In the last 50 years, agricultural intensification has been based on the Green Revolution model characterized by the adoption of industrial synthetic inputs (agrochemicals), seed selection, single-crop systems, specialized livestock breeds, and technology reliant on non-renewable energy sources (Funes-Monzote, 2008; Altieri et al., 2012; Serrano-Tovar, 2014; Preston et al., 2021). The specialized dairy sector in Colombia mainly concentrated in the High Andean region has followed this trend, and stands out as one of the most productive livestock systems in the country (Holmann et al., 2003; Carulla and Ortega, 2016; UPRA, 2020). Besides, this type of intensification in livestock systems implies high costs, inefficiencies, reliance on non-renewable energy, and limited profitability (Holmann et al., 2003; Llanos et al., 2018). This is especially evident in regions characterized by special biophysical conditions, such as the high Andean hillsides. This production model also gives rise to both the direct and indirect environmental consequences observed in conventional intensive systems (Funes-Monzote, 2008), including the loss of biodiversity in a globally significant region (Orme et al., 2005). In contrast, dairy farming in Colombia faces threats from climate change and the demands of globalized markets, which demand increased competitiveness (Carulla and Ortega, 2016; Cadena et al., 2019). This situation leads to a reflection on the sustainability and resilience of dairy farming in the high tropics, within the global discussion derived from the growing demand for animal protein and, at the same time, for environmental services (FAO, 2018; Preston et al., 2021).

The Northern Tropical Andes are considered one of the richest and more biologically diverse regions globally, accounting for more than 100 different ecosystems, 45,000 vascular plant species (20,000 endemic), and 3,400 vertebrate species (1,567 endemic), all within just 1% of the earth’s continental area (Josse et al., 2009). The mountain forests of the Northern Andes hold great significance for conservation efforts, being recognized as one of the world’s six biodiversity hotspots. This region has been a focal point for the diversification of numerous species (Scatena et al., 2010). Additionally, it is also recognized as an area with a high level of endemisms (Tejedor et al., 2012) and threatened species, being the only region on the planet in which these three categories coincide (Orme et al., 2005). The heterogeneity of ecosystems in the Andean region stems from the emergence of the longest mountain range on Earth (Orme, 2007). This diversity is further shaped by factors such as volcanic activity, tectonic shifts, soil formation, and the equatorial climate with year-round rainfall distribution (Guhl, 1959; IGAC, 2015).

In Colombia, the Andes Mountain range divides into three branches, resulting in distinct environments on both the western and eastern flanks. These disparities encompass a wide range of factors, including differences in precipitation levels (varying from 500 to 4,000 mm), average temperatures (ranging from 12° to 18° C), evapotranspiration rates, altitude, and topographic features (Rodríguez et al., 2006). Despite the differences, some characteristics can be generalized in the Andean region, such as the maintenance of a stable temperature throughout the year, albeit with fluctuations in daily maximum and minimum temperatures of up to 20°C (Buytaert et al., 2006), relatively low temperatures comparative to those at lower altitudes in the tropics, steep, sloping topography (>12%), smooth highlands, and fog in higher elevations (Hall et al., 2015). The Andean region experiences the dual influence of the Pacific Ocean to the west and the Orinoco and Amazon River basins to the east. The interplay of oceanic and continental air masses in the region’s intertropical confluence zone results in a bimodal pattern of rainfall (Buytaert et al., 2006). This distinguishes it from other regions in Colombia and provides specific benefits for agricultural endeavors, especially in the production of coffee, tropical fruits, and livestock.

The earliest human settlements in the eastern Andes region of Colombia can be traced to approximately 10,000 to 12,000 B.C., as evidenced by the Tequendama rock shelters (Correal Urrego and Van der Hammen, 1977). Evidence suggests that the Andes ecoregion has been transformed by humans for nearly 9,000 years (Young, 2009). Currently, 70% of the Colombian Andean region is used for agricultural activities, especially livestock, which began to increase its coverage and intensity from the beginning of the 20th century (Etter and Van Wyngaarden, 2000; Murgueitio, 2003). The grasslands expanded in newly converted areas from 65.8% in 1750, to 97.2% after 1970 (Etter and Van Wyngaarden, 2000). These alterations in land use have resulted in the modification of landscapes, leading to significant fragmentation of both altitudinal and longitudinal corridors within the Andean forests, thus affecting biodiversity (Etter and Van Wyngaarden, 2000; Young, 2009).

