adapt stocking bajo nivel de intensificación y carga ganadera rates to the availability of the land's natural ajustada a la disponibilidad alimenticia usando resources, and occasional use of strategic food ocasionalmente suplementación estratégica. Las supplementation; mountain farming system explotaciones (21%) del sistema conservacionista (21%) also relative to small farms, with mainly de sierra y montaña son también de reducida small ruminants and limited use of technology. dimensión, con predominio de pequeños rumian- The third system was a yield targeted system tes y bajo empleo de tecnología. El sistema (30%), corresponding mainly to large cattle productivista (30% de las explotaciones) corres- farms, with greater use of technology and high ponde a ganaderías con predominio de bovinos, levels of food supplementation caused by de mayor dimensión y nivel tecnológico que utili- stocking rates that exceed the land's carrying zan elevados niveles de suplementación pues sus capacity. Typology defined can be used as cargas ganaderas están por encima de la capa- starting point to base technical and economic cidad de la dehesa. Los tipos establecidos pueden characterization of farming systems taking into servir de punto de partida para la caracterización consideration their current and future viability. técnica y económica de los sistemas ganaderos considerando su viabilidad actual y futura. *This study was carried out as part of the Research Project: Prospective study of pastureland in order to INTRODUCTION develop an assessment methodology for extensive farming and its application to agricultural insurance, The dehesa is a system of land use and developed by Consejería de Agricultura (Junta de management based mainly on livestock Andalucía) and Departamento de Producción Ani- farming, as well as forest and agricultural mal (Universidad de Córdoba). farming, in Mediterranean grassland areas

Recibido: 10-11-08. Aceptado: 5-3-09. Arch. Zootec. 59 (228): 577-588. 2010. GARCÍA, PEREA, ACERO, ANGÓN, TORO, RODRÍGUEZ AND GÓMEZ CASTRO which gives rise to an agrosystem in which plans that incorporate the rational and the combination of agrosilvopastoral sustainable management of Andalusian management fosters important environ- dehesa, it is necessary to understand the mental values such as the sustainable use production systems associated with this of land, a balanced landscape and high ecosystem (Gibon et al., 1999). levels of diversity at different levels of Multivariate statistical techniques that integration (Comisión Técnica de la Dehe- quantify relationships of similarity between sa, 2006). structural variables (physical, size, intensifi- There are 5.8 million ha of dehesa in the cation), can be useful for the identification SW Iberian peninsula, 21% of which are in of systems and subsequent classification (Joffre et al., 1999). A large of farms, as proposed by Gibon et al. (1999) proportion of dehesa is in natural parks and and Solano et al. (2000). protected spaces (Junta de Andalucía, The aim of this study is to increase 2005). These very diverse region host a knowledge about the diversity and disper- mixture of agrosilvopastoral production sion of livestock farms in order to group systems, being livestock the primary product them according to their relationships of (Plieninger and Wilbrand, 2001). The farms similarity, both in terms of physical and on the dehesa, generally mixed, aim to extract technical characteristics, and to establish a the widest variety of available resources typology of extensive livestock systems in using as grazers different native species Andalusian dehesa ecosystem. (Milán et al., 2006), which are very well adapted to the harshness of the Medite- MATERIALS AND METHODS rranean climate, the poor quality of the soils and the seasonal nature of food availability Sample selection and survey data (Martín et al., 2001). Hence, livestock grazing collection design. Following the environ- systems are characterized by high levels of mental characterization of dehesas (Junta complexity, diversity and variability, de Andalucía, 2005), these areas are grouped conditioned by technical, social and into 5 geographical zones (table I) covering environmental factors (Martín et al., 2001; 1.2 million ha of dehesa and around 10 000 Escribano et al., 2002). extensive livestock farms (IEA, 2003). A Studies of dehesa systems in Extrema- randomized and stratified sampling was dura have revealed typological biodiversity performed on 206 farms (2% of the (Escribano et al., 2001 and 2002). In population) sampling was representative Andalusia, where conditions are very simi- and proportional to the distribution of lar, the typification of livestock systems zones, species farmed (Iberian pigs, cattle, has not been previously studied. The farms sheep and goats) and production system commonly implement environmentally used (simple and multifunctional). Infor- friendly production practices, like integrated mation was gathered in 2004 through direct and organic livestock farming. Andalusian questionnaires to the farmers and visits to dehesa farms are multifunctional systems, the farms, in accordance with the metho- where pigs commonly graze alongside other dology used by Dobremez and Bousset species (Porras et al., 2002). Most areas of (1995), Frías (1998), Acero (2001) and Milán Andalusian dehesa are within zones that et al. (2003). are protected by the Natura 2000 network Variable determination and database (Junta de Andalucía, 2005), where farms construction. The database generated from play an important role both in the develop- the questionnaires is made up of 34 varia- ment of these areas and protection of the bles (table II), which represents aspects of environment. In order to design strategic classification (4), agroclimatic factors (4),

