Capturing the multidimensionality of land-use agents in a deforestation hotspot

Abstract

Given increasing recognition that strategies to transition to sustainable land use should be context specific, structuring the diversity of land-use agents is important. This is particularly so for the world’s tropical deforestation frontiers, where rapid land-use change, driven by diverse agents, leads to stark social-ecological trade-offs. Focusing on the Argentinean Dry Chaco, a global deforestation hotspot, we employed archetyping to identify key types of land-use agents using data from a questionnaire survey covering three main dimensions: agents’ capital assets (what they have), agents’ activities and management (what they do), and agents’ personal characteristics (who they are). We identified five well-differentiated types of land-use agents: forest-dependent smallholders, semi-subsistence ranchers, crop–livestock farmers, agribusiness farmers, and commercial ranchers. Characterizing these major agent types yielded three main conceptual and methodological insights. First, we reveal considerable heterogeneity of land-use agents in the Argentine Dry Chaco, allowing us to move beyond the common yet oversimplified and dichotomic view of agribusinesses vs. smallholders. Second, the agent typology based on all three dimensions captured the diversity of agents much better than any one-dimensional typology alone, demonstrating the value of richer descriptions of land-use agents. Third, all our agent types share characteristics in some dimensions yet differ in others (e.g., forest-dependent smallholders and crop–livestock farmers were similar in who they are, yet different in what they do), explaining how more simplistic agent descriptions arrive at oversimplified agent types. Overall, our work highlights how archetyping can structure complex human–environment phenomena, diverse land-use agents in our case, for guiding tailored, actor-specific policy interventions.

INTRODUCTION

Land use is a main driver of global environmental change and is tightly connected to many of the most urgent sustainability challenges of our times, including climate change, biodiversity loss, and food security (Jaureguiberry et al. 2022). Understanding how and where land use changes, and how these changes impact on social-ecological systems, is therefore important. A promising avenue for achieving a deeper understanding of land-use change is to focus more closely on the agents driving these changes (Malek et al. 2019). This is particularly urgent for the world’s tropical and subtropical regions, where different forms of industrialized agriculture are expanding (Buchadas et al. 2022, Meyfroidt et al. 2024), often into semi-natural landscapes used and inhabited by a range of traditional and local communities (Cáceres 2015, Levers et al. 2021). Although the expansion of industrialized agriculture provides important commodities and contributes to nation’s economies, the social-ecological trade-offs of these land-use changes are typically negative for both the environment and local people (Mastrangelo and Laterra 2015, Barral et al. 2020, Meyfroidt et al. 2022). As a result, many tropical and subtropical regions are in dire need of sustainability planning (Lambin et al. 2014, Schröder et al. 2021).

Steering social-ecological systems toward sustainability requires a deeper consideration of the people engaged in land use, in other words, the land-use agents (Malek and Verburg 2020). Land-use agents are autonomous individuals with a particular environmental cognition (i.e., values, intentions, attitudes, decision rules) who interact with each other and the ecosystems in the same place and time (Meyfroidt 2013). Understanding and characterizing land-use agents is not trivial, as observed land-use patterns and changes are the aggregated outcome of many individual decisions, made by very different agents, often with diverse cultural backgrounds, capital assets, and motivations (Valbuena et al. 2008, Smajgl et al. 2011). Capturing and representing this complexity is a general challenge in sustainability science (Kuemmerle 2024), yet it is needed to understand land-use decision making and guide the design of conservation and land-use policies (Huber et al. 2024).

A key challenge is how to structure and organize the large diversity of agent characteristics while avoiding overgeneralization and oversimplification (Meyfroidt et al. 2018). The development of archetypes builds on a long-standing tradition of typological approaches, which have been widely applied in social and rural sciences (McKinney 1950, Jollivet 1965, van der Ploeg 1994), and has gained increasing prominence in sustainability science over the past decade (Oberlack et al. 2019). Archetype analysis offers a robust and adaptable approach for identifying recurrent, empirically grounded patterns that are also theoretically informed. Its strength lies in the capacity to integrate diverse methodological perspectives to reveal context-specific mechanisms, while enabling the derivation of generalizable insights aligned with specific research questions (Eisenack et al. 2021). Archetype analyses can be carried out both in a top-down or bottom-up fashion, each with specific advantages (Food and Agriculture Organization (FAO) 2016, Pacheco-Romero et al. 2021). Top-down, deductive approaches such as expert-based categorization of cases (e.g., land-use agents) have the advantage of being more easily linked to theory, being transferable, and making use of diverse types of data (e.g., qualitative and quantitative data), but require a priori knowledge on the number of archetypes (Oberlack et al. 2019, Pratzer et al. 2024). Conversely, bottom-up, inductive approaches such as data-driven clustering can discover how cases structure into archetypes (Tittonell et al. 2020, Vallejos et al. 2020). Importantly, the relative strengths of both approaches can be combined, and comparing top-down and bottom-up archetyping can be highly informative (Pacheco-Romero et al. 2021).

Archetyping will inevitably lead to reduction and generalization, which is why it is important to start from a broad set of characteristics and classification criteria that describe cases (FAO 2016, Martinelli 2011). Land-use agent typologies have usually focused on the capital assets of landholders (what they have) (Tittonell et al. 2020). However, land-use agents differ from one another not only in their capital assets, but also in how they manage those assets, especially, land (what they do), and in their personal characteristics (who they are) (Valbuena et al. 2008). So far, typologies have either focused on the structural or management characteristics of landholdings, resulting in “farm” typologies, or, conversely, on the personal characteristics of landholders, resulting in “farmer” typologies (Huber et al. 2024). Such simple typifications of land-use agents have a limited capacity to explain the behavior of land-use agents and to capture the emergence of social-ecological phenomena at the regional scale (Lokhorst et al. 2014, Mastrangelo 2018).

