Copyright © 2019 by the author(s). Published here under license by the Resilience Alliance. Pérez-Llorente, I., M. Ramírez, J. Paneque-Gálvez, C. Garibay Orozco, and R. González-López. 2019. Unraveling complex relations between forest-cover change and conflicts through spatial and relational analyses. Ecology and Society 24(3):3. https://doi.org/10.5751/ ES-10992-240303 Research Unraveling complex relations between forest-cover change and conflicts through spatial and relational analyses Irene Pérez-Llorente 1, M. Isabel Ramírez 1, Jaime Paneque-Gálvez 1, Claudio Garibay Orozco 1 and Rafael González-López 2 ABSTRACT. Despite the increasing prevalence of forest-cover change and conflicts, most studies have been unable to unravel the complex relations between the two processes. We attribute this failure to methodological limitations. We put forward an alternative approach that combines different datasets (remote sensing, GIS, local narratives, official censuses, newspaper articles), methods (spatial and relational analyses), and scales (subregions, economic sectors, land-based activities) to create a robust explanation of the relations between different intensities of forest-cover change and conflict in the Meseta Purépecha region, central Mexico. This is an important forest region, inhabited by indigenous and mestizo peasants; it has a worldwide reputation for community forestry and is also the epicenter of international avocado production. Forest-cover change is intense and there are recurrent episodes of conflict. We clustered communities in three subregions according to their patterns of forest-cover change. We analyzed the spatial patterns of forest-cover change and conflicts and we characterized the structure and function of the different economic sectors to unravel the nonlinear, interdependent (and sometimes contradictory) relations among these processes. We found that avocado production has differentially shaped the composition and working of society within each subregion, leading to three diverging patterns. Avocado production has provoked conflicts over landownership and over illegal logging in nearby areas. In some areas, a low incidence of conflicts over forest clearance might be explained by high profits, coercion, and violence. We suggest that, by combining spatial and relational analyses, we can integrate and check the congruence of nonequivalent representations from quantitative sources and observant participation at different scales and explain the heterogeneity that processes display across space. Our methodological approach can thus improve our understanding of similar and other complex and uncertain environmental problems elsewhere, especially when accurate or appropriate data are missing. Key Words: central Mexico; complexity theory; environmental conflicts; land-cover/land-use change; uncertainty INTRODUCTION This diversity of narratives may be due to the wide range of Forest-cover change (FCC) and conflicts related to the access and processes of different intensity and scale encompassed by use of forest resources have become increasingly apparent in this conflicts, but also to the use of disciplinary or multidisciplinary century (Koning et al. 2008). The growing literature on the subject methods that may fail to apprehend the extent of the social and has put forward widely diverse theories to explain the relations natural complexity inherent in both FCC and forest conflicts. between FCC and conflicts, and no agreement has been reached Complex processes are subject to uncertainty, emergence, (Rustad et al. 2008). For instance, land tenure insecurity created adaptiveness, nonlinear relations, and “chicken or egg” paradoxes by agrarian conflicts can encourage deforestation (Araujo et al. (Funtowicz and Ravetz 1994, Rosen 1999, Giampietro et al. 2014). 2009, Aldrich et al. 2012). Agrarian conflicts have also been Such complex systems cannot be formalized into a model without associated with deforestation because timber extraction and losing most of their truth, because the definition, observation, (illegal) cash crops provide the means to finance conflict (Glew and management of the research problem implies value and Hudson 2007, Rincón-Ruiz and Kallis 2013, McSweeney et judgement (Rosen 2012). They thus require the use of multiple al. 2014). Others have asserted that deforestation provokes epistemologies and simultaneous nonequivalent descriptions environmental conflicts and resistance movements (Affolderbach (González-López and Giampietro 2017, Saltelli and Giampietro 2011, Gerber and Veuthey 2011, Seghezzo et al. 2011), that 2017). Moreover, the processes underlying FCC and conflicts vary “conservation enclosures” designed to prevent deforestation may depending on the location, resolution, and extent of the analysis also trigger conflicts (Corson 2011, Peluso and Lund 2011, (Veldkamp and Lambin 2001). The use of a regional case study Rocheleau 2015), or that violence and conflicts can lead to the with varying degrees of FCC and conflicts and different scales abandonment of agricultural areas or to “gunpoint or narco- could enhance our understanding of such processes by capturing conservation” (McNeely 2003, Suthakar and Bui 2008, Sánchez- part of their heterogeneity. Cuervo and Aide 2013). Some authors have argued that both FCC Several recent studies have explored the complexity and and conflicts are caused by mutually affecting factors operating heterogeneity of social-ecological changes associated with at multiple scales, such as the expansion of cattle ranching fostered conflicts. For instance, some authors have combined spatial and by increases in global meat demand, the historical agrarian statistical analysis with case studies to explain local land conflicts struggles and violence in the Brazilian Amazon, and the as a result of the interaction of scalar forces that structure resource combination of poverty and drug trafficking in Colombia and access (Simmons 2004), or deforestation and forest-related Mexico (O’Brien 1998, Rodrigues et al. 2009, Hecht and violence as conjoint outcomes of multiscale forest governance Cockburn 2010). (Durán et al. 2011). Hanaček and Rodríguez-Labajos (2018) used 1 Centro de Investigaciones en Geografía Ambiental (CIGA), Universidad Nacional Autónoma de México (UNAM), 2Institute of Environmental Science and Technology (ICTA), Autonomous University of Barcelona, Bellaterra, Spain Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ various theoretical frameworks and network analysis to study the cases. We suggest that this approach can be useful to deal with interrelations among land-use and management changes, cultural the uncertainties inherent in a wide range of complex ecosystem services, and diverse causes, effects, and outcomes of environmental problems anywhere, especially when there is a lack environmental conflicts. Ingalls and Mansfield (2017) applied a of accurate and appropriate data. coupled-systems and resilience approach using diverse causal mechanisms to explain social-ecological transformations in a STUDY AREA region under armed conflict. However, these studies have either The Meseta Purépecha is an elevated plateau ranging from 600 scarcely used spatial or quantitative information or they have to 3818 m above sea level, with abundant volcanoes covered by analyzed it using statistical tools. oak, pine, and fir temperate forests. It is an important commercial forest region in Michoacán, a state that leads Mexico’s resin We use a regional case study to demonstrate an alternative production and ranks third in timber