Archivos de zootecnia vol. 59, núm. 228, p. 578. CHARACTERIZATION OF EXTENSIVE FARMS IN AN ANDALUSIAN DEHESA . Sample ) ) , 23 2.17 ) ) ) Quercus ilex Quercus pyrenaica Quercus ilex pasture and dense 62 2.06 pasture and scrub, dense 27 2.03 Olea europea and with pastures and normal dehesas 89 2.23 Quercus ilex andQuercus faginea Quercus suber and faginea Quercus faginea Predominant type of dehesa n % 2.56 Normal dehesas with pasture and scrub, 5 2.14 23.71 Normal dehesas with 21.42 Mixed dense dehesas ( Dehesa 1727 27.20 Normal dehesas with ). Sierra Sur, Serranía de Ronda, Sierra de Cádiz Peníbetica, 15422 Areas ha % SurSierra and dense dehesas ( Campo de , Antequera, 99630 dehesas ( Andévalo Oriental,Costa, La Sierra Norte, La Vega, Sierra Norte,del Campiña dehesas ( with pasture and scrub ( El Condado, La Sierra, Pedroches, Sierra Morena (Descripción de la muestra y su tipo predominante dehesa según densidad las especies arbóreas predominantes) Junta de Andalucía (2005 . Description of the sample and it predominant type dehesa according to their tree species percentage Table I of wooded area. Ronda La Sierra Sur Zones Campiña Campiña, 91445 Sierra Morena, W Andévalo Occidental, 43 Grazalema and Sierra Morena, EBaja, Campiña 584908 32.01 Open dehesas Source:

Archivos de zootecnia vol. 59, núm. 228, p. 579. GARCÍA, PEREA, ACERO, ANGÓN, TORO, RODRÍGUEZ AND GÓMEZ CASTRO

physical characteristics and size (19) and Table II. Variable codes and description. the intensification of the farms (7). (Codificación y descripción de las variables). The variables average annual rainfall Categories/Zone (AR) and temperature (AT) were obtained 1. Campiña; 2. Grazalema and Ronda; from the closest weather station to each 3. Sierra Sur; 4. West Sierra Morena; farm. Data from 76 stations were used. 5. East Sierra Morena Stocking rates (SR) were determined Production system using the methodology proposed by Mar- S. Simple, only one livestock species; tín et al. (1987) and developed by Escribano M. Multifunctional, several livestock et al. (1996), transforming the stocking rates species on the same farm into livestock units (LU) in order to make Farm specialization comparisons between farms and species 1. Cattle; 2. Pigs; 3. Sheep; 4. Goats; (Pulido and Escribano, 1994, Escribano et 5. Others (poultry, etc.) Vegetation al., 2002 and Acero et al., 2003). The 1. Open dehesas; 2. Normal dehesas; equivalent Livestock Units (LU) are as 3. Dense dehesas; 4. Olive groves follows: For cattle; 1 LU: cow older than 24 months; 0.6: cow between 6 and 24 months, Agro-climatology and 0.4: cow younger than 6 months. For AR Average annual rainfall (mm) sheep and goats: 0.15 breeding females, AT Average annual temperature (ºC) 0.12 adult males and 0.10 animals under a AL Altitude (m) year old. For pigs: 0.5: breeding animals, ASG Average slope gradient (%) 0.027: suckling pigs and 0.3: other pigs. For poultry: 0.014: egg-laying hens and 0.007: Physical and intensification variables TSA Total surface area (ha) chickens bred for meat. FSA Farming and livestock surface area (ha) The methodology proposed by Le POD Percentage of open dehesas (ha/TSA) Houerou and Hoste (1977) and adapted by PDD Percentage of dense dehesas (ha/TSA) Jiménez (1986) was used to determine PND Percentage of normal dehesas (ha/TSA) pasture production (PP). Finally, the RO Percentage of olive groves (ha/TSA) production capacity (CP) of each farm was FP Fodder production (kg DM/FSA) determined using the methodology descri- CP Production capacity of the farm (LU /FSA) max bed by Martín et al. (1986) for pastureland, LU Livestock units (LU) which establishes the maximum stocking CSU Cattle stock units (LU) rate for a farm based on prior estimations of PSU Pig stock units (LU) pasture production and the feeding SSU Sheep stock units (LU) GSU Goat stock units (LU) requirements of the different species of OSU Other stock units (LU of poultry, etc.) livestock on the farm. PC Percentage of cattle (CSU/LU) Variable and factor selection. The most P P Percentage of pigs (PSU/LU) representative primary variables of livestock PS Percentage of sheep (SSU/ LU) farming and the geophysical environment PG Percentage of goats (GSU/LU) were selected (system response variables). PO Percentage of other species (OSU/LU) The selection of variables was done using SR Stocking rate (LU/FSA) the following the steps: CSR Cattle stocking rate (CSU/FSA) - Determine the degree of association PSR Pig stocking rate (PSU/FSA) between variables analyzed, using the total SSR Sheep stocking rate (SSU/FSA) and partial correlation matrix, looking for GSR Goat stocking rate (GSU/FSA) OSR Other species stocking rate (OSU/FSA) original interrelated variables with no linear SUP Daily concentrated supplementation (kg/FSA) dependence (Uriel and Aldás, 2005; Castaldo et al., 2006).