Understanding the diversity of land-use agents involved in land-use changes is particularly urgent for the world’s tropical dry forests and savannas (Pratzer et al. 2024). These ecosystems are often weakly protected and understudied (Schröder et al. 2021), and as a result, many dry forests and savannas have experienced drastic land-use change (Buchadas et al. 2022). Here, we focus on one such region, the Dry Chaco in Argentina, a global deforestation hotspot (Hansen et al. 2013, Vallejos et al. 2015) The combination of rising international demand for agricultural commodities, technological changes (e.g., herbicide-resistant soybean cultivars), and agribusiness-friendly policies stimulated the arrival and expansion of capitalized farmers and ranchers in the Argentine Chaco (Paolasso et al. 2012, Krapovickas and Longhi 2022), creating one of the most drastic social-ecological transformations in the Global South in the 21st century (Goldfarb and van der Haar 2016). Yet, the land-use agent configuration of the Argentine Dry Chaco is highly diverse, as these new actors expand over lands traditionally used by local family farmers and Indigenous communities, who have been carrying out subsistence farming for decades (Morello et al. 2005, Krapovickas and Longhi 2022). Previous studies focused on land-use agents in the Chaco have employed landholder surveys to uncover the heterogeneity of landholders (Mastrangelo et al. 2019), and to capture the different identities, attitudes, and norms of these agents regarding forests and the surrounding landscape (Mastrangelo et al. 2014). Similarly, a few studies have used archetype analyses to identify major types of land-use patterns or social-ecological systems in the Dry Chaco in a top-down fashion, based on aggregate statistics, remote sensing data, or expert consultations (Baldi et al. 2015, Vallejos et al. 2020, Pratzer et al. 2024). What is missing, however, are studies considering a wider range of agent characteristics (Mastrangelo 2018) to produce a bottom-up typology of major land-use agents.

Two research questions that guided our work were: (i) What are the major types of land-use agents in the Argentine Dry Chaco and what are their characteristics? and (ii) How do agent typologies differ when based solely on capital assets, management, or personal attributes compared with a typology that considers all these dimensions together? To address these questions, we classified and characterized agents in the northern Argentine Dry Chaco over an area of 176,000 km², where we collected information on agent characteristics through in-person questionnaire surveys between 2016 and 2022.

METHODS

Study Area

The Argentinian Dry Chaco comprises an area of about 488,000 km², and extends across several provinces (Vallejos et al. 2021). We focused on the northern part of this region, in the provinces of Santiago del Estero, Chaco, Formosa, and Salta (Fig. 1). The Argentine Dry Chaco was originally covered by savannas and xerophytic, semi-deciduous forest, dominated by hardwood trees such as quebrachos (Schinopsis spp.) and algarrobos (Prosopis spp.) The average temperature varies from 13°C in winter to 26°C in summer. Rainfall occurs during the summer (October to March), whereas July and August are the driest months (Morello et al. 2012).

There is a large diversity of land-use agents in the study region competing for land access and use, with different cultural backgrounds, who have historically occupied the territory in successive waves. Historically, Indigenous communities have settled and used the Chaco, practicing mainly hunting and gathering activities and subsistence agriculture. Peasant families and small/medium cattle ranchers settled the region in the 19th and early 20th centuries, practicing forest-based grazing and medium-scaled ranching in some areas. Since the 1990s and especially in the 21st century, the expansion of industrial agriculture by capitalized, large-scale farmers and investors from outside the region has been rampant (Morello et al., 2005, Pratzer et al. 2024), rendering the region one of the most active commodity frontiers in South America (Hansen et al. 2013, le Polain de Waroux et al. 2018).

Survey design and processing of questionnaire

We developed a questionnaire together with an interdisciplinary group of regional researchers and practitioners. The questionnaire covered internal attributes of landholdings, which are aligned with those outlined by FAO (2016). We distinguished them in three dimensions: (i) capital assets (what they have), (ii) management (what they do), and (iii) personal (who they are) (Table 1). Surveys were conducted by different teams of researchers. In total, we conducted 336 semi-structured in-person interviews between 2016 and 2022 (Fig. 1). We interviewed the head of the landholding or the person who makes decisions about the use of the land. The landholdings were selected through a snowball method, consulting key actors in the area. Our sample was intended to cover the variability in environmental conditions of landholdings (indicated by geographical location) and the range of socio-economic characteristics of landholders (indicated by landholding size). We implemented the questionnaire using the Kobo application (kobotoolbox.org) on tablets. Additionally, to enhance data security, we had a paper version of the questionnaire and a voice recorder as a backup.

All survey data were digitized, and the attributes were converted into factorial variables for our analysis, homogenizing the responses into categories (Table 1). Some surveys had missing data due to questions not being answered (in total, we had 739 missing values out of a total of 6,318, equaling a missing data rate of 11.6%). To deal with missing data, we first excluded three attributes that had more than 90 missing values. Next, we excluded questionnaire surveys with more than three unanswered questions (out of a remaining total of 15). This led to the exclusion of 15 interviews. Finally, we used the Mice package in R for imputing the remaining missing data (425 values). This package iteratively imputes values for each variable using predictive models informed by the other variables in the data set. It cycles through all variables multiple times until convergence, generating several complete data sets that reflect the uncertainty of the imputed values (van Buuren and Groothuis-Oudshoorn 2011). This resulted in a database of 336 interviews with complete data for 15 attributes (Table 1).

Archetyping to identify key land-use agents

To identify agent types and organize them into a typology, we performed a hierarchy cluster analysis with the Ward method (Ward 1963) using all the attributes from the final survey dataset (Table 1). For the clustering, we chose the Gower distance, which can combine variables of different natures (numerical and categorical) in the distance calculation. We then created a dendrogram plot to evaluate and visualize the resulting clustering. Additionally, we performed a multiple correspondence analysis (MCA), a descriptive method to reduce the dimensionality of data, using the FactoExtra package in R (Kassambara and Mundt 2016). The numbers of clusters and subclusters were selected considering the dendrogram itself and testing different options in the MCA. We selected the number of classes by testing the best separation in the dendrogram and in the MCA two dimensions. We evaluated the distribution of frequencies of all the variables to characterize each agent type and derived bar charts to plot each variable according to each cluster and subcluster.

To compare unidimensional typologies with the full typology, we repeated the clustering process using only attributes from one dimension. This yielded four typologies: one typology based on the full data set covering all attributes, and one based on “what they have, what they do,” and “who they are ” attributes only. We then assessed the similarity of clusters across typologies by calculating the proportion of agents that were in the same cluster. For example, if a cluster of a one-dimensional typology is equal to a cluster in our three-dimensional typology, a ratio of 1 would be calculated. However, if there was no overlap between clusters, the value was 0. We did this for all cluster combinations and visualized the agreement in an overlap matrix.