Despite the extent of alteration, the paramo ecosystems cover approximately 1,925,410 hectares in the Colombian Andes, of which 746,644 are in National Natural Parks. They host natural habitats characterized by high level of endemism at nearly 90% (Rivera and Rodríguez, 2011). In addition, relicts of Andean forests are still preserved on the highest mountainsides, as well as on the eastern and western flanks, while on the slopes toward the inter-Andean valleys the ecosystems are highly deteriorated, with some areas retaining only 10% of the original ecosystem (Tejedor et al., 2012). Part of this remaining biodiversity is found within cattle farms in which relicts of native forests are preserved (Chaves et al., 2007). These areas represent a crucial focus for restoration and conservation efforts due to the ecoregion’s significance.

The definition of the High Andean region ecosystem varies according to the mountain range and authors’ classification, with altitudes between 2,000 and 3,700 meters above sea level (Rodríguez et al., 2006). Here, references made to the High Andean region correspond to the mountainous areas of the Andes located higher than 2,000 m.a.s.l. and lower than the paramo ecosystem, generally up to 3,200 m.a.s.l., although important local variations are recognized that can expand or reduce the upper limit (Rivera and Rodríguez, 2011; UPRA, 2020).

The mountain forests and part of the lowland paramos were transformed by extractive agricultural and mining human settlements in addition to other civil work uses. Currently, most Andean Mountain land outside protected areas and inhabited regions is dedicated primarily to cattle farming (Etter and Van Wyngaarden, 2000). These farms typically operate according to the predominant conventional model, focusing on specialized or dual-purpose milk production (FEDEGAN, 2021). These are in highlands and slopes over 2,000 meters above sea level with temperatures between 12 and 17°C (Murgueitio, 2008; Carulla and Ortega, 2016; UPRA, 2020). Collectively, they contribute to 32% of the country’s total milk production, equivalent to 2,270 million liters per year (approximately 6.21 million liters per day) (FEDEGAN–Fondo Nacional del Ganado, 2022). This production is obtained in moderate to high-tech grazing systems on predominantly small and medium-sized farms with an average of 3,480 liters/ha/year and 3,689 liters/cow/year (Ministerio de Agricultura y Desarrollo Rural de Colombia, 2020). Livestock farms are grouped into dairy farm regions: the “Cundiboyacense” highland, with 44% of production, Antioquia, with 45%, and Nariño, with 5% (Ministerio de Agricultura y Desarrollo Rural de Colombia, 2020), presenting differences in productivity between regions and production types (FEDEGAN, 2019). Map 1 illustrates the dairy areas of the highland tropics in Colombia.

The intensified specialized dairy farming follows the principles of the Temperate Grazing System (TGS), a model that has been established in southern Australia, New Zealand, select areas of the United States, China, Chile, Mongolia, and the highlands of South America (Seré et al., 1996). This system is characterized by the utilization of enhanced forage, extensive irrigation, high nitrogen fertilization, concentrated feed, and the application of specialized dairy genetics. It results in above-average production levels compared to the national average in Colombia (FEDEGAN, 2015; UPRA, 2020).

This livestock activity is threatened by climate variability which has been manifesting in increasingly extreme patterns (IDEAM et al., 2015), as well as market volatility (FEDEGAN–Fondo Nacional del Ganado, 2022). Considering this, the sustainability of milk production systems in high-tropical regions has become a frequently discussed concern (Durana, 2011; Ruiz et al., 2019; Escobar et al., 2020). In this article, we examine sustainability within the conceptual framework established by Giampietro and Mayumi (2000). Our analysis incorporates field trials, participant observation, interaction with farmers, expert consultations, and a comprehensive literature review. Therefore, we evaluate current production models alongside agroecological practices from silvopastoral systems, considering their ecosystemic context.

Environmental feasibility, as a component of sustainability, is associated to the external limitations that biophysical factors impose on the production system and are beyond human control (Serrano-Tovar, 2014). With regards to conditions in the biophysical environment, one of the current problems of livestock farming in the high tropics is that it is implemented without considering the particularities of mountain ecosystems such as: (i) steep mountainsides susceptible to water erosion and high soil diversity (originating in sedimentary rocks, sediments, pyroclasts, igneous and metamorphic rocks, as well as their combinations), most of which contain chemical, physical, or biological limitations (Malagón, 2003; IGAC, 2015); (ii) variable ranges of rainfall with increasingly noticeable variability (IDEAM et al., 2015); and (iii) low temperatures, limited daylight hours caused by persistent fog cover and diminished oxygen levels at high altitudes (IDEAM et al., 2007). In addition, their heterogeneity is not taken into account, particularly in relation to the differences between more fertile high plateaus, and hillside or slope areas with less productiveness.