Archivos de zootecnia vol. 59, núm. 228, p. 580. CHARACTERIZATION OF EXTENSIVE FARMS IN AN ANDALUSIAN DEHESA

- Statistical description of each variable. Newman-Keuls (SNK) multiple range test. Variance, coefficients of variation, etc. were Statistical analyses were performed using calculated; the variables with the greatest SPSS software, version 12. discriminatory power for the construction of groups were chosen, in accordance with RESULTS AND DISCUSSION. Álvarez and Paz (1998), who recommended the use of variables with coefficients of The size of the farms selected ranged variation higher than 60%. from 6 to 1300 ha and the cattle load fluctuated Factor analysis (FA) was used to reduce between 1 and 5000 livestock units (LU) dimensionality and to indicate the interre- used for meat production. Pasture is lationships between the original variables extremely important in the production (Ness, 1997; Hair et al., 1995; Pérez, 2002). system and contributes around 60-90% of The fit of the data to the factorial model was the animals' energy requirements. verified using Bartlett's test of sphericity and the Kaiser, Meyer and Olkin (KMO) FACTOR SELECTION AND FARM CLASSIFICATION test. Once the model had been verified, Twelve primary variables (FSA, LU, CP, factors were extracted using the principal PP, PS, PG, PS, AR, AT, SR, PC, SUP), that components procedure. In order to facilitate met the criteria detailed in the methodology the interpretation of common factors, a section, and representative of the dehesa Varimax orthogonal rotation was performed farm structure, were chosen. The factor (Uriel and Aldás, 2005). Once the factors analysis determined 8 factors that explain had been extracted, the farms were scored 95% of the total variance. From these, the and classified. first five factors were extracted, which Typification and characterization of accounted for 76% of the accumulated production systems. The farms were variance and presented eigenvalues greater classified using cluster analysis, esta- than one. Table III shows the factors blishing groups of farms that were homoge- selected, the factor loadings of the varia- neous within a group, but heterogeneous bles, the correlation between the variables between different groups (Júdez, 1989; Castel and the resulting factors following axis et al., 2003; Pérez, 2002). The hierarchical rotation. The farms were classified in cluster technique was used in accordance accordance with the score obtained for each with Ward’s method, and the Euclidean of the factors. Distance was used to measure similarity The first factor explains 21% of the total between farms (Castaldo et al., 2006). variance and shows positive saturation with The number of groups was established the variables stocking rate and daily food based on the dendrogram and the distance supplementation per farm (table III). This between clusters (DC) obtained during the factor differentiates the farms according to grouping process (Sharma, 1996). Finally, the degree of intensification. So, 37% of the the conglomerations obtained were vali- farms displayed high levels of intensification dated by verifying the stability of the results and are located to the right of the axis; through comparison with other distance whereas the remaining 63% are on the left of measurements (Squared Euclidean Dis- figure 1 and represent a model with high tance) and other grouping methods levels of extensification. Intensive farms (centroid) (Jonson, 1998; Köbrich et al., are mainly located in the west of Sierra 2003). Once the typology of farms had been Morena and La Campiña; they mostly farm defined, the groups were characterized and cattle and sheep and are predominantly sim- compared using a single-factor analysis of ple farms, with only one productive activity. variance (ANOVA) and the Student- They have large numbers of animals (99 LU)

Archivos de zootecnia vol. 59, núm. 228, p. 581. GARCÍA, PEREA, ACERO, ANGÓN, TORO, RODRÍGUEZ AND GÓMEZ CASTRO

Table III. Factor selection and degree of saturation of the variables. (Selección de factores y grado de saturación con las variables).