RESULTS

Our clustering analyses identified three main agent types and, nested within, five sub-types (Fig. 2). The MCA confirmed the results of the clustering, identifying five distinct groupings of data, clearly differentiated when represented in a reduced-dimensional space (Fig. 3). The decision for three main types and five subtypes thus provided the best-defined clusters: they were well-separated, balanced in size, and showed minimal overlap in the MCA representation. Given the clear separation of our clusters based on agents’ attributes, we refer to the main clusters as archetypes, and the five clusters identified in our second-order clustering as sub-archetypes.

Assessing the characteristics of these archetypes and sub-archetypes based on the attributes contained in our survey and interpreting them through the lens of the three dimensions, allowed us to identify and name these archetypes (Fig. 4). The first archetype consisted of semi-subsistence smallholders, comprising 154 cases. Most semi-subsistence smallholders (71%) face insecure land tenure. Most of these farmers do not use machinery (94%) and rely heavily on family labor (85%). More than half supplement their farm income with off-farm work or receive an additional income. Cattle raising is the predominant activity, supported mostly by natural forage. Many semi-subsistence smallholders also raise small livestock, and half of them collect honey. Around three-quarters of agents in this cluster were local and older than 40 yrs, and almost half had been in the land for >15 years (Append. 1).

This archetype was subsequently split into two sub-archetypes representing different types of semi-subsistence smallholders: forest-dependent smallholders (1a, 98 cases) and semi-subsistence ranchers (1b, 56 cases). Based on what they have, the main difference between these sub-archetypes is that semi-subsistence ranchers have larger landholdings than forest-dependent smallholders. Also, a larger proportion of semi-subsistence ranchers receive an extra income and have extra-family labor, compared with forest-dependent landholders (Fig. 5b,f). Regarding what they do, forest-dependent smallholders mostly use natural forage to raise livestock, whereas semi-subsistence ranchers use mixed sources of livestock forage (Fig. 5i). Also, forest-dependent smallholders produce or collect honey, whereas semi-subsistence ranchers rarely do so (Fig. 5k). There were no strong differences regarding the who they are dimension (Fig. 5l-o).

Our second agent archetype was the crop–livestock farmer (81 cases). Like semi-subsistence smallholders, the landholdings of most crop–livestock farmers were smaller than 1,000 ha (Fig. 5b). Compared with semi-subsistence smallholders, a larger proportion of crop–livestock farmers had secured their land tenure (i.e., recognition of property rights by the State) and used machinery (Fig. 5c, d). Family remains the primary source of labor, although 37% of crop–livestock farmers employ a mix of family and external workers (Fig. 5e). Crop–livestock farmers rely on farm income and develop diversified farming systems, combining cropping and cattle raising, the latter mostly based on natural forage. They also raise small livestock, and a majority either produce or collect honey (Fig. 5i-k). Regarding the who they are dimension, crop–livestock farmers are local landowners, with 75% having been on their land for >30 yrs. Most were older than 40 yrs (Fig. 5m).

Our third archetype was the commercial landholder (101 cases). This group was characterized by larger, commercially oriented landholders, with half of them having >2,500 ha of land. These agents have the highest proportion of titled land and use of machinery. Unlike semi-subsistence smallholders and crop–livestock farmers, the main source of labor is external to the family. Fifty-four percent of the agents in this archetype have an additional source of income. Some of them specialize in crops, typically using high levels of input, and others specialize in cattle production (almost 40% without using natural forage). Small animals are kept by 76% of these agents, but commercial landholders do not engage in honey production. Thirty percent of commercial landholders are extra-local (i.e., from another province or another country), almost half of them (45%) were younger than 40 yrs, and 64% have been in the landholding for <15 yrs. Most commercial landholdings are run by administrators (Append. 1).

The commercial landholder archetype was split into two sub-archetypes: agribusiness farmers (3a, 55 cases) and commercial ranchers (3b, 46 cases). The main difference between these two in terms of the “what they have” dimension was that agribusiness farmers had larger landholdings than commercial ranchers as well as a higher proportion of external labor (Fig. 5b, e). Stronger differences between these two sub-types occurred in the “what they do” dimension: agribusiness farmers specialize in high-input cropping, whereas commercial ranchers on cattle raising (Fig. 5g). Concerning the “who they are” dimension, agribusiness farmers were younger (mostly 20–40 yrs), had spent less time on the landholding (<15 yrs) compared with commercial ranchers, and more often came from another country (27%) or province (25%). Finally, landholdings of most agribusiness farmers were run by administrators, whereas half of the commercial ranchers managed their ranches themselves (Fig. 5l-o).

The agent archetypes that emerged were dependent on the type and number of dimensions used for classifying agents. The degree of overlap between each unidimensional typology and the three-dimensional typology varied (Fig. 6). The unidimensional typology with a larger overlap with our full, three-dimensional typology was the one based on “what they do” attributes (Fig. 6b), especially for the 1a, 1b, and 2 sub-archetypes (73%, 94%, and 88% overlap, respectively, Fig. 6b). The agribusiness farmer and the commercial rancher types of the three-dimensional typology (3a and 3b) were collapsed into a single type in the “what they have” typology, which overlaps 93% with the former and 76% with the latter (Fig. 6a). The agribusiness farmer type also showed strong overlap with one type of the “who they are” typology (Fig. 6c), highlighting the unique personal characteristics of this agent type.

