Sample population

Data for this characterization was obtained from the Sustainable Colombian Cattle Ranching (Ganadería Colombiana Sostenible - GCS) and LivestockPlus (L+) projects. The GCS project consists of 2,011 surveys administered on livestock farms with different production focuses: cow-calf, finishing, dual-purpose, specialized dairy and complete cycle. Selection of the surveyed areas prioritized livestock regions deemed important in terms of environmental attributes, the existence of ecosystems of global importance and proximity of protected areas and significant livestock production areas. The surveyed cattle farms were in twelve departments (number of municipalities in parentheses): Atlántico (13), Bolívar (4), Boyacá (12), Caldas (2), Cesar (10), La Guajira (5), Meta (10), Quindío (9), Risaralda (2), Santander (4), Tolima (6) and Valle del Cauca (7). Three criteria were used to select the farms included in the sample: 1) location in regions prioritized by the GCS project; 2) farm area greater than 2 ha; and 3) owned by Colombian citizen(s). A ten-component questionnaire was applied at each farm: 1) general farm information; 2) livestock composition and management; 3) pasture management practices; 4) livestock productive and reproductive information; 5) animal health; 6) environmental information; 7) social information; 8) organizational information and relationship with external environment; 9) income from livestock; and 10) financial information.

The L+ project involved surveys applied at livestock farms in six municipalities: Cumaral and Restrepo (Piedemonte region), and Puerto Gaitán and Puerto López (Altillanura region) in the department of Meta; and Patía and Mercaderes in the Patía Valley in the department of Cauca. Surveys were conducted at 607 livestock farms in the three regions: Meta Piedmont (150); Meta high plains (147); and dry valley of Patía (310). The questionnaire consisted of eight components: 1) general farm information; 2) farm administrative information; 3) farm land use; 4) technical assistance; 5) productive and reproductive characteristics; 6) association membership; 7) commercial and financial information; and 8) climatic events and environmental practices.

From the total sample of 2,618 cattle farms surveyed between the two projects, 251 farms were identified. Of these, 165 (65.7 %) were cow-calf operations without dairy production and 86 (34.3 %) were cow-calf operations with dairy production. This sample (n= 251) was stratified based on herd size⁸: very small (VSP: 1 to 30 head); small (SP: 31 to 50 head); medium (MP: 51 to 250 head); and large (LP: ≥251 head). The numerical and categorical variables used in the questionnaires from each project were classified into five groups (Table 1).

Statistical analyses

Analysis each of the five components was done with the Factorial Analysis for Mixed Data multivariate technique (FAMD), using the FAMD function of the FactoMineR package for the R statistical package⁹. The FAMD is a factorial method applied to analyze data sets in which a group of individuals is described with quantitative and qualitative variables. The term “mixed” refers to the simultaneous presence of qualitative and quantitative variables as active elements in the sampling units. This method allows simultaneous exploration of these variables by combining principal component analysis (PCA) and multiple correspondence analysis (MCA)¹⁰. The quantitative variables were centered and normalized to Z values, while the qualitative variables were disaggregated into a normalized disjunctive matrix. Starting from mixed samples, this method allows graphic evaluation of the similarities/dissimilarities between productive units (distances) and the correlations between continuous variables¹⁰. Prior to the application of the FAMD, missing data was imputed using the algorithm implemented in the impute FAMD function, which is part of the missMDA package¹¹. The producer category and animal number variables were included in the FAMD as supplementary inputs to exclude their participation in model construction.

General farm information

The configuration of the categorical variables used to characterize the general farm information resulted in an orderly alignment of the producer categories in the first dimension of the FAMD graph (Figure 1(a)). Close associations existed between the LP and MP categories and the machinery and equipment variables (electronic scale, tractor, chainsaw, motor pump, electric pump, drive pen, stable, chute and storehouse). In contrast variables related to non-use of these technologies were to the left of dimension 1, alongside the variables of agroforestry crops and monoculture forest plantations, both associated with the VSP and SP categories. In dimension 2, the presence of small, medium and large species on the farm, and use of a mower and electric fence grouped in the lower portion while the absence and/or non-use of these variables were located in the upper portion; no apparent relationship existed between these variables and producer category.

Six numerical variables were positively related to the first dimension, representing farm size: total farm area; livestock area; improved pasture area; agroforestry crop area; monoculture forest plantation area; and transitory crop area (Figure 1b). These variables were also closely linked to the number of cattle; for instance, the number of buffaloes and undulating area on a farm were negatively associated to dimension 1, meaning that these conditions were more common at small farms. In dimension 2, positively linked variables included number of pigs and hilly area on farm, while the negatively linked variables were number of hens, goats and flat area on farm.

The association of MP and LP with infrastructure (i.e. machinery, equipment and facilities) reflects their greater economic capacity. In addition, cattle producers with larger herds generate more income and profits than smaller producers, allowing them to build facilities and purchase more and better equipment and machinery, all of which are costly inputs in livestock production⁷. Similar circumstances have been reported for Mexico in which larger farms have greater machinery and equipment availability and these were more apt for livestock production¹³.

Larger farms were negatively associated with the undulating area on farm variable but positively associated with the presence of mules and horses (large species). It is noteworthy that the presence and use of other animal species could favor food availability, stability, access and consumption; these are the four dimensions of food security and may benefit producer sustainability¹⁴.

Hilly slopes (>30 %) are not suitable for grazing¹². The percentage of hilly area (slope >60%) on farm was highest in the VSP (25.5 %), followed by the SP (22.0 %), MP (14.2 %) and LP (10.5 %). Grazing livestock on hilly slopes can cause land degradation due to increased erosion associated with livestock trampling¹⁵. This leads to decreased soil filtration capacity and consequent greater surface runoff during rains¹⁶, which translate to lower biomass production and livestock productivity. Also, as pastures degrade net annual GHG emissions tend to rise¹⁷. It is therefore important to estimate GHG emissions on livestock farms to identify their contribution to global GHG emissions and propose appropriate mitigation actions.