Factor Eigenvalue Variance Accumulated Variable Saturation Estimated (%) variance (%) communality

Factor 1 2.56 21.31 21.31 SR 0.98 0.97 SUP 0.94 0.92 Factor 2 2.18 18.20 39.51 CP 0.98 0.98 AR 0.90 0.98 Factor 3 1.80 15.05 54.56 PS -0.92 0.93 PC 0.86 0.83 AT 0.64 0.49 Factor 4 1.48 12.35 66.91 FSA 0.85 0.83 LU 0.80 0.77 P P -0.51 0.70 Factor 5 1.10 9.15 76.06 PG -0.97 0.97 and use a high stocking rate (0.68 LU/ha), inversely related to the proportion of cattle which is sustained through the high use of stock and the average temperature (positive food supplementation (0.29 kg/ha). The more saturation). Figure 1b shows the distri- extensive farms, on the other hand, are bution of the farms in relation to the third located in the west and east of Sierra More- factor: in the upper half, there are mostly na and Grazalema, covering areas of 200 ha, cattle farms (48%) and average temperatures with low stocking rates (0.25 LU/ha). The of 16ºC; the negative values for this factor animals graze on pasture throughout the mostly correspond to sheep (34%) and goat year, and their food is supplemented in (6%) farms with lower average temperatures certain seasons or when there is a shortage (15ºC). Therefore, this factor determines the of natural resources (0.01 kg/ha). substitution of activities and the proportion The second factor explains 18.19% of of land used for cattle and sheep farming. the total variance and presents positive This competitive and substitutive rela- saturation with production capacity and tionship between different activities is in rainfall, classifying the farms according to concordance with the findings of Escribano the potential of the pasture. Hence, high et al. (2002) in Extremadura. scores for this factor indicate farms with a The fourth factor explains 12.35% of the high carrying capacity, whereas low values total variance and presents a positive single correspond to areas with low potential. Fi- polarity with useful land surface area and the gure 1a shows that 68% of the farms are number of livestock units. When comparing located in areas with low carrying capacities, factors 1 and 4 (figure 1c), a proportional in comparison with the 32% that have high relationship is observed, whereby increased carrying capacities. Where the axes cross size corresponds to increased intensification. (F1 vs. F2), 38% of the farms are extensive Large intensive farms have up to 331 ha and with low carrying capacity in comparison 121 LU, with a marked predominance of cattle with 7% of intensive farms that have a high stock (62%). Therefore, this factor defines carrying capacity. the size of the farm. The third factor explains 15.04% of the Finally, the fifth factor explains 9.15% of total variance and is bipolar: the proportion the total variance and displays high of sheep stock (negative saturation) is specificity and negative saturation with goat

Archivos de zootecnia vol. 59, núm. 228, p. 582. CHARACTERIZATION OF EXTENSIVE FARMS IN AN ANDALUSIAN DEHESA stock. Figure 1d shows that most cattle sas, which is verified using the dendrogram (41%) and sheep (34%) farms obtain positive and the distance between the clusters (figu- values for this factor; whereas all goat farms re 2). Table IV shows the percentage score negatively. distribution of the farms in relation to the In line with the findings of Milan et al. category variables: type I farms are typical (2003) intensification and farm specialization of Grazalema and Ronda; type II farms are are the most relevant factors to explain mainly found in Sierra Morena (west and heterogeneity among farms. Of the five east) and type III farms are frequently found factors obtained, four refer to technical in Sierra Morena (west) and also in La Cam- variables and farm specialization: the level piña. of intensification (F1), carrying capacity Type I farms are less intensive, with (F2), the substitution of activities (F3) and complex orography, poor access and a the importance of goat production (F5). The greater presence of dense dehesa and effect of the size of the farm is ranked fourth, Mediterranean woodland. They represented and only explains 12% of the total variance. 40% of the farms. For farms in clusters II and III, the topographical conditions are milder: TYPIFICATION AND CHARACTERIZATION OF there is better access, widespread pasture LIVESTOCK FARMING SYSTEMS and the production systems tend to be more The cluster analysis, based on the five intensive. factors selected, identifies three groups or Even though most farms are still largely production systems in Andalusian dehe- owned and run by families, as we move from