DISCUSSION

Structuring the complexity of human–nature interactions is important to adequately consider it in sustainability research, practice, and policy making. Here, we demonstrate how bottom-up archetype analyses can help to characterize key land-use agents in a global deforestation hotspot in dire need of sustainability planning, the Argentine Chaco, contributing to a growing field of research aiming to characterize key types of actors involved in land-use change, while also filling an important regional gap. Relying on a rich data set of questionnaire surveys, we use hierarchical clustering to build an agent typology based on agents’ capital assets, management activities, and personal characteristics. We found five well-differentiated types of land-use agents: forest-dependent landholder, semi-subsistence rancher, crop–livestock farmer, agribusiness farmer, and commercial rancher (Table 2). Interpreting these archetypes yielded three main insights. First, we reveal considerable heterogeneity of land-use agents in the Argentine Dry Chaco, allowing us to move beyond the common yet oversimplified and dichotomic view of agribusinesses vs. smallholders. Second, the agent typology based on all three dimensions captured the diversity of agents much better than any one-dimensional typology alone, demonstrating the value of richer descriptions of land-use agents. Third, all our agent types share characteristics in some dimensions yet differ in others (e.g., forest-dependent smallholders and crop–livestock farmers were similar in “who they are,” yet different in “what they do”), explaining how more simplistic agent descriptions arrive at oversimplified agent types. These insights are valuable for shaping more nuanced and effective sustainability policies in diverse social-ecological contexts.

Our classification and characterization revealed considerable heterogeneity in land-use agents in the Argentine Chaco. This allowed us to move beyond dichotomous views of land-use agents commonly found in literature and expert analyses of deforestation frontiers, in the Chaco and in other regions. Generally, there has been a tendency to represent the social landscape in binary ways, focusing on the most contrasting agent types: smallholders vs. agribusiness (Baldi et al. 2015, le Polain de Waroux et al. 2018, Kong et al. 2019). This is probably because their impacts on land cover are easily differentiable by remote sensing (Baldi et al. 2015) and because agribusinesses often replace smallholders (le Polain de Waroux et al. 2016). Although our agent typology is not directly transferable to other regions and would require adaptation to different social-ecological contexts, its ability to represent such context specificity is a strength and reveals nuances that a binary agent classification would otherwise overlook. This is the case for crop–livestock farmers, which, in our analysis, emerged as a homogeneous and well-differentiated group (Fig. 3). In the Chaco, these “intermediate” agents who occupy a position between semi-subsistence and commercial farmers, partially sharing attributes and dimensions with both (Table 2), do not appear in regional typologies based on remote-sensing data (Baldi et al. 2015) or expert opinion, even those that capture a wide diversity of actors (Pratzer et al. 2024). Similarly, these intermediate agents are overlooked in other commodity frontier regions as well (Ordway et al. 2017), pointing to a wider need to “put them on the map” and consider them in sustainability assessments.

Characterizing our agents based on a wide range of attributes along our three dimensions allowed us to demonstrate that agents are multidimensional entities (FAO 2016), as archetypes change when different subsets of attributes are used to classify and characterize them (Fig. 6), highlighting the importance of careful variable selection and data collection decisions grounded in hypotheses (FAO 2016, Alvarez et al. 2018). Classifying agents on a single dimension or a few attributes often collapses different agents into the same class, which can be misleading for the design of policy interventions and reduces its effectiveness. On the other hand, it allowed us to capture similarities or differences among agents. Although the typology of agents is based on cross-sectional data and captures a snapshot of the social landscape, our results provide insights into processes of differentiation and convergence among agent types. Semi-subsistence smallholders (sub-archetypes 1a and 1b) and crop–livestock farmers (archetype 2), for example, share personal characteristics and differ in “what they have” and “what they do.” More than two-thirds of these agents are of local origin and have spent most of their lives on the landholding, suggesting strong ties to the land that sustainability policies could build on. However, semi-subsistence smallholders mainly raise livestock on untitled land, and crop–livestock farmers integrate livestock and crops under more secure forms of land tenure. This pattern of similarities and differences may have been generated by a process of social differentiation, whereby agents with a common origin diverged in their trajectories because some of them succeeded in titling their landholdings and diversifying their management systems. In contrast, agribusiness farmers (type 3a) and commercial ranchers (type 3b) are similar in their level of assets but differ in “who they are” and “what they do.” This pattern could be generated by the convergence of agents from different origins (mostly extra-local for 3a and local for 3b) and deploying different production systems (crop-based for 3a and cattle-based for 3b) in a commercial orientation and the management of large landholdings.

Through a simple and replicable methodology, our work provides valuable inputs for design, implementation, and evaluation of interventions aimed at fostering sustainable land use. A strength of this approach lies in its methodological transferability—the archetyping process we carried out here can be readily and easily applied in other social-ecological contexts to generate similar, context-specific insights and avoid oversimplified, top-down typifications. This is particularly relevant given the growing calls for context specificity in policy development and targeting (Kuemmerle 2024). Understanding the diversity of land-use agents and their characteristics allows to design tailored interventions (Huber et al. 2024) as well as to select those with greater potential for effective implementation (e.g., land-use agents with stronger ties to the land, see above). For example, given their orientation toward the commercialization of grain and beef in the international market, the land-use decisions of commercial landholders (archetype 3) could be strongly influenced by supply-chain interventions, such as zero-deforestation laws imposed by importing countries (Lambin et al. 2018), as in the case of the EU Deforestation Regulation (EUDR). Semi-subsistence smallholders, on the other hand, could be targeted by first providing more security tenure, increasing the chances to work with them for long-term sustainability. More secure land tenure can support higher access to credit and financial and technical assistance, all of which can help semi-subsistence smallholders improve their productivity without compromising biodiversity in their landholdings (Mastrangelo et al. 2019, Faingerch et al. 2021). Crop–livestock farmers, in turn, could be strengthened by developing shorter commercialization chains for selling more environmentally friendly products.

Our methodological approach shows potential but also has some limitations. As in any simplification of reality, some agent attributes were missing. This is the case of the ethnicity of the agents, which, not being considered, did not allow us to identify types of agents captured in other typologies, such as Mennonites (Baldi et al. 2015) and Indigenous communities (Pratzer et al. 2024). Other aspects not incorporated are those at levels of organization above the individual, for example, attributes related to social capital and institutions (Pratzer et al. 2024). Geographic dynamics were not considered, even though agricultural expansion across large regions such as the Argentine Dry Chaco (Fig. 1) follows diverse patterns that likely influence agents’ characteristics and their spatial distribution. Adding attributes capturing social and geographic dynamics would have yielded even more context-specific archetypes and sub-archetypes. However, our approach strives to achieve a balance between gaining a nuanced picture of the social landscape of a frontier region and providing a cost-effective and replicable methodology that allows to do this with the limited data and resources often found in similar frontier regions. Despite this, our bottom-up approach can help to improve understanding of the structure and dynamics of social-ecological systems in general, and deforestation frontiers in particular in at least two ways. On the one hand, the definition of agents in terms of the management of a given portion of the territory enables the mapping of agents, the analysis of their spatial distribution, and, thus, the identification of potential for conflicts and environmental risks. On the other hand, collecting data from agents enables longitudinal studies and a deeper analysis of social differentiation and convergence processes over time.

