Dam Impacts in a Time of Globalization: Using Multiple Methods to Document Social and Environmental Change in Rural Honduras
2003; University of Chicago Press; Volume: 44; Issue: S5 Linguagem: Inglês
10.1086/379271
ISSN1537-5382
Autores Tópico(s)Conservation, Biodiversity, and Resource Management
ResumoPrevious articleNext article FreeReportsDam Impacts in a Time of Globalization: Using Multiple Methods to Document Social and Environmental Change in Rural Honduras1WilliamM.LokerWilliamM.LokerDepartment of Anthropology, California State University, Chico, Calif. 959290400, U.S.A. ([email protected]). 19 iii 03Department of Anthropology, California State University, Chico, Calif. 959290400, U.S.A. ([email protected]). 19 iii 03PDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreThe El Cajn region of centralwestern Honduras (fig. 1 and fig. A) is an interesting case for the study of social and environmental change because it is the site of a major hydroelectric dam constructed in the early 1980s that flooded extensive areas of land and affected thousands of people. Daminduced changes coincided and interacted with widespread changes in Honduran rural society. My research on the region addresses several longstanding interests in anthropology: the social causes and consequences of environmental change, the longterm effects of large infrastructure projects, and general processes of social change. The goal of this report is to document the environmental changes caused by the dam, their effects on local people, and subsequent social and economic changes, especially in agricultural land use, within the broader context of a changing political economy.Fig. 1. Location of the study area.View Large ImageDownload PowerPointFig. A. Location of the study area (detail).View Large ImageDownload PowerPointThe interaction of dam impacts with broader politicaleconomic changes complicates our analysis and makes a strictly local account incomplete. These changes, affecting rural Honduras, other areas of Central America, and the Third World in general, are associated with globalization and present serious challenges to rural livelihoods, precipitating a variety of responses from households and communities (Bebbington 1993, 1999; Loker 1999). Globalization in Honduras includes increased incorporation into international markets via trade liberalization, the implementation of structural adjustment packages to effect marketoriented economic reforms, greater economic volatility in terms of prices, inflation, and macroeconomic policies in general (see Thorpe et al. 1995), and an increase in the flow of global cultural ideas and images that is reducing the cultural isolation of rural communities. These broad international and national changes are manifested in the study region in (1) the rise of assembly plant industries (maquiladoras) around San Pedro Sula, providing a relatively accessible migration target and employment alternative for people living in the region (Bickham Mndez and Kopke 1998, Flores and Kennedy 1996); (2) structural adjustment policies that have liberalized markets for agricultural goods and led to currency devaluations and high rates of inflation, contributing to a more volatile economic situation in terms of wages and prices for a variety of commodities and increasing the uncertainty of the economic returns to agriculture (Salgado 1994:2932); (3) volatility and a general downward trend in coffee prices since the collapse in 1989 of the International Coffee Agreement, leading to increased uncertainty in the availability of work and wages paid in this sector; and (4) increased exposure to new cultural images and ideas and to people with broader experience outside the region, including those who have migrated to the United States. All of this has occurred in the context of rapid population growth that has placed increased demographic pressure on agricultural resources.These trends make it difficult to isolate the social and environmental impacts of the dam from more general processes of change. However, information drawn from agricultural census documents and other national and regional studies will be used to compare processes occurring in the El Cajn study area with related changes occurring elsewhere in Honduras to determine the specific effects of the dam. One of the principal arguments advanced here is that the effects of the dam and those of globalization have unexpected and unsettling similarities and synergies such that the dam has greatly accelerated the more general changes associated with globalization.The results reported here are based on 28 months of fieldwork, from 1981 to 1984 and 1994 to 1999. The first period covered the construction of the dam, its closure, and the filling of the reservoir. This report focuses on the results from the second period, drawing on the earlier research primarily as a baseline for assessing change. The methodology employed included (1) indepth, structured interviews carried out with 51 household