(a) (b) 8 5 6 25% 7% 3 4 1 2

Factor 2 0 -1 Factor 3 Factor -2 38% 30% -3 -4 -3-2-1012345 -5 Factor 1 -3 -2 -1 0 1 2 3 4 5 Factor 1

(c) (d) 8 2 6 1 4 0 2 -1

Factor 4 Factor 0

Factor 5 Factor -2 -2 -3 -4 -4 -3-2-1012345 -3-2-10123 Factor 1 Factor 1

Figure 1. Projection of farms according to factorial analysis scores. (Proyección de explota- ciones según puntuación del AF).

Archivos de zootecnia vol. 59, núm. 228, p. 583. GARCÍA, PEREA, ACERO, ANGÓN, TORO, RODRÍGUEZ AND GÓMEZ CASTRO

800

600

400 Distance

200

0

Farms Figure 2. Dendrogram resulting from cluster analysis with three subsystems. (Dendrograma resultante del análisis cluster con tres subsistemas). type I to type III farms, there is an increased al., 2008). Several aspects explain the low presence of hired laborers and the land is relation between the variables linked to pig often leased rather than owned. farming and the typology of the dehesa: In relation to livestock farming activity, Iberian pigs are present in most of Anda- cattle and sheep stock compete for space, lusian pasture lands, making use of the and as the system intensifies, cattle production of acorns to fatten the animals production increases and sheep production on fodder (montanera); hence their decreases. Different cattle production presence does not explain the variability strategies are also observed; Type I farms observed moreover, the intensive pro- mainly focus on rearing; whereas most type duction of suckling pigs in large stables and III farms cover the entire production cycle the permanent stall housing of the mothers (rearing and fattening). mean that this production stage is displaced Sheep and goat stocks appear in less out of the pastoral system of grazing. As a intensive systems (types I and II), reflecting consequence, the pigs only graze in dehesa the food potential of the vegetation caused during the fattening stage (montanera), for by the substitution of pasture for scrub and less than four months; whereas sheep and dense dehesas. Extensive goat farming is cattle are always kept in the pasture; hence absent in type III farms, which tend to focus pigs do not compete with other livestock on milk production in farms without land. activities for space. In Andalusia, Iberian pigs are associated with dehesa and are an extremely important Group I. Conservationist mountain systems economic product; however, this does not This groups represents 21% of farms explain the heterogeneity between farms. and consists of small farms (around 32 LU Similar results were found in dehesa farms and 127 ha). In these systems, rainfall is of Extremadura (Milán et al., 2006; Gaspar et high, almost 1300 mm, and the average annual