Within the sustainability framework used in this document, the viability component refers to the internal conditions of the system: technological, economic, and social factors necessary to maintain its structure, identity, and functions (Serrano-Tovar, 2014). In the context of dairy farming in High Andean region, two contrasting production models can be discerned: extensive and intensive grazing, encompassing a spectrum of production systems that vary in terms of productivity, profitability, and environmental impact (Holmann et al., 2003; Carulla and Ortega, 2016; Cadena et al., 2019; UPRA, 2020). Although attempts have been made, industrialized cattle confinement dairy models such as those in North America, Europe, Argentina, and Uruguay do not persist in Colombia due to economic infeasibility (De Haan et al., 1977; Frossasco et al., 2015).

Desirability or convenience is one of the factors that influence the sustainability of a socioecological system and refers to its alignment with the social values that are expressed through culture, regulations, and institutions (Serrano-Tovar, 2014; Giampietro, 2015). This concept has a subjective component since it depends on the perspective of different actors involved; thus, it must be constructed from a consensus (Serrano-Tovar, 2014).

In the case of dairy farming in the Colombian high tropics, there is consensus that production is desirable. This is due to the significant amount of milk produced (32%) in a proportionally low area (9%), contributing to food security and economy while generating livelihoods and employment for the rural population. This, primarily comprises mainly small and medium-sized farmers (80%), along with the impact on various other participants of the dairy supply chain (Holmann et al., 2003; Carulla and Ortega, 2016; Cadena et al., 2019; UPRA, 2020). However, there is also consensus regarding the adverse effects of production systems, including pollution, the depletion of non-renewable resources, and the alteration of natural ecosystems (Murgueitio, 2008; Ruiz et al., 2019). It is also consensous that the biodiversity in these areas is important for environmental services (Calle, 2020).

The resilience of a socio-ecological system is defined as its ability to assimilate external disturbances, reorganize itself and preserve its structure, functions, and identity (Walker et al., 2004). It depends on the adaptability of individuals and social groups, that is, on the strategies derived from learning and innovation processes developed to assimilate changes in the environment (Salas-Zapata et al., 2012; Nicholls, 2013; León-Siccard, 2014). In agricultural systems, the environmental impacts at different scales can result in increased fragility in agroecosystems and a decrease of their resilience (Funes-Monzote, 2008; Altieri, 2013). In dairy farming in the High Andean region, the most significant challenges in recent years have been changes in climate patterns and the variable market conditions.

Climate change is evident in the form of fewer rainy days throughout the year in certain regions, an increase in atypical dry periods, more frequent frosts outside of typical seasons in the highlands, and longer and more intense rain periods than usual. In climate systems influenced by the Pacific Ocean, such as those in in the Andes, the frequency of the phenomena of El Niño and La Niña has increased significantly in the last two decades (Hurtado and Gonzalez, 2011; IDEAM et al., 2015). For example, in the milk production zone in the Ubaté and Chiquinquirá valley (Cundinamarca – Colombia) Gómez (2014) found that El Niño phenomenon increases the probability of frost by between 40 and 80% in the first dry months of the year. Furthermore, the outcomes of climate simulations show that kikuyu grass (Cenchrus clandestinus) is highly susceptible to frost, resulting in a decrease in milk production yields near 20%.

In seasons of prolonged or more intense rains, fodder production is also affected. This is compounded by poor water management, which exacerbates erosion, often leading to landslides that impact production areas, access roads, and livestock infrastructure. This, in turn, results in additional management and restoration costs. On the other hand, there is a loss in competitiveness with other countries that produce several times the volume of milk in Colombia and receive subsidies from their governments (Carulla and Ortega, 2016). This situation favors industries and large retail outlets, while significantly, and negatively, affecting dairy farmers, especially the smallest ones (Holmann et al., 2003; UPRA, 2020). All these factors combined simultaneously test the production system’s resilience, as illustrated in Figure 2.

Figure 2 shows the trends in the establishment and intensification of livestock systems in highland Andean regions. In the first years, there is a process of deforestation and a decrease in the ecosystem services derived from the High Andean region forest, which drop abruptly in the areas that are transformed into pastures (a). The livestock system with the extensive model (blue lines) expands to new areas and a low level of productivity (blue line) and profitability (blue dotted line) is consolidated (b), which are gradually reduced when soils and pastures are degraded, and climate phenomenon occur (c and d). When an intensification process is carried out with the conventional model, productivity (red line) and profitability (red dotted line) increase (c), but new threats appear with climate change, input expenses, and additionally competitiviness in the globalized market. Under these conditions, productivity cannot be maintained unless external inputs are increased, which in turn affects profitability (d). Ecosystem services continue to deteriorate (b, c, and d).

Recent recommendations promote the transition toward models less dependent on inputs, with greater climate resilience that are concerned with environmental services, social responsibility, and animal welfare (Murgueitio et al., 2016; Gachetá et al., 2018; Mauricio et al., 2019; Escobar et al., 2020; Lentijo et al., 2022; Montoya Uribe et al., 2023).