heads in the largest community in the immediate vicinity of the reservoir, Montauelas (population 869 in 1998), focused on household demography and economy and opinions about the dam; (2) 53 interviews (using the same set of questions) of household heads living in seven different settlements outside Montauelas around the reservoir; (3) a full census of Montauelas and the adjacent community of El Encinal, gathering information on age/sex composition of households and landownership; (4) oral histories of 17 individuals ranging in age from 19 to 83, totaling some 20 hours of taperecorded information; (5) detailed information on the costs of production derived from a milpa plot rented in 1998; (6) observations (recorded in field notes) derived from hundreds of hours of participantobservation and informal interviews focused on land use and environmental change in the study area; and (7) analysis of two sets of 1 : 20,000 blackandwhite aerial photographs, dating from 1981 and 1994, to map the location and extent of vegetation cover in the study area (see Table 1). The vegetation categories defined attempt to capture the varying states of an extremely dynamic (agri)cultural landscape. Particular spaces on this landscape may cycle rapidly through milpa (fields cultivated with maize and beans), pasture, brushy pastures, and fallow/secondary forest. These categories are aggregated at varying points in the analysis to reflect this dynamism and the difficulty of distinguishing these types of vegetation both in the photos and on the ground and to simplify the presentation of research results. The vegetation maps derived from the aerial photographs were digitally scanned, orthorectified, and analyzed with geographic information systems software to identify and quantify changes in vegetation cover in the intervening years. When combined with ethnographic research, analysis of vegetation cover enables us to link environmental change with landuse choices made by local people, an important step forward in documenting and understanding humaninduced environmental change.Table 1 Distribution of Land Cover Types Before and After Construction of the El Cajn DamVegetation TypeABCD(DA)(DB)Total Study Area Before Flooding (1981)Unflooded Portion of Study Area (1981)Flooded Portion of Study Area (1981)Unflooded Portion of Study Area (1994)(Unflooded 1994 Total Before Flooding 1981)(Unflooded 1994 Total After Flooding 1981)(ha)(%)(ha)(%)(ha)(%)(ha)(%)(ha)(%)(ha)(%)Towns/trails230.1230.100490.3+26+0.2+26+0.3Agricultural land Cleared Milpa2250.91110.61141.91911.134+0.2+80+0.5Clean pasture2,89512.21,94511.095015.82,60714.8288+2.6+662+3.8Savannah2691.12691.5002351.334+0.2340.2 Total3,38914.32,32513.11,06417.73,03317.2356+3.0+708+4.1Brush Brushy pasture2,41810.21,3797.81,03917.33,52320.0+ 1,105+9.8+ 2,144+12.2Fallow2,36610.01,3427.61,02417.11,6239.27430.8+281+1.6Cleared riparian1,0684.56183.54507.57864.52820.0+168+1.0Total5,85224.73,33918.92,51341.95,93233.7+80+9.0+ 2,593+14.8Total agricultural9,24139.05,66432.13,57759.68,96550.1276+11.9+ 3,301+18.0Forest Pineoak Thin3,08213.02,28112.980113.33,73921.2+657+8.2+ 1,458+8.3 Closed8,94037.88,32747.161310.23,76621.45,17416.44,56125.7Total12,02250.810,60860.11,41423.57,50542.64,5178.23,10317.5Broadleaf Intact riparian1,0024.25443.14577.61230.78793.54212.4 Broadleaf9203.97874.41332.24142.35061.63732.1 Total1,9228.11,3317.55919.85373.01,3855.17944.5 Total forest13,94758.911,93967.52,00533.38,04245.65,90213.33,89721.9Water and unclassified4611.9340.24277.15583.1+97+1.2+524+2.9Total23,66910017,6601006,00910017,614100 View Table ImageThis research focuses on the interrelationships among the choices people make, the cultural means that both constrain and enable these choices, and the ecosystems that sustain human life and culture. The interactions among action, social structure, and ecosystem are informed by a variety of theoretical approaches, including the literature on farmer decisionmaking (Barlett 1982, Netting 1993), cultural ecology, especially the notion of adaptive strategies and processes (Bennett 1976, 1993), and various of the new ecologies: political ecology (Blaikie and Brookfield 1987, Greenberg and Park 1994, Stonich 1993, Painter and Durham 1995), the ecology of practice (Nyerges 1997), and historical ecology (Bale 1998). These approaches are combined in a hybrid approach that focuses on the social, politicaleconomic, and environmental factors that shape human agency over time in particular places and circumstances (Bebbington and Batterbury 2001).The principal state institution affecting adaptive strategies in the El Cajn region is the Honduran national electrical company, the Empresa Nacional de Energa Elctrica (ENEE), the entity charged with managing the dam and reservoir. While the ENEE is powerful and the dam had tremendous impact on the local environment, local people were still faced with choices and had considerable room for maneuver in coping with that impact. The hybrid approach pursued here