Given widespread evidence for unsustainable land use across the globe, developing and implementing policies to transition to more sustainable land-use futures is critically important. To be successful, such policies must address those that manage land—the land-use agents. Here, we show the importance of recognizing agents as multidimensional entities and highlight the need for and potential of accounting for the diverse factors shaping these land-use agents. Considering the diversity of land-use agents is especially relevant in the global commodity frontier regions, such as in many dry forests and savannas, where land use is changing rapidly, where the interactions among agents are influenced by a range of factors from global to local scales, and where these complex interactions create often stark social-ecological trade-offs (Ordway et al. 2017, le Polain de Waroux et al. 2018, Levers et al. 2021). Our approach demonstrates how key agent types—such as crop–livestock farmers in our case—can be successfully identified, even in relatively data-sparse, dynamic situations such as in the Dry Chaco of Argentina. More generally, our study corroborates how archetype analyses can provide pathways for structuring complexity and diversity in human–environment interactions.

AUTHOR CONTRIBUTIONS

Melina Faingerch: conceptualization, methodology, formal analysis, investigation, writing – original draft, writing – review and editing, visualization. Tobias Kuemmerle: writing – original draft, writing – review and editing, funding acquisition. Matthias Baumann: formal analysis, writing – review and editing, visualization. Marcos Texeira: formal analysis. Matias Mastrangelo: conceptualization, methodology, formal analysis, investigation, writing – original draft, writing – review and editing, visualization, funding acquisition.

ACKNOWLEDGMENTS

We thank all the landholders who answered the questionnaire and all the researchers and technicians participating in the fieldwork. This work is part of the doctoral thesis of the first author at the Programa de Posgrado en Ciencias Agrarias, Facultad de Ciencias Agrarias, Universidad Nacional de Mar del Plata, Argentina. This work received support through the project “Disponibilidad y acceso a especies nativas importantes para la seguridad alimentaria de familias campesinas criollas e indígenas del Impenetrable Chaqueño,” Code: C200 financed by the Ministerio de Ciencia y Tecnología de la Nación from Argentina, by the German Academic Exchange Service (DAAD), as well as by the European Research Council under the European Union’s Horizon 2020 research and innovation programme (grant agreement 101001239 SYSTEMSHIFT, http://hu.berlin/SystemShift). This work contributes to the Global Land Programme (https://glp.earth). We are grateful to the two anonymous reviewers for their valuable feedback and contributions to this manuscript.

Use of Artificial Intelligence (AI) and AI-assisted Tools

The manuscript was reviewed using AI-assisted tools to ensure accuracy in grammar and to enhance the clarity of the language.

DATA AVAILABILITY

The data and code that support the findings of this study are available on request from the corresponding author, MF. None of the data and code are publicly available because they contain information that could compromise the privacy of research participants. Ethical approval for this research study was granted by Humboldt University of Berlin.

Corresponding author:

Melina Faingerch

[email protected]

Table 1. Description of the dimensions, attributes, and categories of our questionnaire survey.

Dimension	Attribute	Description	Categories	Description of the category
What they have	Production type	Orientation of the landholding’s production	Commercial	The main destination of the production is trade
			Semi-subsistence	The main destination of the production is self-consumption or small-scale trade
	Land tenure	Property rights over the land	Titled	All property rights are recognized by the State through the issue of land title
			Possession	Without property rights, or with only some property rights recognized by the State
			Adjudication	In process of recognition of full property rights
			Rented	The form of tenure is the rent of the land for a short term
	Machinery	Use or not of machinery (without discriminating the scale)	Yes
			No
	Labor type	Type of labor on the landholdings	External Labor
			Family Labor
			Both
	Landholding size	Range expressed in hectares	<1000 ha
			1001–2500
			2501–5000
			>5000 ha
	Source of income	Sources of income for the livelihood	No-extra incomes	The main income are from activities related to the farm
			Extra income	Agents with extra income from an activity or source other than the farm (e.g., other economic activity, or a subsidy).
What they do	Main activity	Main activity identified by the landholders	Cattle	Raising cattle and small livestock
			Crop	All types of crops included
			Mixed	More than one main activity
	Management of the crops	For the cropping activity: use of different inputs (pesticides, fertilizers, and/or irrigation)	Without inputs	Cropping without inputs
			High input	Agents who use two or more inputs
			Low input	Agents who use just one input
			No cropping	They do not use inputs because they do not have crops
	Management of the cattle	For the cattle activity: source of forage for feeding	No cattle	Without cattle
			Natural feed	Cattle feed on forest forage
			Cultivated	Cultivated forage in situ and/or ex situ.
			Mixed source with natural feed	Natural forage plus forage ex situ and/or in situ
			Mixed source without natural feed	Feed ex and/or in situ, without natural feed.
	Small livestock	Breeding of small livestock (e.g., goats, pigs, chickens) for production and/or consumption	Yes
			No
	Collection of honey	Collection of wild honey for production and/or consumption	Yes
			No
Who they are	Time on the landholding	Number of years living or working in the landholding	<5 yrs
			6–15 yrs
			16–30 yrs
			31–50 yrs
			> 50 yrs
	Age of the landholder	Age of the landholder surveyed	20–40 yrs
			41–60 yrs
			> 60 yrs
	Origin of the landholder	Where the agent is from	This department
			This province
			Another province
			Another country
	Role of the landholder	Self-perceived role of the landholder. It has no direct link to the form of tenure.	Owner
			Administrator
			Renter

Table 2. Summary of the key characteristics of the five sub-archetypes.