Archivos de zootecnia vol. 59, núm. 228, p. 584. CHARACTERIZATION OF EXTENSIVE FARMS IN AN ANDALUSIAN DEHESA temperature is around 14.8ºC. The farms Table IV. Characteristics of the three included correspond to a model in which the production subsystems based on cluster land is extremely important in the production analysis. (Caracteristicas de los tres subsistemas of food. They are traditionally managed productivos derivados del análisis cluster). farms with limited use of technology. Even though their production capacity is the Cluster highest of the three groups (0-57 LU/ha), IIIIII stocking rates are low, which indicates the Distribution of farms (%) 21 49 30 scant availability of food resources, fundamentally as the result of complex Zones orography. Campiña - 9.7 25.0 These farms are chiefly located in steep Grazalema and Ronda 58.7 3.2 1.9 marginal areas: Antequera and Serranía de Sierra Sur 17.3 - - Ronda in the province of Malaga and Sierra Sierra Morena East - 18.3 9.6 Sierra Morena West 24.0 68.8 63.5 de Cadiz; where dense dehesa is increasingly important (table IV). In this environment, Production system the farms mostly focus on one activity, or Simple 48.3 73.1 76.9 mixed farming with small ruminants, such as Multifunctional 51.7 26.9 23.1 goats, which are better able to use the natu- ral resources, and contribute, through non- Farm specialization productive functions, to a model of social Cattle 20.7 48.4 61.5 economy that is an effective way of Pig 10.3 6.5 2.5 generating self-employment and persuading Sheep 48.3 33.3 35.4 Goat 17.2 10.7 0.5 the population from these areas to remain Other 3.4 1.0 - there, contributing to endogenous develop- ment and sustainability over time (Milán et Vegetation al., 2003). In some of the districts in these Open dehesas (pasture) 51.7 96.7 96.1 areas, farming is the only economic activity, Normal dehesas (scrub) 37.9 3.2 3.9 with no other alternative; hence their Dense dehesas 6.5 - - disappearance would exacerbate its depopu- Olive groves 3.5 - - lation and degradation. Physical and intensification variables Vegetation in these areas is extremely FSA (ha) 127a 95a 376b diverse, generally native and with abundant LU (LU) 32a 35a 139b flora: Mediterranean woodland, carob trees, CSU (LU) 29a 34a 129b wild olives, oaks, cork oaks, fir, Spanish fir, SSU (LU) 27a 31a 109b gall oaks, etc., with a predominance of scrub. GSU (LU) 8a 15a 50b The pasture has scant tree cover and the PC (%) 29.3a 46.6ab 60.3b herbaceous stratum is very highly developed. PP (%) 10.1b 8.9ab 0.8a SR (LU/ha) 0.32a 0.39a 0.52b Group II. Conservationist dehesa system CSR (LU/ha) 0.11a 0.17a 0.33b Group II accounts for 49% of the PSR (LU/ha) 0.023b 0.028b 0.001a population studied and corresponds to small AR (mm/m2) 1287b 645a 616a farms (around 35 LU and 95 ha) in comparison AT (ºC) 14.8a 15.6b 15.9b with the other groups, in particular group CP (LU/ha) 0.57b 0.41a 0.36a III. These farms are mainly located in Sierra SUP (kg /ha) 0.28a 0.20a 0.84b Morena; in East and West Andévalo, Costa and Sierra (Huelva); La Campiña, La Sierra Different letters, in a row, indicates significant differences between clusters (p<0.001). Norte, La Sierra Sur and La Vega (Seville);

Archivos de zootecnia vol. 59, núm. 228, p. 585. GARCÍA, PEREA, ACERO, ANGÓN, TORO, RODRÍGUEZ AND GÓMEZ CASTRO

Campiña Baja, La Sierra and Pedroches (0.52 LU/ha) is higher than dehesa farms of (Cordoba); Campiña del Norte, El Condado Extremadura (Gaspar et al., 2009). and Sierra Morena (Jaen). Average rainfall The farms are mainly located in the west and temperature are 645 mm and 15.5ºC of Sierra Morena (north of Cordoba and respectively. Taking into account the Cabe- Jaen) and in the districts of La Janda, La zas et al. (1986) climatological classification, Campiña and Campo de Gibraltar (Cadiz), in the traditional system corresponds to the areas with an average rainfall of 616 mm, and central group of pasture lands, with average an average temperature of 15.9ºC. This group rainfall of 600-700 mm. represents 30% of the population studied Average stocking rates in group II are and corresponds to specialized production 0.39 LU/ha, lower than the European avera- models that use a certain amount of ge (Colson and Chatellier, 1996) and similar technology. to the stocking rates indicated by Muslera These farms apply productivity based (1992) and Escribano et al. (2002) for Extre- management criteria, with stocking rates madura. above the carrying capacity of the land. As These farms apply conservation manage- a consequence, farming activity depends ment criteria, adjusting the stocking rates to largely on the contribution of food the carrying capacity of the land (0.41 LU/ supplements. These farms tend to focus on ha), using strategic or seasonal food cattle production, with high levels of supplementation in periods of shortage. As intensification and large herds. a consequence, farming activity is adapted In conclusion: five factors explain 76% to the environment and presents low levels of the heterogeneity of dehesa farming in of intensification; the animals are fed using Andalusia. The two main factors are: level the natural resources available through of intensification and carrying capacity. The pasture and fodder. most important factor variables are: stocking These farms mostly focus on the indivi- rate, level of food supplementation, farm’s dual production of sheep or cattle, which is production capacity and average rainfall. determined by the resources produced by Three livestock farming systems are the land and its orography. Combined identified in Andalusian dehesa. The farming of several species is relatively Mountain and Dehesa systems (70% of the infrequent (table IV). livestock farms), apply a conservation management strategy, adjusting stocking Group III. Yield targeted systems rates to the land's carrying capacity or even Group III is made up of large farms (139 below this capacity. The Yield Targeted LU and 376 ha), with values similar to those system (30% of livestock farms), corres- recorded in Extremadura (Escribano et al., ponds to larger farms, which apply criteria 2002) and Portugal (Carvalho and Gama, of intensification, with stocking rates above 1999). However, the average stoking rate the land's carrying capacity.