production (COFOM 2007). approach that combines spatial and relational analyses to better It is also the epicenter of international avocado production, which grasp the relationships between different intensities of FCC and is rapidly expanding owing to high demand and prices in conflicts. By using semantically open accounting methods across international markets. The region owes its name to the indigenous a range of hierarchical levels, this approach allows us to group inhabiting these lands before the Spanish conquest, the coherently integrate and check the congruence of simultaneous, Purépecha, who are still widely present in the area. However, the nonequivalent representations from spatial, quantitative, and demarcation made by the government of Michoacán covers 11 qualitative information (Louie 2009, Rosen 2012, González- municipalities with heterogeneous land-tenure arrangements and López and Giampietro 2017). Relational analysis explores ethnicity (Fig. 1). systems as sets of interdependent relations between functional elements, which express objectives necessary for the stability of Besides the territorial division of municipalities, 51% of the the system, and structural elements, i.e., subparts of a functional Mexican territory was distributed as common property because element that share the same final objective and cooperate to of the agrarian reform in force from 1917 until 1992 (Torres- express the emergent property, making possible their own Mazuera et al. 2018). Over this period, groups of landless peasants reproduction (González-López and Giampietro et al. 2017). could request collective land endowments through the concept of Relational analysis has recently been used to analyze “ejidos,” and indigenous communities could demand the deforestation using typologies of farming systems (Kovacic and restitution of the land that had been taken from them and could Viteri Salazar 2017), but the incorporation of spatial analysis with thereby become agrarian communities (Nuijten 1997). this approach is still in its early stages (González-López and Agricultural land could be parceled and used individually but Giampietro 2017). could not be sold or rented, whereas forests were to remain undivided (Klooster 2003). After a constitutional and agrarian We first carry out spatial analysis to examine the spatial reform in 1992, ejidos and agrarian communities could privatize distribution of different FCC processes (deforestation, forest and divide part or the totality of their collective land among degradation, and forest conversion to avocado orchards) and assembly members, and sell or rent it to external agents (Randall conflicts of a range of expressions (from negotiation to fatal 1996). The study area comprises 41 agrarian communities confrontations), and actors (individuals, communities, or the (161,231 ha), 51 ejidos (63,747 ha), 5 mixed communities, i.e., government) in the Meseta Purépecha region, in central Mexico. localities with both agrarian community and ejido (37,244 ha), We then cluster communities according to such patterns and apply and 131,567 ha of private property (Fig, 1). Unless otherwise relational analysis to explore the composition and working of the stated, herein we will refer to both ejidos and agrarian different economic sectors and land-based activities of each communities as communities, for the sake of simplicity. subregion, combining quantitative information from secondary sources as well as local narratives from observant participation. MATERIALS AND METHODS Using this information, we explore the processes underlying the patterns of relations between FCC and conflicts observed in the Spatial analysis three subregions. The Meseta Purépecha is an economically We employed the land-cover/use maps developed by Mas et al. important forest region, with an internationally recognized (2017) for the state of Michoacán to evaluate FCC processes. They experience of community forestry, and is the epicenter of applied a hybrid semiautomatic classification method that international avocado production (Bofill Poch 2005, Orozco- combined image segmentation, digital classification, and visual Ramírez et al. 2017). It also has a high percentage of indigenous interpretation to SPOT satellite imagery from 2004 and 2014. populations and poverty rates (CDI 2010, CONEVAL 2014), These maps, with a global accuracy of 83.3% and a confidence acute deforestation and forest degradation processes (Mas et al. interval of 3.1%, are the most accurate land-cover/use data 2017), and recurrent episodes of conflict over illegal logging and available for our study area (Mas et al. 2017). Nevertheless, the landownership (Espín Díaz 1986, Vázquez León 2016). global accuracy of our maps is surely higher because our study area lacks land-cover/use areas subject to high spectral confusion, Using relational analysis, we can combine quantitative such as tropical dry forest and shrubland, and because we assessments with the dynamics of social and ecological processes combined the 19 original classes into only 4, thus reducing underlying FCC or environmental conflicts. That way we can omission and commission errors between spectrally similar focus on patterns and relationships rather than on absolute classes. We considered primary forest, secondary forest, avocado numbers and consider alternative explanations, dimensions, and orchards, and no forest. We used the perennial fruit trees class of representations of the study problem (Kovacic 2018). Moreover, our source land-cover/use map as proxy for avocado orchards the use of spatial analysis over a regional scale offers insights into because they represent 96% of the surface covered by perennial patterns and heterogeneities across the territory beyond concrete fruit trees in the study area (SIAP 2016). We computed the Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ Fig. 1. Study area: (A) macro location, (B) land tenure arrangements, and (C) geographical map. following: (1) forest degradation (change from primary to describe the locations and compare the performances of different secondary forest), because this process is an indicator of selective processes carried out for the same final aim, i.e., where and how logging in the area; (2) total deforestation (conversion of primary the system does it (González-López and Giampietro 2018). We and secondary forest to nonforest); and (3) conversion of primary used data from official censuses to characterize the distribution and secondary forest to avocado orchards, which is one of the of population (Appendix1, Table A1.2) and the annual benefits, main processes underlying forest loss in the study area i.e., income or equivalent monetary value of subsistence activities, (Barsimantov and Navia Antezana 2012). (Appendix 1, Table A1.3) across the different economic sectors of each cluster. We considered the paid-work sector (which We documented a total of 389 conflicts. We registered 161 forest includes the primary, secondary, and tertiary sectors); the conflicts between 2004 and 2014 from newspaper articles, household sector (subsistence activities, care work, and academic literature, and interviews with key informants; 133 of dependents); and emigration (permanent and circular) and them were caused by illegal logging and 28 related to ownership emigration flows. We then focused on the land-based activities of the forest. We added to these 28, the 228 landownership (crop production, livestock breeding, and forestry) from the conflicts reported by individuals from official censuses (lack of primary and subsistence