	What they have	What they do	Who they are
1a - Forest-dependent smallholder	They have small semi-subsistence landholdings with insecure land tenure. They mostly have family labor and extra-farm income.	They do cattle ranching and raise small animals in the forest as livestock forage. They also collect honey from the forest.	They are local landowners, more than 40 yrs old, who have been in the landholding since birth.
1b- Semi-subsistence rancher	They have small semi-subsistence landholdings, but larger than forest-dependent smallholders. They have insecure land tenure. The labor is mixed, and they have an extra-farm income.	They do cattle ranching with mixed sources based on forest. They raise small animals and they do not collect honey.	They are local landowners, mostly more than 40 yrs old, who have been in the landholding more than 40 yrs.
2- Crop–livestock farmer	They have semi-subsistence landholdings with the adjudication of the land (a previous step for the titled). They don’t have extra income. Half of them have machinery and mixed labor.	They do crop–livestock production. When they raise crops, it is without inputs. When they do cattle, they use the forest as the main source of forage. They raise small animals and they collect honey.	They are local landowners, with most of them having spent a long time in the landholding. One-third of them have been there since birth. They are mostly greater than 40 yrs old.
3a- Agribusiness farmer	They have large commercial landholdings. They have secure land tenure, machinery, and external labor, and half of them have an extra-farm income.	They do mainly high-input agriculture and mixed activities. When they do cattle, they use mixed sources for forage, without natural feed. Half of them raise small animals.	They are young administrators (<40 yrs old), with less than 15 yrs in the landholding. Half of them are extra-local from other provinces or another country.
3b- Commercial rancher	They have commercial landholdings with large plots. They have machinery, mostly have family labor, and they have an extra income	They do cattle and mixed activities. The cattle are fed with cultivated and mixed forage, but not natural feed. They raise small animals. When they raise crops, they use diverse strategies (high, low, and without input).	Half of them are local landowners, and the other half are administrators, who arrived between 16–30 yrs ago to the landholding. They are in general greater than 40 yrs old.

References

Alvarez, S., C. J. Timler, M. Michalscheck, W. Paas, K. Descheemaeker, P. Tittonell, J. A. Andersson, and J. C. J. Groot. 2018. Capturing farm diversity with hypothesis-based typologies: an innovative methodological framework for farming system typology development. PLoS ONE 13(5):e0194757. https://doi.org/10.1371/journal.pone.0194757

Baldi, G., J. Houspanossian, F. Murray, A. A. Rosales, C. V. Rueda, and E. G. Jobbágy. 2015. Cultivating the dry forests of South America: diversity of land users and imprints on ecosystem functioning. Journal of Arid Environments 123:47-59. https://doi.org/10.1016/j.jaridenv.2014.05.027

Barral, M. P., S. Villarino, C. Levers, M. Baumann, T. Kuemmerle, and M. Mastrangelo. 2020. Widespread and major losses in multiple ecosystem services as a result of agricultural expansion in the Argentine Chaco. Journal of Applied Ecology 57(12):2485-2498. https://doi.org/10.1111/1365-2664.13740

Buchadas, A., M. Baumann, P. Meyfroidt, and T. Kuemmerle. 2022. Uncovering major types of deforestation frontiers across the world’s tropical dry woodlands. Nature Sustainability 5(7):619-627. https://doi.org/10.1038/s41893-022-00886-9

Cáceres, D. M. 2015. Accumulation by dispossession and socio‐environmental conflicts caused by the expansion of agribusiness in Argentina. Journal of Agrarian Change 15(1):116-147. https://doi.org/10.1111/joac.12057

Eisenack, K., C. Oberlack, and D. Sietz. 2021. Avenues of archetype analysis: roots, achievements, and next steps in sustainability research. Ecology and Society 26(2):31. https://doi.org/10.5751/ES-12484-260231

Faingerch, M., M. Vallejos, M. Texeira, and M. E. Mastrangelo. 2021. Land privatization and deforestation in a commodity production frontier. Conservation Letters 14(4):e12794. https://doi.org/10.1111/conl.12794

Food and Agriculture Organization (FAO). 2016. Literature review report and proposal for an international framework for farm typologies (MTF/GLO/372/MUL). FAO, Rome, Italy. https://openknowledge.fao.org/server/api/core/bitstreams/b6f27f8e-5dd2-4005-9a2b-d1eaf8651de4/content

Goldfarb, L., and G. van der Haar. 2016. The moving frontiers of genetically modified soy production: shifts in land control in the Argentinian Chaco. The Journal of Peasant Studies 43(2):562-582. https://doi.org/10.1080/03066150.2015.1041107

Hansen, M. C., P. V. Potapov, R. Moore, M. Hancher, S. A. Turubanova, A. Tyukavina, D. Thau, S. V. Stehman, S. J. Goetz, T. R. Loveland, A. Kommareddy, A. Egorov, L. Chini, C. O. Justice, and J. R. G. Townshend. 2013. High-resolution global maps of 21st-century forest cover change. Science 342(6160):850-853. https://doi.org/10.1126/science.1244693

Huber, R., B. Bartkowski, C. Brown, N. El Benni, J.-H. Feil, P. Grohmann, I. Joormann, H. Leonhardt, H. Mitter, and B. Müller. 2024. Farm typologies for understanding farm systems and improving agricultural policy. Agricultural Systems 213:103800. https://doi.org/10.1016/j.agsy.2023.103800

Jaureguiberry, P., N. Titeux, M. Wiemers, D. E. Bowler, L. Coscieme, A. S. Golden, C. A. Guerra, U. Jacob, Y. Takahashi, J. Settele, S. Díaz, Z. Molnár, and A. Purvis. 2022. The direct drivers of recent global anthropogenic biodiversity loss. Science Advances 8(45):eabm9982. https://doi.org/10.1126/sciadv.abm9982

Jollivet, M. 1965. D’une méthode typologique pour l’étude des sociétés rurales. Revue Française de Sociologie 6:33. https://doi.org/10.2307/3319368

Kassambara, A., and F. Mundt. 2016. factoextra: Extract and visualize the results of multivariate data analyses (p. 1.0.7) [dataset]. https://doi.org/10.32614/CRAN.package.factoextra

Kong, R., J.-C. Diepart, J.-C. Castella, G. Lestrelin, F. Tivet, E. Belmain, and A. Bégué. 2019. Understanding the drivers of deforestation and agricultural transformations in the northwestern uplands of Cambodia. Applied Geography 102:84-98. https://doi.org/10.1016/j.apgeog.2018.12.006