analyzes how the social situations of individuals and groups affect their creative responses to changing circumstances and how institutional frameworks that shape access to resources affect the ability of individuals and households to construct sustainable livelihoods.After a brief discussion of background information on the dam and the social and environmental conditions in the region, I will describe the patterns of environmental change apparent in the aerial photographs and the associated changes in land use and other relationships. Finally, I will explore the implications of this case for our understanding of the interrelated effects of this large infrastructure project and globalization on social and environmental change in the study area.The El Cajn DamThe El Cajn dam is a doublecurvature concrete arch structure 226 meters tall, the highest in Central America and one of the highest in the world (fig. B). It was conceived and planned in the 1970s, when the upward spike in petroleum prices sent oilimporting countries such as Honduras scrambling for alternative sources of energy. Financed largely by loans from the InterAmerican Development Bank and the World Bank, it cost about $800 million to build. The dam created a 92squarekilometer reservoir that flooded about 6,000 hectares of land, and in good years it has generated up to 70% of Hondurass electricity. The project has been beset by management problems, including low levels of water in the reservoir that reduced generating capacity, leaks, and a fire in the generating room. Most relevant to the current study, the project included elaborate plans to resettle the approximately 1,8003,000 people directly displaced by the dam.2 These plans were never fully implemented, and only 47 of the 300 families who lost land and/or houses to flooding were resettled. Most people simply moved upslope, creating a massive and instantaneous increase in demographic pressure on a greatly impoverished land base.Fig. B. El Cajn Dam, 266 m tall, the highest dam in Central America.View Large ImageDownload PowerPointAbout 80% of the population of the study region make their living from agriculture, with the other 20% depending primarily on commerce, construction, and other trades. Some 60% of the population relies on some combination of milpa agriculture (maize and beans) and agricultural labor, sometimes combined with other minor incomeproducing activities. The other agriculturalists (1520%) focus on cattle and coffee, the two highvalue, highstatus activities. These households are the regions large landowners. Coffee is not a prominent crop in the study area, which at less than 900 meters above sea level is not prime coffee land. The small patches of coffee present are generally remnants of rustic, illtended plots of coffee grown in natural forest conditions and are of limited economic importance. Those who derive significant income from coffee generally have modestsized plantations (< 20 hectares) in nearby highland zones. Virtually all the wealthier landowners were born and reside in the region (absentee landownership is rare), are linked by kinship to other local families, and would not be considered very wealthy even by Honduran standards. My 1998 census indicated that only 41 of 151 households (27%) owned land. In summary, the El Cajn region is one in which agriculture remains overwhelmingly important in terms of livelihood and land is very unequally distributed.Ecologically, the El Cajn region is characterized by rugged topographylow but steep hills and narrow river valleys now filled by the reservoir. The climate is seasonally wet and dry, with an average of about 1,500 mm of rainfall, 80% of which falls between May and November. The vegetation of the region is dominated by pineoak forest, scrub vegetation produced by the clearing of forest for agriculture, and small remnants of the broadleaf tropical forest that formerly existed in the nowflooded river bottoms and in upland areas of better soil and the narrow canyons of smaller rivers and creeks (fig. C, D, and E).Fig. C. A typical agricultural landscape in the study area, showing brushy vegetation in foreground (where there was milpa the previous year) and mosaic of cleared land, thin forests, and broadleaf forest new horizon in the background on the left.View Large ImageDownload PowerPointFig. D. Cow grazing in field used as milpa the previous year, showing a year's growth of vegetation.View Large ImageDownload PowerPointFig. E. Gumercindo Anariba cultivating milpa in the less furtile, more highly erodable soils of the pineoak uplands.View Large ImageDownload PowerPoint2In addition to the approximately 1,800 people directly displaced by the dam, about an equal number lost access to land flooded by the reservoir.Environmental ChangeFigures 24 present vegetationcover maps derived from the interpretation of aerial photos for 1981 and 1994 and a map of vegetationcover change. The information presented visually in the maps is summarized in Table 