sectors. Because part of the information titles, land invasions, and agrarian conflicts; Appendix 1, Table used was aggregated at the municipal level, we excluded the A1.1). We characterized illegal logging and landownership municipality of Uruapan, which encompasses communities conflicts according to the actors involved and their expression. belonging to all the clusters identified as well as the city of To sort communities by FCC patterns, we performed a cluster Uruapan, the second largest city of the state, whose structure and analysis with the standardized values of the initial avocado extent, function is quite different from that of a rural community. total forest loss, forest degradation, and spatial coordinates. We also mapped FCC (percentages of forest degradation, total Functional elements are associated with the definition of a goal, deforestation, and forest conversion to avocado) and conflict i.e., what the system does and why (González-López and intensity variables (number of conflicts weighted by the kind of Giampietro 2018). We characterized the functional elements of actors involved and their expression, and per 1000 habitants) to each subregion according to the local narratives to explore the explore their spatial patterns. processes underlying the various patterns of relations between FCC and conflicts. We gathered such information from 6 months Relational analysis of observant participation fieldwork and 59 semistructured Relational analysis explores expected patterns of relations interviews with government officials, foresters, researchers, between structural and functional elements. Structural elements authorities from various communities, and people involved in forest activities in the region. Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ RESULTS Fig. 3. Maps representing: (A) the initial land-cover/use (2004), Spatial patterns of forest-cover change and conflicts (B) forest-cover changes (2004-2014). Data obtained from the The cluster and visual analyses revealed three distinct spatial land-cover/use maps elaborated by Mas et. al (2017). patterns of FCC and conflicts across the study area (Fig. 2). Initial total forest (primary plus secondary) amounted to 58.4% of the land cover in the north, 52.6% in the southeast, and 49.4% in the southwest (Table 1). Initial avocado orchards were mainly restricted to the south, where they represented around 38% of the land cover. The most important FCC was forest degradation (7.0% in the north and 3.5% in the southeast), conversion of secondary forest (6.6%) and primary forest (2.9%) to avocado orchards in the southeast, and secondary forest loss (3.7% in the southeast and 2.4% in the north; Fig. 3, Table 1). Fig. 2. Cluster analysis regarding forest-cover change (FCC) patterns per community. Fig. 4. Number of conflicts per subregion grouped by: (A) causes, (B) actors, and (C) expressions. We registered 389 conflicts in total: 133 were caused by illegal logging and 256 were due to landownership issues (Fig. 4). With respect to the actors, 234 cases involved conflicts between individuals or groups, 85 were conflicts between communities, and 70 confronted communities with state actors. Regarding their expression, 140 involved direct actions, 91 confrontations, 84 The north of the Meseta Purépecha held the highest absolute tensions, 54 complaints, 11 fatal confrontations, and 9 number and the highest proportion of communities with conflict negotiations. because 27 of the 37 communities accounted for 221 of the conflicts. In the southwest, 11 of the 32 communities accounted Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ Table 1. Matrix of percentages of forest-cover changes per Fig. 5. Intensity of conflicts per community regarding their subregion (2004-2014). kind of actors and expressions (A and B) and frequency per 1000 habitants (C and D). 2014 2004 Primary Secondary Avocado No forest Total forest forest North Primary 91.71 7.02 0.40 0.88 32.55 forest Secondary 1.24 96.05 0.34 2.37 25.86 forest Avocado 0.30 0.41 98.66 0.63 1.38 No forest 0.15 0.40 0.33 99.12 40.21 Total 30.24 27.29 1.72 40.76 100.00 Southwest Primary 98.38 0.83 0.53 0.26 28.04 forest Secondary 0.18 97.29 1.80 0.73 21.44 forest Avocado 0.06 0.03 99.54 0.37 38.46 No forest 0.07 0.16 0.62 99.15 12.05 Total 27.66 21.13 38.90 12.32 100.00 Southeast Primary 92.01 3.48 2.95 1.56 29.38 forest Secondary 0.73 88.97 6.59 3.71 23.26 forest Avocado 0.03 0.21 98.91 0.85 39.35 No forest 0.22 0.54 1.91 97.33 8.01 Total 27.23 21.85 41.47 9.45 100.00 Computed using the land-cover/use maps developed by Mas et al. (2017). for 127 conflicts; and in the southeast 8 of the 32 communities accounted for 41 conflicts. The intensity of conflicts regarding their actors and expression and their frequency per 1000 inhabitants showed similar patterns (Fig. 5). Conflicts over illegal logging were especially intense in the north, but their frequency was by comparison lower in many densely populated communities from this region and higher in a few small communities in the The total population involved in land-based activities was south. Landownership conflicts were, in general, less intense and recorded as exceeding 100% because peasant households more evenly distributed than conflicts over illegal logging, but combined various activities (Fig. 7). Waged labor occupied the their frequency was comparatively higher than their expression highest proportion of the population in all subregions. In terms in some communities, especially in the southwest. of relative benefits, this activity ranked second in the north (20%) but its contribution was moderate in the south (< 5%). Structural elements of communities Commercial crop production occupied 14.9% of the peasants in Although in- and out-migration figures were similar in terms of the north, and 23% in the southwest and southeast; it dominated relative population for all subregions, remittances were the the land-based economy in the southwest and provided most of primary source of benefits in the north amounting to 32% of the the land-based benefits in the north. Commercial livestock total versus < 3% in the south (Fig. 6). The proportions of breeding employed similar proportions of peasants across the population within the paid work and household sectors were three subregions (between 11% and 15%). However, in the similar across the whole Meseta Purépecha region (about 36% southeast it provided a higher share of the land-based benefits and 63%, respectively). Moreover, relative population was also (40%). Commercial forest activities involved fewer than 2% of very similar within the subsistence, care work, and dependents peasants and generated less than 2% of benefits in all subregions. subsectors of the household sector. In all subregions, the primary sector held the highest proportion of employed population and The proportions of peasants practicing subsistence crop relative benefits. However, the proportion of benefits from production and of benefits derived from it were substantial in the primary activities varied greatly between the southwest (88%) and north (26% and 15%, respectively), but negligible in other the north (27%). All subregions had similar percentages of subregions, especially in terms of benefits. Subsistence livestock population within the secondary and tertiary sectors, although breeding was the second source of land-based benefits in the these were slightly lower in the southeast. The north held the southeast (35%), and it was also important in the north (10%). highest relative benefits from secondary (11%) and tertiary (20%) Subsistence forest activities were negligible in the three activities. subregions, in terms of both peasants involved and benefits. Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ Fig. 6. Relations between total benefits (TB) and total Fig. 7. Relations between total benefits (TB) and total population (TP) for the various economic sectors within each population (TP) for the commercial and subsistence land-based subregion. activities of each subregion. The high levels of forest degradation observed are probably Functional elements of communities caused by illegal logging, which is usually practiced on a small North subregion: illegal logging scale through the selection of the most commercially valuable Land is scarce in the northern subregion, benefits derived from it trees and does not entail the complete removal of the forest cover. reached a lower proportion of the population than in other Although benefits and population involved in forestry are subregions (Fig. 6), and the area per peasant was the lowest for negligible compared to other productive and subsistence all land-cover/use classes. Our fieldwork suggested that land is activities, its significance is certainly much higher than reported also unevenly distributed and many peasants, especially young because about one-half of the timber industries and volume household heads, lack rights to agricultural and forest plots. extracted in the Meseta Purépecha are illegal (COFOM 2007). Moreover, land has limited productive capacity because of Moreover, the traditional timber demand, associated with family bioclimatic restrictions (frosts and cyclic water scarcity). Maize carpentry and furniture workshops in this region, a nearby paper is the main crop in the north and it is mostly cultivated for mill, and the national construction industry, has been augmented subsistence purposes. Increasing competition due to the North by the demand for wooden packing crates for the export of America Free Trade Agreement, subsidies to producers in the avocados (Espín Díaz 1986, Works and Hadley 2004, Thiébaut USA, and the withdrawal of support to Mexican peasants have 2009). all rendered the sale of this crop profitless (Bello 2009, Thiébaut Agrarian conflicts between communities are also common; 2009). Other economic activities do not offer better alternatives; official figures report about 50 conflicts of this kind in the north benefits per worker from the secondary and tertiary sectors were (Vázquez León 2016) and 250,000 ha in dispute (Jasso-Martínez the lowest of all subregions. Remittances constitute the primary 2010). They result from the historically ambiguous demarcation source of benefits; despite low official figures, about 40% of the of boundaries between communities. However, timber extraction population in some communities currently lives in the United has intensified these landownership conflicts and created new States of America (Leco Tomás et al. 2009). tensions around the forest. Conflicts escalated because some communities who specialized in the production of boards, Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ furniture, and packing crates, exhausted their supplies of Southeast subregion: expansion of avocado orchards commercial-grade timber. Timber businesses forged patronage Benefits per employed inhabitant are quite high in the southeast; relations with illegal loggers, who looted neighboring in the secondary and tertiary sectors they are at least twice as communities, became increasingly organized, armed themselves much as in the other subregions. This is in part due to the presence to confront forest holders, and received support from their of a sugar mill and a poultry farm, which generate substantial extended families. Moreover, since 2007 the regional drug- direct and indirect employment. Benefits from the primary sector trafficking cartel has been extorting illegal sawmills and loggers are also notably high, with livestock breeding being the primary in exchange for protection. This situation reached a tipping point land-based activity in terms of benefits provided; this consists in 2011, when the indigenous community of Cherán rose up mainly of poultry breeding, although revenues from cattle are against illegal logging in its territory and successfully expelled the also important. Commercial crop production has a long tradition cartel (Fuentes Díaz and Paleta Pérez 2015). in this area and ranges from sugarcane to fruit trees, fodder, and some vegetables, all cultivated alongside subsistence maize (SIAP Southwest subregion: consolidation of avocado orchards 2016). The paid work sector of the southwest has the highest average benefits of all three subregions, but its economy is virtually In this subregion, conversion of secondary forest to avocado dominated by avocado cultivation (Fig. 7). This activity orchards is rife. It is indeed one of the hotspots of avocado considerably increases the amount of money flowing into expansion in Michoacán (Mas et al. 2017), which is spreading communities; one hectare of avocado is worth approximately mostly in both directions along an east-west axis following US$8000 whereas one hectare of maize, the usual crop in the study optimal bioclimatic conditions. Some commercial crops such as area, is worth US$800 (SIAP 2016). However, although waged sugarcane and mangoes are being replaced by avocado orchards labor employs most of the people involved in land-based activities, (SIAP 2016) because avocado renders considerably higher profits. most of the benefits are captured by avocado producers, However, avocado orchards are increasingly planted at the landowners, and middlemen. Benefits are thus very unequally expense of primary forest. Forest owners are persuaded to sell distributed (Thiébaut 2009). their land, usually to middlemen who sometimes even burn or throw away the timber, plant avocados, and sell the orchards at a We observed little FCC in this subregion. However, most of the profit (Del Castillo 2008). forest conversion to avocado orchards occurred before our remote sensing dataset. The first commercial avocado orchards were The southeast is composed mainly of ejidos and a small installed in the 1970s with the aid of appropriate infrastructure proportion of private property. Most landowners from ejidos and optimal growing conditions, but since the US approved the secured titles to their individual plots, and land conflicts and importation of avocados in 1997, the orchards have exponentially invasions are less numerous and intense than in other subregions. expanded. The southwest still conserves a considerable amount of forest in which biophysical conditions are not optimal for DISCUSSION avocado growth. Moreover, the successful community forestry Processes underlying forest-cover change and conflicts and company of San Juan Nuevo, in the municipality of Nuevo interrelations between subregions Parangaricutiro, and the presence of two resin plants nearby Spatial and relational analyses revealed three distinct patterns of generate considerable profits, although for a small part of the relations between FCC and conflict across the study area: (1) in population. the north, forest degradation was high and conflicts of varying This subregion suffers a relatively high proportion of conflicts intensity abundant and widespread, especially conflicts over stemming from a lack of land titles and from land invasions. This illegal logging; (2) in the southwest, FCC levels were few but the dispossession was noted during our fieldwork; many small initial avocado area was