Krapovickas, J., and F. Longhi. 2022. Pobrezas, ruralidades y campesinos en el Chaco Argentino a comienzos del siglo XXI. Estudios Rurales 3(4):266. https://doi.org/10.48160/22504001er4.266

Kuemmerle, T. 2024. Moving beyond simplistic representations of land use in conservation. Conservation Letters 17(5):e13055. https://doi.org/10.1111/conl.13055

Lambin, E. F., H. K. Gibbs, R. Heilmayr, K. M. Carlson, L. C. Fleck, R. D. Garrett, Y. le Polain de Waroux, C. L. McDermott, D. McLaughlin, P. Newton, C. Nolte, P. Pacheco, L. L. Rausch, C. Streck, T. Thorlakson, and N. F. Walker. 2018. The role of supply-chain initiatives in reducing deforestation. Nature Climate Change 8(2):109-116. https://doi.org/10.1038/s41558-017-0061-1

Lambin, E. F., P. Meyfroidt, X. Rueda, A. Blackman, J. Börner, P. O. Cerutti, T. Dietsch, L. Jungmann, P. Lamarque, J. Lister, N. F. Walker, and S. Wunder. 2014. Effectiveness and synergies of policy instruments for land use governance in tropical regions. Global Environmental Change 28:129-140. https://doi.org/10.1016/j.gloenvcha.2014.06.007

le Polain de Waroux, Y., M. Baumann, N. I. Gasparri, G. Gavier-Pizarro, J. Godar, T. Kuemmerle, R. Müller, F. Vázquez, J. N. Volante, and P. Meyfroidt. 2018. Rents, actors, and the expansion of commodity frontiers in the Gran Chaco. Annals of the American Association of Geographers 108(1):204-225. https://doi.org/10.1080/24694452.2017.1360761

le Polain de Waroux, Y., R. D. Garrett, R. Heilmayr, and E. F. Lambin. 2016. Land-use policies and corporate investments in agriculture in the Gran Chaco and Chiquitano. Proceedings of the National Academy of Sciences 113(15):4021-4026. https://doi.org/10.1073/pnas.1602646113

Levers, C., A. Romero-Muñoz, M. Baumann, T. De Marzo, P. D. Fernández, N. I. Gasparri, G. I. Gavier-Pizarro, Y. le Polain de Waroux, M. Piquer-Rodríguez, A. Semper-Pascual, and T. Kuemmerle. 2021. Agricultural expansion and the ecological marginalization of forest-dependent people. Proceedings of the National Academy of Sciences 118(44):e2100436118. https://doi.org/10.1073/pnas.2100436118

Lokhorst, A. M., C. Hoon, R. le Rutte, and G. de Snoo. 2014. There is an I in nature: the crucial role of the self in nature conservation. Land Use Policy 39:121-126. https://doi.org/10.1016/j.landusepol.2014.03.005

Malek, Ž., B. Douw, J. van Vliet, E. H. van der Zanden, and P. H. Verburg. 2019. Local land-use decision-making in a global context. Environmental Research Letters 14(8):083006. https://doi.org/10.1088/1748-9326/ab309e

Malek, Ž., and P. H. Verburg. 2020. Mapping global patterns of land use decision-making. Global Environmental Change 65:102170. https://doi.org/10.1016/j.gloenvcha.2020.102170

Martinelli, G. D. 2011. De los conceptos a la construcción de los tipos sociales agrarios. Una mirada sobre distintos modelos y las estrategias metodológicas. Revista Latinoamericana de Metodología de la Investigación Social (ReLMIS) 2:24-43. https://dialnet.unirioja.es/servlet/articulo?codigo=5275929

Mastrangelo, M. E. 2018. Aproximaciones al estudio del comportamiento de los productores agropecuarios en el Chaco Seco. Ecología Austral 28(2):418-434. https://doi.org/10.25260/EA.18.28.2.0.644

Mastrangelo, M. E., M. C. Gavin, P. Laterra, W. L. Linklater, and T. L. Milfont. 2014. Psycho-social factors influencing forest conservation intentions on the agricultural frontier. Conservation Letters 7(2):103-110. https://doi.org/10.1111/conl.12033

Mastrangelo, M. E., and P. Laterra. 2015. From biophysical to social-ecological trade-offs: integrating biodiversity conservation and agricultural production in the Argentine Dry Chaco. Ecology and Society 20(1):20. https://doi.org/10.5751/ES-07186-200120

Mastrangelo, M. E., Z. Sun, L. Seghezzo and D. Müller. 2019. Survey-based modeling of land-use intensity in agricultural frontiers of the Argentine dry Chaco. Land Use Policy 88:104183. https://doi.org/10.1016/j.landusepol.2019.104183

McKinney, J. C. 1950. The role of constructive typology in scientific sociological analysis. Social Forces 28(3):235-240. https://doi.org/10.2307/2572006

Meyfroidt, P. 2013. Environmental cognitions, land change, and social-ecological feedbacks: an overview. Journal of Land Use Science 8(3):341-367. https://doi.org/10.1080/1747423X.2012.667452

Meyfroidt, P., D. Abeygunawardane, M. Baumann, A. Bey, A. Buchadas, C. Chiarella, V. Junquera, A. Kronenburg García, T. Kuemmerle, Y. le Polain de Waroux, E. Oliveira, M. Picoli, S. Qin, V. Rodriguez García, and P. Rufin. 2024. Explaining the emergence of land-use frontiers. Royal Society Open Science 11(7):240295. https://doi.org/10.1098/rsos.240295

Meyfroidt, P., A. de Bremond, C. M. Ryan, E. Archer, R. Aspinall, A. Chhabra, G. Camara, E. Corbera, R. DeFries, S. Díaz, J. Dong, E. C. Ellis, K.-H. Erb, J. A. Fisher, R. D. Garrett, N. E. Golubiewski, H. R. Grau, J. M. Grove, H. Haberl, and E. K. H. J. zu Ermgassen. 2022. Ten facts about land systems for sustainability. Proceedings of the National Academy of Sciences 119(7):e2109217118. https://doi.org/10.1073/pnas.2109217118