BIBLIOGRAPHY

Acero, R. 2001. Modelos avanzados de gestión y of Animal Production and natural resources optimización de la producción caprina extensi- utilization in the Mediterranean Mountain Areas. va en la provincia de Jaén. Doctoral Thesis. Ioannina. 5-7 June. University of Cordoba. Álvarez, R. and Paz, R. 1998. Metodología para la Acero, R., Martos, J., García, A., Luque, M., Herre- tipificación de la producción lechera de caprinos ra, M. and Peña, F. 2003. Characterization of en Santiago del Estero. Argentina. Arch. Zootec., extensive goat systems through factorial 47: 649-658. analysis. Proc. of 12th International Symposium Cabezas, J., Núñez, E. y Escudero, J.C. 1986.

Archivos de zootecnia vol. 59, núm. 228, p. 586. CHARACTERIZATION OF EXTENSIVE FARMS IN AN ANDALUSIAN DEHESA

Dinámica de las precipitaciones en los sistemas farms in the Spanish rangelands (dehesas): de dehesa en la provincia de Badajoz. En: MAPA Typologies according to livestock management (Ed.) Conservación y desarrollo de las dehesas and economic indicators. Small Ruminant Res., portuguesa y española. Madrid. 74: 52-63. Carvalho, L.M. e Gama, L.T. 1999. Caracterizaçao Gaspar, P., Mesías, F.J., Escribano, M. and Pulido, dos sistemas de produçao de algunas raças F. 2009. Assesing the technical efficiency of bovinas autóctonas portuguesas (Barrosã, extensive livestock farming systems in Mirandesa, Maronesa, Alentejana, Mertolenga). Extremadura. Spain. Livest. Sci., 121: 7-14. Proc. of 9º Congresso de Zootecnia. Porto, 10- Gibon, A., Sibbald, A.R., Flamant, J.C., Lhoste, P., 14 November. Revilla, R., Rubino, R. and Sorensen, J.T. 1999. Castaldo, A., Acero, R., Martos, J., Valerio, D., Livestock farming systems in and its García, A., Pamio, J. y Perea, J. 2006. Tipología potential contribution for managing towards de los sistemas de producción de engorde sustainability in livestock farming. Livest. Prod. bovino en la Pampa Argentina. Arch. Zootec., Sci., 61: 121-137. 55: 183-193. Hair, J.F., Anderson, R.E., Tathakm, R.L. and Black, Castel, J.M., Mena, Y., Delgado-Pertíñez, M., W.C. 1995. Multivariate data analysis with Camúñez, J., Basalto, J., Caravaca, F., Guzmán- readings. Prentice Hall. New Jersey. Guerrero, J.L. and Alcalde, M.J. 2003. IEA. 2003. Instituto de Estadística de Andalucía. Characterization of semi-extensive goat Censo Agrario. Sevilla. production systems in southern Spain. Small Jiménez, J. 1986. Una aproximación metodológica Ruminant Res. 47: 133-143. de un sistema de evaluación de la productividad Comisión Técnica de la Dehesa. 2006. Pacto An- potencial de un territorio de dehesa del SO daluz por la Dehesa. Boja de 4 de abril de 2006. peninsular. I Jornada Técnica sobre Conserva- Sevilla. ción y Desarrollo de la Dehesa Ibérica. Badajoz. Colson, F. et Chatellier, V. 1996. Les exploitations Joffre, R., Rambal, S. and Ratte, J.P. 1999. The bovines françaises sont parmi les plus dehesa system of southern Spain and Portugal extensives de l'Union européenne. INRA Prod. as a natural ecosystem mimic. Agroforest. Syst., Anim., 9: 273-284. 45: 57-79. Dobremez, L. et Bousset, J.P. 1995. Rendre compte Johnson, D.E. 1998. Applied multivariate methods de la diversité des exploitations agricoles. Une for data analysis. Brooks Cole Publishing démarche d'analyse par exploration conjointe Company. New York. de sources statistiques, comptables et technico- Júdez, L., Chaib Filho, H. and Pérez Hugalde, C. économiques. Cemagref. France. 1989. A statistical analysis of the evolution of Escribano, M., Pulido, F., Rodríguez de Ledesma, the size of farms in Spain. Investigación Agra- A. and Mesías, F.J. 1996. Determination of ria, Economía, 4: 155-174. energetic resources that cover cattle food needs Junta de Andalucía. 2005. Dehesas de Andalucía. in dehesa systems. Arch. Zootec., 45: 379-393. Caracterización Ambiental. Sevilla. Escribano, M., Rodríguez de Ledesma, A., Mesías, Köbrich, C., Rehman, T. and Khan, M. 2003. F.J. y Pulido, F. 2001. Tipología de sistemas Typification of farming systems for constructing adehesados. Arch. Zootec., 50: 411-414. representative farm models: two illustrations of Escribano, M., Rodríguez de Ledesma, A., Mesías, the application of multi-variate analyses in Chile F.J. y Pulido, F. 2002. Niveles de cargas gana- and Pakistan. Agr. Syst., 76: 141-157. deras en la dehesa Extremeña. Arch. Zootec. Le Houerou, H.N. and Hoste, C.H. 1977. Rangeland 51: 315–326. production and annual rainfall relations in the Frías, J.J. 1998. Situación actual y perspectivas mediterranean basin and in the african sahelo- de conservación de las razas caprinas en sudanian zone. J. Range Manage., 30: 181- peligro de extinción en la provincia de Jaén. 189. Doctoral Thesis. University of Cordoba. Martín, M., Espejo, M., Plaza, J. y López, T. 1986. Gaspar, P., Escribano, M., Mesías, F.J., Rodríguez Metodología para la determinación de la carga de Ledesma, A. and Pulido, F. 2008. Sheep ganadera de pastos extensivos. INIA. Madrid.