1. The reservoir flooded about 6,000 hectares of land, of which about 60% were in some sort of agricultural use. Before flooding there had been 9,241 hectares of agricultural land (milpa, pasture, fallow, etc.); immediately thereafter there were 5,664 hectares, representing a loss of 40%. By 1994, however, there were 8,965 hectares of agricultural land, nearly the same as before the dam. Pineoak forest was reduced from 12,022 hectares in 1981 to 7,505 hectares in 1994. Flooding was responsible for 1,414 hectares of this loss while the rest was cleared for agriculture between 1984 and 1994. Broadleaf forest declined from 1,922 hectares in 1981 to 537 hectares in 1994, a reduction of nearly 60%. The reservoir flooded about 590 hectares of broadleaf forest, and in the ensuing ten years an additional 794 hectares of broadleaf forest were cleared for agricultural uses (milpa and pasture).Fig. 2. Land cover in the El Cajn region, 1981.View Large ImageDownload PowerPointFig. 3. Land cover in the El Cajn region, 1994.View Large ImageDownload PowerPointFig. 4. Land cover change in the El Cajn region, 198194.View Large ImageDownload PowerPointAs a proportion of the landscape, agricultural land increased from about 39% in 1981 to slightly over 50% in 1994 while forest declined from 59% to about 46%. Much of the land cover classified as forest is actually a hybrid land use involving grazing (thin pineoak forest) or shade for coffee (broadleaf forest). In 1981, 74% of the pineoak forest was closed forest; in 1994 just 50% was closed. Thin pineoak forest (modified by logging, selective clearing, and/or burning to maintain a more open habitat for grazing) grew from 24% to 50% of the total area in this vegetation class. If we include it as part of the agricultural landscape, the area devoted to agriculture is closer to 74% of the total.In summary, the El Cajn dam flooded a significant portion of the regions agricultural land. In response, from 1984 to 1994 people replaced, almost hectareforhectare, the agricultural land lost to the reservoir. In terms of total area, pineoak forests suffered the greatest decline, but in terms of percentage, speciesrich broadleaf forests suffered the most, probably leading to local extinctions of flora and fauna in this ecotype.Land Use and Social ChangeThe filling of the El Cajn reservoir dramatically affected the productive capacity of the local environment. As we have seen, the dam flooded approximately 40% of the agricultural land in the study area, and one of the principal adaptive priorities of local people was to recoup this loss through accelerated clearance of land for agriculture. That they succeeded in recovering the lost land in the aggregate does not mean that all individuals or households had their access to land and livelihood restored. The lands cleared after 1981 were generally of lower quality than the nowflooded alluvial lands, and 88% of the residents interviewed in 1994 said that land was scarcer or much scarcer than ten years before. In subsequent years, people often reported that cattlemen had either grabbed or bought up land to devote to pasture after the creation of the reservoir.3 Here is an excerpt from my field notes of June 21, 1998, on this subject: As C puts it, after criticizing the high milpa rents, in which he basically accused ganaderos of renting out lands that were tierras nacionales [government lands], he said, I am not asking for a handout, I just want a place to work and feed my family, with resentment in his voice. Those interviewed mentioned that large landowners, who were more likely to have valid legal documents to demonstrate landownership, used the compensation they received from the state when their lands were flooded to buy unflooded lands (a notion supported by analysis of the compensation paid by the ENEE [see Loker 1998]). The net result was a severe attenuation of access to land for the vast majority of households. Older individuals also remarked that lands that had formerly been secondary forest or milpa lands were now sown with pasture grasses, which made them more difficult to use for growing maize (field notes, June 24, 1998).4 People have coped with the loss of prime agricultural land through extensification (expansion of agriculture into previously uncultivated areas), intensification (the rapid adoption of chemical inputs, especially herbicides and fertilizers, in the context of a change from shifting to nearly continuous cultivation), and migration (some to jobs in the maquiladora plants around San Pedro Sula, some to the United States).The average size of milpas has declined 28%, from 2.5 ha in 1983 to 1.8 ha in 1998, a statistically significant change (p < 0.05). Average yield has declined 15%from 1,440 kg/ha in 1983 to 1,220 kg/ha 1998 (p < 0.05)but this actually underestimates the magnitude of the impact of the dam for two reasons. One is that the average yield before construction of the dam on riverbottom lands (all of which were flooded) was 1,831 