considerable, and conflicts were mostly landowners had lost or sold their land to investors and big over landownership; and (3) the southeast was undergoing producers because of the high investment costs and lagged return intensive forest conversion to avocado crops and forest on investments. Moreover, packing companies set up by the degradation but there were almost no conflicts. avocado elites or foreign capital have monopolized the We suggest that these diverging patterns correspond to different commercialization of avocado and have fixed prices at their processes and temporalities. In the northern subregion, poverty convenience (Ornelas 2018). Avocado production has also and land scarcity might have pushed the population toward proceeded through coercion and violence (Maldonado Aranda alternative sources of income. The importance of remittances 2014, Ornelas 2018). Since 2007, the regional drug-trafficking points to constrained economic activities in this subregion. cartel has been purchasing or stealing orchards to launder money Moreover, the prevalence of subsistence activities suggests that and has virtually controlled the market by extorting producers land provides some households with strategic resources (food and and packing companies through privileged access to official fuel) that could not be acquired in the market owing to insufficient production data and a “silver or lead” policy, i.e., accept the bribe money. Migration and benefits from remittances have historically or risk being killed (Ornelas 2018). Some communities from the alleviated the population pressure and economic constraints. southwest indeed participated in the vigilante groups that took However, according to local informants, increasing timber arms in 2013 to expel the organized crime from several demand has turned illegal logging into the best alternative for municipalities of the state (Flannery 2017). In contrast, conflicts landless and unemployed young peasants because they do not over deforestation are scarce. have to leave their communities and a single tree can be sold for about US$200, roughly the monthly salary of a laborer in nearby cash crops fields. The population has thus been able to survive at Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ the expense of looted common forest resources, but this has Spatial and relational analyses to unravel complex relations in exacerbated historical agrarian conflicts and provoked new ones socio-natural systems over the appropriation of others’ timber. By combining spatial and relational analyses, we were able to find contradictions and interpret diverse information to create a robust In the southwest subregion, large FCCs have occurred prior to explanation of the processes underlying FCC and conflicts in the the present study. Logging might indeed had facilitated this three subregions, as well as the complex relations among them. process, because avocados were usually planted over secondary The methodological approach was useful in identifying patterns forest owing to lower labor requirements (Barsimantov and Navia of relations more than exact quantities, and it also partially Antezana 2012). The consolidation of avocado orchards has overcame the limitations regarding data availability and seemingly provoked land grabbing and the concentration of the consequent uncertainty. In our case, these limitations were industry around big producers. These grievances are mostly because of reliance on data from different sources, temporalities, confronted individually, probably because this subregion is sample sizes, and levels of aggregation, and to the absence of composed mainly of private properties or ejido land, where many appropriate variables to explain some of the processes underlying peasants secured individual titles to their parcels. In contrast, both FCC and conflicts (especially migration and subsistence FCC might generate little opposition because profits are high and activities). Moreover, the comparison of data from different employment opportunities are generated by avocado growing, sources suggests considerable gaps in the information in the and because individual land titles make landowners fully agriculture, livestock, and forestry census (our main source of responsible for their land-use decisions. However, violence from information on land-based activities), especially for the north crime cartels might have deterred organized responses against subregion. These gaps reflect the informal character of many forest clearance. land-tenure arrangements, timber extraction, and processing In the southeast, the high incidence of FCC might be related to activities (COFOM 2007) and, within the rural population, the the process of avocado expansion. Avocados are sometimes general mistrust of public officials (Mathews 2008). planted inside cleared forests to hide illegal land-use changes, and Despite the limitations of our quantitative data, the use of this could explain the observed forest degradation, although complementary information from local narratives and a regional secondary forest loss probably corresponds to recently planted case study has illustrated the methodological approach we have orchards because young avocado trees cannot be detected by our developed, an approach that may help to unravel complex remote sensing dataset. Although there is massive conversion of relations in socio-natural systems. Recent studies have analyzed secondary forests, avocado orchards are increasingly growing at the relation between land-use change and environmental conflicts the expense of primary forests because high returns compensate by using complexity approaches (Hanaček and Rodríguez- for the investment in forest clearing. Cattle might also have some Labajos 2018) or spatial analysis (Simmons 2004, Durán et al. effect on FCC because forage crops occupied the fourth largest 2011). However, the combination of the two approaches is still in area (SIAP 2016), and the proportion of pasture was higher than its early stages (Ingalls and Mansfield 2017). By integrating spatial in the other subregions. Higher levels of benefits from other and relational analyses, we found diverse patterns of relations activities and initial benefits from the avocado growth have so far between FCC and conflicts in the Meseta Purépecha that were promoted social stability. However, coercion and violence from nonetheless interrelated and driven by the process of avocado big avocado producers and drug cartels might increase production. landownership conflicts as in the southwest. The importance of the forest cover for water retention together with the intensive Using relational analysis, we decomposed the Meseta Purépecha water requirements of avocado production could also create region into three subregions that differed in their constituent additional tensions in the whole region in the future (Friedmann components, analyzed their interactions and interdependent and McNair 2008). relations, and answered various questions to obtain a robust explanation of the FCC and conflicts observed in the three These FCCs and conflict patterns, although diverse, are subregions. Whereas structural elements provide information nonetheless interrelated. The whole territory of the Meseta about external constraints such as availability of natural Purépecha has been reorganized by avocado production and its resources, functional elements provide insights into the multiple secondary activities (Thiébaut 2009). These activities socioeconomic factors that impose internal constraints over have differentially