The number of the evaluated cow-calf operations that also engaged in agricultural activities or grew some crop was low in all four producer categories (Table 2). This contrasts with small producers in the state of Veracruz, Mexico, where 85% combined livestock activity with agriculture¹⁸. It is to be expected that the overall crop area on the farms was low. Most cow-calf operations in Colombia are extensive systems¹⁹, and the current results are therefore more comparable to those for extensive and semi-extensive livestock systems in Mexico²⁰, in which the area used to grow crops was reported as less than 20 %. Joint agricultural and livestock production helps to guarantee food security. It can also contribute to climate change adaptation and mitigation by, for example, incorporating short-cycle or transitory crops into grazing systems²¹.

Technical component

Productive and reproductive information

Among the categorical variables for productive and reproductive information (Figure 4(a)), the productive and reproductive practices variables tended to group on the right side of dimension 1: weighing at birth and weaning, controlled mating, artificial insemination, embryo transfer, palpation of cows, reproductive examination of bulls, weighing of heifers for first service and separation of the dry cow lot, in addition to existence of a birth pasture, artificial insemination equipment and inseminator. Those variables quantifying the absence of these practices, activities, facilities and/or equipment grouped to the left side of dimension 1.

In dimension 2, production practice variables such as weighing by measuring tape, breeding male calves and finishing animals grouped at the bottom. Variables representing the absence of these practices grouped in the upper portion, as well as calculation of intercalving interval and weighing with scales. These groupings suggest that the MP and LP implemented better productive and reproductive practices than did the VSP and SP. Use of better reproductive and productive practices by the MP and LP may translate into higher productivity. These results are similar to the reported absence at small producers of practices such as weighing heifers for first service, reproductive examinations, use of artificial insemination, availability of birth pastures, separation of the dry cow lot and calculation of the intercalving interval²³. These reproductive practices are more common in larger more specialized livestock systems with greater financial resources, and manifest in higher rates of productivity⁷.

Weighing cattle is important for assessing growth, planning feeding to maximize yields and take full advantage of available feed resources, properly managing technical and financial records, and implementing monitoring, medication, genetic improvement and reproductive management²⁷. More of the MP and LP producers weighed their animals than did the SP and VSP, a fact which may be linked to the greater availability of scales at larger farms. Barimetry facilitates estimation of approximate animal live weight through body metrics and formulas²⁷. These should therefore be promoted and applied at VSP and SP farms, where scales are rare, and thus provide them with greater control of yield rates and potentially increase their productivity.

The variables of weaning age, number of milked cows, number of births and total milk production (L farm^-1 day^-1) were positively related to each other and to dimension 2 (Figure 4(b)). One measure that could improve herd reproductive performance is early weaning, which reduces cow energy requirements, resulting in weight gain, improved body condition, increased pregnancy rates and better overall reproductive performance²⁸. However, this requires adequate nutritional management of weaned calves to ensure sufficient quantities of good quality energy and protein, which can incur additional costs for producers²⁹.

In the present analysis the number of animals was negatively linked to mortality rates (Figure 4(b)). The opposite was reported for livestock systems in Costa Rica where mortality rates tended to be lower at smaller farms (<30 head)³⁰. In small farms the death of one animal has a proportionally greater effect than it would at larger farms. No matter farm size, a lower mortality rate can result in greater profitability and competitiveness, and consequent higher income.

Preweaning daily weight gain (DWG), and weights at birth and weaning were higher in the MP and LP. Higher birth weight offspring are reported to grow more rapidly and experience lower mortality, which would support the reported positive correlation between birth weight and weaning weight in beef cattle³¹. However, some studies indicate that offspring with the highest birth weight do not necessarily reach the highest weaning weight^32,33. Nonetheless, a greater preweaning DWG does result in higher weight at weaning, which, through sale of heavier animals for finishing, can result in greater profitability³².

Environmental information

In this component no separation from the centroid was observed (Figure 5), suggesting that the different producer categories did not differ in terms of environmental practices. Promoting broader implementation of good environmental practices on farms requires strategies that generate productive and economic benefits for producers in addition to environmental benefits such as climate change mitigation and adaptation. For example, the conservation of trees and shrubs in pastures and silvopastoral systems is a production strategy adopted in livestock systems in the Latin American tropics. It contributes to reducing extreme temperatures, conserving wildlife, controlling water quality in watersheds, capturing atmospheric carbon, mitigating climate change, as well as improving milk and meat production and livestock profitability³⁴.

The wastewater generated in livestock systems usually originates from any on-site dwellings and the stables. Few (<40%) of the farms included in the study employed wastewater treatment systems meaning any liquid waste was discharged directly into water sources and soils. Those farms with treatment systems largely used septic tanks and biodigesters, which are primary and secondary low nutrient removal processes. Therefore, direct discharges and effluents from treatment systems could generate eutrophication in water bodies in response to nitrogen (N) and phosphorus (P) from cattle excreta. Tertiary treatment systems that increase nutrient removal from discharge are vital to reduce eutrophication phenomena.

Non-biodegradable solid waste is one of the main environmental impacts of the cattle production chain in Colombia, but a lack of quantitative studies prevents accurate estimation of its magnitude³⁵. Incineration and burial were the main forms of solid waste management and were used at more than 70% of the farms. Incineration generates GHG emissions that increases potential global warming, while burial without proper sanitary controls can contaminate aquifers. Characterization of cattle producer solid waste processing capacity is an important variable for quantitatively assessing the degree of potential negative impacts generated at farms and for establishing management strategies to minimize them.