kg/ha. The other is that yields reported for 1998 include milpas on which fertilizer was used; the average yield of unfertilized lands in 1998 was only 850 kg/ha. None of the farmers interviewed in 1983 used fertilizer, while 82% of those interviewed in 1998 did. Thus, if we compare the average yield of unfertilized lands in 1983 (1,440 kg/ha) with the average yield of unfertilized milpa in 1998 (850 kg/ha), the result is a 40% decline. This comparison supports the frequently voiced claim that fertilizer is absolutely necessary to maintain acceptable yields because the land is tired.The land is tired because, increasingly, it is under continuous cultivation. When not under cultivation, it is in pasture; rarely is it left fallow long enough for forest to regenerate. In 1983, before the dam, the average fallow period for upland milpas was five to ten years. Now agricultural land goes rapidly through a cycle of milpa cultivation, grazing immediately after harvest for a variable period of time, and a brief fallow of two to three years or, alternatively, from milpa to grazing and back to milpa again without fallow, often through land rental from the cattleman to the landless campesino. Forest regeneration is a critical aspect of successful shifting cultivation as a means of restoring soil fertility and suppressing weeds (Nair 1987, Ewel et al. 1981, Szott et al. 1987). Multiyear cropping increased from 57% of milpas in 1983 to 78% of milpas in 1998. Continuous cultivation has been accompanied by increased use of chemical inputs in an attempt to maintain acceptable yields and combat persistent weeds, including aggressive pasture grasses.Agricultural census data indicate that there has been a general trend of fertilizer adoption in Honduras during the period covered here. In the late 1970s about 5% of Honduran farmers used chemical fertilizers, and this figure rose to 25% in 198788 and 35% in 1993 (Baumeister and Wattel 1996). Jensen (1998), reporting on research carried out in Santa Barbara, Honduras, in 1994, indicated that 45% of his sample used fertilizer and 50% used herbicides. As we have seen, fertilizer use in the El Cajn region, absent in the 1980s, had increased by 1998 to well above these figures. Herbicide use increased from 65% of farmers in 1983 to 100% in 1998, with producers citing the same reason as Jensens (1998:118) respondents: the weed invasion associated with shorter fallows. The adoption of fertilizer and herbicides in the El Cajn region is driven by land scarcity induced by flooding and a social system that forces a largely landless farming population to rent (degraded) pastures for milpa production. Chemical fertilizers and herbicides are necessary to maintain adequate levels of production on these lands. Thus the construction of the dam has accelerated trends found elsewhere in the country.Land scarcity and rising costs of production are felt differentially and most acutely by the poorest and the landless. Larger landowners may not face scarcity, but they do face higher production costs (640 personhours/ha in contrast to 400 before the dam) and demands for land to rent from dispossessed kin and neighbors. Land rental has increased from 28% to 34% between 1983 and 1998, while the incidence of cultivation on government lands, the most common form of land access in 1983, has declined from 32% to none. This is another trend documented for Honduras as a whole in which the El Cajn region represents an extreme. Nationally, the percentage of farms of government lands fell from 34% in 1952 to 33% in 1974 and 23% in 1993, while rental land increased from 9% in 1952 to 23% in 1974 then declined to 17% in 1993 (Salgado 1996:9596). On a regional level, approximately 18% of the plots studied by Jensen in 1993 in Santa Barbara were rented and 21% were located on ejidal (common) lands5 (Jensen 1998:119). Between 1983 and 1998, El Cajns government lands disappeared. Some were privatized by fencing with barbed wire, and some were flooded. The loss of these lands represents a closure of the commons in which poorer households are losing access to land as tenure becomes more formalized and restrictive. This process is not unique to the El Cajn region and could be considered linked to the marketbased reforms and liberalization typical of globalization. Here the process was catalyzed and accelerated by daminduced flooding.The main social effect of the dam and its reservoir has been to increase the insecurity, marginalization, and vulnerability of poorer households. These social changes have increased dependency and intensified patronclient ties. The social response of the poor to this situation has been to seek access to productive resources from their betteroff neighborskin and/or patrons who rent milpa plots to the landless. Other strategies that strengthen the dependence of the poor on the wealthier households include informal moneylending (at rates of 10% per month) and the pawning of future maize harvests at deep discounts as poorer