shaped the structure and function of each natural systems (González-López and Giampietro 2017). subregion and the interrelations between them. The avocado Relational analysis thus allows critical examination of narratives boom soon dominated the economy in the southwest. High profits that link FCC and conflicts by using different representations of from avocado have prompted conversion of forest and other the study problem. This approach deals with the complexity commercial crops into avocado orchards in the southeast in inherent in both processes, although this implies a lack of response to optimal growing conditions; at the same time, a definitive answers about causality. Moreover, the integration of successful community forestry company has conserved a wide information from interviews shows patterns and relations area of forest that is less suitable for avocado growing. Sawmills between processes that quantitative data might hide. The use of in the north provide below-cost packing crates from illegal timber, spatial analysis greatly enhances this approach, especially and subsistence activities help supply cheap labor for the nearby regarding land-use processes because it accounts for the orchards. Thus, benefits along the commercialization chain are heterogeneities displayed by such processes across space. being extracted at the expense of human resources and the forests in the north (Reis 2019). Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ CONCLUSION Bofill Poch, S. 2005. El bosque político: los avatares de la We have developed a methodological approach to unravel the construcción de una comunidad modelo, San Juan Nuevo, underlying factors driving FCC and conflict, as well as the Michoacán, 1981-2001. El Colegio de Michoacán, Universidad complex relations between the two processes. Our approach de Barcelona, Zamora, Mexico. combines spatial and relational analyses to integrate and check Comisión Forestal del Estado de Michoacán (COFOM). 2007. the congruence of nonequivalent representations from Programa de desarrollo forestal sustentable del Estado de quantitative sources and observant participation at various scales, Michoacán 2030. Tomo I. Comisión Forestal del Estado de and to explain the heterogeneity that processes display through Michoacán, Morelia, Mexico. [online] URL: http://www.conafor. space. We show that relations between FCC and conflict are not gob.mx:8080/documentos/docs/43/6046La%20producci%C3%B3n% linear and depend on a combination of factors that hold 20de%20resina%20de%20pino%20en%20M%C3%A9xico.pdf interdependent, complex, and sometimes contradictory relations with the patterns observed. We suggest that this methodological Comisión Forestal del Estado de Michoacán (COFOM). 2007. approach can yield a more comprehensive and accurate Programa de desarrollo forestal sustentable del Estado de representation of a wide range of complex environmental Michoacán 2030. Tomo I. Comisión Forestal del Estado de problems in any geographic or social context, or when the lack of Michoacán, Morelia, Mexico. [online] URL: http://www.conafor. accurate and appropriate data for capturing different dimensions gob.mx:8080/documentos/docs/43/6046La%20producci%C3%B3n% of the study problem creates uncertainties. 20de%20resina%20de%20pino%20en%20M%C3%A9xico.pdf Comisión Nacional Forestal (CONAFOR). 2013. La producción Responses to this article can be read online at: de resina de pino en México. Comisión Nacional Forestal, Mexico http://www.ecologyandsociety.org/issues/responses. City, Mexico. [online] URL: https://www.gob.mx/cms/uploads/ attachment/file/80999/Estudio_-_La_produccion_de_resina_de_­ php/10992 pino_en_Mexico.pdf Comisión Nacional Para el Desarrollo de los Pueblos Indígenas (CDI). 2010. Catálogo de localidades indígenas 2010. Comisión Acknowledgments: Nacional Para el Desarrollo de los Pueblos Indígenas, Mexico This work was funded through the project PAPIIT IN301215 from City, Mexico. [online] URL: http://www.cdi.gob.mx/localidades2010- DGAPA-UNAM, the national scholarship program from gobmx/ CONACYT (number 558247/296508), and the scholarship Consejo Nacional de Evaluación de la Política de Desarrollo CONACYT-SENER Sustentabilidad Energética (number 409781). Social (CONEVAL). 2014. Medición de la pobreza en México Part of the data came from the Microdata Laboratory from INEGI 2010, a escala municipal. Consejo Nacional de Evaluación de la and we acknowledge the kindness of its staff. We are indebted to all Política de Desarrollo Social, Mexico City, Mexico. [online] URL: the interviewees and community members for their time, knowledge, https://www.coneval.org.mx/Medicion/Paginas/consulta_pobrez­ and hospitality during fieldwork. We are also grateful to Ann Grant a_municipal.aspx for her revision and language edition of the manuscript. Corson, C. 2011. Territorialization, enclosure and neoliberalism: non-state influence in struggles over Madagascar’s forests. Journal LITERATURE CITED of Peasant Studies 38(4):703-726. http://doi.org/10.1080/030661­ Affolderbach, J. 2011. Environmental bargains: power struggles 50.2011.607696 and decision making over British Columbia’s and Tasmania’s old- growth forests. Economic Geography 87(2):181-206. http://dx.doi. Del Castillo, A. 2008. Michoacán: la fiebre del “oro verde.” org/10.1111/j.1944-8287.2011.01107.x Público-Milenio, 14 December. [online] URL: http://verdebandera. mx/michoacan-la-fiebre-del-oro-verde/ Aldrich, S., R. Walker, C. Simmons, M. Caldas, and S. Perz. 2012. Contentious land change in the Amazon’s arc of deforestation. Durán, E., D. B. Bray, A. Velázquez, and A. Larrazábal. 2011. Annals of the Association of American Geographers 102 Multi-scale forest governance, deforestation, and violence in two (1):103-128. http://dx.doi.org/10.1080/00045608.2011.620501 regions of Guerrero, Mexico. World Development 39(4):611-619. http://doi.org/10.1016/j.worlddev.2010.08.018 Araujo, C., C. A. Bonjean, J.-L. Combes, P. Combes Motel, and E. J. Reis. 2009. Property rights and deforestation in the Brazilian Espín Díaz, J. L. 1986. Tierra fría, tierra de conflictos en Amazon. Ecological Economics 68(8-9):2461-2468. http://dx.doi. Michoacán. El Colegio de Michoacán, Zamora, México. org/10.1016/j.ecolecon.2008.12.015 Flannery, N. P. 2017. Mexico’s avocado army: how one city stood Barsimantov, J., and J. Navia Antezana. 2012. Forest cover change up to the drug cartels. Guardian 18 May. [online] URL: https:// and land tenure change in Mexico’s avocado region: is community www.theguardian.com/cities/2017/may/18/avocado-police-tancitaro- forestry related to reduced deforestation for high value crops? mexico-law-drug-cartels Applied Geography 32(2):844-853. http://dx.doi.org/10.1016/j. Friedmann, H., and A. McNair. 2008. Whose rules rule? apgeog.2011.09.001 Contested projects to certify ‘local production for distant Bello, W. F. 2009. Eroding the Mexican countryside. Pages 39-53 consumers.’ Journal of Agrarian Change 8(2-3):408-434. https:// in W. F. Bello, editor. The food wars. Verso, London, UK. doi.org/10.1111/j.1471-0366.2008.00175.x Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ Fuentes Díaz, A., and G. Paleta Pérez. 2015. Violencia y Kovacic, Z., and O. Viteri Salazar. 2017. The lose-lose autodefensas comunitarias en Michoacán, México. Iconos. predicament of deforestation through subsistence farming: Revista de Ciencias Sociales (53):171-186. unpacking agricultural expansion in the Ecuadorian Amazon. Journal of Rural Studies 51:105-114. http://doi.org/10.1016/j. Funtowicz, S., and J. R. Ravetz. 