Meyfroidt, P., R. Roy Chowdhury, A. de Bremond, E. C. Ellis, K.-H. Erb, T. Filatova, R. D. Garrett, J. M. Grove, A. Heinimann, T. Kuemmerle, C. A. Kull, E. F. Lambin, Y. Landon, Y. le Polain de Waroux, P. Messerli, D. Müller, J. Ø. Nielsen, G. D. Peterson, V. Rodriguez García, and P. H. Verburg. 2018. Middle-range theories of land system change. Global Environmental Change 53:52-67. https://doi.org/10.1016/j.gloenvcha.2018.08.006

Morello, J., S. D. Matteucci, A. F. Rodriguez, and M. E. Silva. 2012. Ecorregiones y complejos ecosistémicos argentinos. Facultad de Arquitectura Desarrollo y Urbanismo, University of Buenos Aires, Buenos Aires, Argentina. https://www.researchgate.net/profile/Silvia-Matteucci-2/publication/268447092_Ecorregiones_y_complejos_ecosistemicos_Argentinos/links/598333be0f7e9b2ac353f62e/Ecorregiones-y-complejos-ecosistemicos-Argentinos.pdf

Morello, J., W. Pengue, and A. Rodriguez. 2005. Etapas de uso de los recursos y desmantelamiento de la biota del Chaco. Fronteras 4:1-17.

Oberlack, C., D. Sietz, E. Bürgi Bonanomi, A. de Bremond, J. Dell’Angelo, K. Eisenack, E. C. Ellis, G. Epstein, M. Giger, A. Heinimann, C. Kimmich, M. T. Kok, D. Manuel-Navarrete, P. Messerli, P. Meyfroidt, T. Václavík, and S. Villamayor-Tomas. 2019. Archetype analysis in sustainability research: meanings, motivations, and evidence-based policy making. Ecology and Society 24(2):26. https://doi.org/10.5751/ES-10747-240226

Ordway, E. M., G. P. Asner, and E. F. Lambin. 2017. Deforestation risk due to commodity crop expansion in sub-Saharan Africa. Environmental Research Letters 12(4):044015. https://doi.org/10.1088/1748-9326/aa6509

Pacheco-Romero, M., T. Kuemmerle, C. Levers, D. Alcaraz-Segura, and J. Cabello. 2021. Integrating inductive and deductive analysis to identify and characterize archetypical social-ecological systems and their changes. Landscape and Urban Planning 215:104199. https://doi.org/10.1016/j.landurbplan.2021.104199

Paolasso, P., J. Krapovickas, and N. I. Gasparri. 2012. Deforestación, expansión agropecuaria y dinámica demográfica en el Chaco seco argentino durante la década de los noventa. Latin American Research Review 47(1):35-63. https://doi.org/10.1353/lar.2012.0009

Pratzer, M., P. Meyfroidt, M. Antongiovanni, R. Aragon, G. Baldi, S. Czaplicki Cabezas, C. A. de la Vega-Leinert, S. Dhyani, J.-C. Diepart, P. D. Fernandez, S. T. Garnett, G. I. Gavier Pizarro, T. Kalam, P. Koulgi, Y. le Polain de Waroux, S. Marinaro, M. Mastrangelo, D. Mueller, R. Mueller, and T. Kuemmerle. 2024. An actor-centered, scalable land system typology for addressing biodiversity loss in the world’s tropical dry woodlands. Global Environmental Change 86:102849. https://doi.org/10.1016/j.gloenvcha.2024.102849

Schröder, J. M., L. P. Ávila Rodríguez, and S. Günter. 2021. Research trends: tropical dry forests: the neglected research agenda? Forest Policy and Economics 122:102333. https://doi.org/10.1016/j.forpol.2020.102333

Smajgl, A., D. G. Brown, D. Valbuena, and M. G. A. Huigen. 2011. Empirical characterisation of agent behaviours in socio-ecological systems. Environmental Modelling and Software 26(7):837-844. https://doi.org/10.1016/j.envsoft.2011.02.011

Tittonell, P., O. Bruzzone, A. Solano-Hernández, S. López-Ridaura, and M. H. Easdale. 2020. Functional farm household typologies through archetypal responses to disturbances. Agricultural Systems 178:102714. https://doi.org/10.1016/j.agsy.2019.102714

Valbuena, D., P. H. Verburg, and A. K. Bregt. 2008. A method to define a typology for agent-based analysis in regional land-use research. Agriculture, Ecosystems and Environment 128(1-2):27-36. https://doi.org/10.1016/j.agee.2008.04.015

Vallejos, M., S. Aguiar, G. Baldi, M. E. Mastrángelo, F. Gallego, M. Pacheco-Romero, D. Alcaraz-Segura, and J. M. Paruelo. 2020. Social-ecological functional types: connecting people and ecosystems in the Argentine Chaco. Ecosystems 23(3):471-484. https://doi.org/10.1007/s10021-019-00415-4

Vallejos, M., G. H. Camba Sans, S. Aguiar, M. E. Mastrángelo, and J. M. Paruelo. 2021. The law is spider’s web: an assessment of illegal deforestation in the Argentine Dry Chaco ten years after the enactment of the “Forest Law”. Environmental Development 38:100611. https://doi.org/10.1016/j.envdev.2021.100611

Vallejos, M., J. N. Volante, M. J. Mosciaro, L. M. Vale, M. L. Bustamante, and J. M. Paruelo. 2015. Transformation dynamics of the natural cover in the Dry Chaco ecoregion: a plot level geo-database from 1976 to 2012. Journal of Arid Environments 123:3-11. https://doi.org/10.1016/j.jaridenv.2014.11.009

van Buuren, S., and K. Groothuis-Oudshoorn. 2011. mice: Multivariate imputation by chained equations in R. Journal of Statistical Software 45(3):1-67. https://doi.org/10.18637/jss.v045.i03

van der Ploeg, J. D. 1994. Styles of farming: an introductory note on concepts and methodology. Pages 7-31 in J. D. van der Ploeg and A. Long, editors. Practice and perspectives of endogenous rural development. Van Gorcum, Assen, The Netherlands.

Ward, J. H. 1963. Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association 58(301):236-244. https://doi.org/10.1080/01621459.1963.10500845