families struggle to scrape together the cash necessary to rent land, buy inputs, and deal with daily contingencies. This further indebts and obligates the landless poor to the betteroff local households. In exchange for access to land, the poor are basically on call to work for the patron for low wages. The result is an economic system increasingly inimical to the interests of the poor majority.3The state of land titling in the study area is highly confused. Local land records are in disarray and have not been updated since the mid1980s. People claim to own land on the basis of informal titles and land sales that are not formally registered and therefore of questionable legal validity. The Ley de Modernizacin Agrcola (Agricultural Modernization Law), whose major provision includes formalization of land titles, has had no discernible effect in the study area (see Salgado 1996:11819).4Because of the difficulty of distinguishing recently sown milpas from pastures in the photographs, this trend is difficult to document solely through aerial photograph interpretation. Only detailed ethnographic research brought this trend to light.5Ejidal lands and government lands are not in the same legal category, but both imply rentfree use rights. These categories are combined in Salgado (1996).MigrationIn the face of increasingly precarious livelihoods, one of the most common individual and household responses has been outmigration. Parents may encourage a daughter of young working age (over about 14 years old) to seek work in the city or encourage a son or daughter to make the trip north to the United States, always with the intention that he or she will send money home or accumulate cash to assist in setting up an independent household and acquiring land or other assets upon return. Migration is not new or necessarily damaging: physical mobility often opens up routes to social mobility. But when the decision to migrate is forced on a household or individual by lack of economic opportunities at home and often involves great personal risk, migration must be seen in a more negative light (Bebbington 1993:910).Outmigration has increased since the dam was constructed. Nearly 80% of respondents feel that migration is affecting the area, and 71% have a close relative who has migrated. My 1998 census produced a list of 348 people who had left in the past ten years. Major targets of migration are San Pedro Sula (Hondurass industrial capital) and Tegucigalpa. Significant numbers are also migrating to the United States. Nearly 50% of those interviewed have relatives in the United States, and about 16% of the households report remittances from these relatives as a significant source of income. Quantitative data on migration patterns before the dam are unavailable, but my observations lead me to believe that migration is much more common now than in the past. Conducting research in Copn, Honduras, in 2001, I asked the same questions regarding migration. The Copn and El Cajn samples are similarly rural and predominantly agricultural and are experiencing the same nationallevel politicaleconomic forces. In Copn, only 21% of respondents indicated that migration was affecting their town, only 25% of those interviewed had close relatives in the United States, and about 8% of households received remittances. That all these migration indicators are higher in El Cajn suggests that the dam has greatly accelerated migration in the region.The maquiladoras of San Pedro Sula provide migrants muchneeded employment, but wages are very low, making it difficult for people to accumulate any savings. They provide jobs to people with little education, especially young women, who in the past had few prospects after finishing primary school beyond early childbearing and marriage. Working conditions are harsh: hours are long, and there have been numerous cases of abusive behavior on the part of employers. Protection of worker health and safety is often inadequate (Flores and Kennedy 1996). All of the problems enumerated here were confirmed in interviews with local people from the El Cajn region who had worked in maquilas, who also complained of crime and the difficulties of raising children in the disorder of the periurban fringe.Migrants to the United States come from a variety of backgrounds, including the relatively welloff and the very poor. Two of the most important aspects of migration are the demonstration effects of the benefits of having a child migrate to the United States and the channels of information and images opened up by returning migrants who share their experiences when they return. In one case, a man with three children in the United States now owns a truck paid for by his children and is breaking into the potentially lucrative trade in cattle on the basis of remittances. The son of one of the betteroff shopkeepers has sent money back to buy some of the best remaining agricultural land near town for when he returns in a few years. Returning mig
Referência(s)