1994. Emergent complex systems. jrurstud.2017.02.002 Special Issue Complexity: Fad or Future? 26(6):568-582. https:// doi.org/10.1016/0016-3287(94)90029-9 Leco Tomás, C., A. Kido-Cruz, and R. Molina Martínez. 2009. Migración indígena a Estados Unidos. Purhépechas en Burnsville Gerber, J.-F., and S. Veuthey. 2011. Possession versus property in Norte Carolina. UMSNH, ININEE, CIMEXUS CIC, FH, a tree plantation socioenvironmental conflict in Southern Secretaría del Migrante, Morelia, México. Cameroon. Society and Natural Resources 24(8):831-848. http:// doi.org/10.1080/08941920.2010.521887 Louie, A. H. 2009. More than life itself: a synthetic continuation in relational biology. Ontos Verlag, Frankfurt, Germany. https:// Giampietro, M., R. J. Aspinall, J. Ramos-Martin, and S. G. F. doi.org/10.1515/9783110321944 Bukkens. 2014. Resource accounting for sustainability assessment: the nexus between energy, food, water and land use. Routledge, Maldonado Aranda, S. 2014. “You don’t see any violence here London, UK. https://doi.org/10.4324/9781315866895 but it leads to very ugly things”: forced solidarity and silent violence in Michoacán, Mexico. Dialectical Anthropology 38 Glew, L., and M. D. Hudson. 2007. Gorillas in the midst: the (2):153-171. http://doi.org/10.1007/s10624-014-9335-4 impact of armed conflict on the conservation of protected areas in sub-Saharan Africa. Oryx 41(2):140-150. https://doi. Mas, J.-F., R. Lemoine-Rodríguez, R. González-López, J. López- org/10.1017/S0030605307001755 Sánchez, A. Piña-Garduño, and E. Herrera-Flores. 2017. Land use/land cover change detection combining automatic processing González-López, R., and M. Giampietro. 2017. Multi-scale and visual interpretation. European Journal of Remote Sensing 50 integrated analysis of charcoal production in complex social- (1):626-635. http://doi.org/10.1080/22797254.2017.1387505 ecological systems. Frontiers in Environmental Science 5:54. http:// doi.org/10.3389/fenvs.2017.00054 Mathews, A. S. 2008. State making, knowledge, and ignorance: translation and concealment in Mexican forestry institutions. González-López, R., and M. Giampietro. 2018. Relational American Anthropologist 110(4):484-494. https://doi.org/10.1111/ analysis of the oil and gas sector of Mexico: implications for j.1548-1433.2008.00080.x Mexico’s energy reform. Energy 154:403-414. http://doi. org/10.1016/j.energy.2018.04.134 McNeely, J. A. 2003. Biodiversity, war, and tropical forests. Journal of Sustainable Forestry 16(3-4):1-20. http://doi. Hanaček, K., and B. Rodríguez-Labajos. 2018. Impacts of land- org/10.1300/J091v16n03_01 use and management changes on cultural agroecosystem services and environmental conflicts - a global review. Global McSweeney, K., E. A. Nielsen, M. J. Taylor, D. J. Wrathall, Z. Environmental Change 50:41-59. http://doi.org/10.1016/j. Pearson, O. Wang, and S. T. Plumb. 2014. Drug policy as gloenvcha.2018.02.016 conservation policy: narco-deforestation. Science 343(6170):489-490. https://doi.org/10.1126/science.1244082 Hecht, S. B., and A. Cockburn. 2010. The fate of the forest: developers, destroyers, and defenders of the Amazon. Updated Nuijten, M. 1997. Agrarian reform and the ejido in Mexico; edition. University of Chicago Press, Chicago, Illinois, USA. illegality within the framework of the law. Pages 72-104 in R. Kuppe and R. Potz, editors. Natural resources, environment, and Ingalls, M. L., and D. Mansfield. 2017. Resilience at the periphery: legal pluralism. Martinus Nijhoff, The Hague, The Netherlands. insurgency, agency and social-ecological change under armed conflict. Geoforum 84:126-137. http://doi.org/10.1016/j. O’Brien, K. 1998. Sacrificing the forest: environmental and social geoforum.2017.06.012 struggles in Chiapas. Westview, Boulder, Colorado, USA. https:// doi.org/10.4324/9780429497476 Jasso-Martínez, I. J. 2010. Las demandas agrarias en el discurso de las organizaciones purépechas de Michoacán: problemas sin Ornelas, R. G. 2018. Organized crime in Michoacán: rent-seeking resolver. Ra Ximhai 6(2):229-241. https://doi.org/10.35197/ activities in the avocado export market. Politics and Policy 46 rx.06.02.2010.06.ij (5):759-789. http://doi.org/10.1111/polp.12270 Klooster, D. 2003. Campesinos and Mexican forest policy during Orozco-Ramírez, Q., M. Astier, and S. Barrasa. 2017. the twentieth century. Latin American Research Review 38 Agricultural land use change after NAFTA in Central West (2):94-126. https://doi.org/10.1353/lar.2003.0018 Mexico. Land 6(4):66. http://doi.org/10.3390/land6040066 Koning, R., D. Capistrano, Y. Yasmi, and P. Cerutti. 2008. Forest- Peluso, N. L., and C. Lund. 2011. New frontiers of land control: related conflict. Impacts, links, and measures to mitigate. Rights introduction. Journal of Peasant Studies 38(4):667-681. http://doi. and Resources Initiative, Washington, D.C., USA. http://www. org/10.1080/03066150.2011.607692 apafri.org/activities/Bhutan2013/publication/m8%20forest%20related% Randall, L. 1996. Reforming Mexico’s agrarian reform. M.E. 20conflict.pdf Sharpe, Armonk, New York, USA. Kovacic, Z. 2018. Conceptualizing numbers at the science-policy Reis, N. 2019. A farewell to urban/rural bias: peripheral finance interface. Science, Technology, and Human Values 43 capitalism in Mexico. Journal of Peasant Studies 46:702-728. (6):1039-1065. https://doi.org/10.1177/0162243918770734 http://doi.org/10.1080/03066150.2017.1395856 Ecology and Society 24(3): 3 https://www.ecologyandsociety.org/vol24/iss3/art3/ Rincón-Ruiz, A., and G. Kallis. 2013. Caught in the middle, Vázquez León, L. 2016. Multiculturalism as a juridical weapon: Colombia’s war on drugs and its effects on forest and people. the use and abuse of the concept of ‘Pueblo Originario’ in agrarian Geoforum 46:60-78. https://doi.org/10.1016/j.geoforum.2012.12.009 conflicts in Michoacán, Mexico. Pages 35-73 in D. Lehmann, editor. The crisis of multiculturalism in Latin America. Palgrave Rocheleau, D. E. 2015. Networked, rooted and territorial: green Macmillan, New York, New York, USA. grabbing and resistance in Chiapas. Journal of Peasant Studies 42 (3-4):695-723. http://doi.org/10.1080/03066150.2014.993622 Veldkamp, A., and E. F. Lambin. 2001. Predicting land-use change. Agriculture, Ecosystems and Environment 85(1):1-6. Rodrigues, A. S. L., R. M. Ewers, L. Parry, C. Souza, Jr., A. https://doi.org/10.1016/S0167-8809(01)00199-2 Veríssimo, and A. Balmford. 2009. Boom-and-bust development patterns across the Amazon deforestation frontier. Science 324 Works, M. A., and K. S. Hadley. 2004. The cultural context of (5933):1435-1437. http://doi.org/10.1126/science.1174002 forest degradation in adjacent Purépechan communities, Michoacán, Mexico. Geographical Journal 170(1):22-38. https:// Rosen, R. 1999. Essays on life itself. Columbia University Press, doi.org/10.1111/j.0016-7398.2004.05024.x New York, New York, USA. Rosen, R. 2012. Anticipatory systems. Philosophical, mathematical, and methodological foundations. Vol. 1. Springer, New York, New York, USA. Rustad, S. C. A., J. K. Rød, W. Larsen, and N. P. Gleditsch. 2008. Foliage and fighting: forest resources and the onset, duration, and location of civil war. Political Geography 27(7):761-782. http:// doi.org/10.1016/j.polgeo.2008.09.004 Saltelli, A., and M. Giampietro. 2017. What is wrong with evidence based policy, and how can it be improved? Futures 91:62-71. https://doi.org/10.1016/j.futures.2016.11.012 Sánchez-Cuervo, A. M., and T. M. Aide. 2013. Consequences of the armed conflict, forced human displacement, and land abandonment on forest cover change in Colombia: a multi-scaled analysis. Ecosystems 16(6):1052-1070. http://doi.org/10.1007/ s10021-013-9667-y Seghezzo, L., J. N. Volante, J. M. Paruelo, D. J. Somma, E. C. Buliubasich, H. E. Rodríguez, S. Gagnon, and M. Hufty. 