Archivos de zootecnia vol. 59, núm. 228, p. 587. GARCÍA, PEREA, ACERO, ANGÓN, TORO, RODRÍGUEZ AND GÓMEZ CASTRO

Martín, M., Espejo, M., Plaza, J. y López, T. 1987. Statgraphics. Pearson Educación S.A. Madrid. Cálculo de la carga ganadera en la dehesa. En: Plieninger, T. and Wilbrand, C. 2001. Land use, Campos, P., Martín, M. (Eds.). Conservación y biodiversity conservation, and rural development desarrollo de las dehesas portuguesa y espa- in the dehesas of Cuatro Lugares, Spain. ñola. SGT-MAPA. Madrid. Agroforest. Syst., 51: 23-34. Martín, M., Escribano, M., Mesías, F.J., Rodríguez Porras Tejeiro, C., González Sánchez, A., Brun de Ledesma, A. y Pulido, F. 2001. Sistemas Esquiliche, P., Sánchez Rodríguez, C. y Sán- extensivos de producción animal. Arch. Zootec., chez Núñez, R. 2002. Estudio Técnico-econó- 50: 465-489. mico de explotaciones de dehesa en Andalucía. Milán, M.J., Arnalte, E. and Caja, G. 2003. Economic Proc. 6 Jornadas Internacionales de la Socie- profitability and typology of Ripollesa breed dad Española de Ovinotecnia y Caprinotecnia. sheep farms in Spain. Small Ruminant Res., 49: Valencia. 10-13 mayo. 97-105. Pulido, F. and Escribano, M. 1994. The dehesa Milán, M.J., Bartolomé, J., Quintanilla, R., García- system. Analysis of typical dehesas of south Cachán, M.D., Espejo, M., Herráiz, P.L., Sán- best of Spain. In: Environmental and land use chez-Recio, J.L. and Piedrafita, J. 2006. issues in the mediterranean basin. Wissenchaft Structural characterization and typology of beef Verlag. Vauk. Kiel. cattle farms of Spanish wooded rangelands Sharma, S. 1996. Applied multivariate techniques. (dehesas). Livest. Prod. Sci., 99: 197-209. John Wiley & Sons. New York. Muslera, E. 1992. La dehesa: características pro- Solano, C., Bernués, A., Rojas, F., Joaquín, N., ductivas y alternativas ganaderas para su apro- Fernández, W. and Herrero, M. 2000. vechamiento. FIG-Zafra. Badajoz. Relationships between management intensity Ness, M. 1997. Multivariate analysis in marketing and structural and social variables in dairy and research. In: Padberg, D.I. Ritson, C. (Eds.). dual-purpose systems in Santa Cruz, Bolivia. Agro-Food marketing. Albisu CAB Internacio- Agroforest. Syst., 65: 159-177. nal, Oxon. Uriel, E. and Aldás, J. 2005. Análisis multivariado Pérez, C. 2002. Estadística práctica con aplicado. Thomson. Madrid.

Archivos de zootecnia vol. 59, núm. 228, p. 588.