2011. Native forests and agriculture in Salta (Argentina): conflicting visions of development. Journal of Environment and Development 20(3):251-277. http://doi.org/10.1177/1070496511416915 Servicio de Información Agroalimentaria y Pesquera (SIAP). 2016. Estadística de producción agrícola. Servicio de Información Agroalimentaria y Pesquera, Mexico City, Mexico. [online] URL: http://infosiap.siap.gob.mx/gobmx/datosAbiertos.php Simmons, C. S. 2004. The political economy of land conflict in the eastern Brazilian Amazon. Annals of the Association of American Geographers 94(1):183-206. http://doi.org/10.1111/ j.1467-8306.2004.09401010.x Suthakar, K., and E. N. Bui. 2008. Land use/cover changes in the war-ravaged Jaffna Peninsula, Sri Lanka, 1984-early 2004. Singapore Journal of Tropical Geography 29(2):205-220. https:// doi.org/10.1111/j.1467-9493.2008.00329.x Thiébaut, V. 2009. La libéralisation économique et la restructuration des territoires ruraux au Mexique: L’exemple de la culture de l’avocat dans l’État de Michoacán. Cahiers de géographie du Québec 53(148):63-81. https://doi.org/10.7202/038142ar Torres-Mazuera, G., J. Fernández Mendiburu, and C. Gómez Godoy. 2018. Informe sobre la jurisdicción agraria y los derechos de los pueblos indígenas y campesinos en México. Fundación para el Debido Proceso, Washington, D.C., USA. [online] URL: http:// www.dplf.org/sites/default/files/informe_jurisdiccion_agraria_ve­ rsion_final.pdf Appendix 1. Variables used in spatial and relational analyses Table A1.1. Variables used for characterizing forest conflicts Variable name Description Actors in conflict Categories: individuals or groups; communities; government Causes of conflict Categories: illegal logging; landownership Intensities of conflict Categories and examples: negotiation (conciliation); complaint (legal complaints); tension (threats, protracted and latent conflicts, lack of land titles); direct action (protests, roadblocks, kidnappings, land invasions); confrontation (violent clashes, sometimes with injured people, land disputes); killing (violent clashes that result in deaths) Weighted conflicts Sum of the forest conflicts multiplied by their intensity (1=negotiation, 2=complaints, 3=tension, 4=direct actions, 5=confrontations, 6=deadly confrontations) and actors (2=individuals or groups, 4=communities, 6=government) for each community † Conflicts were documented from secondary sources (newspaper articles, academic literature, and interviews with key informants) and from the 2007 Agricultural, livestock and forestry census. 1 Table A1.2. Variables used for characterizing population in relational analyses Category Source Description Migration PC Population involved in permanent or circular migration (sum of migration subcategories) Permanent PC Population that emigrated to the United States of America between 2005 and 2010 and remained there during the census survey Circular PC Population that resided in United States of America or in another Mexican state in June 2005 but were back during the census survey Immigration PC Population born in another state Total PC Population residing habitually in the locality Paid work PC Economically active population (population over 12 years old employed or that looked for a job in the week before the survey) Tertiary EC Population employed in the tertiary sector Secondary EC Population employed in the secondary sector Primary ALF Population employed in the primary sector (sum of primary sector subcategories) Agriculture ALF Population involved in commercial agriculture (sum of agricultural households that sell all their crops plus half of the ones that sell part of their crops) Livestock ALF Population involved in commercial livestock breeding (half of the peasant households that own livestock) Forest ALF Population engaged in commercial forest activities (sum of the peasant households involved in timber processing, logging or resin tapping) Labor ALF Population hired in the primary sector Household PC Population that belongs to the household sector (sum of economically inactive population plus population between 0 and 12 years old) Subsistence ALF Population involved in subsistence activities (population retired and involved in housework minus the total economically inactive population) Agriculture ALF Population that grow crops for self-consumption (sum of agricultural households that consume all their crops plus half of the ones that consume part of them) Livestock ALF Population that breed animals for self-consumption (half of the peasant households that own livestock) Forest ALF Population that extract forest products for domestic purposes (number of peasant households that collect firewood) Care work PC Population over 12 years engaged in housework Dependents PC Dependent population (population retired plus the population between 0 and 12 years old) † PC stands for 2010 Population census, gathered from published data at locality level; EC for 2009 Economic census, gathered from published data at municipal level; and ALF for 2007 Agricultural, livestock and forestry census, gathered from microdata at locality level. All censuses come from the National Institute of Statistics and Geography (INEGI) 2 Table A1.3. Variables used for characterizing benefits in relational analysis Category Source Description Migration CONAPO Benefits received as remittances in 2004 Total EC, SIAP, ALF Annual benefits (sum of remittances plus benefits from the paid work and subsistence sectors) Paid work EC, SIAP, ALF Annual benefits plus salaries from the paid work sector (sum of benefits from all the economic sectors) Tertiary EC Annual net benefits plus salaries from the tertiary sector Secondary EC Annual net benefits plus salaries from the secondary sector Primary SIAP, ALF Annual net benefits plus salaries from the primary sector Agriculture SIAP Annual gross benefits from commercial agriculture (sum of the value of all produce except corn) Livestock SIAP Annual gross benefits from commercial livestock breeding (value of all livestock products except standing livestock) Forest ALF Annual gross benefits from timber and resin extraction. Computed using 7400 MXN as the price per tone of resin1, and 1227 MXN as the price of timber per cubic meter2 Labor ALF Annual benefits obtained as salaries in the primary sector. Computed using 4000 MXN as the monthly salary of a laborer Subsistence SIAP, ALF Annual monetary value from subsistence activities (sum of the subsistence categories) Agriculture SIAP Annual monetary value from subsistence agriculture. We assumed that all the corn cultivated was for self-consumption Livestock SIAP Annual monetary value from subsistence livestock breeding. We equated subsistence livestock breeding to the standing livestock. We only used bovine cattle in Figure 7 Forest ALF Annual monetary value from firewood collection. We used 100 MXN as the price per cubic meter of firewood3 † CONAPO stands for National Population Council, from the Bank of Mexico, gathered from published data from 2014 at municipal level; EC for 2009 Economic census, from INEGI, gathered from published data at municipal level; SIAP for Agri-food and Fisheries Information Service from SAGARPA, gathered from published data from 2007 at municipal level; and ALF for 2007 Agricultural, livestock and forestry census, from INEGI, gathered from microdata at locality level. 1 CONAFOR (2013), 2CONAFOR (2010), 3COFOM (2007) 3